Research ArticleBiometry Refractive Errors and theResults of Cataract Surgery ALarge Sample Study

Ivan Hernandez-Lopez 1 Sahily Estrade-Fernandez 2 Taimı Cardenas-Dıaz 1

and Alfo Jose Batista-Leyva 2

1Instituto Cubano de Oftalmologıa ldquoRamon Pando Ferrerrdquo La Habana 11500 Cuba2Departamento de Fısica Atomica yMolecular Instituto Superior de Tecnologıas y Ciencias Aplicadas Universidad de la HabanaLa Habana 10400 Cuba

Correspondence should be addressed to Alfo Jose Batista-Leyva abatistainsteccu

Received 12 March 2021 Accepted 9 April 2021 Published 28 April 2021

Academic Editor Michele Figus

Copyright copy 2021 Ivan Hernandez-Lopez et alis is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

e statistical characteristics of biometry and refractive error in a large sample of cataractous Cuban patients are presentedcomparing between sexes and age groups All patients were studied at the Cuban Institute of Ophthalmology ldquoRamon PandoFerrerrdquo La Habana e sample consists of 28252 eyes of 25068 patients subjected to cataract surgery during the time periodbetween 2006 and 2019 eir biometry was obtained using IOL Master devices also visual acuity refraction and corneal powerwere registered After surgery the visual acuity and refraction were measurede refractive prediction error was determined Forpatients with both eyes registered anisometropia was also calculated Age and sex were used to segment the datae preoperativebiometric parameters show highly significant differences between sexes with male eyes being longer and with a deeper anteriorchamber but with a thinner lens Also keratometry shows highly significant differences with female eyes being steeper than maleBefore surgery both sexes have myopic eyes as average with males being more myopic than females (plt 0001) After surgery theaverage spherical equivalent is minus036 D and female eyes are more myopic than males (plt 0001) Visual acuity with and withoutdistance correction has a significant increase after surgery ese results are of importance not only for Cuba but also for othercountries with a large Cuban population andor similar ethnic composition such as the USA (particularly the south of Florida)Spain and many countries in Latin America

1 Introduction

Latest advances in surgical techniques and intraocular lensdesigns have notably improved the postoperative refractiveresults of modern cataract surgery increasing patientsrsquoexpectations of good visual outcome without the use ofspectacles is outcome depends on the accurate predictionof the power of the implanted intraocular lens (IOL) whichin turn depends mainly on preoperative biometry data [1]Optical biometers have become the gold standard for ocularbiometry resulting in eye measurements being more reliableand allowing the evaluation of new parameters [2] IOLMaster 700 based on swept source optical coherence to-mography (SS-OCT) is more effective in obtaining

biometric measurements in eyes with posterior subcapsularand dense nuclear cataracts than its predecessor series basedon partial coherence interferometry (PCI) However anexcellent agreement has been demonstrated between thesetwo optical technologies regarding axial length anteriorchamber depth and steepest and flattest keratometric values[3] In a recent comparative study it was found that despiteIOLMaster 700 having a superior acquisition rate thanIOLMaster 500 both devices showed good agreement andfound statistical differences were clinically negligible [4]Meanwhile knowing ocular biometric characteristics of apopulation is an important factor to be considered in de-signing intraocular lenses and improving IOL calculationmethods for selecting the most appropriate lenses for

HindawiJournal of OphthalmologyVolume 2021 Article ID 9918763 16 pageshttpsdoiorg10115520219918763

patients [5 6] e relation of ocular biometry with diabeticretinopathy has been reviewed highlighting the importanceof further biometric studies to elucidate the influence of highmyopia and ocular biometry in the protective mechanismsof the eye [7] Hence several studies describing ocularbiometric characteristics of different populations have beenconducted across the world [6 7] In an extensive review theprevalence of myopia in populations of different continentsis studied finding larger number of myopic eyes in Asianpopulations than in other regions of the world [8]

Several studies deal with the differences among gendersin the biometric parameters In Refs [9 10] authors foundsignificantly longer eyes and deeper anterior chambers inmen than in women in British and Chinese adult pop-ulations respectively ese findings are consistent with theresults of a Latino study [11] A recent review [6] found thistrend in a number of different studies Regarding averagecorneal power differences associated with sex vary amongstudies In Ref [12] studying an African population authorsdid not find differences between sexes in astigmatism eresults of different studies are difficult to compare due to thedifferences among them (ethnicity of the sample samplesize measurement method) All these factors highlight theimportance of studies associated with a given population

Published studies of ocular biometric parameters inCuban population are scarce and their casuistic have in-cluded no more than three hundred eyes [13ndash16] To ourknowledge there has not been published before a largesample study of biometric values for the Cuban populationldquoRamon Pando Ferrerrdquo Cuban Institute of Ophthalmology isa national reference center where thousands of patientsfrom all island provinces are admitted for cataract surgeryevery year For that reason it was chosen as the setting by thepresent research work in order to provide the largest andmore representative hospital-based population study ofocular biometric characteristics and refractive visual out-comes of cataract patients in Cuba e results have im-portance not only to Cuba but also to other countries with alarge Cuban population or similar ethnic composition

2 Materials and Methods

A retrospective study of 28252 eyes (female 17811 male10441) of 25068 patients (female 15745 male 9323) whounderwent uneventful cataract surgery from January 2006 toDecember 2019 was performed at the ldquoRamon PandoFerrerrdquo Cuban Institute of Ophthalmology in Havana Cubais study was approved by Institutional Scientific andEthical Committees from the Ramon Pando Ferrer CubanInstitute of Ophthalmology and the Higher Institute ofTechnology and Applied Sciences (InSTEC Spanish acro-nym) University of Havana and conformed to the tenets ofthe Declaration of Helsinki

Preoperative ocular biometric parameters includingaxial length (AL) anterior chamber depth (ACD) lensthickness (LT) and corneal power (steepest keratometry SKflattest keratometry FK and average keratometry Kav)using optical biometers (IOL Master 302 500 or 700version 114 Carl Zeiss Meditec AG Jena Germany) were

recorded as well as visual acuity (uncorrected distanceUDVA1 and corrected distance CDVA1) and refraction(sphere SPH1 cylinder CYL1 axis AX1 and sphericalequivalent SE1) Uncorrected distance (UDVA2) and cor-rected distance visual acuity (CDVA2) and refraction(sphere SPH2 cylinder CYL2 axis AX2 and sphericalequivalent SE2) after surgery were also registered SE isdefined as sphere plus half the cylinder e refractiveprediction error (PE) defined as the postoperative sphericalequivalent minus the predicted refraction was calculatedDemographic data such as age and sex were likewisedocumented Ethnicity is not registered in Cuban protocolsso its influence cannot be analyzed [17]

Axial length was divided in four groups short (lt22mm)medium (22ndash2399mm) long (24ndash26mm) and very long(gt26mm) eyes ere are reported different criteria formyopia and hyperopia [10 18] In this paper we classified asmyopic those eyes with a refractive spherical value underndash 050 diopters (D) and hyperopic eyes when these valueswere over + 050 D Astigmatism was classified as ldquowith-the-rulerdquo (wtr) when the steep meridian on the corneal surfacewas in the range 60ndash120deg or ldquoagainst-the-rulerdquo (atr) whenthe steep meridian was 0ndash30deg or 150ndash180deg e remainingcases of astigmatism were classified as oblique Keratometricastigmatism is defined as the difference of SKndashFK

e dependence of the number of patients having a givencylinder with the cylinder value (the distribution of cylindervalues) is used to fit an exponential model e model is

y A exp minusx

χ1113896 1113897 (1)

In Equation (1) x is the cylinder in diopters A is the pre-exponential factor which depends on the size of the sampley is the number of patients with the cylinder in a giveninterval with x in the center e (271828) is the base of thenatural logarithms and χ is the decaying factor whichdepends on how fast the number of patients with a givencylinder decreases with the increase of ite larger the χ theslower the decrease of y is In this sense χ characterizes thedistribution of the magnitude considered

e distribution of implanted IOL and refractive errorswere also analyzed obviously excluding from the sample thepatients who became aphakic

In patients with both eyes measured refractive aniso-metropia was considered as SE differences ge10 D betweeneyes We also considered the anisometropia in the spheree distribution of the number of eyes with a given mag-nitude of the difference is used to fit the same exponentialmodel of Equation (1) e influence of AL and ACD inanisometropia is also studied

e distribution of ocular biometric characteristics vi-sual acuity and refractive results were analyzed both acrossthe entire study population and segmented by sex (males andfemales) and age (lt40 40ndash59 60ndash80 and gt80 years old)

Statistical analyses were performed using the free soft-ware R v 363 e results are expressed as the parametermean (with a confidence interval of 95) median standarddeviation (SD) variance (Var) skewness kurtosis and the

2 Journal of Ophthalmology

interquartile range (IQR) defined as the difference betweenthe third quartile and the first one Q3ndashQ1 Skewness (c) is ameasure of the asymmetry of the distribution of the con-sidered random variable around its mean A normal dis-tribution (and every symmetric distribution) has zeroskewness A negative skew (left-skewed) indicates that theleft tail is more important in the distribution than the rightone while a positive skew (right-skewed) indicates thecontrary Kurtosis (κ) is a measure of the tailedness of thedistribution of the random variable it is related with theprobability of production of outliers according to the dis-tribution It is commonly compared with the kurtosis of anormal distribution (3) saying that a distribution is platy-kurtic if its kurtosis is smaller than three and leptokurtic if itis larger e difference between the kurtosis of a givendistribution and the normal kurtosis is named excess kur-tosis All these magnitudes characterize the random variabledistribution

Differences between two levels of a variable wereassessed for significance using the Studentrsquos t-test for un-paired samples given as t and p values ose values ofplt 005 were considered statistically significant e dif-ference between several levels is elucidated with an analysisof variance test ese criteria were used for age- and sex-related differences

A KolmogorovndashSmirnov goodness of fit test was used todetermine if the studied parameters were normally dis-tributed or not

3 Results and Discussion

31 Results Table 1 shows the selected magnitudes and theirunits as well as the number of eyes that were included andexcluded from the results for each one Note that the numberof eyes used in the analysis of the different magnitudeschanges e reasons for the differences are diverse

(i) emeasurement was not registered in the database(ii) IOL Master 302 and 500 does not measure the lens

thickness(iii) e patient became aphakic(iv) Preoperative refraction was not completed for un-

known reasons(v) Patients did not attend postoperative refraction

appointment

ough the number of patients with registered LT is muchless than the total number of eyes this number is large enoughto yield sound statistics e statistical distribution of the in-volvedmagnitudes was assessed calculating theirmomenta andrelated values meanmedian standard deviation (σ) skewnessand kurtosis In Table 2 the values of momenta and relatedquantities for the different random variables considered areshown In Table S1 of Supplementary Material the 95confidence intervals of the magnitudes (expressed as the meanplusmn196σ

n

radic σ the standard deviation and n the number of

patients) and the Interquartile Range are presented eKolmogorovndashSmirnov test showed that none of the variables isnormally distributed

e segmentation by age groups is shown in Figure S1(Supplementary Materials) Figure S1(a) shows a histogramof the distribution of age in intervals of five years Averageage is 6870 years with a standard deviation of 1089 yearse distribution is left skewed as expected and platykurticwith a kurtosis close to three Figure S1(b) shows the dis-tribution among age groups e larger percent is as usualin the age group between 60 and 80 years

Figure 1 shows the ALrsquos distribution It has two inter-esting characteristics Firstly it is right-skewed meaningthat the right tail is longer In the second place it is highlyleptokurtic with an excess kurtosis of 546 meaning that thedistribution has fat tails e combination of these twocharacteristics indicates that there are more extra-long eyes(ALgt 30mm 272 eyes) than extra-short ones (ALlt 20mm36 eyes) e average AL is 2352mm (minimum 1498mmmaximum 3740mm σ 159mm) Figure S2 shows theaxial length distribution of the sample segmented by lengthgroups Note the predominant group is in the normal lengthrange

Figure 2 presents the distribution of ACD Average ACDis 302mm (σ 044mm) with a small skewness (033) and asmall kurtosis (163) which is reflected in the thin tails thatcan be appreciated from the figure Figure S3 shows ahistogram of LT values e crystalline thickness has anaverage value of 455mm (σ 052mm) It is possible to seethat the distribution is highly symmetric (c asymp 0) and slightlyleptokurtic (κ asymp 3)

A histogram of Kav is shown in Figure 3 Its average valueis 4404 D (σ 177 D) e distribution of values of SK andFK is shown in Figures S4(a) and S4(b) respectively eaverage values are 4461 (σ 186 D) and 4347 D (σ 180 D)

e preoperative SE has a bimodal distribution as shownin Figure 4(a) e two modes are -150 D (myopic) and +150D (hypermetropic) e average preoperative SE is minus113 Dwith a standard deviation of 422 D Figure S5(a) is a histogramrepresenting the distribution of the preoperative sphere It isevident that the preoperative sphere is also bimodal (modesminus15 D and +15 D average of minus052 D) influencing thedistribution of the SEe distribution of the SE after surgery isdifferent as can be seen in Figure 4(b) In this case the dis-tribution is unimodal with an average value of minus036 D(σ 109 D)

Figure S5(b) is the histogram of SPH2 that has an averagevalue of 041 D (σ 106 D)

e distribution of cylinders has a maximum in 075 D anddecreases exponentially (accordingly with Equation (1)) as thecylinder increases as can be seen for CYL1 in Figure S6(a)edecay rate is χ 092 D with an Adjusted R2 09995 edecrease of CYL2 is also exponential (see Figure S6(b))decaying slower (decay rate of 136 D) and R2 09945

Figure 5 shows the distribution of eyes classified asmyopic emmetropic and hyperopic before and after sur-gery e distribution of refractive defects changes aftersurgery e relative proportion and the absolute number ofemmetropic eyes increase notably e number and percentof hyperopic and myopic eyes considerably decreases Notethat the number of eyes measured after surgery is larger thanbefore (see Table 1)

Journal of Ophthalmology 3

Regarding the type of astigmatism Figure 6 indicatesthat the number and percent of eyes showing atr astigmatismincreases after the surgery e number and proportion ofwtr astigmatism diminish while the number and the pro-portion of eyes with oblique astigmatism increase slightlyafter surgery Note that the number of eyes used for de-termining AX1 and AX2 is different (see Table 1)

Keratometric astigmatism has a distribution shown inFigure 7 is histogram tells us that more than half of thepatients have keratometric astigmatism below 1 D eaverage value is 115 D with a standard deviation of 096

D e distribution is right-skewed and highly leptokurticFigure S7 shows that the distribution of keratometricastigmatism has an exponential shape with a decaying factorof 0858 D

Figure 8 shows the distribution of IOL of the lensesimplanted to the patients It is easy to see that the morefrequently implanted lenses have a power between 20 D and22 D (12759 lenses 453) Figure S8 shows a histogram ofuncorrected distance (UD) and corrected distance (CD)visual acuity before and after surgery Before surgery mostof the patients have an uncorrected visual acuity below 03

Table 1 Definition of the used variables their units and the amount of eyes included and excluded

Variables Def (units) Included ExcludedAGElowast Age (years) 25068 0AL Axial length (mm) 28240 12ACD Anterior chamber depth (mm) 27923 229LT Lens thickness (mm) 4395 23857SPH1 Sphere of the preoperative refraction (D) 22818 5434CYL1 Cylinder of the preoperative refraction (D) 22818 5434AX1 Preoperative axis (degree) 22818 5434SE1 Preoperative spherical equivalent (D) 22818 5434SK Power of the steepest meridian (D) 28248 4FK Power of the flattest meridian (D) 28248 4Kav Average corneal power (D) 28248 4UDVA1 Preoperative uncorrected distance visual acuity 28252 0CDVA1 Preoperative corrected distance visual acuity 28252 0IOL IOL power (D) 28252 0SPH2 Sphere of the postoperative refraction (D) 25589 2663CYL2 Cylinder of the postoperative refraction (D) 25589 2663AX2 Postoperative axis (degrees) 25589 2663SE2 Postoperative spherical equivalent (D) 25589 2663UDVA2 Postoperative uncorrected visual acuity 28252 0CDVA2 Postoperative best corrected visual acuity 28252 0PE Prediction Error (D) 25589 2663lowastIn the AGE field only the number of patients is considered

Table 2 Statistical parameters of the considered variables

Variables Mean Median SD Var Skewness KurtosisAGE (years) 6870 7000 1089 11861 minus106 232AL (mm) 2352 2328 159 252 218 846ACD (mm) 302 302 044 019 033 163LT (mm) 455 455 052 027 minus024 341SPH1 (D) minus052 minus025 417 1737 minus090 427CYL1 (D) 122 100 090 081 148 404AX1 (deg) 8278 8500 4396 193254 minus002 002SE1 (D) minus113 minus088 422 1784 minus095 424SK (D) 4461 4464 186 347 minus053 583FK (D) 4347 4354 180 323 minus084 532Kav (D) 4404 4407 177 312 minus073 542UDVA1 016 010 014 002 183 398CDVA1 040 040 027 007 054 -041IOL (D) 2065 2100 435 1889 minus218 886SPH2 (D) 041 050 106 113 072 2254CYL2 (D) 152 125 108 117 140 298AX2 (deg) 9362 9500 3850 148239 minus027 066SE2 (D) minus036 minus025 109 103 039 2890UDVA2 046 040 028 008 039 minus081CDVA2 077 090 028 008 minus111 016PE (D) 028 034 099 098 minus006 1944

4 Journal of Ophthalmology

while the corrected visual acuity is below 06 in a majority ofthe cases After surgery on the contrary 47 of the patientshave uncorrected distance visual acuity above 05 and 64

have corrected distance visual acuity above 08 As thecataract has a major influence on visual acuity the post-operative results are more important In Figure 9 thepostoperative uncorrected and corrected distance visualacuity are compared Figure 9(a) compares the cumulativepercent of both quantities while Figure 9(b) shows thedifference between them in terms of Snellen lines

Figure 10 shows the distribution of the absolute value ofPE in different intervals In the figure both the number andpercent of eyes in each interval are presented Around 75of eyes reach a postoperative refraction that differs less thanone diopter of the predicted one Figure S9 shows thedistribution of PE with a mean value of 028 D (σ 099 D)e presence of refractive surprises is signaled by the highvalue of kurtosis is magnitude and its dependence withthe formula used to calculate the prediction refraction(including the optimization of the constants used) as well asthe lens model are so important that they deserve a separatestudy It will be conducted in a forthcoming paperFigure S10 shows the distribution of anisometropia beforesurgery Figure S10(a) represents the distribution of theabsolute value of the difference SPH1OS-SPH1OD and the fitof an exponential model (Equation (1)) to the experimentaldata Figure S10(b) is similar but the magnitude is related tothe absolute value of the difference SE1OS-SE1OD Overall ofthe 1863 pair of eyes having preoperative data 1038 (557)present anisometropia of the sphere while 959 (515)present anisometropia of the spherical equivalent Bothdistributions have an exponential decay (R2gt 099) withsimilar decaying factors (116 and 111 D respectively)

e effects of the surgery on the anisometropia areevident by analyzing Figure S11 Figure S11(a) shows thedistribution of the SPH2rsquos anisometropia andFigure S11(b) shows the distribution of the anisometropiain SE2 Of the 2632 measured pairs of eyes 636 (242)show anisometropia in the sphere and 773 (294) in thespherical equivalent Not only the proportion of affectedeyes diminishes but also the prevalence of large aniso-metropias decreases It is evident in the values of thedecaying factor (071 and 0620 D respectively)

An additional study of the anisometropia in the biometricparameters of the mate eyes was performed e difference |ALOSndashALOD| has an average value of 024mm (σ 045mmIQR 017mm) When the same calculation is performed withthe ACD |ACDOSndashACDOD| has an average value of 018mmσ 032mm and IQR 017mm

Table 3 reports some statigraphs of different magni-tudes segmented by sex Female eyes are shorter and haveshallower anterior chambers than male eyes e averageLT of both sexes are statistically different with femalelenses thicker than the male ones ese factors combinedimply that men have deeper vitreous chambers than fe-males e power of the femalersquos corneas is consistentlybigger than that of males It is so not only for the averagepower but also for the extreme keratometries

e average value of SE1 is more myopic in malepatients than in female ones After surgery though thedifferences between the average SE of males and femalesdiminish the values are significantly different CYL1 and

5000

4000

3000

Patie

nts

2000

1000

015 20 25

AL (mm)30 35

Av = 2352mmσ = 159mmγ = 218κ = 846

Figure 1 Histogram of the axial length with 05mm bins In theinset the main statigraphs of the distribution are shown

2500

3000

2000

1500

Patie

nts

1000

500

01 2 3

ACD (mm)4 5 6

Av = 302mmσ = 044mmγ = 033κ = 163

Figure 2 Histogram of the anterior chamber depth with bins of02mm In the inset the main statigraphs of the distribution areshown

3500

4000

3000

2500

2000

Patie

nts

1500

1000

500

030 35 40

Kav (D)45 50 55

Av = 4404Dσ = 177Dγ = ndash073κ = 542

Figure 3 Histogram of the average corneal power with 05 D binsIn the inset the main statigraphs of the distribution are shown

Journal of Ophthalmology 5

+ 15ndash 153000

2500

2000

1500Pa

tient

s

1000

500

0ndash25 ndash20 ndash15 ndash10

SE1 (D)ndash5 0 5 10 15

Av = ndash113Dσ = 422Dγ = ndash095κ = 424

(a)

6000

5000

4000

3000

2000

1000

0ndash10 ndash5 0

SE2 (D)5 10 15

Av = ndash036Dσ = 109Dγ = 039κ = 2890

(b)

Figure 4 Histograms of SE with 1 D bins (a) Preoperative it is possible to note a bimodal distribution (b) Postoperative Comparing bothfigures note the effect of the surgical procedure In the insets the main statigraphs of the distributions are shown

17500

636

696

243172

121 132

15000

12500

10000

Patie

nts

7500

5000

2500

0a t r w t r

Axis orientationo b l

PrePost

Figure 6 Proportion of type of astigmatism preoperative and postoperative wtr with the rule atr against the rule obl oblique Note thatthe total number of patients preoperative and postoperative are different (see Table 1)

14000

460369

PrePost

121

515

419

116

12000

10000

Patie

nts

8000

6000

4000

2000

0Myopic Emmetropic

RefractionHyperopic

Figure 5 Classification of the refraction preoperative and postoperative Note that the total number of patients in the preoperative andpostoperative stages is different (see Table 1)

6 Journal of Ophthalmology

CYL2 are different between sexes though in the post-stagethe differences between the distributions are lessimportant

Consistent with the biometric findings female eyes needin average more powerful implanted lenses than men estandard deviation skewness and kurtosis of the

9000

7500

6000222

323 324

87

27 17

4500Patie

nts

3000

1500

0lt05 05ndash099 1ndash199

Keratometric astigmatism (D)2ndash299 3ndash399 ge4

Figure 7 Histogram of the keratometric astigmatism

100

80

60

Cum

ulat

ive

of e

yes

40

UC postopCD postop

20

066 68 69 612 615Cumulative Snellen visual acuity (6x or better)

624 630 660

(a)

40

444

138 128

002

29030

20

10

03 or gtworst

2 worst 1 worst

Difference between UDVA and CDVA(Snellen lines)

Same 1 or gtbetter

o

f eye

s

22852 eyes1-month postop

UDVA same or better than CDVA 2902UDVA within 1 line of CDVA 4182

(b)

Figure 9 Distribution of the postoperative visual acuity (both uncorrected distance UD and corrected distance CD) (a)e percent of eyeswith a given visual acuity (b) Differences between UD and CD in terms of Snellen lines

4000

3000

Patie

nts

2000

1000

00 5 10 15

IOL power (D)20 25 30 35

Av = 2065Dσ = 435Dγ = ndash218κ = 886

Figure 8 Histogram of the implanted lens power with 1 D bin In the inset the values of the main statigraphs are shown

Journal of Ophthalmology 7

distributions segmented by sex are most of times very closeimplying that both distributions are only displaced havingsimilar shapes

Figure 11 presents the percent distribution of refractionbefore and after surgery segmented by sex e distributionbefore surgery shows that for males the majority of eyes(60) are myopic while for females the prevalence of

myopia is larger than the prevalence of hyperopia but thedifference is not as large Both sexes present a small numberof emmetropic eyes slightly above 10 After surgery theproportion of emmetropic eyes increases up to 50 di-minishing the percent differences among sexes in the threesegments Regarding anisometropia both sexes present thesame prevalence (p 045)

IntervalAccumulated

25000

30000

20000

15000

Patie

nts

10000

5000

0le025 025ndash050 050ndash100

|PE| (D)gt100

224 205

429322

751

249

100

Figure 10 Distribution of the absolute value of the PE in intervals and its cumulative values

Table 3 Gender segmentation of the data

Variables Sex Mean Median SD Var Skew Kurt Number t-test p value

Age F 6877 7000 1070 11449 minus094 181 17811 078 043M 6867 7000 1093 11951 minus124 309 10441

AL F 2331 2302 163 267 235 866 17802 3091 lt0001M 2388 2372 143 205 230 1062 10437

ACD F 295 295 042 018 033 140 17628 3166 lt0001M 313 313 044 020 031 222 10294

LT F 457 458 050 025 minus023 207 2681 258 001M 453 450 055 031 minus022 470 1714

SK F 4487 4488 183 336 minus060 602 17807 3091 lt0001M 4417 4418 183 334 minus048 673 10440

FK F 4374 4377 177 315 minus096 652 17807 3431 lt0001M 4300 4305 174 303 minus077 466 10440

Kav F 4431 4433 174 303 minus081 583 17807 3383 lt0001M 4359 4364 172 294 minus073 608 10440

SE1 F minus091 minus038 436 1903 minus114 426 12989 1005 lt0001M minus152 minus150 403 1622 minus061 439 7498

SE2 F minus037 minus038 103 106 036 2966 14269 345 lt0001M minus032 minus025 102 103 021 2828 8296

CYL1 F minus118 minus100 089 079 minus161 478 12989 707 lt0001M minus127 minus100 092 085 minus137 343 7498

CYL2 F minus149 minus125 107 116 minus145 321 14269 336 lt0001M minus154 minus125 109 119 minus140 315 8296

IOL F 2104 2200 457 2088 minus223 818 17810 2062 lt0001M 1999 2000 385 1480 minus232 1199 10441

PE F 027 032 103 106 minus032 2716 14269 175 008M 029 034 100 099 014 1993 8296A Studentrsquos t test for unpaired samples was performed in order to compare the means of each variable among sexes pndashvaluelt 005 is considered as highlysignificant Nonsignificant values are shown in boldface

8 Journal of Ophthalmology

Table 4 reports the segmentation by age of the biometricmagnitudes as well as the refraction (pre- and post-) and PEfor the selected age groups A one-way ANOVA test wasperformed to detect differences between mean values of thedifferent groups e bottom row of the table gives theF-values obtained as well as the real power of the test evalues per age group of AL LT ACD and SE1 present highlysignificant differences between all groups Table 5 shows thebehavior of the average keratometry e age group of 41ndash60presents differences with all other groups that do not presentdifferences among them PE on the contrary has a differentbehavior explained in Table 6 SE2 does not present dif-ferences among age groups

Note that in all these magnitudes the standard deviationand other statigraphs (not shownhere) change between groups

32 Discussion e results of the present hospital-basedpopulation study constitute to our knowledge the largestand more representative investigation on ocular biometricparameters of Cuban cataract patients using optical biom-etry In addition extensive data on refractive errors andvisual outcomes distribution are also provided by this work

Missing information of some variables in numerouscases due to aforementioned dissimilar reasons makes itmandatory to exclude certain amount of analyzed eyes forspecific parameters Since available data of eachmagnitude isenormous these exclusions did not hinder from obtainingrepresentative results of every one of them

Mean age in our study (6870 years) is consistent with thefrequent cataract association to senescence and its left-skewed distribution showing that the larger percent in theage group between 60 and 80 years behaves similar as vastlyreported on literature in all latitudes [19 20] Most previousCuban studies [13 15 21] agree with our finding e av-erage age in Ref [22] is smaller than the value found in thisstudy It also happens in Ref [16] dedicated to study younghyperopes patients Ref [14] does not report age data

e standard value of AL of the human eye is inter-nationally taken to be around 24mm in adulthood re-gardless of sex or race whereas AL tends to be longer inmyopic and shorter in hypermetropic eyes comparing to thatof emmetropic [23] e right-skewed distribution is inaccordance with the predominance of myopia in the studiedsample AL average values are variables in the literature emean AL observed in this paper is larger than those of Refs[19 22 24] in which ultrasound low-coherence reflec-tometry and partial coherence interferometry are usedrespectively Conversely Refs [2 25] using low-coherencereflectometry and partial coherence interferometry re-spectively show larger average values Previous researches inCuba report larger [26] much larger [15] shorter [27] andmuch shorter [16] average AL when compared with ourresults It should be noted that the analyzed periods in allthese works include only one year or less and that thesample size is much smaller than ours explaining the dis-persion in their results Specifically in the case of Ref [15]more than 93 of patients hadmyopic astigmatismwhile theresult of Ref [16] is in accordance with his studyacutes designwhere all the patients were hyperopes

Average ACD in our work (302mm) is shallower thanreported in Latino populations [11] and to a lesser extentthose of Refs [28 29] in Asian populations On the otherhand Hashemi et al [24] report an ACD average evenshallower than ours using the biometer LENSTARBioG-raph (WaveLight AG Erlangen Germany) Differencesbetween these two biometers may affect the comparativeanalysis since ACD value in Lenstar refers to the distancebetween corneal endothelium and lens anterior capsulewhile the distance for this variable in IOL Master is takenfrom corneal epithelium to lens anterior capsuleat is whyLenstar uses the sum of central pachymetry and ACD valuesfor IOL calculation [27] erefore this issue has to beconsidered when comparing Lenstar and IOL Master asCuban researchers did in a previous study [27] showingvalues closer to ours with no significant differences between

7000484

604

377

Fem-preMale-pre

Fem-postMale-post

353

130108

509

524

408

266113

122

6000

5000

4000

Patie

nts

3000

2000

1000

0Myopic Emmetropic

RefractionHyperopic

Figure 11 Classification of the refraction preoperative and postoperative segmented by sex

Journal of Ophthalmology 9

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

interquartile range (IQR) defined as the difference betweenthe third quartile and the first one Q3ndashQ1 Skewness (c) is ameasure of the asymmetry of the distribution of the con-sidered random variable around its mean A normal dis-tribution (and every symmetric distribution) has zeroskewness A negative skew (left-skewed) indicates that theleft tail is more important in the distribution than the rightone while a positive skew (right-skewed) indicates thecontrary Kurtosis (κ) is a measure of the tailedness of thedistribution of the random variable it is related with theprobability of production of outliers according to the dis-tribution It is commonly compared with the kurtosis of anormal distribution (3) saying that a distribution is platy-kurtic if its kurtosis is smaller than three and leptokurtic if itis larger e difference between the kurtosis of a givendistribution and the normal kurtosis is named excess kur-tosis All these magnitudes characterize the random variabledistribution

Differences between two levels of a variable wereassessed for significance using the Studentrsquos t-test for un-paired samples given as t and p values ose values ofplt 005 were considered statistically significant e dif-ference between several levels is elucidated with an analysisof variance test ese criteria were used for age- and sex-related differences

A KolmogorovndashSmirnov goodness of fit test was used todetermine if the studied parameters were normally dis-tributed or not

3 Results and Discussion

31 Results Table 1 shows the selected magnitudes and theirunits as well as the number of eyes that were included andexcluded from the results for each one Note that the numberof eyes used in the analysis of the different magnitudeschanges e reasons for the differences are diverse

(i) emeasurement was not registered in the database(ii) IOL Master 302 and 500 does not measure the lens

thickness(iii) e patient became aphakic(iv) Preoperative refraction was not completed for un-

known reasons(v) Patients did not attend postoperative refraction

appointment

ough the number of patients with registered LT is muchless than the total number of eyes this number is large enoughto yield sound statistics e statistical distribution of the in-volvedmagnitudes was assessed calculating theirmomenta andrelated values meanmedian standard deviation (σ) skewnessand kurtosis In Table 2 the values of momenta and relatedquantities for the different random variables considered areshown In Table S1 of Supplementary Material the 95confidence intervals of the magnitudes (expressed as the meanplusmn196σ

n

radic σ the standard deviation and n the number of

patients) and the Interquartile Range are presented eKolmogorovndashSmirnov test showed that none of the variables isnormally distributed

e segmentation by age groups is shown in Figure S1(Supplementary Materials) Figure S1(a) shows a histogramof the distribution of age in intervals of five years Averageage is 6870 years with a standard deviation of 1089 yearse distribution is left skewed as expected and platykurticwith a kurtosis close to three Figure S1(b) shows the dis-tribution among age groups e larger percent is as usualin the age group between 60 and 80 years

Figure 1 shows the ALrsquos distribution It has two inter-esting characteristics Firstly it is right-skewed meaningthat the right tail is longer In the second place it is highlyleptokurtic with an excess kurtosis of 546 meaning that thedistribution has fat tails e combination of these twocharacteristics indicates that there are more extra-long eyes(ALgt 30mm 272 eyes) than extra-short ones (ALlt 20mm36 eyes) e average AL is 2352mm (minimum 1498mmmaximum 3740mm σ 159mm) Figure S2 shows theaxial length distribution of the sample segmented by lengthgroups Note the predominant group is in the normal lengthrange

Figure 2 presents the distribution of ACD Average ACDis 302mm (σ 044mm) with a small skewness (033) and asmall kurtosis (163) which is reflected in the thin tails thatcan be appreciated from the figure Figure S3 shows ahistogram of LT values e crystalline thickness has anaverage value of 455mm (σ 052mm) It is possible to seethat the distribution is highly symmetric (c asymp 0) and slightlyleptokurtic (κ asymp 3)

A histogram of Kav is shown in Figure 3 Its average valueis 4404 D (σ 177 D) e distribution of values of SK andFK is shown in Figures S4(a) and S4(b) respectively eaverage values are 4461 (σ 186 D) and 4347 D (σ 180 D)

e preoperative SE has a bimodal distribution as shownin Figure 4(a) e two modes are -150 D (myopic) and +150D (hypermetropic) e average preoperative SE is minus113 Dwith a standard deviation of 422 D Figure S5(a) is a histogramrepresenting the distribution of the preoperative sphere It isevident that the preoperative sphere is also bimodal (modesminus15 D and +15 D average of minus052 D) influencing thedistribution of the SEe distribution of the SE after surgery isdifferent as can be seen in Figure 4(b) In this case the dis-tribution is unimodal with an average value of minus036 D(σ 109 D)

Figure S5(b) is the histogram of SPH2 that has an averagevalue of 041 D (σ 106 D)

e distribution of cylinders has a maximum in 075 D anddecreases exponentially (accordingly with Equation (1)) as thecylinder increases as can be seen for CYL1 in Figure S6(a)edecay rate is χ 092 D with an Adjusted R2 09995 edecrease of CYL2 is also exponential (see Figure S6(b))decaying slower (decay rate of 136 D) and R2 09945

Figure 5 shows the distribution of eyes classified asmyopic emmetropic and hyperopic before and after sur-gery e distribution of refractive defects changes aftersurgery e relative proportion and the absolute number ofemmetropic eyes increase notably e number and percentof hyperopic and myopic eyes considerably decreases Notethat the number of eyes measured after surgery is larger thanbefore (see Table 1)

Journal of Ophthalmology 3

Regarding the type of astigmatism Figure 6 indicatesthat the number and percent of eyes showing atr astigmatismincreases after the surgery e number and proportion ofwtr astigmatism diminish while the number and the pro-portion of eyes with oblique astigmatism increase slightlyafter surgery Note that the number of eyes used for de-termining AX1 and AX2 is different (see Table 1)

Keratometric astigmatism has a distribution shown inFigure 7 is histogram tells us that more than half of thepatients have keratometric astigmatism below 1 D eaverage value is 115 D with a standard deviation of 096

D e distribution is right-skewed and highly leptokurticFigure S7 shows that the distribution of keratometricastigmatism has an exponential shape with a decaying factorof 0858 D

Figure 8 shows the distribution of IOL of the lensesimplanted to the patients It is easy to see that the morefrequently implanted lenses have a power between 20 D and22 D (12759 lenses 453) Figure S8 shows a histogram ofuncorrected distance (UD) and corrected distance (CD)visual acuity before and after surgery Before surgery mostof the patients have an uncorrected visual acuity below 03

Table 1 Definition of the used variables their units and the amount of eyes included and excluded

Variables Def (units) Included ExcludedAGElowast Age (years) 25068 0AL Axial length (mm) 28240 12ACD Anterior chamber depth (mm) 27923 229LT Lens thickness (mm) 4395 23857SPH1 Sphere of the preoperative refraction (D) 22818 5434CYL1 Cylinder of the preoperative refraction (D) 22818 5434AX1 Preoperative axis (degree) 22818 5434SE1 Preoperative spherical equivalent (D) 22818 5434SK Power of the steepest meridian (D) 28248 4FK Power of the flattest meridian (D) 28248 4Kav Average corneal power (D) 28248 4UDVA1 Preoperative uncorrected distance visual acuity 28252 0CDVA1 Preoperative corrected distance visual acuity 28252 0IOL IOL power (D) 28252 0SPH2 Sphere of the postoperative refraction (D) 25589 2663CYL2 Cylinder of the postoperative refraction (D) 25589 2663AX2 Postoperative axis (degrees) 25589 2663SE2 Postoperative spherical equivalent (D) 25589 2663UDVA2 Postoperative uncorrected visual acuity 28252 0CDVA2 Postoperative best corrected visual acuity 28252 0PE Prediction Error (D) 25589 2663lowastIn the AGE field only the number of patients is considered

Table 2 Statistical parameters of the considered variables

Variables Mean Median SD Var Skewness KurtosisAGE (years) 6870 7000 1089 11861 minus106 232AL (mm) 2352 2328 159 252 218 846ACD (mm) 302 302 044 019 033 163LT (mm) 455 455 052 027 minus024 341SPH1 (D) minus052 minus025 417 1737 minus090 427CYL1 (D) 122 100 090 081 148 404AX1 (deg) 8278 8500 4396 193254 minus002 002SE1 (D) minus113 minus088 422 1784 minus095 424SK (D) 4461 4464 186 347 minus053 583FK (D) 4347 4354 180 323 minus084 532Kav (D) 4404 4407 177 312 minus073 542UDVA1 016 010 014 002 183 398CDVA1 040 040 027 007 054 -041IOL (D) 2065 2100 435 1889 minus218 886SPH2 (D) 041 050 106 113 072 2254CYL2 (D) 152 125 108 117 140 298AX2 (deg) 9362 9500 3850 148239 minus027 066SE2 (D) minus036 minus025 109 103 039 2890UDVA2 046 040 028 008 039 minus081CDVA2 077 090 028 008 minus111 016PE (D) 028 034 099 098 minus006 1944

4 Journal of Ophthalmology

while the corrected visual acuity is below 06 in a majority ofthe cases After surgery on the contrary 47 of the patientshave uncorrected distance visual acuity above 05 and 64

have corrected distance visual acuity above 08 As thecataract has a major influence on visual acuity the post-operative results are more important In Figure 9 thepostoperative uncorrected and corrected distance visualacuity are compared Figure 9(a) compares the cumulativepercent of both quantities while Figure 9(b) shows thedifference between them in terms of Snellen lines

Figure 10 shows the distribution of the absolute value ofPE in different intervals In the figure both the number andpercent of eyes in each interval are presented Around 75of eyes reach a postoperative refraction that differs less thanone diopter of the predicted one Figure S9 shows thedistribution of PE with a mean value of 028 D (σ 099 D)e presence of refractive surprises is signaled by the highvalue of kurtosis is magnitude and its dependence withthe formula used to calculate the prediction refraction(including the optimization of the constants used) as well asthe lens model are so important that they deserve a separatestudy It will be conducted in a forthcoming paperFigure S10 shows the distribution of anisometropia beforesurgery Figure S10(a) represents the distribution of theabsolute value of the difference SPH1OS-SPH1OD and the fitof an exponential model (Equation (1)) to the experimentaldata Figure S10(b) is similar but the magnitude is related tothe absolute value of the difference SE1OS-SE1OD Overall ofthe 1863 pair of eyes having preoperative data 1038 (557)present anisometropia of the sphere while 959 (515)present anisometropia of the spherical equivalent Bothdistributions have an exponential decay (R2gt 099) withsimilar decaying factors (116 and 111 D respectively)

e effects of the surgery on the anisometropia areevident by analyzing Figure S11 Figure S11(a) shows thedistribution of the SPH2rsquos anisometropia andFigure S11(b) shows the distribution of the anisometropiain SE2 Of the 2632 measured pairs of eyes 636 (242)show anisometropia in the sphere and 773 (294) in thespherical equivalent Not only the proportion of affectedeyes diminishes but also the prevalence of large aniso-metropias decreases It is evident in the values of thedecaying factor (071 and 0620 D respectively)

An additional study of the anisometropia in the biometricparameters of the mate eyes was performed e difference |ALOSndashALOD| has an average value of 024mm (σ 045mmIQR 017mm) When the same calculation is performed withthe ACD |ACDOSndashACDOD| has an average value of 018mmσ 032mm and IQR 017mm

Table 3 reports some statigraphs of different magni-tudes segmented by sex Female eyes are shorter and haveshallower anterior chambers than male eyes e averageLT of both sexes are statistically different with femalelenses thicker than the male ones ese factors combinedimply that men have deeper vitreous chambers than fe-males e power of the femalersquos corneas is consistentlybigger than that of males It is so not only for the averagepower but also for the extreme keratometries

e average value of SE1 is more myopic in malepatients than in female ones After surgery though thedifferences between the average SE of males and femalesdiminish the values are significantly different CYL1 and

5000

4000

3000

Patie

nts

2000

1000

015 20 25

AL (mm)30 35

Av = 2352mmσ = 159mmγ = 218κ = 846

Figure 1 Histogram of the axial length with 05mm bins In theinset the main statigraphs of the distribution are shown

2500

3000

2000

1500

Patie

nts

1000

500

01 2 3

ACD (mm)4 5 6

Av = 302mmσ = 044mmγ = 033κ = 163

Figure 2 Histogram of the anterior chamber depth with bins of02mm In the inset the main statigraphs of the distribution areshown

3500

4000

3000

2500

2000

Patie

nts

1500

1000

500

030 35 40

Kav (D)45 50 55

Av = 4404Dσ = 177Dγ = ndash073κ = 542

Figure 3 Histogram of the average corneal power with 05 D binsIn the inset the main statigraphs of the distribution are shown

Journal of Ophthalmology 5

+ 15ndash 153000

2500

2000

1500Pa

tient

s

1000

500

0ndash25 ndash20 ndash15 ndash10

SE1 (D)ndash5 0 5 10 15

Av = ndash113Dσ = 422Dγ = ndash095κ = 424

(a)

6000

5000

4000

3000

2000

1000

0ndash10 ndash5 0

SE2 (D)5 10 15

Av = ndash036Dσ = 109Dγ = 039κ = 2890

(b)

Figure 4 Histograms of SE with 1 D bins (a) Preoperative it is possible to note a bimodal distribution (b) Postoperative Comparing bothfigures note the effect of the surgical procedure In the insets the main statigraphs of the distributions are shown

17500

636

696

243172

121 132

15000

12500

10000

Patie

nts

7500

5000

2500

0a t r w t r

Axis orientationo b l

PrePost

Figure 6 Proportion of type of astigmatism preoperative and postoperative wtr with the rule atr against the rule obl oblique Note thatthe total number of patients preoperative and postoperative are different (see Table 1)

14000

460369

PrePost

121

515

419

116

12000

10000

Patie

nts

8000

6000

4000

2000

0Myopic Emmetropic

RefractionHyperopic

Figure 5 Classification of the refraction preoperative and postoperative Note that the total number of patients in the preoperative andpostoperative stages is different (see Table 1)

6 Journal of Ophthalmology

CYL2 are different between sexes though in the post-stagethe differences between the distributions are lessimportant

Consistent with the biometric findings female eyes needin average more powerful implanted lenses than men estandard deviation skewness and kurtosis of the

9000

7500

6000222

323 324

87

27 17

4500Patie

nts

3000

1500

0lt05 05ndash099 1ndash199

Keratometric astigmatism (D)2ndash299 3ndash399 ge4

Figure 7 Histogram of the keratometric astigmatism

100

80

60

Cum

ulat

ive

of e

yes

40

UC postopCD postop

20

066 68 69 612 615Cumulative Snellen visual acuity (6x or better)

624 630 660

(a)

40

444

138 128

002

29030

20

10

03 or gtworst

2 worst 1 worst

Difference between UDVA and CDVA(Snellen lines)

Same 1 or gtbetter

o

f eye

s

22852 eyes1-month postop

UDVA same or better than CDVA 2902UDVA within 1 line of CDVA 4182

(b)

Figure 9 Distribution of the postoperative visual acuity (both uncorrected distance UD and corrected distance CD) (a)e percent of eyeswith a given visual acuity (b) Differences between UD and CD in terms of Snellen lines

4000

3000

Patie

nts

2000

1000

00 5 10 15

IOL power (D)20 25 30 35

Av = 2065Dσ = 435Dγ = ndash218κ = 886

Figure 8 Histogram of the implanted lens power with 1 D bin In the inset the values of the main statigraphs are shown

Journal of Ophthalmology 7

distributions segmented by sex are most of times very closeimplying that both distributions are only displaced havingsimilar shapes

Figure 11 presents the percent distribution of refractionbefore and after surgery segmented by sex e distributionbefore surgery shows that for males the majority of eyes(60) are myopic while for females the prevalence of

myopia is larger than the prevalence of hyperopia but thedifference is not as large Both sexes present a small numberof emmetropic eyes slightly above 10 After surgery theproportion of emmetropic eyes increases up to 50 di-minishing the percent differences among sexes in the threesegments Regarding anisometropia both sexes present thesame prevalence (p 045)

IntervalAccumulated

25000

30000

20000

15000

Patie

nts

10000

5000

0le025 025ndash050 050ndash100

|PE| (D)gt100

224 205

429322

751

249

100

Figure 10 Distribution of the absolute value of the PE in intervals and its cumulative values

Table 3 Gender segmentation of the data

Variables Sex Mean Median SD Var Skew Kurt Number t-test p value

Age F 6877 7000 1070 11449 minus094 181 17811 078 043M 6867 7000 1093 11951 minus124 309 10441

AL F 2331 2302 163 267 235 866 17802 3091 lt0001M 2388 2372 143 205 230 1062 10437

ACD F 295 295 042 018 033 140 17628 3166 lt0001M 313 313 044 020 031 222 10294

LT F 457 458 050 025 minus023 207 2681 258 001M 453 450 055 031 minus022 470 1714

SK F 4487 4488 183 336 minus060 602 17807 3091 lt0001M 4417 4418 183 334 minus048 673 10440

FK F 4374 4377 177 315 minus096 652 17807 3431 lt0001M 4300 4305 174 303 minus077 466 10440

Kav F 4431 4433 174 303 minus081 583 17807 3383 lt0001M 4359 4364 172 294 minus073 608 10440

SE1 F minus091 minus038 436 1903 minus114 426 12989 1005 lt0001M minus152 minus150 403 1622 minus061 439 7498

SE2 F minus037 minus038 103 106 036 2966 14269 345 lt0001M minus032 minus025 102 103 021 2828 8296

CYL1 F minus118 minus100 089 079 minus161 478 12989 707 lt0001M minus127 minus100 092 085 minus137 343 7498

CYL2 F minus149 minus125 107 116 minus145 321 14269 336 lt0001M minus154 minus125 109 119 minus140 315 8296

IOL F 2104 2200 457 2088 minus223 818 17810 2062 lt0001M 1999 2000 385 1480 minus232 1199 10441

PE F 027 032 103 106 minus032 2716 14269 175 008M 029 034 100 099 014 1993 8296A Studentrsquos t test for unpaired samples was performed in order to compare the means of each variable among sexes pndashvaluelt 005 is considered as highlysignificant Nonsignificant values are shown in boldface

8 Journal of Ophthalmology

Table 4 reports the segmentation by age of the biometricmagnitudes as well as the refraction (pre- and post-) and PEfor the selected age groups A one-way ANOVA test wasperformed to detect differences between mean values of thedifferent groups e bottom row of the table gives theF-values obtained as well as the real power of the test evalues per age group of AL LT ACD and SE1 present highlysignificant differences between all groups Table 5 shows thebehavior of the average keratometry e age group of 41ndash60presents differences with all other groups that do not presentdifferences among them PE on the contrary has a differentbehavior explained in Table 6 SE2 does not present dif-ferences among age groups

Note that in all these magnitudes the standard deviationand other statigraphs (not shownhere) change between groups

32 Discussion e results of the present hospital-basedpopulation study constitute to our knowledge the largestand more representative investigation on ocular biometricparameters of Cuban cataract patients using optical biom-etry In addition extensive data on refractive errors andvisual outcomes distribution are also provided by this work

Missing information of some variables in numerouscases due to aforementioned dissimilar reasons makes itmandatory to exclude certain amount of analyzed eyes forspecific parameters Since available data of eachmagnitude isenormous these exclusions did not hinder from obtainingrepresentative results of every one of them

Mean age in our study (6870 years) is consistent with thefrequent cataract association to senescence and its left-skewed distribution showing that the larger percent in theage group between 60 and 80 years behaves similar as vastlyreported on literature in all latitudes [19 20] Most previousCuban studies [13 15 21] agree with our finding e av-erage age in Ref [22] is smaller than the value found in thisstudy It also happens in Ref [16] dedicated to study younghyperopes patients Ref [14] does not report age data

e standard value of AL of the human eye is inter-nationally taken to be around 24mm in adulthood re-gardless of sex or race whereas AL tends to be longer inmyopic and shorter in hypermetropic eyes comparing to thatof emmetropic [23] e right-skewed distribution is inaccordance with the predominance of myopia in the studiedsample AL average values are variables in the literature emean AL observed in this paper is larger than those of Refs[19 22 24] in which ultrasound low-coherence reflec-tometry and partial coherence interferometry are usedrespectively Conversely Refs [2 25] using low-coherencereflectometry and partial coherence interferometry re-spectively show larger average values Previous researches inCuba report larger [26] much larger [15] shorter [27] andmuch shorter [16] average AL when compared with ourresults It should be noted that the analyzed periods in allthese works include only one year or less and that thesample size is much smaller than ours explaining the dis-persion in their results Specifically in the case of Ref [15]more than 93 of patients hadmyopic astigmatismwhile theresult of Ref [16] is in accordance with his studyacutes designwhere all the patients were hyperopes

Average ACD in our work (302mm) is shallower thanreported in Latino populations [11] and to a lesser extentthose of Refs [28 29] in Asian populations On the otherhand Hashemi et al [24] report an ACD average evenshallower than ours using the biometer LENSTARBioG-raph (WaveLight AG Erlangen Germany) Differencesbetween these two biometers may affect the comparativeanalysis since ACD value in Lenstar refers to the distancebetween corneal endothelium and lens anterior capsulewhile the distance for this variable in IOL Master is takenfrom corneal epithelium to lens anterior capsuleat is whyLenstar uses the sum of central pachymetry and ACD valuesfor IOL calculation [27] erefore this issue has to beconsidered when comparing Lenstar and IOL Master asCuban researchers did in a previous study [27] showingvalues closer to ours with no significant differences between

7000484

604

377

Fem-preMale-pre

Fem-postMale-post

353

130108

509

524

408

266113

122

6000

5000

4000

Patie

nts

3000

2000

1000

0Myopic Emmetropic

RefractionHyperopic

Figure 11 Classification of the refraction preoperative and postoperative segmented by sex

Journal of Ophthalmology 9

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

Regarding the type of astigmatism Figure 6 indicatesthat the number and percent of eyes showing atr astigmatismincreases after the surgery e number and proportion ofwtr astigmatism diminish while the number and the pro-portion of eyes with oblique astigmatism increase slightlyafter surgery Note that the number of eyes used for de-termining AX1 and AX2 is different (see Table 1)

Keratometric astigmatism has a distribution shown inFigure 7 is histogram tells us that more than half of thepatients have keratometric astigmatism below 1 D eaverage value is 115 D with a standard deviation of 096

D e distribution is right-skewed and highly leptokurticFigure S7 shows that the distribution of keratometricastigmatism has an exponential shape with a decaying factorof 0858 D

Figure 8 shows the distribution of IOL of the lensesimplanted to the patients It is easy to see that the morefrequently implanted lenses have a power between 20 D and22 D (12759 lenses 453) Figure S8 shows a histogram ofuncorrected distance (UD) and corrected distance (CD)visual acuity before and after surgery Before surgery mostof the patients have an uncorrected visual acuity below 03

Table 1 Definition of the used variables their units and the amount of eyes included and excluded

Variables Def (units) Included ExcludedAGElowast Age (years) 25068 0AL Axial length (mm) 28240 12ACD Anterior chamber depth (mm) 27923 229LT Lens thickness (mm) 4395 23857SPH1 Sphere of the preoperative refraction (D) 22818 5434CYL1 Cylinder of the preoperative refraction (D) 22818 5434AX1 Preoperative axis (degree) 22818 5434SE1 Preoperative spherical equivalent (D) 22818 5434SK Power of the steepest meridian (D) 28248 4FK Power of the flattest meridian (D) 28248 4Kav Average corneal power (D) 28248 4UDVA1 Preoperative uncorrected distance visual acuity 28252 0CDVA1 Preoperative corrected distance visual acuity 28252 0IOL IOL power (D) 28252 0SPH2 Sphere of the postoperative refraction (D) 25589 2663CYL2 Cylinder of the postoperative refraction (D) 25589 2663AX2 Postoperative axis (degrees) 25589 2663SE2 Postoperative spherical equivalent (D) 25589 2663UDVA2 Postoperative uncorrected visual acuity 28252 0CDVA2 Postoperative best corrected visual acuity 28252 0PE Prediction Error (D) 25589 2663lowastIn the AGE field only the number of patients is considered

Table 2 Statistical parameters of the considered variables

Variables Mean Median SD Var Skewness KurtosisAGE (years) 6870 7000 1089 11861 minus106 232AL (mm) 2352 2328 159 252 218 846ACD (mm) 302 302 044 019 033 163LT (mm) 455 455 052 027 minus024 341SPH1 (D) minus052 minus025 417 1737 minus090 427CYL1 (D) 122 100 090 081 148 404AX1 (deg) 8278 8500 4396 193254 minus002 002SE1 (D) minus113 minus088 422 1784 minus095 424SK (D) 4461 4464 186 347 minus053 583FK (D) 4347 4354 180 323 minus084 532Kav (D) 4404 4407 177 312 minus073 542UDVA1 016 010 014 002 183 398CDVA1 040 040 027 007 054 -041IOL (D) 2065 2100 435 1889 minus218 886SPH2 (D) 041 050 106 113 072 2254CYL2 (D) 152 125 108 117 140 298AX2 (deg) 9362 9500 3850 148239 minus027 066SE2 (D) minus036 minus025 109 103 039 2890UDVA2 046 040 028 008 039 minus081CDVA2 077 090 028 008 minus111 016PE (D) 028 034 099 098 minus006 1944

4 Journal of Ophthalmology

while the corrected visual acuity is below 06 in a majority ofthe cases After surgery on the contrary 47 of the patientshave uncorrected distance visual acuity above 05 and 64

have corrected distance visual acuity above 08 As thecataract has a major influence on visual acuity the post-operative results are more important In Figure 9 thepostoperative uncorrected and corrected distance visualacuity are compared Figure 9(a) compares the cumulativepercent of both quantities while Figure 9(b) shows thedifference between them in terms of Snellen lines

Figure 10 shows the distribution of the absolute value ofPE in different intervals In the figure both the number andpercent of eyes in each interval are presented Around 75of eyes reach a postoperative refraction that differs less thanone diopter of the predicted one Figure S9 shows thedistribution of PE with a mean value of 028 D (σ 099 D)e presence of refractive surprises is signaled by the highvalue of kurtosis is magnitude and its dependence withthe formula used to calculate the prediction refraction(including the optimization of the constants used) as well asthe lens model are so important that they deserve a separatestudy It will be conducted in a forthcoming paperFigure S10 shows the distribution of anisometropia beforesurgery Figure S10(a) represents the distribution of theabsolute value of the difference SPH1OS-SPH1OD and the fitof an exponential model (Equation (1)) to the experimentaldata Figure S10(b) is similar but the magnitude is related tothe absolute value of the difference SE1OS-SE1OD Overall ofthe 1863 pair of eyes having preoperative data 1038 (557)present anisometropia of the sphere while 959 (515)present anisometropia of the spherical equivalent Bothdistributions have an exponential decay (R2gt 099) withsimilar decaying factors (116 and 111 D respectively)

e effects of the surgery on the anisometropia areevident by analyzing Figure S11 Figure S11(a) shows thedistribution of the SPH2rsquos anisometropia andFigure S11(b) shows the distribution of the anisometropiain SE2 Of the 2632 measured pairs of eyes 636 (242)show anisometropia in the sphere and 773 (294) in thespherical equivalent Not only the proportion of affectedeyes diminishes but also the prevalence of large aniso-metropias decreases It is evident in the values of thedecaying factor (071 and 0620 D respectively)

An additional study of the anisometropia in the biometricparameters of the mate eyes was performed e difference |ALOSndashALOD| has an average value of 024mm (σ 045mmIQR 017mm) When the same calculation is performed withthe ACD |ACDOSndashACDOD| has an average value of 018mmσ 032mm and IQR 017mm

Table 3 reports some statigraphs of different magni-tudes segmented by sex Female eyes are shorter and haveshallower anterior chambers than male eyes e averageLT of both sexes are statistically different with femalelenses thicker than the male ones ese factors combinedimply that men have deeper vitreous chambers than fe-males e power of the femalersquos corneas is consistentlybigger than that of males It is so not only for the averagepower but also for the extreme keratometries

e average value of SE1 is more myopic in malepatients than in female ones After surgery though thedifferences between the average SE of males and femalesdiminish the values are significantly different CYL1 and

5000

4000

3000

Patie

nts

2000

1000

015 20 25

AL (mm)30 35

Av = 2352mmσ = 159mmγ = 218κ = 846

Figure 1 Histogram of the axial length with 05mm bins In theinset the main statigraphs of the distribution are shown

2500

3000

2000

1500

Patie

nts

1000

500

01 2 3

ACD (mm)4 5 6

Av = 302mmσ = 044mmγ = 033κ = 163

Figure 2 Histogram of the anterior chamber depth with bins of02mm In the inset the main statigraphs of the distribution areshown

3500

4000

3000

2500

2000

Patie

nts

1500

1000

500

030 35 40

Kav (D)45 50 55

Av = 4404Dσ = 177Dγ = ndash073κ = 542

Figure 3 Histogram of the average corneal power with 05 D binsIn the inset the main statigraphs of the distribution are shown

Journal of Ophthalmology 5

+ 15ndash 153000

2500

2000

1500Pa

tient

s

1000

500

0ndash25 ndash20 ndash15 ndash10

SE1 (D)ndash5 0 5 10 15

Av = ndash113Dσ = 422Dγ = ndash095κ = 424

(a)

6000

5000

4000

3000

2000

1000

0ndash10 ndash5 0

SE2 (D)5 10 15

Av = ndash036Dσ = 109Dγ = 039κ = 2890

(b)

Figure 4 Histograms of SE with 1 D bins (a) Preoperative it is possible to note a bimodal distribution (b) Postoperative Comparing bothfigures note the effect of the surgical procedure In the insets the main statigraphs of the distributions are shown

17500

636

696

243172

121 132

15000

12500

10000

Patie

nts

7500

5000

2500

0a t r w t r

Axis orientationo b l

PrePost

Figure 6 Proportion of type of astigmatism preoperative and postoperative wtr with the rule atr against the rule obl oblique Note thatthe total number of patients preoperative and postoperative are different (see Table 1)

14000

460369

PrePost

121

515

419

116

12000

10000

Patie

nts

8000

6000

4000

2000

0Myopic Emmetropic

RefractionHyperopic

Figure 5 Classification of the refraction preoperative and postoperative Note that the total number of patients in the preoperative andpostoperative stages is different (see Table 1)

6 Journal of Ophthalmology

CYL2 are different between sexes though in the post-stagethe differences between the distributions are lessimportant

Consistent with the biometric findings female eyes needin average more powerful implanted lenses than men estandard deviation skewness and kurtosis of the

9000

7500

6000222

323 324

87

27 17

4500Patie

nts

3000

1500

0lt05 05ndash099 1ndash199

Keratometric astigmatism (D)2ndash299 3ndash399 ge4

Figure 7 Histogram of the keratometric astigmatism

100

80

60

Cum

ulat

ive

of e

yes

40

UC postopCD postop

20

066 68 69 612 615Cumulative Snellen visual acuity (6x or better)

624 630 660

(a)

40

444

138 128

002

29030

20

10

03 or gtworst

2 worst 1 worst

Difference between UDVA and CDVA(Snellen lines)

Same 1 or gtbetter

o

f eye

s

22852 eyes1-month postop

UDVA same or better than CDVA 2902UDVA within 1 line of CDVA 4182

(b)

Figure 9 Distribution of the postoperative visual acuity (both uncorrected distance UD and corrected distance CD) (a)e percent of eyeswith a given visual acuity (b) Differences between UD and CD in terms of Snellen lines

4000

3000

Patie

nts

2000

1000

00 5 10 15

IOL power (D)20 25 30 35

Av = 2065Dσ = 435Dγ = ndash218κ = 886

Figure 8 Histogram of the implanted lens power with 1 D bin In the inset the values of the main statigraphs are shown

Journal of Ophthalmology 7

distributions segmented by sex are most of times very closeimplying that both distributions are only displaced havingsimilar shapes

Figure 11 presents the percent distribution of refractionbefore and after surgery segmented by sex e distributionbefore surgery shows that for males the majority of eyes(60) are myopic while for females the prevalence of

myopia is larger than the prevalence of hyperopia but thedifference is not as large Both sexes present a small numberof emmetropic eyes slightly above 10 After surgery theproportion of emmetropic eyes increases up to 50 di-minishing the percent differences among sexes in the threesegments Regarding anisometropia both sexes present thesame prevalence (p 045)

IntervalAccumulated

25000

30000

20000

15000

Patie

nts

10000

5000

0le025 025ndash050 050ndash100

|PE| (D)gt100

224 205

429322

751

249

100

Figure 10 Distribution of the absolute value of the PE in intervals and its cumulative values

Table 3 Gender segmentation of the data

Variables Sex Mean Median SD Var Skew Kurt Number t-test p value

Age F 6877 7000 1070 11449 minus094 181 17811 078 043M 6867 7000 1093 11951 minus124 309 10441

AL F 2331 2302 163 267 235 866 17802 3091 lt0001M 2388 2372 143 205 230 1062 10437

ACD F 295 295 042 018 033 140 17628 3166 lt0001M 313 313 044 020 031 222 10294

LT F 457 458 050 025 minus023 207 2681 258 001M 453 450 055 031 minus022 470 1714

SK F 4487 4488 183 336 minus060 602 17807 3091 lt0001M 4417 4418 183 334 minus048 673 10440

FK F 4374 4377 177 315 minus096 652 17807 3431 lt0001M 4300 4305 174 303 minus077 466 10440

Kav F 4431 4433 174 303 minus081 583 17807 3383 lt0001M 4359 4364 172 294 minus073 608 10440

SE1 F minus091 minus038 436 1903 minus114 426 12989 1005 lt0001M minus152 minus150 403 1622 minus061 439 7498

SE2 F minus037 minus038 103 106 036 2966 14269 345 lt0001M minus032 minus025 102 103 021 2828 8296

CYL1 F minus118 minus100 089 079 minus161 478 12989 707 lt0001M minus127 minus100 092 085 minus137 343 7498

CYL2 F minus149 minus125 107 116 minus145 321 14269 336 lt0001M minus154 minus125 109 119 minus140 315 8296

IOL F 2104 2200 457 2088 minus223 818 17810 2062 lt0001M 1999 2000 385 1480 minus232 1199 10441

PE F 027 032 103 106 minus032 2716 14269 175 008M 029 034 100 099 014 1993 8296A Studentrsquos t test for unpaired samples was performed in order to compare the means of each variable among sexes pndashvaluelt 005 is considered as highlysignificant Nonsignificant values are shown in boldface

8 Journal of Ophthalmology

Table 4 reports the segmentation by age of the biometricmagnitudes as well as the refraction (pre- and post-) and PEfor the selected age groups A one-way ANOVA test wasperformed to detect differences between mean values of thedifferent groups e bottom row of the table gives theF-values obtained as well as the real power of the test evalues per age group of AL LT ACD and SE1 present highlysignificant differences between all groups Table 5 shows thebehavior of the average keratometry e age group of 41ndash60presents differences with all other groups that do not presentdifferences among them PE on the contrary has a differentbehavior explained in Table 6 SE2 does not present dif-ferences among age groups

Note that in all these magnitudes the standard deviationand other statigraphs (not shownhere) change between groups

32 Discussion e results of the present hospital-basedpopulation study constitute to our knowledge the largestand more representative investigation on ocular biometricparameters of Cuban cataract patients using optical biom-etry In addition extensive data on refractive errors andvisual outcomes distribution are also provided by this work

Missing information of some variables in numerouscases due to aforementioned dissimilar reasons makes itmandatory to exclude certain amount of analyzed eyes forspecific parameters Since available data of eachmagnitude isenormous these exclusions did not hinder from obtainingrepresentative results of every one of them

Mean age in our study (6870 years) is consistent with thefrequent cataract association to senescence and its left-skewed distribution showing that the larger percent in theage group between 60 and 80 years behaves similar as vastlyreported on literature in all latitudes [19 20] Most previousCuban studies [13 15 21] agree with our finding e av-erage age in Ref [22] is smaller than the value found in thisstudy It also happens in Ref [16] dedicated to study younghyperopes patients Ref [14] does not report age data

e standard value of AL of the human eye is inter-nationally taken to be around 24mm in adulthood re-gardless of sex or race whereas AL tends to be longer inmyopic and shorter in hypermetropic eyes comparing to thatof emmetropic [23] e right-skewed distribution is inaccordance with the predominance of myopia in the studiedsample AL average values are variables in the literature emean AL observed in this paper is larger than those of Refs[19 22 24] in which ultrasound low-coherence reflec-tometry and partial coherence interferometry are usedrespectively Conversely Refs [2 25] using low-coherencereflectometry and partial coherence interferometry re-spectively show larger average values Previous researches inCuba report larger [26] much larger [15] shorter [27] andmuch shorter [16] average AL when compared with ourresults It should be noted that the analyzed periods in allthese works include only one year or less and that thesample size is much smaller than ours explaining the dis-persion in their results Specifically in the case of Ref [15]more than 93 of patients hadmyopic astigmatismwhile theresult of Ref [16] is in accordance with his studyacutes designwhere all the patients were hyperopes

Average ACD in our work (302mm) is shallower thanreported in Latino populations [11] and to a lesser extentthose of Refs [28 29] in Asian populations On the otherhand Hashemi et al [24] report an ACD average evenshallower than ours using the biometer LENSTARBioG-raph (WaveLight AG Erlangen Germany) Differencesbetween these two biometers may affect the comparativeanalysis since ACD value in Lenstar refers to the distancebetween corneal endothelium and lens anterior capsulewhile the distance for this variable in IOL Master is takenfrom corneal epithelium to lens anterior capsuleat is whyLenstar uses the sum of central pachymetry and ACD valuesfor IOL calculation [27] erefore this issue has to beconsidered when comparing Lenstar and IOL Master asCuban researchers did in a previous study [27] showingvalues closer to ours with no significant differences between

7000484

604

377

Fem-preMale-pre

Fem-postMale-post

353

130108

509

524

408

266113

122

6000

5000

4000

Patie

nts

3000

2000

1000

0Myopic Emmetropic

RefractionHyperopic

Figure 11 Classification of the refraction preoperative and postoperative segmented by sex

Journal of Ophthalmology 9

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

while the corrected visual acuity is below 06 in a majority ofthe cases After surgery on the contrary 47 of the patientshave uncorrected distance visual acuity above 05 and 64

have corrected distance visual acuity above 08 As thecataract has a major influence on visual acuity the post-operative results are more important In Figure 9 thepostoperative uncorrected and corrected distance visualacuity are compared Figure 9(a) compares the cumulativepercent of both quantities while Figure 9(b) shows thedifference between them in terms of Snellen lines

Figure 10 shows the distribution of the absolute value ofPE in different intervals In the figure both the number andpercent of eyes in each interval are presented Around 75of eyes reach a postoperative refraction that differs less thanone diopter of the predicted one Figure S9 shows thedistribution of PE with a mean value of 028 D (σ 099 D)e presence of refractive surprises is signaled by the highvalue of kurtosis is magnitude and its dependence withthe formula used to calculate the prediction refraction(including the optimization of the constants used) as well asthe lens model are so important that they deserve a separatestudy It will be conducted in a forthcoming paperFigure S10 shows the distribution of anisometropia beforesurgery Figure S10(a) represents the distribution of theabsolute value of the difference SPH1OS-SPH1OD and the fitof an exponential model (Equation (1)) to the experimentaldata Figure S10(b) is similar but the magnitude is related tothe absolute value of the difference SE1OS-SE1OD Overall ofthe 1863 pair of eyes having preoperative data 1038 (557)present anisometropia of the sphere while 959 (515)present anisometropia of the spherical equivalent Bothdistributions have an exponential decay (R2gt 099) withsimilar decaying factors (116 and 111 D respectively)

e effects of the surgery on the anisometropia areevident by analyzing Figure S11 Figure S11(a) shows thedistribution of the SPH2rsquos anisometropia andFigure S11(b) shows the distribution of the anisometropiain SE2 Of the 2632 measured pairs of eyes 636 (242)show anisometropia in the sphere and 773 (294) in thespherical equivalent Not only the proportion of affectedeyes diminishes but also the prevalence of large aniso-metropias decreases It is evident in the values of thedecaying factor (071 and 0620 D respectively)

An additional study of the anisometropia in the biometricparameters of the mate eyes was performed e difference |ALOSndashALOD| has an average value of 024mm (σ 045mmIQR 017mm) When the same calculation is performed withthe ACD |ACDOSndashACDOD| has an average value of 018mmσ 032mm and IQR 017mm

Table 3 reports some statigraphs of different magni-tudes segmented by sex Female eyes are shorter and haveshallower anterior chambers than male eyes e averageLT of both sexes are statistically different with femalelenses thicker than the male ones ese factors combinedimply that men have deeper vitreous chambers than fe-males e power of the femalersquos corneas is consistentlybigger than that of males It is so not only for the averagepower but also for the extreme keratometries

e average value of SE1 is more myopic in malepatients than in female ones After surgery though thedifferences between the average SE of males and femalesdiminish the values are significantly different CYL1 and

5000

4000

3000

Patie

nts

2000

1000

015 20 25

AL (mm)30 35

Av = 2352mmσ = 159mmγ = 218κ = 846

Figure 1 Histogram of the axial length with 05mm bins In theinset the main statigraphs of the distribution are shown

2500

3000

2000

1500

Patie

nts

1000

500

01 2 3

ACD (mm)4 5 6

Av = 302mmσ = 044mmγ = 033κ = 163

Figure 2 Histogram of the anterior chamber depth with bins of02mm In the inset the main statigraphs of the distribution areshown

3500

4000

3000

2500

2000

Patie

nts

1500

1000

500

030 35 40

Kav (D)45 50 55

Av = 4404Dσ = 177Dγ = ndash073κ = 542

Figure 3 Histogram of the average corneal power with 05 D binsIn the inset the main statigraphs of the distribution are shown

Journal of Ophthalmology 5

+ 15ndash 153000

2500

2000

1500Pa

tient

s

1000

500

0ndash25 ndash20 ndash15 ndash10

SE1 (D)ndash5 0 5 10 15

Av = ndash113Dσ = 422Dγ = ndash095κ = 424

(a)

6000

5000

4000

3000

2000

1000

0ndash10 ndash5 0

SE2 (D)5 10 15

Av = ndash036Dσ = 109Dγ = 039κ = 2890

(b)

Figure 4 Histograms of SE with 1 D bins (a) Preoperative it is possible to note a bimodal distribution (b) Postoperative Comparing bothfigures note the effect of the surgical procedure In the insets the main statigraphs of the distributions are shown

17500

636

696

243172

121 132

15000

12500

10000

Patie

nts

7500

5000

2500

0a t r w t r

Axis orientationo b l

PrePost

Figure 6 Proportion of type of astigmatism preoperative and postoperative wtr with the rule atr against the rule obl oblique Note thatthe total number of patients preoperative and postoperative are different (see Table 1)

14000

460369

PrePost

121

515

419

116

12000

10000

Patie

nts

8000

6000

4000

2000

0Myopic Emmetropic

RefractionHyperopic

Figure 5 Classification of the refraction preoperative and postoperative Note that the total number of patients in the preoperative andpostoperative stages is different (see Table 1)

6 Journal of Ophthalmology

CYL2 are different between sexes though in the post-stagethe differences between the distributions are lessimportant

Consistent with the biometric findings female eyes needin average more powerful implanted lenses than men estandard deviation skewness and kurtosis of the

9000

7500

6000222

323 324

87

27 17

4500Patie

nts

3000

1500

0lt05 05ndash099 1ndash199

Keratometric astigmatism (D)2ndash299 3ndash399 ge4

Figure 7 Histogram of the keratometric astigmatism

100

80

60

Cum

ulat

ive

of e

yes

40

UC postopCD postop

20

066 68 69 612 615Cumulative Snellen visual acuity (6x or better)

624 630 660

(a)

40

444

138 128

002

29030

20

10

03 or gtworst

2 worst 1 worst

Difference between UDVA and CDVA(Snellen lines)

Same 1 or gtbetter

o

f eye

s

22852 eyes1-month postop

UDVA same or better than CDVA 2902UDVA within 1 line of CDVA 4182

(b)

Figure 9 Distribution of the postoperative visual acuity (both uncorrected distance UD and corrected distance CD) (a)e percent of eyeswith a given visual acuity (b) Differences between UD and CD in terms of Snellen lines

4000

3000

Patie

nts

2000

1000

00 5 10 15

IOL power (D)20 25 30 35

Av = 2065Dσ = 435Dγ = ndash218κ = 886

Figure 8 Histogram of the implanted lens power with 1 D bin In the inset the values of the main statigraphs are shown

Journal of Ophthalmology 7

distributions segmented by sex are most of times very closeimplying that both distributions are only displaced havingsimilar shapes

Figure 11 presents the percent distribution of refractionbefore and after surgery segmented by sex e distributionbefore surgery shows that for males the majority of eyes(60) are myopic while for females the prevalence of

myopia is larger than the prevalence of hyperopia but thedifference is not as large Both sexes present a small numberof emmetropic eyes slightly above 10 After surgery theproportion of emmetropic eyes increases up to 50 di-minishing the percent differences among sexes in the threesegments Regarding anisometropia both sexes present thesame prevalence (p 045)

IntervalAccumulated

25000

30000

20000

15000

Patie

nts

10000

5000

0le025 025ndash050 050ndash100

|PE| (D)gt100

224 205

429322

751

249

100

Figure 10 Distribution of the absolute value of the PE in intervals and its cumulative values

Table 3 Gender segmentation of the data

Variables Sex Mean Median SD Var Skew Kurt Number t-test p value

Age F 6877 7000 1070 11449 minus094 181 17811 078 043M 6867 7000 1093 11951 minus124 309 10441

AL F 2331 2302 163 267 235 866 17802 3091 lt0001M 2388 2372 143 205 230 1062 10437

ACD F 295 295 042 018 033 140 17628 3166 lt0001M 313 313 044 020 031 222 10294

LT F 457 458 050 025 minus023 207 2681 258 001M 453 450 055 031 minus022 470 1714

SK F 4487 4488 183 336 minus060 602 17807 3091 lt0001M 4417 4418 183 334 minus048 673 10440

FK F 4374 4377 177 315 minus096 652 17807 3431 lt0001M 4300 4305 174 303 minus077 466 10440

Kav F 4431 4433 174 303 minus081 583 17807 3383 lt0001M 4359 4364 172 294 minus073 608 10440

SE1 F minus091 minus038 436 1903 minus114 426 12989 1005 lt0001M minus152 minus150 403 1622 minus061 439 7498

SE2 F minus037 minus038 103 106 036 2966 14269 345 lt0001M minus032 minus025 102 103 021 2828 8296

CYL1 F minus118 minus100 089 079 minus161 478 12989 707 lt0001M minus127 minus100 092 085 minus137 343 7498

CYL2 F minus149 minus125 107 116 minus145 321 14269 336 lt0001M minus154 minus125 109 119 minus140 315 8296

IOL F 2104 2200 457 2088 minus223 818 17810 2062 lt0001M 1999 2000 385 1480 minus232 1199 10441

PE F 027 032 103 106 minus032 2716 14269 175 008M 029 034 100 099 014 1993 8296A Studentrsquos t test for unpaired samples was performed in order to compare the means of each variable among sexes pndashvaluelt 005 is considered as highlysignificant Nonsignificant values are shown in boldface

8 Journal of Ophthalmology

Table 4 reports the segmentation by age of the biometricmagnitudes as well as the refraction (pre- and post-) and PEfor the selected age groups A one-way ANOVA test wasperformed to detect differences between mean values of thedifferent groups e bottom row of the table gives theF-values obtained as well as the real power of the test evalues per age group of AL LT ACD and SE1 present highlysignificant differences between all groups Table 5 shows thebehavior of the average keratometry e age group of 41ndash60presents differences with all other groups that do not presentdifferences among them PE on the contrary has a differentbehavior explained in Table 6 SE2 does not present dif-ferences among age groups

Note that in all these magnitudes the standard deviationand other statigraphs (not shownhere) change between groups

32 Discussion e results of the present hospital-basedpopulation study constitute to our knowledge the largestand more representative investigation on ocular biometricparameters of Cuban cataract patients using optical biom-etry In addition extensive data on refractive errors andvisual outcomes distribution are also provided by this work

Missing information of some variables in numerouscases due to aforementioned dissimilar reasons makes itmandatory to exclude certain amount of analyzed eyes forspecific parameters Since available data of eachmagnitude isenormous these exclusions did not hinder from obtainingrepresentative results of every one of them

Mean age in our study (6870 years) is consistent with thefrequent cataract association to senescence and its left-skewed distribution showing that the larger percent in theage group between 60 and 80 years behaves similar as vastlyreported on literature in all latitudes [19 20] Most previousCuban studies [13 15 21] agree with our finding e av-erage age in Ref [22] is smaller than the value found in thisstudy It also happens in Ref [16] dedicated to study younghyperopes patients Ref [14] does not report age data

e standard value of AL of the human eye is inter-nationally taken to be around 24mm in adulthood re-gardless of sex or race whereas AL tends to be longer inmyopic and shorter in hypermetropic eyes comparing to thatof emmetropic [23] e right-skewed distribution is inaccordance with the predominance of myopia in the studiedsample AL average values are variables in the literature emean AL observed in this paper is larger than those of Refs[19 22 24] in which ultrasound low-coherence reflec-tometry and partial coherence interferometry are usedrespectively Conversely Refs [2 25] using low-coherencereflectometry and partial coherence interferometry re-spectively show larger average values Previous researches inCuba report larger [26] much larger [15] shorter [27] andmuch shorter [16] average AL when compared with ourresults It should be noted that the analyzed periods in allthese works include only one year or less and that thesample size is much smaller than ours explaining the dis-persion in their results Specifically in the case of Ref [15]more than 93 of patients hadmyopic astigmatismwhile theresult of Ref [16] is in accordance with his studyacutes designwhere all the patients were hyperopes

Average ACD in our work (302mm) is shallower thanreported in Latino populations [11] and to a lesser extentthose of Refs [28 29] in Asian populations On the otherhand Hashemi et al [24] report an ACD average evenshallower than ours using the biometer LENSTARBioG-raph (WaveLight AG Erlangen Germany) Differencesbetween these two biometers may affect the comparativeanalysis since ACD value in Lenstar refers to the distancebetween corneal endothelium and lens anterior capsulewhile the distance for this variable in IOL Master is takenfrom corneal epithelium to lens anterior capsuleat is whyLenstar uses the sum of central pachymetry and ACD valuesfor IOL calculation [27] erefore this issue has to beconsidered when comparing Lenstar and IOL Master asCuban researchers did in a previous study [27] showingvalues closer to ours with no significant differences between

7000484

604

377

Fem-preMale-pre

Fem-postMale-post

353

130108

509

524

408

266113

122

6000

5000

4000

Patie

nts

3000

2000

1000

0Myopic Emmetropic

RefractionHyperopic

Figure 11 Classification of the refraction preoperative and postoperative segmented by sex

Journal of Ophthalmology 9

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

+ 15ndash 153000

2500

2000

1500Pa

tient

s

1000

500

0ndash25 ndash20 ndash15 ndash10

SE1 (D)ndash5 0 5 10 15

Av = ndash113Dσ = 422Dγ = ndash095κ = 424

(a)

6000

5000

4000

3000

2000

1000

0ndash10 ndash5 0

SE2 (D)5 10 15

Av = ndash036Dσ = 109Dγ = 039κ = 2890

(b)

Figure 4 Histograms of SE with 1 D bins (a) Preoperative it is possible to note a bimodal distribution (b) Postoperative Comparing bothfigures note the effect of the surgical procedure In the insets the main statigraphs of the distributions are shown

17500

636

696

243172

121 132

15000

12500

10000

Patie

nts

7500

5000

2500

0a t r w t r

Axis orientationo b l

PrePost

Figure 6 Proportion of type of astigmatism preoperative and postoperative wtr with the rule atr against the rule obl oblique Note thatthe total number of patients preoperative and postoperative are different (see Table 1)

14000

460369

PrePost

121

515

419

116

12000

10000

Patie

nts

8000

6000

4000

2000

0Myopic Emmetropic

RefractionHyperopic

Figure 5 Classification of the refraction preoperative and postoperative Note that the total number of patients in the preoperative andpostoperative stages is different (see Table 1)

6 Journal of Ophthalmology

CYL2 are different between sexes though in the post-stagethe differences between the distributions are lessimportant

Consistent with the biometric findings female eyes needin average more powerful implanted lenses than men estandard deviation skewness and kurtosis of the

9000

7500

6000222

323 324

87

27 17

4500Patie

nts

3000

1500

0lt05 05ndash099 1ndash199

Keratometric astigmatism (D)2ndash299 3ndash399 ge4

Figure 7 Histogram of the keratometric astigmatism

100

80

60

Cum

ulat

ive

of e

yes

40

UC postopCD postop

20

066 68 69 612 615Cumulative Snellen visual acuity (6x or better)

624 630 660

(a)

40

444

138 128

002

29030

20

10

03 or gtworst

2 worst 1 worst

Difference between UDVA and CDVA(Snellen lines)

Same 1 or gtbetter

o

f eye

s

22852 eyes1-month postop

UDVA same or better than CDVA 2902UDVA within 1 line of CDVA 4182

(b)

Figure 9 Distribution of the postoperative visual acuity (both uncorrected distance UD and corrected distance CD) (a)e percent of eyeswith a given visual acuity (b) Differences between UD and CD in terms of Snellen lines

4000

3000

Patie

nts

2000

1000

00 5 10 15

IOL power (D)20 25 30 35

Av = 2065Dσ = 435Dγ = ndash218κ = 886

Figure 8 Histogram of the implanted lens power with 1 D bin In the inset the values of the main statigraphs are shown

Journal of Ophthalmology 7

distributions segmented by sex are most of times very closeimplying that both distributions are only displaced havingsimilar shapes

Figure 11 presents the percent distribution of refractionbefore and after surgery segmented by sex e distributionbefore surgery shows that for males the majority of eyes(60) are myopic while for females the prevalence of

myopia is larger than the prevalence of hyperopia but thedifference is not as large Both sexes present a small numberof emmetropic eyes slightly above 10 After surgery theproportion of emmetropic eyes increases up to 50 di-minishing the percent differences among sexes in the threesegments Regarding anisometropia both sexes present thesame prevalence (p 045)

IntervalAccumulated

25000

30000

20000

15000

Patie

nts

10000

5000

0le025 025ndash050 050ndash100

|PE| (D)gt100

224 205

429322

751

249

100

Figure 10 Distribution of the absolute value of the PE in intervals and its cumulative values

Table 3 Gender segmentation of the data

Variables Sex Mean Median SD Var Skew Kurt Number t-test p value

Age F 6877 7000 1070 11449 minus094 181 17811 078 043M 6867 7000 1093 11951 minus124 309 10441

AL F 2331 2302 163 267 235 866 17802 3091 lt0001M 2388 2372 143 205 230 1062 10437

ACD F 295 295 042 018 033 140 17628 3166 lt0001M 313 313 044 020 031 222 10294

LT F 457 458 050 025 minus023 207 2681 258 001M 453 450 055 031 minus022 470 1714

SK F 4487 4488 183 336 minus060 602 17807 3091 lt0001M 4417 4418 183 334 minus048 673 10440

FK F 4374 4377 177 315 minus096 652 17807 3431 lt0001M 4300 4305 174 303 minus077 466 10440

Kav F 4431 4433 174 303 minus081 583 17807 3383 lt0001M 4359 4364 172 294 minus073 608 10440

SE1 F minus091 minus038 436 1903 minus114 426 12989 1005 lt0001M minus152 minus150 403 1622 minus061 439 7498

SE2 F minus037 minus038 103 106 036 2966 14269 345 lt0001M minus032 minus025 102 103 021 2828 8296

CYL1 F minus118 minus100 089 079 minus161 478 12989 707 lt0001M minus127 minus100 092 085 minus137 343 7498

CYL2 F minus149 minus125 107 116 minus145 321 14269 336 lt0001M minus154 minus125 109 119 minus140 315 8296

IOL F 2104 2200 457 2088 minus223 818 17810 2062 lt0001M 1999 2000 385 1480 minus232 1199 10441

PE F 027 032 103 106 minus032 2716 14269 175 008M 029 034 100 099 014 1993 8296A Studentrsquos t test for unpaired samples was performed in order to compare the means of each variable among sexes pndashvaluelt 005 is considered as highlysignificant Nonsignificant values are shown in boldface

8 Journal of Ophthalmology

Table 4 reports the segmentation by age of the biometricmagnitudes as well as the refraction (pre- and post-) and PEfor the selected age groups A one-way ANOVA test wasperformed to detect differences between mean values of thedifferent groups e bottom row of the table gives theF-values obtained as well as the real power of the test evalues per age group of AL LT ACD and SE1 present highlysignificant differences between all groups Table 5 shows thebehavior of the average keratometry e age group of 41ndash60presents differences with all other groups that do not presentdifferences among them PE on the contrary has a differentbehavior explained in Table 6 SE2 does not present dif-ferences among age groups

Note that in all these magnitudes the standard deviationand other statigraphs (not shownhere) change between groups

32 Discussion e results of the present hospital-basedpopulation study constitute to our knowledge the largestand more representative investigation on ocular biometricparameters of Cuban cataract patients using optical biom-etry In addition extensive data on refractive errors andvisual outcomes distribution are also provided by this work

Missing information of some variables in numerouscases due to aforementioned dissimilar reasons makes itmandatory to exclude certain amount of analyzed eyes forspecific parameters Since available data of eachmagnitude isenormous these exclusions did not hinder from obtainingrepresentative results of every one of them

Mean age in our study (6870 years) is consistent with thefrequent cataract association to senescence and its left-skewed distribution showing that the larger percent in theage group between 60 and 80 years behaves similar as vastlyreported on literature in all latitudes [19 20] Most previousCuban studies [13 15 21] agree with our finding e av-erage age in Ref [22] is smaller than the value found in thisstudy It also happens in Ref [16] dedicated to study younghyperopes patients Ref [14] does not report age data

e standard value of AL of the human eye is inter-nationally taken to be around 24mm in adulthood re-gardless of sex or race whereas AL tends to be longer inmyopic and shorter in hypermetropic eyes comparing to thatof emmetropic [23] e right-skewed distribution is inaccordance with the predominance of myopia in the studiedsample AL average values are variables in the literature emean AL observed in this paper is larger than those of Refs[19 22 24] in which ultrasound low-coherence reflec-tometry and partial coherence interferometry are usedrespectively Conversely Refs [2 25] using low-coherencereflectometry and partial coherence interferometry re-spectively show larger average values Previous researches inCuba report larger [26] much larger [15] shorter [27] andmuch shorter [16] average AL when compared with ourresults It should be noted that the analyzed periods in allthese works include only one year or less and that thesample size is much smaller than ours explaining the dis-persion in their results Specifically in the case of Ref [15]more than 93 of patients hadmyopic astigmatismwhile theresult of Ref [16] is in accordance with his studyacutes designwhere all the patients were hyperopes

Average ACD in our work (302mm) is shallower thanreported in Latino populations [11] and to a lesser extentthose of Refs [28 29] in Asian populations On the otherhand Hashemi et al [24] report an ACD average evenshallower than ours using the biometer LENSTARBioG-raph (WaveLight AG Erlangen Germany) Differencesbetween these two biometers may affect the comparativeanalysis since ACD value in Lenstar refers to the distancebetween corneal endothelium and lens anterior capsulewhile the distance for this variable in IOL Master is takenfrom corneal epithelium to lens anterior capsuleat is whyLenstar uses the sum of central pachymetry and ACD valuesfor IOL calculation [27] erefore this issue has to beconsidered when comparing Lenstar and IOL Master asCuban researchers did in a previous study [27] showingvalues closer to ours with no significant differences between

7000484

604

377

Fem-preMale-pre

Fem-postMale-post

353

130108

509

524

408

266113

122

6000

5000

4000

Patie

nts

3000

2000

1000

0Myopic Emmetropic

RefractionHyperopic

Figure 11 Classification of the refraction preoperative and postoperative segmented by sex

Journal of Ophthalmology 9

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

CYL2 are different between sexes though in the post-stagethe differences between the distributions are lessimportant

Consistent with the biometric findings female eyes needin average more powerful implanted lenses than men estandard deviation skewness and kurtosis of the

9000

7500

6000222

323 324

87

27 17

4500Patie

nts

3000

1500

0lt05 05ndash099 1ndash199

Keratometric astigmatism (D)2ndash299 3ndash399 ge4

Figure 7 Histogram of the keratometric astigmatism

100

80

60

Cum

ulat

ive

of e

yes

40

UC postopCD postop

20

066 68 69 612 615Cumulative Snellen visual acuity (6x or better)

624 630 660

(a)

40

444

138 128

002

29030

20

10

03 or gtworst

2 worst 1 worst

Difference between UDVA and CDVA(Snellen lines)

Same 1 or gtbetter

o

f eye

s

22852 eyes1-month postop

UDVA same or better than CDVA 2902UDVA within 1 line of CDVA 4182

(b)

Figure 9 Distribution of the postoperative visual acuity (both uncorrected distance UD and corrected distance CD) (a)e percent of eyeswith a given visual acuity (b) Differences between UD and CD in terms of Snellen lines

4000

3000

Patie

nts

2000

1000

00 5 10 15

IOL power (D)20 25 30 35

Av = 2065Dσ = 435Dγ = ndash218κ = 886

Figure 8 Histogram of the implanted lens power with 1 D bin In the inset the values of the main statigraphs are shown

Journal of Ophthalmology 7

distributions segmented by sex are most of times very closeimplying that both distributions are only displaced havingsimilar shapes

Figure 11 presents the percent distribution of refractionbefore and after surgery segmented by sex e distributionbefore surgery shows that for males the majority of eyes(60) are myopic while for females the prevalence of

myopia is larger than the prevalence of hyperopia but thedifference is not as large Both sexes present a small numberof emmetropic eyes slightly above 10 After surgery theproportion of emmetropic eyes increases up to 50 di-minishing the percent differences among sexes in the threesegments Regarding anisometropia both sexes present thesame prevalence (p 045)

IntervalAccumulated

25000

30000

20000

15000

Patie

nts

10000

5000

0le025 025ndash050 050ndash100

|PE| (D)gt100

224 205

429322

751

249

100

Figure 10 Distribution of the absolute value of the PE in intervals and its cumulative values

Table 3 Gender segmentation of the data

Variables Sex Mean Median SD Var Skew Kurt Number t-test p value

Age F 6877 7000 1070 11449 minus094 181 17811 078 043M 6867 7000 1093 11951 minus124 309 10441

AL F 2331 2302 163 267 235 866 17802 3091 lt0001M 2388 2372 143 205 230 1062 10437

ACD F 295 295 042 018 033 140 17628 3166 lt0001M 313 313 044 020 031 222 10294

LT F 457 458 050 025 minus023 207 2681 258 001M 453 450 055 031 minus022 470 1714

SK F 4487 4488 183 336 minus060 602 17807 3091 lt0001M 4417 4418 183 334 minus048 673 10440

FK F 4374 4377 177 315 minus096 652 17807 3431 lt0001M 4300 4305 174 303 minus077 466 10440

Kav F 4431 4433 174 303 minus081 583 17807 3383 lt0001M 4359 4364 172 294 minus073 608 10440

SE1 F minus091 minus038 436 1903 minus114 426 12989 1005 lt0001M minus152 minus150 403 1622 minus061 439 7498

SE2 F minus037 minus038 103 106 036 2966 14269 345 lt0001M minus032 minus025 102 103 021 2828 8296

CYL1 F minus118 minus100 089 079 minus161 478 12989 707 lt0001M minus127 minus100 092 085 minus137 343 7498

CYL2 F minus149 minus125 107 116 minus145 321 14269 336 lt0001M minus154 minus125 109 119 minus140 315 8296

IOL F 2104 2200 457 2088 minus223 818 17810 2062 lt0001M 1999 2000 385 1480 minus232 1199 10441

PE F 027 032 103 106 minus032 2716 14269 175 008M 029 034 100 099 014 1993 8296A Studentrsquos t test for unpaired samples was performed in order to compare the means of each variable among sexes pndashvaluelt 005 is considered as highlysignificant Nonsignificant values are shown in boldface

8 Journal of Ophthalmology

Table 4 reports the segmentation by age of the biometricmagnitudes as well as the refraction (pre- and post-) and PEfor the selected age groups A one-way ANOVA test wasperformed to detect differences between mean values of thedifferent groups e bottom row of the table gives theF-values obtained as well as the real power of the test evalues per age group of AL LT ACD and SE1 present highlysignificant differences between all groups Table 5 shows thebehavior of the average keratometry e age group of 41ndash60presents differences with all other groups that do not presentdifferences among them PE on the contrary has a differentbehavior explained in Table 6 SE2 does not present dif-ferences among age groups

Note that in all these magnitudes the standard deviationand other statigraphs (not shownhere) change between groups

32 Discussion e results of the present hospital-basedpopulation study constitute to our knowledge the largestand more representative investigation on ocular biometricparameters of Cuban cataract patients using optical biom-etry In addition extensive data on refractive errors andvisual outcomes distribution are also provided by this work

Missing information of some variables in numerouscases due to aforementioned dissimilar reasons makes itmandatory to exclude certain amount of analyzed eyes forspecific parameters Since available data of eachmagnitude isenormous these exclusions did not hinder from obtainingrepresentative results of every one of them

Mean age in our study (6870 years) is consistent with thefrequent cataract association to senescence and its left-skewed distribution showing that the larger percent in theage group between 60 and 80 years behaves similar as vastlyreported on literature in all latitudes [19 20] Most previousCuban studies [13 15 21] agree with our finding e av-erage age in Ref [22] is smaller than the value found in thisstudy It also happens in Ref [16] dedicated to study younghyperopes patients Ref [14] does not report age data

e standard value of AL of the human eye is inter-nationally taken to be around 24mm in adulthood re-gardless of sex or race whereas AL tends to be longer inmyopic and shorter in hypermetropic eyes comparing to thatof emmetropic [23] e right-skewed distribution is inaccordance with the predominance of myopia in the studiedsample AL average values are variables in the literature emean AL observed in this paper is larger than those of Refs[19 22 24] in which ultrasound low-coherence reflec-tometry and partial coherence interferometry are usedrespectively Conversely Refs [2 25] using low-coherencereflectometry and partial coherence interferometry re-spectively show larger average values Previous researches inCuba report larger [26] much larger [15] shorter [27] andmuch shorter [16] average AL when compared with ourresults It should be noted that the analyzed periods in allthese works include only one year or less and that thesample size is much smaller than ours explaining the dis-persion in their results Specifically in the case of Ref [15]more than 93 of patients hadmyopic astigmatismwhile theresult of Ref [16] is in accordance with his studyacutes designwhere all the patients were hyperopes

Average ACD in our work (302mm) is shallower thanreported in Latino populations [11] and to a lesser extentthose of Refs [28 29] in Asian populations On the otherhand Hashemi et al [24] report an ACD average evenshallower than ours using the biometer LENSTARBioG-raph (WaveLight AG Erlangen Germany) Differencesbetween these two biometers may affect the comparativeanalysis since ACD value in Lenstar refers to the distancebetween corneal endothelium and lens anterior capsulewhile the distance for this variable in IOL Master is takenfrom corneal epithelium to lens anterior capsuleat is whyLenstar uses the sum of central pachymetry and ACD valuesfor IOL calculation [27] erefore this issue has to beconsidered when comparing Lenstar and IOL Master asCuban researchers did in a previous study [27] showingvalues closer to ours with no significant differences between

7000484

604

377

Fem-preMale-pre

Fem-postMale-post

353

130108

509

524

408

266113

122

6000

5000

4000

Patie

nts

3000

2000

1000

0Myopic Emmetropic

RefractionHyperopic

Figure 11 Classification of the refraction preoperative and postoperative segmented by sex

Journal of Ophthalmology 9

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

distributions segmented by sex are most of times very closeimplying that both distributions are only displaced havingsimilar shapes

Figure 11 presents the percent distribution of refractionbefore and after surgery segmented by sex e distributionbefore surgery shows that for males the majority of eyes(60) are myopic while for females the prevalence of

myopia is larger than the prevalence of hyperopia but thedifference is not as large Both sexes present a small numberof emmetropic eyes slightly above 10 After surgery theproportion of emmetropic eyes increases up to 50 di-minishing the percent differences among sexes in the threesegments Regarding anisometropia both sexes present thesame prevalence (p 045)

IntervalAccumulated

25000

30000

20000

15000

Patie

nts

10000

5000

0le025 025ndash050 050ndash100

|PE| (D)gt100

224 205

429322

751

249

100

Figure 10 Distribution of the absolute value of the PE in intervals and its cumulative values

Table 3 Gender segmentation of the data

Variables Sex Mean Median SD Var Skew Kurt Number t-test p value

Age F 6877 7000 1070 11449 minus094 181 17811 078 043M 6867 7000 1093 11951 minus124 309 10441

AL F 2331 2302 163 267 235 866 17802 3091 lt0001M 2388 2372 143 205 230 1062 10437

ACD F 295 295 042 018 033 140 17628 3166 lt0001M 313 313 044 020 031 222 10294

LT F 457 458 050 025 minus023 207 2681 258 001M 453 450 055 031 minus022 470 1714

SK F 4487 4488 183 336 minus060 602 17807 3091 lt0001M 4417 4418 183 334 minus048 673 10440

FK F 4374 4377 177 315 minus096 652 17807 3431 lt0001M 4300 4305 174 303 minus077 466 10440

Kav F 4431 4433 174 303 minus081 583 17807 3383 lt0001M 4359 4364 172 294 minus073 608 10440

SE1 F minus091 minus038 436 1903 minus114 426 12989 1005 lt0001M minus152 minus150 403 1622 minus061 439 7498

SE2 F minus037 minus038 103 106 036 2966 14269 345 lt0001M minus032 minus025 102 103 021 2828 8296

CYL1 F minus118 minus100 089 079 minus161 478 12989 707 lt0001M minus127 minus100 092 085 minus137 343 7498

CYL2 F minus149 minus125 107 116 minus145 321 14269 336 lt0001M minus154 minus125 109 119 minus140 315 8296

IOL F 2104 2200 457 2088 minus223 818 17810 2062 lt0001M 1999 2000 385 1480 minus232 1199 10441

PE F 027 032 103 106 minus032 2716 14269 175 008M 029 034 100 099 014 1993 8296A Studentrsquos t test for unpaired samples was performed in order to compare the means of each variable among sexes pndashvaluelt 005 is considered as highlysignificant Nonsignificant values are shown in boldface

8 Journal of Ophthalmology

Table 4 reports the segmentation by age of the biometricmagnitudes as well as the refraction (pre- and post-) and PEfor the selected age groups A one-way ANOVA test wasperformed to detect differences between mean values of thedifferent groups e bottom row of the table gives theF-values obtained as well as the real power of the test evalues per age group of AL LT ACD and SE1 present highlysignificant differences between all groups Table 5 shows thebehavior of the average keratometry e age group of 41ndash60presents differences with all other groups that do not presentdifferences among them PE on the contrary has a differentbehavior explained in Table 6 SE2 does not present dif-ferences among age groups

Note that in all these magnitudes the standard deviationand other statigraphs (not shownhere) change between groups

32 Discussion e results of the present hospital-basedpopulation study constitute to our knowledge the largestand more representative investigation on ocular biometricparameters of Cuban cataract patients using optical biom-etry In addition extensive data on refractive errors andvisual outcomes distribution are also provided by this work

Missing information of some variables in numerouscases due to aforementioned dissimilar reasons makes itmandatory to exclude certain amount of analyzed eyes forspecific parameters Since available data of eachmagnitude isenormous these exclusions did not hinder from obtainingrepresentative results of every one of them

Mean age in our study (6870 years) is consistent with thefrequent cataract association to senescence and its left-skewed distribution showing that the larger percent in theage group between 60 and 80 years behaves similar as vastlyreported on literature in all latitudes [19 20] Most previousCuban studies [13 15 21] agree with our finding e av-erage age in Ref [22] is smaller than the value found in thisstudy It also happens in Ref [16] dedicated to study younghyperopes patients Ref [14] does not report age data

e standard value of AL of the human eye is inter-nationally taken to be around 24mm in adulthood re-gardless of sex or race whereas AL tends to be longer inmyopic and shorter in hypermetropic eyes comparing to thatof emmetropic [23] e right-skewed distribution is inaccordance with the predominance of myopia in the studiedsample AL average values are variables in the literature emean AL observed in this paper is larger than those of Refs[19 22 24] in which ultrasound low-coherence reflec-tometry and partial coherence interferometry are usedrespectively Conversely Refs [2 25] using low-coherencereflectometry and partial coherence interferometry re-spectively show larger average values Previous researches inCuba report larger [26] much larger [15] shorter [27] andmuch shorter [16] average AL when compared with ourresults It should be noted that the analyzed periods in allthese works include only one year or less and that thesample size is much smaller than ours explaining the dis-persion in their results Specifically in the case of Ref [15]more than 93 of patients hadmyopic astigmatismwhile theresult of Ref [16] is in accordance with his studyacutes designwhere all the patients were hyperopes

Average ACD in our work (302mm) is shallower thanreported in Latino populations [11] and to a lesser extentthose of Refs [28 29] in Asian populations On the otherhand Hashemi et al [24] report an ACD average evenshallower than ours using the biometer LENSTARBioG-raph (WaveLight AG Erlangen Germany) Differencesbetween these two biometers may affect the comparativeanalysis since ACD value in Lenstar refers to the distancebetween corneal endothelium and lens anterior capsulewhile the distance for this variable in IOL Master is takenfrom corneal epithelium to lens anterior capsuleat is whyLenstar uses the sum of central pachymetry and ACD valuesfor IOL calculation [27] erefore this issue has to beconsidered when comparing Lenstar and IOL Master asCuban researchers did in a previous study [27] showingvalues closer to ours with no significant differences between

7000484

604

377

Fem-preMale-pre

Fem-postMale-post

353

130108

509

524

408

266113

122

6000

5000

4000

Patie

nts

3000

2000

1000

0Myopic Emmetropic

RefractionHyperopic

Figure 11 Classification of the refraction preoperative and postoperative segmented by sex

Journal of Ophthalmology 9

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

Table 4 reports the segmentation by age of the biometricmagnitudes as well as the refraction (pre- and post-) and PEfor the selected age groups A one-way ANOVA test wasperformed to detect differences between mean values of thedifferent groups e bottom row of the table gives theF-values obtained as well as the real power of the test evalues per age group of AL LT ACD and SE1 present highlysignificant differences between all groups Table 5 shows thebehavior of the average keratometry e age group of 41ndash60presents differences with all other groups that do not presentdifferences among them PE on the contrary has a differentbehavior explained in Table 6 SE2 does not present dif-ferences among age groups

Note that in all these magnitudes the standard deviationand other statigraphs (not shownhere) change between groups

32 Discussion e results of the present hospital-basedpopulation study constitute to our knowledge the largestand more representative investigation on ocular biometricparameters of Cuban cataract patients using optical biom-etry In addition extensive data on refractive errors andvisual outcomes distribution are also provided by this work

Missing information of some variables in numerouscases due to aforementioned dissimilar reasons makes itmandatory to exclude certain amount of analyzed eyes forspecific parameters Since available data of eachmagnitude isenormous these exclusions did not hinder from obtainingrepresentative results of every one of them

Mean age in our study (6870 years) is consistent with thefrequent cataract association to senescence and its left-skewed distribution showing that the larger percent in theage group between 60 and 80 years behaves similar as vastlyreported on literature in all latitudes [19 20] Most previousCuban studies [13 15 21] agree with our finding e av-erage age in Ref [22] is smaller than the value found in thisstudy It also happens in Ref [16] dedicated to study younghyperopes patients Ref [14] does not report age data

e standard value of AL of the human eye is inter-nationally taken to be around 24mm in adulthood re-gardless of sex or race whereas AL tends to be longer inmyopic and shorter in hypermetropic eyes comparing to thatof emmetropic [23] e right-skewed distribution is inaccordance with the predominance of myopia in the studiedsample AL average values are variables in the literature emean AL observed in this paper is larger than those of Refs[19 22 24] in which ultrasound low-coherence reflec-tometry and partial coherence interferometry are usedrespectively Conversely Refs [2 25] using low-coherencereflectometry and partial coherence interferometry re-spectively show larger average values Previous researches inCuba report larger [26] much larger [15] shorter [27] andmuch shorter [16] average AL when compared with ourresults It should be noted that the analyzed periods in allthese works include only one year or less and that thesample size is much smaller than ours explaining the dis-persion in their results Specifically in the case of Ref [15]more than 93 of patients hadmyopic astigmatismwhile theresult of Ref [16] is in accordance with his studyacutes designwhere all the patients were hyperopes

Average ACD in our work (302mm) is shallower thanreported in Latino populations [11] and to a lesser extentthose of Refs [28 29] in Asian populations On the otherhand Hashemi et al [24] report an ACD average evenshallower than ours using the biometer LENSTARBioG-raph (WaveLight AG Erlangen Germany) Differencesbetween these two biometers may affect the comparativeanalysis since ACD value in Lenstar refers to the distancebetween corneal endothelium and lens anterior capsulewhile the distance for this variable in IOL Master is takenfrom corneal epithelium to lens anterior capsuleat is whyLenstar uses the sum of central pachymetry and ACD valuesfor IOL calculation [27] erefore this issue has to beconsidered when comparing Lenstar and IOL Master asCuban researchers did in a previous study [27] showingvalues closer to ours with no significant differences between

7000484

604

377

Fem-preMale-pre

Fem-postMale-post

353

130108

509

524

408

266113

122

6000

5000

4000

Patie

nts

3000

2000

1000

0Myopic Emmetropic

RefractionHyperopic

Figure 11 Classification of the refraction preoperative and postoperative segmented by sex

Journal of Ophthalmology 9

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

these two biometers In contrast Song et al [4] observed thatACD measured by IOL Master 500 was shallower than thatmeasured by IOL Master 700 or Lenstar LS900 (Haag SteitAG Koeniz Switzerland)) In Songrsquos study the ACD meanvalue is similar to ours when measured by IOL Master 500while values from IOL Master 700 and Lenstar are deeperis result corresponds with that of the number of eyesmeasured using PCI (IOL Master 302 and 500) in oursample is much bigger than those measured by SS-OCT Aformer study conducted in the Cuban Institute of Oph-thalmology [26] using optic (IOL Master 500) and im-mersion ultrasonic biometry shows similar results to oursWe found no ACD mean values reported in other Cubanpopulation-based studies

Accommodation must be considered for an accuratemeasurement of lens thickness Nonstandardized accom-modative stimuli used for research and the wide resolutionvariability among the available techniques to measure thismagnitude explain the difference in lens thickness reportedin literature However many studies including this areconducted without recording accommodative responseassuming that the subject was accommodating accurately tothe target [30 31] In addition considering those patientsunder 40 years old in the present study the nonpresbyopeones represent only 145 of the sample and all patients hadbeen diagnosed with cataract making improbably that lensthickness would be significantly affected by accommodationSince PCI-based biometers are not able to offer this mag-nitude lens thickness was measured in 4395 eyes by usingexclusively the SS-OCT-based IOL Master 700 which offersthe best resolution approximately 10ndash20 μm

e average LT of 455mm is higher than that reportedin studies that used an A-scan ultrasound which has lowerresolution than the SS-OCT for studying Latino [11 19] andelderly Chinese populations [10] Hashemi et al [24] andFerreira et al [2] using both low-coherence optical reflec-tometry (Lenstar) observed much smaller LT values As faras we know a previous report of this magnitude has not beenpublished on Cuban patients

Keratometry average value and its distribution in ourstudy are close to those observed in European population-based studies using partial-coherence interferometry [20]and optical low-coherence reflectometry [2]

In the Asian population Song et al [4] show similarkeratometry average value when using Lenstar and IOLMaster 700 and a bit bigger when using IOL Master 500while Wickremasinghe et al [18] show much close values to

ours using an autorefractorndashkeratometer (Canon RK-5Autorefractor Keratometer Canon Inc Tokyo Japan)Cuban researchers [15 32] using autorefractorndashkera-tometers have previously reported slightly lower valuesKeratometry measurements may vary when different eval-uation methods are used as it was demonstrated in Ref [33]reporting a significantly lower corneal power when mea-surements were obtained with autorefractorndashkeratometersthan those measured by IOL Master

Relating to FK and SK the average values are 4347 Dand 4462 D respectivelye Auckland Cataract Study [34]in a defined New Zealand population found a near value foraverage flattest keratometry and a rather lower value for thesteepest one Likewise in a preceding Cuban study [27] FKand SK mean values were lower than ours using bothLenstar and IOL Master 500

e bimodal distribution of SPH1 with bothmyopic andhypermetropic peaks has a notorious influence in thesimilar bimodal behavior of SE1 distribution SE1 shows amyopic mean value ough correlation between myopiaand cataract is not a purpose of this paper this myopic SE1may be related in some extent to the presence of nuclearcataract taking into account the previously reported asso-ciation between these variables [35]

Similar myopic SE1 is reported in Ref [36] In Refs[15 22] the spherical component shows a similar influenceon the prevalence of ametropias in Cuban samples SE2mean value within emmetropia range (plusmn05 D) is the ex-pected outcome and it is similar to those reported in lit-erature [34]

Before analyzing astigmatism behavior it is important toremark that keratometry is not commonly measured fol-lowing cataract surgery at the postoperative stage in Cubaunless it is required for investigative purposes erefore

Table 4 Segmentation by age of the biometric parameters spherical equivalents (SE1 and SE2) and PE

Agegroup

AL (mm) mean(sd)

LT (mm) mean(sd)

ACD (mm) mean(sd)

Kav (D) mean(sd)

SEpre (D) mean(sd)

SEpost (D) mean(sd)

PE (D) mean(sd)

le40 2461 (299) 387 (058) 339 (055) 4417 (237) minus343 (854) minus044 (135) 013 (153)41ndash60 2395 (245) 431 (052) 318 (045) 4383 (218) minus205 (611) minus037 (126) 032 (136)61ndash79 2344 (132) 458 (050) 300 (042) 4408 (166) minus103 (356) minus033 (103) 028 (107)ge80 2328 (099) 473 (052) 285 (041) 4413 (156) 006 (278) minus035 (118) 020 (177)F 2042 (1) 959 (1) 4978 (1) Table 5 392 (1) No diff (1) Table 6e results of the analysis of variance are reported using the F-value with the actual power of the test in parentheses e analysis of Kav and PE isparticularized in Tables 5 and 6 respectively

Table 5 Kav41ndash60 61ndash79 ge80

le40 plt 0001 No (p 032) No (p 072)41ndash60 plt 0001 plt 000161ndash79 No (p 012)

Table 6 Prediction error41ndash60 61ndash79 ge80

le40 plt 005 plt 005 No (p 045)41ndash60 plt 0001 plt 000161ndash79 No (p 018)

10 Journal of Ophthalmology

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

only refractive astigmatism is available after surgery in thisretrospective study In our series both corneal and refractiveaverage astigmatism is larger than 1 D before surgeryHowever taking into account that most of total astigmatismcomes from the cornea and lenticular astigmatism iseliminated with extraction of the cataract only the cornealastigmatism is considered when planning cataract surgery

e distribution of CYL1 (see Figure S6(a)) for valuesabove 1 D is exponential In the inset of this figure the fit ofan exponential model to the experimental values is showne high value of R2 (09995) indicates that the model ex-plains almost all the variations observed Below this value amarked decrement in the number of patients is foundAttebo et al [37] found a similar behavior though they donot make a fitting of a model to the data ey also found adecrement in the number of cases for values of cylinderbelow 05 D As they note it may be associated with apreference of not to report these small cylinders as par-ticipants could usually read the same number of letterswithout it

A similar dependence is found for CYL2 (Figure S6(b))e exponential model explains more than 99 of thevariation of the cylinder e main difference between bothdistributions is the value of the decaying factor χ While forCYL1 it is 092 D for CYL2 it is 136 D indicating a slowerdecrementis is a signal that as a result of the incision therefractive astigmatism increases is is confirmed by theaverage values of both magnitudes reported in Table 2

Keratometric astigmatism behaves similar to that of thepreoperatory refractive one in both cases almost 90 of theeyes have a value below 2 D Figure S7 represents thedistribution of the keratometric astigmatism e expo-nential fitting shown in the inset has a similar decrement asthat presented in Figure S6(a) Note that the decaying factorof the keratometric astigmatism (0858 D) is close to thedecaying factor of the preoperative refractive astigmatism(092 D) indicating the close relation between bothmagnitudes

e increment of astigmatism after surgery may beexplained by the lack of toric lenses and the nonsystematicpractice of anti-astigmatic incisions by the surgeons due tononavailability of graduated scalpels in our environment Atthe Cuban Institute of Ophthalmology most of the surgeonshabitually perform the principal corneal incision duringphacoemulsification on temporal or supero-temporal sitessince incisions on these locations are less astigmatogenicSome others treat astigmatism up to 15 D by making theprincipal corneal incision on the steepest meridian duringcataract surgery In some cases this may be technicallydifficult and may require less comfortable positioning andconsequently it is frequently avoided by surgeons having anegative impact on postoperative astigmatism In a studyconducted in the United Kingdom where nontoric standardmonofocal IOLs were implanted in 99 percent of the eyes asimilar trend toward greater postoperative astigmatism wasobserved [38] Despite the nondisposable toric IOL orgraduated scalpels for limbal releasing incision (LRI) aformer Cuban study showed a mean surgical inducedastigmatism (SIA) of 061 D after phacoemulsification [39]

whereas SIA resulted in higher than 1 D when superiorsclera-corneal incisions were performed [40]

ATR-astigmatism represents by far the major per-centage at preoperative stage Similar behavior has beenobserved in Refs [2 5 41] and previous Cuban researches[15 22]

e preoperative refractive defect with a higher preva-lence is myopia Since the strong correlation between axiallength and ocular refraction have been formerly confirmed[42] the aforementioned presence of more extra-long eyesthan extra-short ones observed in this studyrsquos sample may bea contributing factor for this fact In addition the probablylarger lens thickness related to the condition of cataractdiagnosis for all patients selected in this study with respectto the noncataractous population may be related to anincreased lens refractive power Olsen et al [42] observedthat the stronger the lens the more myopic the eye Al-though the present paper did not explicitly analyze the lensrsquorefractive power this magnitude is significantly associatedwith higher lens thickness [43] e result of the presentstudy is coincident with the predominance of myopic eyesreported in the preoperative setting by Fernandez et al [15]and Santos et al [21] in former Cuban studies

At the postoperative stage the percent of emmetropiceyes increases notably leading to an improvement in bothUDVA and CDVA is is the expected outcome con-sidering that the purpose of modern cataract surgery is notlimited to remove the opacified crystalline lens but to offerthe best visual and refractive outcomes without the use ofspectacles Comparing with previous Cuban studies ourresults are better than that reported by Galindo Zamoraet al [14] Conversely Santos et al [21] show a majorpercentage of eyes with better UDAV and CDVA than thepresent paper but the studied sample was quite smallerthan ours ese good visual outcomes are much relatedwith refractive PE distribution showing a postoperativerefraction that differs less than one diopter of the predictedone e average PE is smaller than those of Refs [44ndash46]However those authors found better results than ourswhen analyzing the percentage of eyes with average re-fractive prediction error within plusmn050 D e high excess ofkurtosis of this parameter can explain this behavior and itis in accordance with the almost 25 of refractive surprisesobserved in this paper A previous Cuban study shows asmaller percentage of eyes with average PE in a rangeinferior to 050 D using third-generation formulas inpatients who underwent uneventful cataract surgery from2007 to 2011 [47]

Anisometropia in both SPH and SE is present in morethan 50 of patients whose preoperative data of both eyeswere registered is value is quite higher than those re-ported in previous Cuban studies [15 22] It must be notedthat in those studies anisometropia was considered whenthe difference between the spherical components of botheyes was higher than 25 D which explains the discordancewith respect to our results However Akhgary et al [48]analyzing patients in a wide range of age (1ndash80 years old)observed that anisometropia was higher than in afore-mentioned Cuban studies but still pretty smaller than ours

Journal of Ophthalmology 11

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

e high percentage of anisometropia observed in thispaper may be related to the predominance of patients be-yond 60 years old (more than 80) e marked incrementin the prevalence of anisometropic refractive errors later inlife has been demonstrated [49] An extensive review aboutanisometropia [50] concluded that although asymmetriccataract development or unilateral cataract extraction hasbeen suggested as a cause of anisometropia in older adultsthe more probable origin for this ametropia in senescencemay be related to a regression of neural control of binocularvision e exponential decay in its distribution shows thatthe prevalence of large differences between both eyes beforesurgery (decaying factor around 11 D) is alleviated bysurgery giving a smaller decaying factor (le071 D) Averagevalue of the difference |ALOSndashALOD| is significantly bigger inpatients with anisometropia than in those who do not sufferfrom this refractive disturbance while the average of dif-ference |ACDOSndashACDOD| is not is result suggests apositive correlation between the unequal axial length growthof both eyes and anisometropia Coincident results havebeen observed in a study that examined the prevalence ofanisometropia and interocular differences of ocular bio-metric parameters among elderly female twins showing thathigher anisometropia concerning both spherical refractionand spherical equivalent were associated with more myopicrefraction and longer AL [51] Unlike the present studysome authors [52 53] have found a positive correlationbetween the anterior chamber depth difference and aniso-metropia specifically in hyperopic anisometropic eyes epredominance of myopia in our sample may explain thenoncoincident findings concerning ACD since no divisionwas made between anisomyopes and anisohyperopes

When comparing different magnitudes segmented bysex females have shorter eyes shallower ACD and steepercorneas Coincident results have been reported in severalstudies in Asian [9 18 24 54] European [2] and Latino[11 19] populations However Lee et al [25] found that theassociation of sex with AL and ACD were no longer sig-nificant after adjustment for height suggesting that theusually smaller stature of women may be the real cause ofsmaller eyes Many other studies have reported the associ-ation between height and AL [55] One study [56] stated thatsex-differences with respect to AL and ACD were still sig-nificant in multivariate analyses controlling for staturesuggesting that sex may be an independent determinant ofAL Likewise in Ref [2] it was concluded that the greatestpredictor of ocular biometrics is gender However thesegender differences related to AL are not significant inemmetropic subjects as reported by another study [23] Anaverage flatter cornea observed in males is similar to reportsin literature but unlike AL and ACD it seems to be notinfluenced by height adjustments [25]

Lens thickness is significantly bigger in females Otherauthors have reported similar results [19 54] In contrast an-other study [57] found that males had slightly greater thick-nesses than females but the mean difference was not significantIn the Los Angeles Latino Eye Study [11] a significant gender-related difference in lens thickness was found although it wasnot statistically significant after adjusting for height

e abovementioned findings are the reason for thegreater average value of the IOL power implanted to females

Refractive error is probably a consequence of a mismatchof the biometric parameters of the eyes (ie AL K ACD LT)and the refractive power of these structures [53 58] AL (orvitreous chamber depth when AL is broken down into itscomponents) has been remarked as the most importantcontributor parameter in determining refractive error fol-lowed by the corneal power of the eye [11] ese authorsalso found that other variables had a smaller but significantcontributing effect on refractive error such as lens opales-cence and the two other components of axial length lensthickness and anterior chamber depth Correspondingly thedeeper anterior chamber and vitreous chamber and con-gruently the longer AL may explain the preponderance ofpreoperative myopia in males considering both the distri-bution and the average spherical equivalent Other studies inthe Chinese population report that women tend to be morehyperopic [58 59] However in the present paper myopiaalso exhibits higher prevalence than hyperopia in femalesbut the difference is not as large as that observed in malesOn the other hand other studies found nonsignificant sexdifference in refractive error [10]

Because all of the aforementioned biometric parametersfor each individual are considered for IOL calculation inorder to achieve a better visual and refractive outcome thepercent differences of myopia and hyperopia between sexesdiminish while emmetropia increases after cataract surgerye average SE of both males and females is closer to zeroand the difference between them diminishes after surgeryalthough the values remain significantly different e rel-evance of these gender differences in ocular biometry hasbeen considered in formulating some of the newer gener-ation IOL calculation formulas [60 61] Anisometropia hasnot been associated with gender as an independent factorInter-gender difference of anisometropia is not significant inthis study coinciding with previous reports in the Cubanpopulation [15 22]

Biometric parameters show significant differences whenage is considered e younger the patients the longer ALand deeper the ACD while LT tends to be thicker in olderpatients One study analyzing 750 cataractous eyes in an USpopulation [62] showed similar correlations Analogousresults are also reported in Iranian [24] Australian [57] andLos Angeles Latino populations [11] for these three mag-nitudes However another study [10] found that AL did notchange with age although ACD showed a decrease Kera-tometry particularly shows no difference among ages exceptfor the 41ndash60 group which significantly differs from theother groups is result is different from that reported inRef [57] where no significant change was observed foranterior corneal radii of curvature related to age He et al[10] observed that corneal power did not change with ageeither

Regarding refractive error and age SE1 shows a shift tohyperopia as patients get older is outcome is coincidentwith the refraction patterns described in Ref [63] in whichrefraction is stable from 20 s to 40 s and then moves in thehypermetropic direction Likewise a former study in Cuba

12 Journal of Ophthalmology

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

[34] found that ametropia with a hypermetropic componentwasmore frequent in the older age groups while those with amyopic component predominated in the younger agegroups A study in a Latino population [11] shows a coin-cident result In Ref [10] the spherical equivalent tended tobecome hyperopic at 60 years and shifted toward myopia at75 years Similarly the Blue Mountains Eye Study [63]observed that a myopic shift in refraction occurred mostfrequently in older (ge70 years) than younger (lt70 years)participants Another study [44] shows that in the50ndash59 year age group and the 60ndash69 year age group therewas a hyperopic shift of 041 D and 034 D respectively

4 Conclusions

In summary this paper describes the distribution of ocularbiometric characteristics visual acuity and refractive resultsin a large sample of the Cuban population

e preoperative biometric parameters show highlysignificant differences between sexes and age with male andyounger eyes being longer and with a deeper anteriorchamber but with a thinner lens Keratometry (KS KF andKav) shows highly significant sex differences being steeperin female eyes than in males however age-related differ-ences are not significant except for the 41ndash60 group Beforesurgery both sexes have myopic eyes on average with maleeyes being more myopic than female eyes Preoperativespherical equivalent tends to hyperopia with age Aftersurgery the average spherical equivalent is minus036 D andfemale eyes are more myopic than males (plt 0001) epercent of emmetropic eyes increases notably leading to asignificant improvement in both UDVA and CDVA evi-dencing the tremendous impact of cataract surgery in re-ducing preoperative ametropias

ese results are of interest not only in Cuba but also inregions with a large Cuban and Cuban descendant pop-ulation such as several places in the USA (particularly thesouth of Florida) Spain and several counties of LatinAmerica is would also be important in other countrieswith similar ethnic composition (a mixing of Europeansfrom the Iberian Peninsula and Black Africans) in the Ca-ribbean region and South America

Data Availability

e data are available upon request to the correspondingauthor

Disclosure

is work was performed as part of the employment of theauthors at the Instituto Cubano de Oftalmologıa (I H-L andTC-D) and the Universidad de La Habana (AJ B-L and SE-F)

Conflicts of Interest

e authors declare that there are no conflicts of interestregarding the publication of this paper

Supplementary Materials

Table S1 95 confidence intervals and interquartile range ofthe selected magnitudes Figure S1 age distribution of thesample (a) Histogram with a bin size of five years showingin the inset the main statigraphs of the distribution (b) Agegroups and the percent of eyes belonging to each one FigureS2 number of eyes in the different length ranges short(lt22mm) medium (22ndash239mm) large (24ndash26mm) andvery large (gt26mm) e percent of eyes in each group ispresented Figure S3 histogram of the lens thickness with02mm bin size In the inset the main statigraphs of thedistribution are shown Figure S4 histogram of the kera-tometries with 05 D bins (a) Steepest keratometry (b)Flattest keratometry In the insets the main statigraphs ofboth distributions are shown Figure S5 histograms of thesphere of the refraction with 05 D bins (a) Preoperativesphere (b) Postoperative sphere In the insets the mainstatigraphs of both distributions are shown Figure S6histograms of the refractive cylinder with 05 D bins (a)Preoperative (b) Postoperative In the insets the fitting of anexponential model (Equation (1)) to the data is shown evalue of the decaying factor χ is given with its standard errorin parentheses Figure S7 keratometric astigmatism thedistribution function and the fit of an exponential model tothe data e value of the decaying factor χ is given with itsstandard error in parentheses Figure S8 visual acuityuncorrected distance (UD) and corrected (CD) before andafter surgery Figure S9 distribution of PE In the inset somestatigraphs of the distribution are shown Figure S10 ani-sometropy in the preoperative stage (a) In the sphere (b) Inthe spherical equivalent In the insets of both graphs theresults of the fitting of an exponential model (Equation (1))to the data are shown e decaying factor is presented withits standard error in parentheses Figure S11 anisometropyin the postoperative stage (a) In the sphere (b) In thespherical equivalent In the insets of both graphs the resultsof the fitting of an exponential model (Equation (1)) to thedata are shown e decaying factor is presented with itsstandard error in parentheses (Supplementary Materials)

References

[1] P C Hoffmann and W W Hutz ldquoAnalysis of biometry andprevalence data for corneal astigmatism in 23 239 eyesrdquoJournal of Cataract and Refractive Surgery vol 36 no 9pp 1479ndash1485 2010

[2] T B Ferreira K J Hoffer F Ribeiro P Ribeiro andJ G OrsquoNeill ldquoOcular biometric measurements in cataractsurgery candidates in Portugalrdquo PLoS One vol 12 no 10Article ID e0184837 2017

[3] A Akman L Asena and S G Gungor ldquoEvaluation andcomparison of the new swept source OCT-based IOLMaster700 with the IOLMaster 500rdquo British Journal of Ophthal-mology vol 100 no 9 pp 1201ndash1205 2016

[4] J S Song D Y Yoon J Y Hyon and H S Jeon ldquoCom-parison of ocular biometry and refractive outcomes using IOLMaster 500 IOL Master 700 and Lenstar LS900rdquo KoreanJournal of Ophthalmology vol 34 no 2 pp 126ndash132 2020

Journal of Ophthalmology 13

[5] Q Huang Y Huang Q Luo and W Fan ldquoOcular biometriccharacteristics of cataract patients in western Chinardquo BMCOphthalmology vol 18 no 1 p 99 2018

[6] Z S Gaurisankar G A Rijn J E E Lima et al ldquoCorrelationsbetween ocular biometrics and refractive error a systematicreview and meta-analysisrdquo Acta Ophthalmologica vol 97no 8 pp 735ndash743 2019

[7] M He H Chen and W Wang ldquoRefractive errors ocularbiometry and diabetic retinopathy a comprehensive reviewrdquoCurrent Eye Research vol 46 no 2 pp 151ndash158 2020

[8] J C Sherwin and D A Mackey ldquoUpdate on the epidemiologyand genetics of myopic refractive errorrdquo Expert Review ofOphthalmology vol 8 no 1 pp 63ndash87 2013

[9] P J Foster D C Broadway S Hayat et al ldquoRefractive erroraxial length and anterior chamber depth of the eye in Britishadults the EPIC-Norfolk Eye Studyrdquo British Journal ofOphthalmology vol 94 no 7 pp 827ndash830 2010

[10] M He W Huang Y Li Y Zheng Q Yin and P J FosterldquoRefractive error and biometry in older Chinese adults theLiwan eye studyrdquo Investigative Opthalmology amp Visual Sci-ence vol 50 no 11 pp 5130ndash5136 2009

[11] C Shufelt S Fraser-Bell M Ying-Lai M Torres andR Varma ldquoRefractive error ocular biometry and lensopalescence in an adult population the Los Angeles LatinoEye Studyrdquo Investigative Opthalmology amp Visual Sciencevol 46 no 12 pp 4450ndash4460 2005

[12] T Bagus K Alberto M Muteba and A Makgotloe ldquoAnalysisof corneal biometry in a black South African populationrdquoAfrican Vision and Eye Health vol 78 no 1 p a495 2019

[13] A B Polanco Fontela Y Dıaz Hernandez L AveleiraTamayo K Correa Barzaga and M Perez GarcıaldquoCaracterizacion de pacientes operados de catarata conimplante de lente intraocular Bayamo 2013rdquo MULTIMEDRev Medica Granma vol 18 no 1 pp 50ndash66 2014 httpwwwrevmultimedsldcuindexphpmtmarticleview231075

[14] I B Galindo Zamora K R Fernandez Ferrer E HernandezPentildea Y Gonzalez Iglesias P Chang Chao and L R DıazAlfonso ldquoCalculo biometrico y resultados refractivos Estudiode 250 casos operados de cataratardquo MediSur vol 10 no 1pp 22ndash26 2012 httpscielosldcuscielophpscript=sci_arttextamppid=S1727-897X2012000100004amplng=esampnrm=isoamptlng=es

[15] J A Fernandez Soler R del C Garcıa Perez O M MarintildeoHidalgo and J A Cobas Gonzalez ldquoCaracterizacion de lasametropıas atendidas en Consulta de Cirugıa Refractiva delCentro Oftalmologico de Holguınrdquo Correo Cientıfico Medicovol 19 no 2 pp 233ndash245 2015 httpscielosldcuscielophpscript=sci_arttextamppid=S1560-43812015000200006amplng=esampnrm=isoamptlng=en

[16] E Rojas Alvarez I Miranda Hernandez Y Ruiz Rodrıguezand J Gonzalez Sotero ldquoExtraccion de cristalino transparenteen pacientes hipermetropesrdquo Revista Cubana de Oftalmologıavol 24 no 1 pp 40ndash45 2011 httpwwwrevoftalmologiasldcuindexphpoftalmologiaarticleview6

[17] J C Fernandez-Alvarez I Hernandez-Lopez P P Cruz-Cobas T Cardenas-Dıaz and A J Batista-Leyva ldquoUsing amultilayer perceptron in intraocular lens power calculationrdquoJournal of Cataract amp Refractive Surgery vol 45 no 12pp 1753ndash1761 2019

[18] S Wickremasinghe P J Foster D Uranchimeg et al ldquoOcularbiometry and refraction in Mongolian adultsrdquo InvestigativeOpthalmology amp Visual Science vol 45 no 3 pp 776ndash7832004

[19] M S Barlatey W Koga-Nakamura V Moreno-LondontildeoM Takane-Imay and C Gonzalez-Gonzalez ldquoDistribution ofaxial length and related factors in an adult population ofMexico Cityrdquo Revista Mexicana de Oftalmologıa vol 93no 5 pp 233ndash237 2019

[20] N E Knox Cartwright R L Johnston P D JaycockD M Tole and J M Sparrow ldquoe Cataract National Datasetelectronic multicentre audit of 55 567 operations whenshould IOLMaster biometric measurements be recheckedrdquoEye vol 24 no 5 pp 894ndash900 2010

[21] J O Santos J E O Santos S M M Noda L A L Herreraand J C M Domınguez ldquoResultados visuales y refractivos enpacientes operados de catarata en el Hospital ldquoAbelSantamarıa Cuadradordquordquo Universidad Medica Pinarentildeavol 14 no 1 pp 36ndash44 2018 httpwwwrevgalenosldcuindexphpumparticleview268

[22] L Curbelo Cunill J Hernandez Silva E Machado Fernandezet al ldquoFrecuencia de ametropıasrdquo Frecuencia de Ametropıasvol 18 no 1 2005 httpscielosldcupdfoftv18n1oft06105pdf

[23] A Roy ldquoVariation of axial ocular dimensions with age sexheight BMI -and their relation to refractive statusrdquo Journal ofClinical and Diagnostic Research vol 9 no 1 pp AC01ndashAC04 2015

[24] H Hashemi M Khabazkhoob M Miraftab et al ldquoedistribution of axial length anterior chamber depth lensthickness and vitreous chamber depth in an adult populationof Shahroud Iranrdquo BMC Ophthalmology vol 12 no 1 p 502012

[25] K E Lee B E K Klein R Klein Z Quandt and T Y WongldquoAssociation of age stature and education with ocular di-mensions in an older white populationrdquo Archives of Oph-thalmology vol 127 no 1 pp 88ndash93 2009

[26] E Montero Diaz M Serpa Valdes Y Cuan Aguilar PerezCandelaria E de la C I Hernandez Lopez and M Vidal delCastillo ldquoEfectividad de la biometrıa de inmersion para elcalculo del poder dioptrico de la lente intraocularrdquo RevistaCubana de Oftalmologıa vol 27 no 3 pp 350ndash358 2014httpwwwrevoftalmologiasldcuindexphpoftalmologiaarticleview310

[27] I Miranda Hernandez J R Hernandez Silva M Rio TorresY Ruiz Rodrıguez J Del Amo Freire and A Bisnubia VargasldquoEvaluacion del equipo de interferometria optica de coher-encia parcial Lenstar en la biometrıa ocularrdquo Revista Cubanade Oftalmologıa vol 23 pp 665ndash677 2010 httpscielosldcuscielophpscript=sci_arttextamppid=S0864-21762010000400001

[28] J G Yu J Zhong Z M Mei F Zhao N Tao and Y XiangldquoEvaluation of biometry and corneal astigmatism in cataractsurgery patients from Central Chinardquo BMC Ophthalmologyvol 17 no 1 p 56 2017

[29] C-W Pan T-Y Wong L Chang et al ldquoOcular biometry inan Urban Indian population the Singapore Indian Eye study(SINDI)rdquo Investigative Opthalmology amp Visual Sciencevol 52 no 9 pp 6636ndash6642 2011

[30] K Richdale M A Bullimore and K Zadnik ldquoLens thicknesswith age and accommodation by optical coherence tomog-raphyrdquo Ophthalmic and Physiological Optics vol 28 no 5pp 441ndash447 2008

[31] J M Morales Avalos and A J Torres Moreno ldquoEstudiobiometrico ocular en una poblacion adulta del estado deSinaloardquo Revista Mexicana de Oftalmologıa vol 90 no 1pp 14ndash17 2016

14 Journal of Ophthalmology

[32] Y E Garcıa V T Torriente Z M Legon and I T CasadoldquoCalculo del poder de la lente intraocular mediante biometrıaultrasonicardquo Revista Cubana de Oftalmologıa vol 26 no 3pp 399ndash409 2013 httpwwwrevoftalmologiasldcuindexphpoftalmologiaarticleview207

[33] M Garza-Leon P de la Parra-Colın and T Barrientos-Gutierrez ldquoEstudio comparativo de la medicion del podercorneal central usando el queratometro manual el IOLMastery el tomografo Siriusrdquo Revista Mexicana de Oftalmologıavol 90 no 3 pp 111ndash117 2016

[34] A F Riley T Y Malik C N GrupchevaM J Fisk J P Craigand C N McGhee ldquoe Auckland Cataract Study co-morbidity surgical techniques and clinical outcomes in apublic hospital servicerdquo British Journal of Ophthalmologyvol 86 no 2 pp 185ndash190 2002

[35] N A Brown and A R Hill ldquoCataract the relation betweenmyopia and cataract morphologyrdquo British Journal of Oph-thalmology vol 71 no 6 pp 405ndash414 1987

[36] A Sawada A Tomidokoro M Araie A Iwase andT Yamamoto ldquoRefractive errors in an elderly Japanesepopulationrdquo Ophthalmology vol 115 no 2 pp 363ndash3702008

[37] K Attebo R Q Ivers and P Mitchell ldquoRefractive errors in anolder populationrdquo Ophthalmology vol 106 no 6pp 1066ndash1072 1999

[38] A C Day M Dhariwal M S Keith et al ldquoDistribution ofpreoperative and postoperative astigmatism in a large pop-ulation of patients undergoing cataract surgery in the UKrdquoBritish Journal of Ophthalmology vol 103 no 7 pp 993ndash1000 2019

[39] A Gonzalez Pentildea L Ortega Dıaz and E Perez CandelarialdquoAstigmatismo inducido en la cirugıa de catarata por tecnicade facoemulsificacionrdquo Revista Cubana de Oftalmologıavol 24 no 1 pp 30ndash39 2011 httpscielosldcupdfoftv24n1oft03111pdf

[40] R Montejo Valdes EM Ballate Nodales M MarquezFernandez E Fernandez Perez and Y Ortega BallateldquoFactores pronosticos del astigmatismo corneal inducido enpacientes operados de catarata por la tecnica tunelizadardquoRevista Cubana de Oftalmologıa vol 21 no 2 2008 httpscielosldcuscielophpscript=sci_arttextamppid=S0864-21762008000200005amplng=es

[41] S Feizi M Naderan V Ownagh and F SadeghpourldquoDistribution of the anterior posterior and total cornealastigmatism in healthy eyesrdquo International Ophthalmologyvol 38 no 2 pp 481ndash491 2018

[42] T Olsen A Arnarsson H Sasaki K Sasaki and F JonassonldquoOn the ocular refractive components the Reykjavik EyeStudyrdquo Acta Ophthalmologica Scandinavica vol 85 no 4pp 361ndash366 2007

[43] R Iribarren I G Morgan V Nangia and J B JonasldquoCrystalline lens power and refractive errorrdquo InvestigativeOpthalmology amp Visual Science vol 53 no 2 pp 543ndash5502012

[44] K Darcy D Gunn S Tavassoli J Sparrow and J X KaneldquoAssessment of the accuracy of new and updated intraocularlens power calculation formulas in 10 930 eyes from the UKNational Health Servicerdquo Journal of Cataract and RefractiveSurgery vol 46 no 1 pp 2ndash7 2020

[45] R B Melles J X Kane T Olsen andW J Chang ldquoUpdate onintraocular lens calculation formulasrdquo Ophthalmologyvol 126 no 9 pp 1334-1335 2019

[46] T T V Roberts C Hodge G Sutton and M LawlessldquoComparison of Hill-radial basis function Barrett Universal

and current third generation formulas for the calculation ofintraocular lens power during cataract surgeryrdquo Clinical ampExperimental Ophthalmology vol 133 no 12 pp 1431ndash14362017

[47] M Carracedo H Silva D Imalvet S Garcıa et al ldquoPer-sonalizacion de las constantes en las formulas de calculo de lalente intraocularrdquo Revista Cubana de Oftalmologıa vol 25no 2 pp 1ndash15 2012

[48] M Akhgary M Ghassemi-Broumand M Aghazadeh Amiriand M T Seyed ldquoPrevalence of strabismic binocularanomalies amblyopia and anisometropia Rehabilitationfaculty of Shahid Beheshti Medical Universityrdquo Journal ofOptometry vol 4 no 3 pp 110ndash114 2011

[49] G Haegerstrom-Portnoy M E Schneck J A Brabyn andL A Lott ldquoDevelopment of refractive errors into old agerdquoOptometry and Vision Science vol 79 no 10 pp 643ndash6492002

[50] S J Vincent M J Collins S A Read and L G CarneyldquoMyopic anisometropia ocular characteristics and aetio-logical considerationsrdquo Clinical and Experimental Optometryvol 97 no 4 pp 291ndash307 2014

[51] O Parssinen M Kauppinen J Kaprio and T RantanenldquoAnisometropia of ocular refractive and biometric measuresamong 66- to 79-year-old female twinsrdquo Acta Oph-thalmologica vol 94 no 8 pp 768ndash774 2016

[52] N Singh J Rohatgi and V Kumar ldquoA prospective study ofanterior segment ocular parameters in anisometropiardquo Ko-rean Journal of Ophthalmology vol 31 no 2 pp 165ndash1712017

[53] S Warrier H M Wu H S Newland et al ldquoOcular biometryand determinants of refractive error in rural Myanmar theMeiktila eye studyrdquo British Journal of Ophthalmology vol 92no 12 pp 1591ndash1594 2008

[54] T Y Wong P J Foster T P Ng J M Tielsch G J Johnsonand S K L Seah ldquoVariations in ocular biometry in an adultChinese population in Singapore the Tanjong Pagar surveyrdquoInvestigative Ophthalmology amp Visual Science vol 42 no 1pp 73ndash80 2001

[55] H M Wu A Gupta H S Newland D Selva T Aung andR J Casson ldquoAssociation between stature ocular biometryand refraction in an adult population in rural Myanmar theMeiktila eye studyrdquo Clinical amp Experimental Ophthalmologyvol 35 no 9 pp 834ndash839 2007

[56] L S Lim S-M Saw V S E Jeganathan et al ldquoDistributionand determinants of ocular biometric parameters in an Asianpopulation the Singapore Malay eye studyrdquo InvestigativeOpthalmology amp Visual Science vol 51 no 1 pp 103ndash1092010

[57] D A Atchison E L Markwell S Kasthurirangan J M PopeG Smith and P G Swann ldquoAge-related changes in opticaland biometric characteristics of emmetropic eyesrdquo Journal ofVision vol 8 no 4 pp 1ndash20 2008

[58] C-Y Cheng W-M Hsu J-H Liu S-Y Tsai and P ChouldquoRefractive errors in an elderly Chinese population in taiwanthe shihpai eye studyrdquo Investigative Opthalmology amp VisualScience vol 44 no 11 pp 4630ndash4638 2003

[59] L Xu J Li T Cui et al ldquoRefractive error in urban and ruraladult Chinese in Beijingrdquo Ophthalmology vol 112 no 10pp 1676ndash1683 2005

[60] K J Hoffer and G Savini ldquoClinical results of the hoffer H-5formula in 2707 eyes first 5th-generation formula based ongender and racerdquo International Ophthalmology Clinicsvol 57 no 4 pp 213ndash219 2017

Journal of Ophthalmology 15

[61] B J Connell and J X Kane ldquoComparison of the Kane formulawith existing formulas for intraocular lens power selectionrdquoBMJ Open Ophthalmology vol 4 pp 1ndash6 2019

[62] R Jivrajka M C Shammas T Boenzi M Swearingen andJ H Shammas ldquoVariability of axial length anterior chamberdepth and lens thickness in the cataractous eyerdquo Journal ofCataract and Refractive Surgery vol 34 no 2 pp 289ndash2942008

[63] J Panchapakesan E Rochtchina and P Mitchell ldquoMyopicrefractive shift caused by incident cataract the Blue Moun-tains Eye Studyrdquo Ophthalmic Epidemiology vol 10 no 4pp 241ndash247 2003

16 Journal of Ophthalmology