foot length

8/2/2019 Foot Length

1/5

umerous fetal structures measured sono-graphically correlate well with gestationalage. Fetal biometry is useful for screening

and diagnosing growth d isturbances as well as struc-tural and genetic abnormalities. One measurementthat has been suggested to be useful is fetal footlength, but growth of the fetal foot has not beenexamined thoroughly. Nearly eight decades ago,Streeter1 reported a strong relationship between fetalfoot length and menstrual age in a series of fetal

pathologic specimens. That study and others2 havevalidated the practice of using foot length for accu-rate gestational age assessment in the postmortemfetus. More recently, Mercer and coworkers3 demon-strated that sonographic fetal foot lengths werehighly correlated with gestational age beyond thefirst trimester. Existing literature has concentrated onfetal foot size in the normally grown fetus,36 whilelittle is known regarding the effect of fetal growthdisturbances on foot length. Given the fact that

Received August 17, 1999, from the Division of Maternal-FetalMedicine (N.B.M., J.C.S., D.A.M., E.R.G., A.M.V.), Division ofEpidemiology and Biostatistics (C.V.A.), and Center for Perinatal

Health Initiatives (C.V.A., J.C.S.), Department of Obstetrics,Gynecology, and Reproductive Sciences, UMDNJRobert WoodJohnson Med ical School/ St. Peters University Hospital, NewBrunsw ick, New Jersey. Revised man uscript accepted for publicationNovember 2, 1999.

Address correspondence to Natalie B. Meirowitz, MD, Division ofMaternal-Fetal Medicine, Departmen t of Obstetrics and Gynecology,St. Peters University Hospital, 254 Easton Avenue, MOB 4th Floor,New Brunswick, NJ 08903-0591. Reprints not available.

The Center for Perinatal Health Initiatives is supported , in part, bya grant (#029553) from the Robert Wood Johnson Foundation, NJ.Opinions, views, and conclusions expressed in this paper are those ofthe authors and not those of Robert Wood Johnson Foundation.

ABBREVIATIONS

SGA, Small-for-gestational-age; EFW, Estimated fetal weight;LGA, Large-for-gestational-age; H/ A, Head to abd ominal cir-cumference

2000 by the Am erican Institute of Ultrasou nd in Med icine J Ultrasoun d Med 19:201205, 2000 0278-4297/ 00/ $3.50

N

Foot Length in Fetuses withAbnormal Grow th

Natalie B. Meirowitz, MD, Cand e V. Ananth , PhD, MPH, John C. Smu lian, MD, MPH,

David A. McLean, MD, Edwin R. Guzman, MD, Anthony M. Vintzileos, MD

Sonographic fetal foot length is highly predictive ofgestational age. In order to assess the reliability of thisparameter in predicting gestational age in cases ofabnormal fetal growth , we examined fetal foot lengthin small- and large-for-gestational-age fetuses. Anomogram of foot length versus gestational agebetween 15 and 37 weeks was constructed usingcross-sectional data obtained from 5372 singletonfetuses. Fetal foot length s for small-for-gestational-agefetuses (estimated fetal weight below the 10th per-centile) and large-for-gestational-age fetuses (abovethe >90th p ercentile) fetuses were plotted against thefoot length nomogram in order to d etermine the num -ber of small-for-gestational-age fetuses and large-for-

gestational-age fetuses with foot lengths below the10th and above the 90th percentiles, respectively. Ofthe 586 small-for-gestational-age fetuses, 355 (60.6%)had foot lengths below the 10th percentile on thenomogram. When foot lengths from large-for-gestational-age fetuses were plotted on the foot lengthnomogram, 29.4% (219 of 744) had measurementsabove the 90th percentile. Fetal foot length can beinfluenced by growth restriction as well as states ofaccelerated fetal growth. Our findings imply thatthere are limitations to the use of fetal foot length forgestational age assessment, particularly in fetuseswith growth abnormalities. KEY WORDS: Fetus, footlength; Foot, length , fetal; Growth abn ormalities, fetal.


2/5

pathologists currently use the fetal foot for accurategestational age assessment, it is reasonable to inves-tigate whether this practice also is valid in theantepartu m period, wh en a fetal growth abnormalityis suspected. Therefore, this study was designed toevaluate the effect of growth disturbances on sono-graph ic fetal foot lengths in both the SGA and LGAfetuses.

MATERIAL AND METHODS

The study p opu lation was identified by reviewing acomputerized database of sonographic foot lengthmeasurements recorded prospectively at our institu-tion from October 1994 to April 1998. All sono-graph ic examinations were p erformed at St. Peter s

University Hospital, New Brunswick, NJ, usingeither an ATL Ultramark H DI 9 or 3000 (Advan cedTechnology Laboratories, Bothell, WA) or an Acuson128 (Acuson, Mountain View, CA). Institutionalreview board app roval was obtained for performingthis stud y.

Fetal Foot Length N omog ram

A nomogram of fetal foot length versus gestationalage for singleton fetuses was constructed for theperiod between 15 and 37 weeks gestation usingavailable measurements. Inclusion of a fetus in the

nomogram required concordance within 2 weeksbetween menstrual and ultrasonographic dating atthe initial sonographic examination. Gestational agewas recorded as completed weeks (based on men-strual dating). The last foot length measurementobtained before delivery was chosen for analysis,and this contributed a single observation for eachfetus. Fetuses with abnormalities involving theextremities, such as m icromelia, talipes, polyd actyly,and arthrogryposis, were exclud ed. Foot length w asmeasured in centimeters in the plantar view from theheel to the first or second toe (whichever was longer)using on-screen calipers. The relationship between

foot length and gestational age was d etermined.

Identification of Fetuses w ith

Growth Abnormalities

SGA fetuses were identified by first generating anomogram of EFW versus gestational age for ourpopulation. This EFW nomogram was constructedusing 18,628 obstetrical ultrasonographic examina-tions from 10,741 singleton fetuses obtained in ourun it du ring the same study period. EFWs (in grams)were determined from 15 to 42 weeks gestation

using head circumference, bipar ietal diameter, femurlength, and abdominal circumference measure-ments.7 For all fetuses analyzed, sonographic biome-try at the initial examination was within 2 weeks ofmenstrual gestational age. Since multiple fetalweight estimations w ere available for m any fetuses,the nom ogram w as constructed after accounting forthis nonindep endence using the generalized estimat-ing equations procedure.8 SGA fetuses were identi-fied as those with EFWs below the 10th percentile,and LGA fetuses were those having EFWs above the90th percentile on our popu lation nom ogram.

In order to distinguish fetuses with pathologicgrowth abnormalities from those who may havebeen constitutionally small or large for gestationalage, we chose to further categorize the growth pat-tern of SGA and LGA fetuses using H/ A ratios. First

a nomogram of H/ A ratio versus gestational agefrom 12 to 42 weeks was constructed based on theavailable 18,628 ultrasonographic examinations.Then both SGA and LGA fetuses were classified assymmetrically or asymmetrically on the basis ofH/ A ratios. SGA fetuses with H/ A ratios above the90th percentile for gestational age were defined assymmetrically small, and those with H / A ratios at orbelow the 90th percentile for gestational age weredefined as asymmetrically small. Asymmetricallylarge fetuses were defined as those with H/ A ratiosbelow the 10th percentile, and symmetrically largefetuses were those with H/ A ratios at or above the

10th p ercentile for gestational age.Fetal foot lengths of SGA and LGA fetuses were

plotted against the foot length nomogram for fetusesin our general population in order to determine theaffect of growth disturbances on this sonographicmeasurement. Only the foot length measurementobtained closest to delivery for each SGA and LGAfetus was used. The num ber of symm etric and asym -metric SGA fetuses with foot lengths less than the5th, 10th, and 50th percentiles and the number ofsymm etric and asymm etric LGA fetuses with footlengths above the 50th, 90th, and 95th percentileswere determined. A subgroup analysis was per-

formed u sing only SGA and LGA fetuses beyond 24weeks gestation. This cutoff was chosen as the earli-est gestational age at w hich grow th abn ormalities arelikely to be detected and are clinically relevant.

Statistical Analysis

Multivariable linear regression models were estab-lished in ord er to derive the smoothed curves for the5th, 10th, 50th, 90th, and 95th percentiles for allnomograms, e.g., foot length versus gestational age,EFW versus gestational age, and H/ A ratio versus

202 FOOT LENGTH IN FETUSES J Ultrasound Med 19:201205, 2000


3/5

gestational age. The smoothed percentiles were gen-erated by fitting linear regression models based onthe restricted cubic spline smoothing procedure.9

Prior to fitting the regression models, the depend entvariables (foot length and H/ A ratio) were checkedfor normality. In addition, the data were checked forthe presence of extreme outliers (beyond four stan-dard deviations from the mean), and coefficients ofvariation (expressed as a ratio of mean to the stan-dard deviation) were compu ted. Chi-square test wasused to compare the number of symmetric versusasymm etric SGA fetuses with foot lengths below the5th, 10th, and 50th percentiles and the number ofsymm etric versus asymm etric LGA fetuses with footlengths above th e 50th, 90th, and 95th p ercentiles. AP value of < 0.05 was considered statistically signifi-cant. All statistical analyses were performed using

the SAS system (SAS Institute, Cary, NC).

RESULTS

Our patient population was 85% Caucasian (otherthan Hispanic) and 10% African American, with theremaining patients being of Hispanic origin and otherminorities. Approximately 45% of the populationwere nulliparous women and 31% were clinicpatients.

Fetal Foot Length N omog ram

Of the 10,741 singleton fetuses who underw ent sono-graph ic examinations during th e study period, 5372fetuses had foot length measurements between 15and 37 weeks gestation. The median nu mber (range)of foot length measurements available at each gesta-tional week was 108 (15 to 1160). The relationshipbetween foot length and gestational age (based onmenstrual dating) was best described by a linearregression model (R2 = 0.93, P < 0.0001; Fig. 1) withgestational age modeled using cubic splines withfive knots (located at 17, 19, 20, 22, and 30 weeks).Table 1 presents the predicted values for the 5th,10th, 50th, 90th, and 95th percentiles of foot length,the number of fetuses examined, and the coefficientof variation in foot length m easurements at each ges-

tational week between 15 and 37 weeks.

SGA Fetuses

Overall, 586 (10.9%) of 5372 fetuses with foot lengthmeasurements had an EFW below the 10th p ercentilefor gestational age and therefore were classified asSGA. The median gestational age (range) for thisgroup was 21 (15 to 36) weeks. One hundred andtwenty-nine (22.0%) of these 586 fetuses w ere classi-fied as asymmetrically small, and 457 (78.0%) were

203J Ultrasound Med 19:201205, 2000 MEIROWITZ ET AL

Figure 1 Fetal foot length percentiles (5th, 10th, 50th, 90th, and 95th) by gestational age between 15 and 37 weeks.


4/5

classified as symmetr ically small. When foot lengthsof all SGA fetuses were plotted against the foot lengthnom ogram, 60.6% (355 of 586) fell below the 10th per-centile (Table 2). The percentages of asymmetric andsymm etric SGA fetuses with foot lengths below the10th percentile (55.0% versus 62.1%; P = 0.28) weresimilar. As shown in Table 2, a subgroup analysis ofSGA fetuses who were beyond 24 weeks gestation(median gestational age [range] of 27 [24 to 37] weeks)yielded similar results. Fifty-seven of 99 fetuses(57.6%) had measurements below the 10th percentile.

LGA Fetuses

Seven hu nd red and forty-four of 5372 (13.8%) fetuseswith foot length measurements were classified asLGA based on an EFW above the 90th p ercentile forgestational age. The median gestational age (range)for this group was 19 (15 to 36) weeks. Of the 744LGA fetuses, 25.3% (n = 188) were classified asasymmetric and 74.7% (n = 556) were classified assymm etric. When foot lengths of LGA fetuses wereplotted against the foot length nomogram, 29.4%(219 of 744) fell above the 90th percentile (Table 3).The percentages of asymmetric and symm etric LGAfetuses with foot lengths above the 90th percentilewere not significantly different regard less of whetherthe analysis was restricted to LGA fetuses of greaterthan 24 weeks gestation (median gestational age[range] of 26 [24 to 36] weeks). How ever, symmetr ic

LGA fetuses were significantly more likely to havefoot lengths above the 50th percentile in the analysisof both all LGA fetuses and those above 24 weeksgestation (Table 3).

DISCUSSION

Fetal foot length nomograms were originally estab-lished using pathologic specimens, using measure-ments obtained from both fresh and fixed specimens.1

Postmortem measurements of the fetus may beaffected by the method of tissue fixation and the time

interval between delivery and fixation.1

Formalin fixedtissues will swell to a variable degree, depending onthe age of the specimen, and will gradually shrink totheir original size after several months.1 Specimens thathave not been fixed rapidly may be subject to dryingartifacts. Nomograms established using postmortemspecimens also may include a disproportionate num-ber of abnormal (i.e., anomalous, growth restricted)fetuses. Despite these concerns, Mercer and coworkers3

demonstrated that prenatal ultrasonographic measure-ments of the fetal foot were in close agreement withStreeters 1920 postmortem data.

Using both pathologic and sonograp hic data, previ-ous investigators16 have suggested that fetal footlength is a reliable predictor of gestational age.However, the results of this investigation illustratethat fetal foot length can be influenced by growthrestriction as well as states of accelerated fetal growth.Our find ings imply that there are limitations to the useof fetal foot length for gestational age assessment, par-ticularly in fetuses with growth abnormalities.

The effects on fetal foot length that are observedwith variations in fetal growth are likely to be attrib-utable to the fact that measurements of the fetal foot

incorporate both soft tissue and bone. Soft tissuestores of subcutaneous fat are reduced in cases ofgrowth restriction and increased in situations ofaccelerated growth. It therefore makes sense thatfetal foot length is relatively smaller in the SGA fetusand larger in the LGA fetus. Since the foot is notentirely soft tissue, the effects of abnormal fetalgrowth on foot length are probably attenuated by therelatively unaffected foot bones. Thus, over 95% ofSGA fetuses have foot lengths below the 50th per-centile but on ly 60.6% are below the 10th p ercentile.Similarly, LGA fetuses are very likely (greater than

204 FOOT LENGTH IN FETUSES J Ultrasound Med 19:201205, 2000

Table 1: Fetal Foot Length Percent iles by Gestationa lAge*

Fetal Foot Length Percentiles

Gestational Age (Smoothed)

(weeks) N CV (%) 5th 10th 50th 90th 95th

15 18 12.7 1.4 1.5 1.8 2.2 2.3

16 146 10.4 1.6 1.7 2.1 2.5 2.6

17 375 9.7 1.9 2.0 2.4 2.8 2.9

18 613 9.8 2.2 2.3 2.7 3.1 3.2

19 1160 8.9 2.5 2.6 3.0 3.3 3.4

20 929 9.3 2.8 2.9 3.2 3.6 3.7

21 552 8.5 3.1 3.2 3.5 3.9 4.0

22 360 8.9 3.4 3.5 3.9 4.2 4.3

23 222 8.1 3.7 3.8 4.2 4.6 4.7

24 177 7.0 4.0 4.1 4.5 4.9 5.0

25 125 7.1 4.3 4.4 4.8 5.1 5.2

26 123 7.0 4.6 4.7 5.1 5.4 5.5

27 108 6.3 4.8 4.9 5.3 5.7 5.8

28 74 5.4 5.1 5.2 5.6 5.9 6.029 66 6.2 5.3 5.4 5.8 6.2 6.3

30 65 5.2 5.6 5.7 6.1 6.4 6.5

31 62 5.7 5.8 5.9 6.3 6.7 6.8

32 65 5.3 6.0 6.1 6.5 6.9 7.0

33 39 4.4 6.3 6.4 6.8 7.1 7.2

34 37 6.8 6.5 6.6 7.0 7.4 7.5

35 24 6.2 6.8 6.9 7.3 7.6 7.7

36 15 5.5 7.0 7.1 7.5 7.9 8.0

37 17 5.3 7.3 7.4 7.7 8.1 8.2

*Values for percentiles are in centimeters. N, Number of fetuses;

CV, coefficient of variation.


5/5

90%) to have foot lengths above the 50th percentile,but only 29.4% are above the 90th p ercentile.

The methods we have used to provide normativedata on sonographic fetal foot length have severaladvan tages over previously published reports. Earlierstudies3,4,6 employed polynomial terms for the vari-ables (foot length and gestational age) in their models.Although this approach would depict nonlinear rela-tionships between the variables, it imposes curvaturesthat may not be valid. The restricted cubic splinemodel used to generate our foot length nomogram

allows flexible modeling by not restricting the relation-ship between the independent and dependent vari-ables. Our nomogram was constructed using a patientpopulation several orders of magnitude larger thanthose used by previous studies,36 and our popu lationappears to have a broader ethnic composition.Although our values for the 50th percentile are almostidentical to those of Mercer and colleagues,3 our outerpercentiles are different, perhaps more reflective ofpopulation norms. Previously published studies havenot presented information regarding coefficient ofvariation for ultrasonographic measurements of fetalfoot length.36 Our data demonstrated larger coeffi-

cients of variation at early gestational ages comparedwith later gestational ages. This was an unexpectedfinding, because foot length is easier to measure dur-ing the second trimester than at later gestational ages.The most plausible explanation for this finding is thatin early gestation foot length is so small that minorvariations in measurements (for technical reasons) willresult in a large coefficient of variation.

The relatively consistent findings for fetuses ana-lyzed beyond 24 weeks of gestation indicates thatfetal foot length may be affected by abnormal growtheven prior to the third trimester. Although we

observed that a statistically greater nu mber of sym-metric LGA fetuses were above the 50th percentilethan asymm etric LGA fetuses, the magnitud e of thedifference suggests that fetal foot length would notbe helpful for categorizing the typ e of LGA growth.Finally, our findings suggest that the practice ofusing fetal foot length to assess gestational age in theantepartu m p eriod should be temp ered, especially inthe p resence of susp ected fetal growth abnormalities.

REFERENCES

1. Streeter GL: Weight, sitting height, head size, foot length,and m enstrual age of the hum an embryo. Contrib EmbryolCarnegie Inst 11:143, 1920

2. Huxley AK: Comparability of gestational age valuesderived from diaphyseal length and foot length fromknow n forensic foetal remains. Med Sci Law 38:42, 1998

3. Mercer BM, Sklar S, Shariatmad ar A, et al: Fetal foot lengthas a predictor of gestational age. Am J Obstet Gynecol156:350, 1987

4. Goldstein I, Reece EA, Hobbins JC: Sonographic appear-ance of the fetal heel ossification centers and foot lengthmeasurements provide independent markers for gesta-tional age estimation . Am J Obstet Gyn ecol 159:923, 1988

5. Hata T, Senoh D, Hata K, et al: Mathematical modeling offetal foot growth: Use of the Rossavik growth mod el. Am JPerinatol 13:155, 1996

6. Platt LD, Medearis AL, DeVore GR, et al: Fetal foot length:The relationship to menstrual age and fetal measurementsin the second trimester. Obstet Gynecol 71:526, 1988

7. Had lock FP, Harr ist RB, Carpenter RJ, et al: Sonograp hicestimation of fetal w eight. Rad iology 150:535, 1984

8. Liang K-Y, Zeger SL: Longitudinal data analysis using gen-eralized linear m odels. Biometrika 73: 13, 1986

9. Durrelman S, Simon R: Flexible regression models withcubic sp lines. Stat Med 8:551, 1989

205J Ultrasound Med 19:201205, 2000 MEIROWITZ ET AL

Table 2: Fetal Foot Length Percentiles for SGA Fetuses

FFL Asymmetric Symmetric

Percentile All SGA SGA SGA P Value*

All FetusesNumber 586 129 457

95th 21 (21.2%) 4 (12.1%) 17 (25.8%) 0.12

>90th 33 (33.3%) 9 (27.3%) 24 (36.4%) 0.37

>50th 88 (88.9%) 25 (75.8%) 63 (95.5%) 0.003

*P value refers to comparisons between asymmetric and symmetric

categories. FFL, Fetal foot length.

foot length

Documents