ª 2013 by the Academy of Nutrition and Dietetics.

RESEARCH

Research and Professional Briefs

An Isocaloric Low Glycemic Index Diet ImprovesInsulin Sensitivity in Women With Polycystic OvarySyndromeSuzanne Barr, PhD, RD; Sue Reeves, PhD; Kay Sharp, PhD; Yvonne M. Jeanes, PhD, RD

ARTICLE INFORMATION

Article history:Accepted 11 June 2013

Keywords:Polycystic ovary syndromeGlycemic indexInsulin resistance

Copyright ª 2013 by the Academy of Nutritionand Dietetics.2212-2672/$36.00http://dx.doi.org/10.1016/j.jand.2013.06.347

ABSTRACTPolycystic ovary syndrome (PCOS) is a common endocrine disorder affecting 5% to 10%of women worldwide. Approximately half of women with PCOS are lean, yet may stillpresent with central obesity and metabolic disturbances. Low-glycemic index (GI) di-etary intervention studies have demonstrated improvements in insulin sensitivity ininsulin-resistant populations; however, there is little evidence of this effect in womenwith PCOS. This research aimed to determine the efficacy of an isocaloric low-GI dietaryintervention on insulin sensitivity, independent of weight change, in womenwith PCOS.A nonrandomized 12-week low-GI dietary intervention, preceded by a 12-week habitualdiet control phase and proceeded by a 12-week follow-up phase was conducted. Dietaryintake, body composition, and metabolic risk markers were determined at baseline,after completion of the habitual diet control phase, and after the low-GI dietary inter-vention. Twenty-six participants were recruited at baseline, 22 commenced and 21participants completed the low-GI dietary intervention phase. The primary outcomewas change in insulin sensitivity. Secondary outcomes included assessment of changesto lipids, body composition, and estimated macronutrient intake. Repeated measuresanalysis of variance with Bonferroni correctionwere used to detect changes to outcomesacross study timepoints. Twenty-one women with PCOS with mean (� standard devi-ation) age of 32.1�6.7 years completed the 12-week low-GI dietary intervention. Asexpected, no significant changes occurred during the 12-week habitual diet controlphase. However, during the dietary intervention phase, dietary GI decreased from 54.5�3.5 to 48.6�5.1 (P<0.001) with a concurrent small reduction in saturated fat intake(12.4%�3% to 11.7%�3% contribution from energy, P¼0.03), despite no specific recom-mendations to modify fat intake. Measures of insulin sensitivity and nonesterified fattyacid improved after intervention (P¼0.03 and P¼0.01, respectively). This is the firststudy to implement an isocaloric low-GI diet in women with PCOS and findings maycontribute to the limited research in this area.J Acad Nutr Diet. 2013;-:---.

POLYCYSTIC OVARY SYNDROME (PCOS) AFFECTSapproximately 5% to 10% of women of reproductiveage.1,2 The most widely accepted diagnostic criteriarequires the presence of two of the three following

features: (a) oligo- or anovulation, (b) clinical and/orbiochemical signs of hyperandrogenism, and (c) polycysticovaries and exclusion of other etiologies.3 Insulin resistanceis a common feature of PCOS, with as many as 80% of womenwith PCOS considered insulin resistant4; compensatory hy-perinsulinaemia is proposed as one of the key contributorsto the pathogenesis of the syndrome,5,6 which leads to a sub-sequent increase in type 2 diabetes and cardiovascular dis-ease (CVD) risk.7,8

Approximately 30% to 70% of womenwith PCOS worldwideare overweight or obese,9 with weight management advisedas the primary strategy for those with a body mass index(BMI) more than 25. As little as 5% weight loss has been

shown to improve reproductive, metabolic, and clinicalmarkers.10-12 However, many women with PCOS are consid-ered lean13 and, thus, weight-management strategies are notappropriate. Despite being ideal weight, these women oftenpresent with clinical symptoms and elevated metabolic riskfactors compared with matched control subjects14; therefore,it is important to establish appropriate dietary strategies forsymptomatic lean women. Strategies that improve clinicaland biochemical outcomes independent of weight loss wouldbe of additional benefit to overweight or obese women withPCOS for whom weight loss remains challenging.There is a paucity of good-quality studies aimed at opti-

mizing dietary intake in this population. A recent reviewhighlighted that diets aimed at reducing weight in PCOS canbe of clinical benefit, but there is no consensus about theoptimum dietary composition to facilitate clinical improve-ments.15 Moran and colleagues also highlighted the need for

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 1

RESEARCH

assessment of the effect of isocaloric diets, independent ofchanges in weight in PCOS.15 Numerous studies havedemonstrated the benefits of reducing dietary glycemic index(GI) for the improvement of insulin sensitivity in otherchronic diseases,16,17 but there remains no clear consensusabout the benefits of low-GI diets in managing the clinicaland biochemical manifestations of PCOS, with limited trialsconducted in this area.18-20 The primary aim of the currentstudy was to determine the efficacy of an isocaloric low-GIdietary intervention on changes to insulin sensitivity inwomen with PCOS. Secondary outcomes were to assesschanges to lipids, body composition, and estimated macro-nutrient intake.

METHODSParticipantsWomen with PCOS in the United Kingdom were recruited viaa self-help PCOS charity and by local advertisement. Partici-pant consent was obtained and eligibility was determined bya screening questionnaire. Participants were required to bepremenopausal, at least 18 years of age, and to self-reporttheir means and date of PCOS diagnosis. Each participant’sgeneral practitioner (GP) was then asked to confirm PCOSstatus via an “opt out” system, whereby nonreturn of thereply slip provided was taken to confirm that the participantwas eligible. The researchers did not stipulate or requestconfirmation of the diagnostic criteria utilized by GPs, andbecause participants may have been diagnosed with PCOS forsome time prior to involvement in the study, it is possiblethat participants were diagnosed using criteria other than thelatest diagnostic criteria.3

Participants were excluded if they were unable to providesupporting evidence of PCOS diagnosis, were pregnant,breastfeeding, or had any medical condition known to inde-pendently influence body composition or biochemistry (eg,thyroid disturbances, Cushings disease, cancer, HIV/AIDS,renal disease, liver disease, familial hyperlipidaemia, diabetesmellitus [type 1 and 2]), were receiving any prescriptionmedication (with the exception of oral contraceptives), orwere receiving treatment for, or currently dealing with, anuntreated eating disorder. Oral contraceptive use was recor-ded for later consideration in data analysis. Presence ofhirsutism was defined according to the validated Ferriman-Gallway scoring system,21 and acne defined according tothe validated Cardiff Acne Disability Index (CADI).22

Eighty-seven women initially expressed an interest intaking part in the study, with 25 of those considered ineli-gible. Reasons for exclusion were use of metformin (n¼14),use of orlistat (n¼1), diagnosis of type 2 diabetes mellitus(n¼3), current breastfeeding (n¼2), age older than 40 years(n¼3), current eating disorder (n¼1), and unwillingness fortheir GP to be contacted to confirm diagnosis (n¼1).

Study OverviewParticipants were invited to the research center on five oc-casions throughout the 36-week study period. The studyincluded a habitual diet control phase (weeks 0 to 12), a low-GI dietary intervention phase (weeks 12 to 24), and a follow-up phase (weeks 24 to 36) in which no dietary interventionor support was provided. At each of the five visits (weeks 0,12, 18, 24, and 36) anthropometrics (weight, height, BMI,

2 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

body fat percentage, lean body mass percentage, waist andhip circumferences) and blood pressure were monitored.Macronutrient intakes were also assessed at these timepointsusing an unweighed food diary. Biochemical measures wereassessed at all visits except Week 18. The night before eachvisit to the research center (except Week 18), participantswere asked to ensure their evening meal contained between30 to 50 grams of carbohydrate to be consumed within aspecified time frame and then remain fasted from 8:00 PM. Itis unlikely that instructions regarding the evening mealwould have influenced the food diary data because this onlyaccounted for one meal across a 7-day period.

Anthropometric and Biochemical MeasurementsHeight, weight, and waist circumference were assessed ac-cording to standardized protocols.23 Total body adiposity wasalso measured under standardized conditions24 by tetra-polar bioelectrical impedance analysis using the Bodystat4000 multifrequency quadscan. Fasting and 2-hour glucoseand insulin levels, fasting total cholesterol, high-density li-poprotein (HDL) cholesterol, triglycerides (TAG), and nones-terified fatty acid (NEFA) were measured and low-densitylipoprotein (LDL) cholesterol was calculated using the Frie-dewald’s equation.25 Plasma glucose, lipids, and NEFA wereanalyzed using an enzymatic colorimetric method using theInstrumentation Laboratory 650. Selected hormones, namelyluteinizing hormone (LH), sex hormone binding globulin(SHBG), total testosterone (T), and follicle stimulating hor-mone (FSH) were analyzed to characterize PCOS status atbaseline. All hormones were measured using enzyme-linkedimmunoassay technique with consumables and reagentssupplied by DRG Diagnostics (DRG Instruments GmbH). A75-gram oral glucose-tolerance test was done and plasmaglucose and serum insulin were measured at 0 and 120 mi-nutes to determine fasting insulin sensitivity and glucosetolerance. The updated nonlinear homeostatic modelassessment (HOMA2) was used to calculate insulin resis-tance, beta cell function, and insulin sensitivity using theOxford Diabetes Trials Unit calculator (www.dtu.ox.ac.uk).The HOMA2 estimates correspond well to nonesteady stateestimates of beta cell function and insulin sensitivity derivedfrom the hyperinsulinemic and hyperglycemic clamp.26

Dietary InterventionThe principle dietary modification during the low-GI dietaryintervention phase was individualized advice to replace high-GI and medium-GI foods with low-GI foods. The re-placements were based on an exchange system ofsubstitutions; for example, white bread replaced with whole-grain bread. No specific advice was given to modify type andamount of fat or protein intake nor the amount of carbohy-drate. Information was provided and explained in the contextof UK Department of Health dietary guidelines using the“Eatwell Plate” model.27 Before the main intervention trial, asmall pilot study in five women without PCOS was done toinform the delivery of the low-GI dietary intervention. Thediet was designed to be isocaloric; therefore, participantswere encouraged to maintain their usual energy intake andmacronutrient intake, in addition to maintaining habitualphysical activity levels throughout the study. Macronutrientintake was assessed prospectively by a 7-day unweighed food

-- 2013 Volume - Number -

RESEARCH

diary at weeks 0, 12, and 24 and by a 3-day unweighed fooddiary at weeks 18 and 36. Physical activity was assessedretrospectively using a short physical activity questionnaireadapted for this study.During weeks 0 to 12 participants acted as their own

controls by following their habitual diet. On completion ofthe control phase (Week 12), an individual consultation witha registered dietitian was held and instructions on how toimplement a low-GI diet were provided. The same registereddietitian provided group education and supporting literatureon two occasions during the dietary intervention phase(weeks 12 and 18). Participants were provided with a low-GIdiet information handbook, low-GI diet recipe book, andlow-GI foodstuffs (for example: bread, breakfast cereals,cereal bars, and pasta). Only foods with a tested low-GI dietvalue were provided and participants were encouraged toself-select foodstuffs most suited to their usual preferences.Weekly motivational electronic mails were sent to all par-ticipants, which contained information and guidance onfollowing the low-GI diet, seasonal recipes, and providedthe opportunity to address any queries. After 6 weeks offollowing the low-GI diet, anthropometric measures wererepeated, the 3-day food diary was reviewed, verbal rein-forcement of the low-GI diet and encouragement to maintainusual activity levels was given by the registered dietitian, andfurther low-GI foodstuffs were provided. At Week 24, aftercompletion of the dietary intervention phase, no explicitguidance was provided to the participants for future dietaryintake. Patients were advised that they could either continuewith the low-GI changes or return to their habitual diet aspreferred. The aim of the follow-up was to ascertain partici-pant willingness to continue with the low-GI diet without thecontinual support of a registered dietitian or the provision offoods.

Dietary AnalysisFood diaries were coded by a trained nutritionist, data wereentered into a computerized dietary analysis package(Dietplan6.3, Forestfield Software), and daily absolute andpercentage macronutrient intakes were estimated. GI valuesfor foods containing carbohydrate were obtained from theinternational glycemic index tables,28 the updated values,29

the University of Sydney online database, and additional GIvalues published for UK products30 according to standardmethodology.31 Mean daily dietary glycemic load (GL) wascalculated using the formula GL¼GI�carbohydrate(g)/100,using the amount of carbohydrate per portion of foodconsumed, and daily GL was then calculated as the product ofthese values. The international GI tables do not provide GIvalues for a number of foods; therefore, if GI values were notavailable (such as for UK brands), suitable alternatives werechosen according to previously published methods31 anddocumented for transparency.

Data AnalysesStatistical analyses were done using the Statistical Packagefor the Social Sciences (SPSS) software for Windows version17 (SPSS Inc). All available data for participants whocompleted the 12-week dietary intervention were includedfor analyses. One participant dropped out during this phaseand therefore that subject’s data were not available for

-- 2013 Volume - Number -

inclusion at the end of the intervention. The Kolmogorov-Smirnov test was used to assess normality for all contin-uous data. Repeated measures analysis of variance withBonferroni post hoc correction or nonparametric equivalentwas used to assess changes across timepoints (eg, beforeand after dietary intervention). Results are presented asmean�standard deviation (SD) for continuous data. Thecutoff for statistical significance was set at the 95% level(P�0.05). The sample size was determined by power calcu-lations using the software package G*Power 3. It was calcu-lated that a sample size of 18 would provide more than 80%statistical power at a power level of P<0.05 to detect a1 mmol/L change in fasting insulin (a component of HOMA2).Full ethical approval was granted by University and NationalResearch Ethics Service (REC:07/H0803/236).

RESULTS AND DISCUSSIONTwenty-six women attended an initial visit at the researchcenter. Baseline characteristics are described in Table 1.Mean�SDBMIwas 29.0�5.9, (range: 18.7-41.6). Approximatelyhalf (48%) of the participants were considered overweight, 24%were lean, and 29%were obese. During the habitual diet controlphase (weeks 0 to 12), no significant changes were found inmacronutrient intake, body composition, or biochemicalmarkers (Table 2), indicating lifestyle habits of the participantshad remainedunchanged. Fourparticipantsdroppedoutduringthe control phase, with one moving from the local area andthree stating lack of time to commit to the remainder of thestudy.Oneadditionalparticipantdroppedoutduring the low-GIdietary interventionphase, also stating lackof time toattend thefinal twovisits. Therefore, 21 participants completed the low-GIdietary intervention phase (weeks 12 to 24). A further threeparticipants dropped out before the follow-up appointment atWeek 36.Mean hormone values for FSH, SHBG, and total T were

within reference ranges. LH levels were raised, consideredcharacteristic of women with PCOS. Ninety-six percent ofparticipants had fasting insulin levels within the referencerange at baseline (<25 mIU/mL).32

Compared with habitual control-phase levels, a significantincrease was demonstrated in insulin sensitivity (72.8�32.0%vs 61.1�24.9%) (P¼0.03) after dietary intervention (Figure,panel A). NEFA levels decreased after the intervention(0.48�0.21 mmol/L vs 0.56�0.25 mmol/L) (P¼0.01) and HDLcholesterol showed a small reduction (1.67�0.5 mmol/L to1.60�0.35 mmol/L) (P¼0.05). There were no changes to mea-sures of glycemia, other lipids, weight, or waist circumferenceafter the intervention phase. A significant decrease in dietaryGI (54.7�4.6 at Week 12 to 48.6�4.6 at Week 24) (P<0.001)was achieved in 21 women who completed the 12-week di-etary intervention phase (Figure, panel B). Between thosetaking the combined oral contraceptive pill (n¼4) and thosetaking no medication (n¼17), no differences were found tometabolic risk markers, body composition, or macronutrientintake following low-GI dietary intervention (P>0.05).At Week 36, 18 women attended the follow-up appoint-

ment (12 weeks after the end of the dietary intervention).Mean dietary GI was 52.6, a significant increase (P<0.001)from Week 24 (48.6), and similar to habitual level (54.6)(P>0.05). Dietary GL was also similar at follow-up comparedwith habitual level (P>0.05). At the follow-up visit, body

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 3

Table 1. Baseline characteristics of women with polycystic ovary syndrome prior to low glycemic index dietary intervention(n¼26)

Characteristic Sample size (n)

��������mean�standard deviation��������!Age (y) 31.5�6.9 26

Mean BMIa 29.0�5.9 26

Ethnicity �����������������%�����������������!White 84.6 22

Black 7.7 2

Asian 3.8 1

Mixed ethnic origin 3.8 1

Weight status

Underweight (BMI<18.5) 3 1

Healthy weight (BMI 18.5-24.9) 24 6

Overweight (BMI 25-29.9) 46 12

Obese (BMI �30) 27 7

Menstrual regularity

Regular mensesb 49 13

Irregular menses 40 10

Amenorrhea 11 3

Clinical symptoms

Hirsutismc 69 18

Acned 27 7

Insulin resistant (according to HOMA2-IR)e 85 22

No medication 80 21

Taking combined oral contraceptive 20 5

aBMI¼body mass index.bRegular menses were defined by self-report as “regular, once per month.”cDefined using Ferriman-Gallway score.21dDefined using Cardiff Acne Disability Index score.22eHOMA2-IR¼homeostatic model assessment, insulin resistance.

RESEARCH

composition and biochemical parameters were not signifi-cantly different from habitual baseline levels, nor followingintervention (P>0.05). Percentage macronutrient intakeswere similar to original habitual levels. These findingsdemonstrate a reduction in dietary GI after an isocaloric di-etary intervention with no changes to body weight. Smallchanges occurred to biochemical measures, suggestingimprovement to insulin sensitivity and NEFA in this smallgroup of women with PCOS. Larger randomized trials thatinclude a long-term follow-up are warranted in this area tofurther explore these findings.Habitual dietary GI and GL values at baseline were consistent

withmeandietaryGIof56.5 in thegeneralUKpopulation.33Totalcarbohydrateandfiber intakedidnotchangesignificantlyduringthe study period and therefore changes to metabolic markerscannotbeattributed tochanges incarbohydrateorfiberquantity.However, percentage energy intake of saturated fatty acid (SFA)decreased significantly during the dietary intervention, despite

4 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

no information on alteration to dietary fat intake provided.Therefore, it cannot be dismissed that changes to saturated fattyacid (SFA) intake may have resulted in direct effects on insulinsensitivity in these participants, as noted in previous studiesevaluating the effect of dietary fat on insulin sensitivity.34

Habitual macronutrient intakes at baseline were similar tothose reported in observational studies of women withPCOS.35,36 Estimated baseline percentage and absolute en-ergy intake from total fat and saturated fat intakes seemed tobe greater than dietary recommendations and carbohydrateintake was less than UK Department of Health recommen-dations for health of 35%E (percentage contribution to energyintake), 10%E, and 50%E, respectively37 (Table 2). Although asecondary aim of the study, estimated macronutrient intakessuggest that women with PCOS would benefit from clearerdietary guidance on appropriate dietary strategies to meetdietary recommendations and ultimately aim to improvelong-term health.

-- 2013 Volume - Number -

Table 2. Changes in biochemical, anthropometric, and dietary measures following low glycemic index dietary intervention in women with polycystic ovary syndrome

Habitual baseline level(Week 0)

Pre-intervention(Week 12)

End of intervention(Week 24)a P value

Biochemical measures (n¼21) (n¼21) (n¼19) ������������������������

mean�standard deviationb������������������������!

Fasting glucose (mmol/L) 5.14�0.32 5.24�0.22 5.16�0.30 0.31

2-hour glucose (mmol/L) 5.49�1.08 5.22�1.04 5.57�1.47 0.11

TAGc (mmol/L) 1.15�0.51 1.29�0.51 1.26�0.57 0.29

Total cholesterol (mmol/L) 4.84�0.73 4.89�0.79 4.83�0.72 0.59

HDL-Cd (mmol/L) 1.61�0.47 1.67�0.5* 1.60�0.35* 0.05*

LDL-Ce (mmol/L) 2.71�0.55 2.63�0.55 2.67�0.61 0.71

NEFAf (mmol/L) 0.68�0.27 0.56�0.25* 0.48�0.21* 0.01*

Fasting insulin (mIU/mL) 13.3�5.4 14.6�6.3 12.5�5.2 0.07

2-hour insulin (mIU/mL) 42.3�21.8 40.4�25.4 41.2�26.1 0.26

HOMA2-ßg (%) 128.2�32.1 132.5�40.5 118.7�28.7 0.06

HOMA2-ISh (%) 69.2�34.1 61.1�24.9* 72.8�32.0* 0.03*

HOMA2-IRi 1.71�0.69 1.89�0.79 1.63�0.66 0.08

Body composition: (n¼21) (n¼21) (n¼21) ������������������������mean�standard deviationb������������������������!

Weight (kg) 80.4�15.0 80.8�14.5 80.3�14.2 0.33

BMIj 29.0�5.3 29.1�5.0 29.0�4.9 0.45

Waist circumference (cm) 95�14 94�12 93�12 0.36

Hip circumference (cm) 111�12 111�11 110�11 0.73

Waist:hip ratio 0.85�0.06 0.85�0.06 0.84�0.05 0.37

Body fat (%) 36.0�8.2 37.2�6.7 36.8�7.0 0.20

Systolic blood pressure (mm Hg) 119�9 116�9 118�9 0.15

Diastolic blood pressure (mm Hg) 72�9 70�8 71�7 0.50

Macronutrient intake (n¼21) (n¼21) (n¼21) ������������������������mean�standard deviationb������������������������!

Energy (kcal) 2,030�386 2,068�385 1,970�489 0.91

Dietary glycemic index 54.6�3.5 54.7�4.6* 48.6�4.6* <0.001*

Dietary glycemic load 10.3�3.7 10.4�3.0 9.9�5.1 0.68(continued on next page)

RESEA

RCH

--

2013Volum

e-

Num

ber-

JOURNALOFTH

EACADEM

YOFNUTR

ITION

AND

DIETETIC

S5

Table 2. Changes in biochemical, anthropometric, and dietary measures following low glycemic index dietary intervention in women with polycystic ovary syndrome(continued)

Habitual baseline level(Week 0)

Pre-intervention(Week 12)

End of intervention(Week 24)a P value

Carbohydrate (%) 43�8 47�8 46�8 0.24

Fiber (NSP)k (g) 16.2�4.1 15.7�4.9 15.0�6.3 0.29

Total sugar (g) 104�37 118�43 100�54 0.49

Sucrose (g) 42�18 50�26 36�22 0.22

Total dietary fat (%) 38�7 36�6 34�5 0.17

Saturated fatty acid (%) 12.5�3 12.4�3* 11.7�3* 0.03*

Monounsaturated fatty acid (%) 13.9�4 11.8�3 11.6�2 0.26

Polyunsaturated fatty acid (%) 6.7�2 5.8�2 5.4�1 0.06

P:S ratiol 0.58�0.26 0.50�0.22 0.48�0.16 0.24

Protein (%) 16�4 15�4 17�4 0.36

aBlood samples unable to be obtained from two participants at Week 24 due to difficulties in venipuncture.bMeans and standard deviations between weeks 12 and 24 were compared by using Bonferroni post hoc test. No significant differences were observed between weeks 0 and 12.cTAG¼triacylglycerols.dHDL-C¼high-density lipoprotein cholesterol.eLDL¼low-density lipoprotein.fNEFA¼nonesterified fatty acid.gHOMA2-IS¼homeostatic model assessment, insulin sensitivity.hHOMA2-IR¼homeostatic model assessment, insulin resistance.iHOMA2-ß¼homeostatic model assessment, beta cell activity.jBMI¼body mass index.kNSP¼nonstarch polysaccharide.lP:S ratio¼polyunsaturated fatty acid: saturated fatty acid ratio.*Significant at P<0.05 after analysis of variance analysis.

RESEA

RCH

6JO

URNALOFTH

EACADEM

YOFNUTR

ITION

AND

DIETETIC

S-

-2013

Volume

-Num

ber-

Figure. (A) Percentage change in insulin sensitivity (HOMA2-IS) and insulin resistance (HOMA2-IR) among women with polycysticovary syndrome participating in an isocaloric low glycemic index (GI) diet intervention. *P<0.05 (insulin sensitivity [HOMA2-IS]),error bars represent the 95% confidence intervals. (B) Percentage change dietary GI among women with polycystic ovary syndromeparticipating in an isocaloric low-GI diet intervention. *P<0.001, error bars represent the 95% confidence intervals. Week 0:GI¼54.6�3.5; Week 12: GI¼54.7�4.6; Week 24: GI¼48.6�4.6.

RESEARCH

The study was not powered to detect changes to LDLcholesterol or triglycerides, and no changes to theseparameters were noted after dietary intervention. A findingof potential concern was the small reduction to HDLcholesterol after dietary intervention; however, significancewas borderline and levels remained within normal referencerange (0.89 to 2.05 mmol/L).32 NEFA levels significantlydecreased after dietary intervention, a particular benefit inwomen with PCOS because high levels of NEFA are associ-ated with dyslipidaemia and increased risk of type 2 diabetesand CVD.38 No previous low-GI studies in women with PCOShave reported improvements in NEFA, but this effecthas been demonstrated in low-GI studies of non-PCOSpopulations.39

A particular strength of this study was the nonprescriptivemode of delivery and recommendation for ad libitummacronutrient intake. As expected in accordance with the

-- 2013 Volume - Number -

protocol, no changes were noted to anthropometric measures.Therefore, possible improvements in metabolic risk factorsin a period of weight stability could confer benefit to bothlean women with PCOS and to overweight or obese womenin conjunction with established weight-loss strategies.A short self-reported physical activity questionnaire was

completed by participants during the study, but no objectivemeasures of physical activity such as accelerometry wereused. This would be of interest to explore and include as apotential confounding variable in future studies. A limitationof this study was that the diagnostic criteria used by the GPwas unknown to the researchers. In addition, althoughdropout was relatively low during the 12-week low-GI di-etary intervention phase (approximately 5%), it should also beacknowledged that dropout during the initial control periodwas approximately 20% and approximately 30% in the follow-up phase. This may be as a result of the more intensive

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 7

RESEARCH

weekly contact with participants during the 12-week dietaryintervention phase.Improvements in health between lean and overweight

women would also have been of interest to compare, but dueto the small sample size it was not possible to explorechanges to metabolic markers between weight categories.Future larger intervention studies should aim to recruit leanwomen with the aim of improving data on metabolic riskprofile in this subgroup.Low-GI diets are currently popular with both registered

dietitians and patients in the management of PCOS,40 andmay be useful for the lean PCOS population for whomweightloss is not indicated, or in overweight patients for whomweight loss remains challenging. This study suggested smallincreases to insulin sensitivity in women with PCOS areachievable through dietary modification in the short-term.Reduction in SFA after dietary intervention meant it wasnot possible to distinguish between the effect of SFA or GI oninsulin sensitivity and is an area that warrants furtherinvestigation in future trials. Low-GI diets may also be a moreachievable and acceptable means for women with PCOS tomodify dietary intake, whether used in isolation or inconjunction with energy restriction.Findings demonstrate possible improvements in metabolic

risk factors from a low-GI diet in a period of weight stability;however, because of the relatively small sample size of thisstudy, it is not possible to fully inform clinical practice fromthese results. This study therefore adds to the limited evi-dence base for the potential benefits of a low-GI diet toimprove insulin sensitivity of women with PCOS of varyingweight status. Findings warrant further investigation vialarger, long-term randomized clinical trials in either hypo-caloric or isocaloric conditions to fully elucidate the effects ofa low-GI diet in PCOS.

References1. Azziz R, Carmina E, Dewailly D, Androgen Excess Society, et al. Po-

sition statement: Criteria for defining polycystic ovary syndrome as apredominantly hyperandrogenic syndrome: An androgen excesssociety guideline. J Clin Endocrinol Metab. 2006;91(11):4237-4245.

2. Lindholm A, Andersson L, Eliasson M, Bixo M, Sundstrom-Poromaa I.Prevalence of symptoms associated with polycystic ovary syndrome.Int J Gynecol Obstet. 2008;102:39-43.

3. Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus WorkshopGroup. Revised 2003 consensus on diagnostic criteria and long termhealth risks related to polycystic ovary syndrome. Fertil Steril.2004;81(1):19-25.

4. Carmina E, Lobo RA. Use of fasting blood to assess the prevalene ofinsulin resistance in women with polycystic ovary syndrome. FertilSteril. 2004;82:661-665.

5. Carmina E. Metabolic syndrome in polycystic ovary syndrome.Minerva Ginecol. 2006;58(2):109-114.

6. Balen A. The pathophysiology of polycystic ovary syndrome: Tryingto understand PCOS and its endocrinology. Best Pract Res Clin ObstetGynaecol. 2004;18(5):685-706.

7. Sir-Petermann T. Polycystic ovary syndrome, a pathway to type 2diabetes. Nutrition. 2005;21:1160-1163.

8. Westerveld H, Hoogendoorn M, de Jong AW, Goverde AJ, Fauser BC,Dallinga-Thie GM. Cardiometabolic abnormalities in the polycysticovary syndrome: pharmacotherapeutic insights. Pharmacol Ther.2008;119(3):223-241.

9. Yucel A, Noyan V, Sagsoz N. The association of serum androgens andinsulin resistance with fat distribution in polycystic ovary syndrome.Eur J Obstet Gynecol Reprod Biol. 2006;126(1):81-89.

10. Crosignani P, Vegetti W, Somigliana E, Gessati A, Ragni G. Overweightand obese anovulatory patients with polycystic ovaries: Parallel

8 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

improvements in anthropometric indicies, ovarian physiology andfertility rate induced by diet. Hum Reprod. 2003;18(9):1928-1932.

11. Moran L, Noakes M, Clifton PM, Tomlinson L, Galletly C, Norman RJ.Dietary composition in restoring reproductive and metabolic phys-iology in overweight women with polycystic ovary syndrome. J ClinEndocrinol Metab. 2003;88(2):812-819.

12. Stamets K, Taylor DS, Kunselman A, Demers LM, Pelkman CL,Legro RS. A randomized trial of the effects of two types of short-termhypocaloric diets on weight loss in women with polycystic ovarysyndrome. Fertil Steril. 2004;81(3):630-637.

13. Faloia E, Canibus P, Gatti C, et al. Body composition, fat distributionand metabolic characteristics in lean and obese women with poly-cystic ovary syndrome. J Endocrinol Invest. 2004;27(5):424-429.

14. Holte J, Bergh T, Berne C, Lithell H. Serum lipoprotein lipid profile inwomen with the polycystic ovary syndrome: Relation to anthropo-metric, endocrine and metabolic variables. Clin Endocrinol (Oxf).1994;41(4):463-471.

15. Moran LJ, Ko H, Misso M, et al. Dietary composition in thetreatment of polycystic ovary syndrome: A systematic review toinform evidence-based guidelines. J Acad Nutr Diet. 2013;113(4):520-545.

16. Wolever T, Jenkins DJ, Vuksan V, et al. Beneficial effect of a low gly-caemic index diet in type 2 diabetes. Diabet Med. 1992;9(5):451-458.

17. Barclay A, Petocz P, McMillan-Price J, et al. Glycemic index, glycemicload, and chronic disease riskeA meta-analysis of observationalstudies. Am J Clin Nutr. 2008;87(3):627-637.

18. Marsh K, Brand-Miller J. The optimal diet for women with polycysticovary syndrome? Br J Nutr. 2005;94(2):154-165.

19. Marsh K, Steinbeck KS, Atkinson FS, Petocz P, Brand-Miller JC.Effect of a low glycemic index compared with a conventionalhealthy diet on polycystic ovary syndrome. Am J Clin Nutr.2010;92(2):83-92.

20. Atomio W, Read A, Golding M, et al. Local recruitment experiencein a study comparing the effectiveness of a low glycaemic indexdiet with a low calorie healthy eating approach at achieving weightloss and reducing the risk of endometrial cancer in women withpolycystic ovary syndrome (PCOS). Contemp Clin Trials. 2009;30(5):451-620.

21. Ferriman D, Gallwey JD. Clinical assessment of body hair growth inwomen. J Clin Endocrinol Metab. 1961;21(11):1440-1446.

22. Motley R, Finlay AL. Practical use of a disability index in the routinemanagement of acne. Clin Exp Dermatol. 1992;17(1):1-3.

23. Norgan N. Laboratory and field measurements of body composition.Public Health Nutr. 2005;8(7):1108-1122.

24. Kushner R. Bioelectrical impedance analysis: A review of principlesand applications. J Am Coll Nutr. 1992;11(2):199-209.

25. Friedewald W, Levy RI, Fredrickson DS. Estimation of the concen-tration of low-density lipoprotein cholesterol in plasma, without useof the preparative ultracentrifuge. Clin Chem. 1972;18(6):499-502.

26. Levy J, Matthews DR, Hermans MP. Correct Homeostasis ModelAssessment (HOMA2) Evaluation uses the computer program. Dia-betes Care. 1998;21(12):2191-2192.

27. The Food Standards Agency F. ‘Eatwell’ plate www.eatwellgovuk/healthydiet/eatwellplate/. 2007. Accessed August 16, 2011.

28. Foster-Powell K, Holt S, Brand-Miller J. International table of glycemicindex and glycemic load values: 2002. Am J Clin Nutr. 2002;76(1):5-56.

29. Atkinson F, Foster-Powell K, Brand-Miller JC. International tables ofglycemic index and glycemic load values: 2008. Diabetes Care.2008;31(12):2281-2283.

30. Aston L, Gambell JM, Lee DM, Bryant SP, Jebb SA. Determination ofthe glycaemic index of various staple carbohydrate-rich foods in theUK diet. Eur J Clin Nutr. 2008;62(2):279-285.

31. Olendzki B, Ma Y, Culver AL, et al. Methodology for adding glycemicindex and glycemic load values to 24-hour dietary recall database.Nutrition. 2006;22(11-12):1087-1095.

32. Holt RIG, Hanley NA. Chapter 4. Investigations in endocrinology. In:Essential Endocrinology and Diabetes. 5th ed. Oxford, UK: BlackwellPublishing; 2007.

33. Van Bakel M, Kaaks R, Feskens EJ, Rohrmann S, Welch AA. Dietaryglycaemic index and glycaemic load in the European ProspectiveInvestigation into Cancer and Nutrition. Eur J Clin Nutr. 2009;63(4 suppl):188S-205S.

-- 2013 Volume - Number -

RESEARCH

34. Galgani J, Uuauy RD, Aguirre CA, Diaz EO. Effect of the dietary fatquality on insulin sensitivity. Br J Nutr. 2008;100(3):471-479.

35. Barr S, Hart K, Reeves S, Sharp K, Jeanes YM. Habitual dietary intake,eating pattern and physical activity of women with polycystic ovarysyndrome. Eur J Clin Nutr. 2011;65(10):1126-1132.

36. Wright C, Zborowski JV, Talbott EO, McHugh-Pemu K, Youk A. Di-etary intake, physical activity, and obesity in women with polycysticovary syndrome. Int J Obes. 2004;28(8):1026-1032.

37. Department of Health. Dietary Reference Values for Food Energy andNutrients for the United Kingdom. Report on Health and Social Subjects41. London: The Stationary Office HSMO; 1991.

-- 2013 Volume - Number -

38. Carlsson M, Wessman Y, Almgren P, Groop L. High levels of nones-terified fatty acids are associated with increased familial risk ofcardiovascular disease. Arterioscler Thromb Vasc Biol. 2000;20(6):1588-1594.

39. Rizkalla S, Taghrid L, Laromiguiere M, et al. Improved plasma glucosecontrol, whole-body glucose utilization, and lipid profile on a low-glycemic index diet in type 2 diabetic men: A randomizedcontrolled trial. Diabetes Care. 2004;27(8):1866-1872.

40. Jeanes Y, Barr S, Smith K, Hart KH. Dietary management of womenwith polycystic ovary syndrome in the United Kingdom: The role ofdietitians. J Hum Nutr Diet. 2009;22(6):551-558.

AUTHOR INFORMATIONS. Barr is a research dietitian, Diabetes & Nutritional Sciences Division, King’s College London, London, UK; at the time of the study, she was a PhDstudent, Department of Life Sciences, University of Roehampton, London, UK. S. Reeves is a programme convener and registered nutritionist,Department of Life Sciences, University of Roehampton, London, UK. K. Sharp is retired; at the time of the study, she was a programme convener,Department of Life Sciences, University of Roehampton, London, UK. Y. M. Jeanes is a senior lecturer and registered dietitian, Department of LifeSciences, University of Roehampton, London, UK.

Address correspondence to: Suzanne Barr, PhD, RD, Diabetes & Nutritional Sciences Division, King’s College London, School of Medicine, Franklin-Wilkins Bldg, 150 Stamford St, London, SE1 9NH E-mail: suzanne.barr@kcl.ac.uk

STATEMENT OF POTENTIAL CONFLICT OF INTERESTNo potential conflict of interest was reported by the authors.

FUNDING/SUPPORTY. M. Jeanes and S. Barr received internal funding from University of Roehampton, London, UK, to conduct the entirety of the research.

JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 9