A Randomized Controlled Trial Investigating the Effects of a Low-Glycemic Index Diet on Pregnancy Outcomes in Gestational Diabetes Mellitus

Jimmy Chun Yu Louie, MNUTRDIET, APD; Tania P. Markovic, MBBS, PHD; Nimalie Perera, MBBS; Deborah Foote, BSC DIPNUTRDIET, APD; Peter Petocz, PHD; Glynis P. Ross, MBBS; Jennie C. Brand-Miller, PHD


Diabetes Care. 2011;34(11):2341-2346. 

In This Article

Research Design and Methods

This study was a two-arm parallel randomized controlled trial based at the Diabetes Antenatal Clinic of the Royal Prince Alfred Hospital, Camperdown, Australia. With the exception of the study dietitian (J.C.Y.L.), who provided the dietary education, all study personnel and participants were blinded to dietary assignment.

Subject Recruitment, Randomization, and Stratification

Women aged 18–45 years diagnosed with GDM by a 75-g oral glucose tolerance test at 20–32 weeks' gestation, with an otherwise healthy singleton pregnancy, were eligible for the study. GDM diagnosis was based on the modified Australasian Diabetes in Pregnancy Society (ADIPS) criteria: fasting blood glucose level (BGL) ≥5.5 mmol/L, 1-h BGL ≥10.0 mmol/L, or 2-h BGL ≥8.0 mmol/L. Most women were tested at 26–32 weeks, but testing occurred earlier in those at high risk. Women who had special dietary requirements (including vegetarianism/veganism), preexisting diabetes, or pregnancy achieved by assisted reproduction and those who smoked or consumed alcohol during pregnancy were excluded. A total of 482 women were approached between September 2008 and November 2010, of whom 99 met the inclusion criteria and agreed to participate. The enrolled subjects were centrally randomized to study diet by computer-generated random numbers, stratified by BMI (BMI <30 vs. ≥30 kg/m2) and weeks of gestation (<28 or ≥28 weeks). The allocation sequence was unpredictable and concealed from the recruiter. Participants received routine GDM care regardless of dietary assignment, including instructions to monitor BGL before breakfast and 1 h after meals. The treating endocrinologist (T.P.M. or N.P.) reviewed the subjects every 2–4 weeks prior to 36 weeks and then every week until delivery. Insulin treatment was commenced if the mean fasting BGL or 1-h postprandial BGL in the preceding week exceeded 5.2 and 7.5 mmol/L, respectively.

Demographics and Dietary Assessment

At enrollment, demographic information, family history of diabetes, and ethnicity were recorded. Subjects were asked to recall their prepregnancy weight, were weighed, and were asked to complete a 3-day food record (including 2 weekdays and 1 weekend day) at baseline and again at 36–37 weeks' gestation. A two-dimensional food model booklet was provided to the subjects to assist in portion size estimation. Last recorded weight before delivery was obtained from the medical record.

Dietary Interventions

Subjects were randomized to one of two healthy diets of similar protein (15–25%), fat (25–30%), and carbohydrate (40–45%) content—one with an LGI (target GI ≤50) and the other with a high-fiber content and moderate GI, similar to the Australian population average (HF) (target GI ~60).[13–15] Both study diets provided all essential nutrients for pregnancy other than iron and iodine, which were supplemented as appropriate by the treating endocrinologist. The baseline 3-day food diary provided information on baseline dietary composition and served as the basis of individualized dietary counseling. Sample menus and their nutritional analyses are given in Supplementary Table A1.

Subjects attended at least three face-to-face visits with the study dietitian, scheduled to coincide with regular antenatal visits. A 24-h recall of all food and drink intake was conducted during each session to assess compliance. In the case of noncompliance, suitable alternative foods were encouraged. Food sample baskets containing key foods for the assigned diet were provided to promote product recognition and dietary adherence. The content of the sample baskets is listed in Supplementary Table A2.

Data Collection

Subjects provided blood samples at baseline and ~36 weeks' gestation. Pregnancy outcomes, including birth weight, infant length, infant head circumference, and the need for emergency caesarean section, were obtained from the electronic medical records system. Gestational age was based on the last menstrual period and early pregnancy ultrasound. Birth weight centile was calculated using a macro program for Microsoft Excel (available from http://www.gestation.net) that adjusted for ethnic differences.[16] The calculated birth weight centile was used to categorize the infant as small for gestational age (birth weight <10th centile), normal, or LGA (birth weight >90th centile). Ponderal index, an estimate of neonatal adiposity, was calculated as birth weight in kg × infant length (m)−3. Macrosomia was defined as birth weight >4 kg.

Nutritional Analysis and Assessment of Compliance

The study dietitian entered the food records into Australian nutrition analysis software based on AUSNUT2001 (FoodWorks Professional 2009; Xyris Software, Brisbane, Australia). The GI of individual food items was assigned according to a published method.[17] Dietary glycemic load (GL) was calculated as follows: ∑ GI × available carbohydrate of each food in a day/100. Dietary GI was calculated as follows: (dietary GL/total daily available carbohydrate) × 100. Subjects were deemed compliant if their final dietary GI was ≤50 in the LGI group and >50 in the HF group.

Power Calculation

Based on previous data, the study was designed to provide 80% statistical power to detect an ~260 g difference in birth weight, with 60 subjects in each group. Recruitment was halted at 99 subjects because the SD in birth weight among the study population was smaller than expected. In the primary analysis, the observed SD of 416 g in birth weight provided 80% power to detect a group difference of 246 g in birth weight or an ~17% point difference in birth weight centile.

Statistical Analyses

A biostatistician blinded to the diet allocation performed the statistical analyses. The primary analysis included all women randomized who attended at least one dietary education session but excluded those with preterm delivery (<37 weeks; n = 4; two from each group) regardless of compliance. All statistical analyses were performed with SPSS (version 19; IBM Australia, St. Leonards, Australia). Results for continuous data are reported as mean ± SD or mean ± SEM, and categorical data (e.g., need for insulin) are reported as percentage. Pearson χ2 test was used to test for differences between groups for categorical data, and continuous data were tested using one-way ANOVA. A paired t test was used to assess within-group changes from baseline to final outcomes.

The study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects/patients were approved by the Human Research Ethics Committee of the Sydney South West Area Health Service (Royal Prince Alfred Hospital Zone). Informed consent was obtained from all subjects in this study.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: