FTO Genotype and 2-Year Change in Body Composition and Fat Distribution in Response to Weight-Loss Diets


Xiaomin Zhang; Qibin Qi; Cuilin Zhang; Frank B. Hu; Frank M. Sacks; Lu Qi


Diabetes. 2012;61(11):3005-3011. 

In This Article

Research Design and Methods

Study Population

The POUNDS LOST Trial was conducted from October 2004 through December 2007 at two sites as follows: Harvard School of Public Health and Brigham & Women's Hospital in Boston, Massachusetts; and the Pennington Biomedical Research Center of Louisiana State University System, Baton Rouge, Louisiana. The design and sample collection have been described previously in detail.[22] In brief, the study population was composed of 811 overweight or obese (BMI ranged from 25 to 40 kg/m2) participants aged 30–70 years. Major criteria for exclusion were the presence of diabetes or unstable cardiovascular disease, the use of medications that affect body weight, and insufficient motivation as assessed by interview and questionnaire. Among the 742 participants who were genotyped successfully, 61% were women, 80% were white, 15% were black, 3% were Hispanic, and 2% were Asian or other ethnic groups by self-report. The participants were assigned randomly to one of four diets constituting a two-by-two factorial design; the target percentages of energy derived from fat, protein, and carbohydrate in the four diets were 20, 15, and 65%; 20, 25, and 55%; 40, 15, and 45%; and 40, 25, and 35%. After 2 years, 645 participants (80% of total population) completed the trial. The study was approved by the human subjects committee at each institution and by a data and safety monitoring board appointed by the National Heart, Lung, and Blood Institute. All participants provided written informed consent.


In the morning before breakfast, body weight and waist circumference (WC) were measured on 2 days at baseline; 6, 12, and 18 months; and 2 years. BMI was calculated as weight by height squared (kg/m2). A dual-energy X-ray absorptiometry (DEXA) scan was performed on 50% of a random sample from the total study participants (n = 424), including 242 (57.1%) women, using a Hologic QDR 4500A (Waltham, MA).[23] Total fat mass (kg), total fat-free mass (FFM; kg), whole body total percentage of fat mass (FM%) and percentage of trunk fat were obtained once at baseline, 6 months, and 2 years. Computed tomography (CT) was used in 50% of a random sample from those participants who had DEXA scans, resulting in a sample of 25% of the total participants (n = 195), including 113 (58.2%) women. Total adipose tissue (TAT) mass, visceral adipose tissue (VAT) mass, deep subcutaneous adipose tissue (DSAT) mass, and superficial adipose tissue (SAT) mass within the abdomen were measured by standard methods,[24] once at baseline, 6 months, and 2 years. Because of radiation exposure, premenopausal women would not subject themselves to CT scans. A series of eight single-slice images were obtained every 10 cm from 2 below and 5 above the fourth and fifth lumbar vertebrae interspaces. These contiguous cross-sectional images were analyzed, and then the total volume was calculated from the individual slices. In this analysis, we only included data at baseline, 6 months, and 2 years with all the outcomes because the DEXA and CT scans were only performed at these three time points.


DNA was extracted from the buffy coat fraction of centrifuged blood using the QIAmp Blood Kit (Qiagen, Chatsworth, CA). Single nucleotide polymorphism (SNP) rs1558902 was selected because it had emerged as the top variant of FTO locus for BMI and WC in recent obesity-related GWAS.[25,26] The SNP was genotyped successfully in 742 of 811 total participants and 603 of 645 participants who completed the trial using the OpenArray SNP Genotyping System (BioTrove, Woburn, MA). Of the 424 participants who received DEXA scans, 391 were genotyped at baseline, and 224 participants who completed the trial were genotyped. Of the 195 participants who received CT scans, 175 were genotyped at baseline and 105 participants who completed the trial were genotyped. The genotype success rate was 99% in available DNA samples. Replicated quality control samples (10%) were included in every genotyping plate with greater than 99% concordance.[27] The allele frequency in two major ethnic groups (white and black) was compatible with Hardy-Weinberg equilibrium (P > 0.05).

Statistical Analysis

The primary outcomes were changes in body weight and WC. Secondary outcomes were changes in body composition including total fat mass, FFM, FM% and percentage of trunk fat, and fat distribution (TAT, VAT, SAT, and DSAT). Data were pooled from the diets for the two factorial comparisons: low protein versus high protein and low fat versus high fat.[22] Because the majority of the study population were white (80%), we also analyzed the main effects and interactions among white participants separately. The Hardy-Weinberg equilibrium and comparison of categorical variables at baseline were assessed with χ2 test. Differences in continuous variables at baseline were tested using ANCOVA, with adjustment for age, sex, and ethnicity. The main effects of genotype and diet intervention on outcome changes at 6 months and 2 years were analyzed using general linear regression models, with adjustment for covariates including age, sex, ethnicity, carbohydrate, the baseline value for the respective outcome, and baseline BMI. We excluded individuals with missing measures at each time point in the analysis. Moreover, to analyze the potential interactions between genotype and diet intervention, an interaction product term of genotype-diet was included in the models. In a secondary analysis, we used linear mixed models, with time as a repeated measurement factor, to test genetic associations with the trajectory of changes in outcomes in the participants who provided measurements at baseline, 6 months, and 2 years in each of four diet groups over the 2-year intervention by including genotype-time interaction terms. Because an additive genetic effect was reported in the original large-scale GWAS in which the SNP was identified,[25,26] additive models were analyzed for genotype. All reported P values were two-sided and a P value of 0.05 was considered statistically significant. All data were analyzed with SAS version 9.1 (SAS Institute, Inc., Cary, NC).