Genetic Predisposition to Central Obesity and Risk of Type 2 Diabetes: Two Independent Cohort Studies

The dramatic increase in the global incidence of obesity has been accompanied by an increase in incident type 2 diabetes (T2D) (1). Anthropometric measures capturing abdominal adiposity, such as waist circumference or waist-to-hip ratio (WHR), have been related to diabetes risk independent of BMI across various ethnic groups (2–5). Compelling evidence shows that accumulation of visceral fat plays a pivotal role in the etiology of diabetes by overexposing the liver to free fatty acids, which subsequently result in insulin resistance and hyperinsulinemia (6). However, whether central obesity plays a causal role in the development of T2D remains unclear. A meta-analysis of the WHR genome-wide association studies (GWAS) comprising >100,000 individuals of European ancestry established a comprehensive genetic profile for modulating body fat distribution independent of overall adiposity (7). Information on the association of genetic predisposition to central obesity with risk of T2D is limited.

In the current study, we calculated the genetic score based on 14 well-established single nucleotide polymorphisms (SNPs) for WHR (7) as a proxy of genetic predisposition to central obesity. We examined the association of this genetic score with T2D risk in women and men of European ancestry from two prospective cohorts: the Nurses’ Health Study (NHS) and the Health Professionals Follow-up Study (HPFS).

The NHS is a prospective cohort study of 121,700 female registered nurses aged 30–55 years at study inception in 1976 when all participants completed a mailed questionnaire about medical history and lifestyle (8). Between 1989 and 1990, 32,826 women provided blood samples. The HPFS is a prospective cohort study of 51,529 U.S. male health professionals who were aged 40–75 years at study inception in 1986 (9). Between 1993 and 1999, 18,159 men provided blood samples. In both cohorts, medical and lifestyle information has been collected biennially by self-administered questionnaires since inception. Both studies were approved by the human research committee at the Brigham and Women’s Hospital (Boston, MA), and all participants provided written informed consent.

Participants for the current study were selected among those with a blood sample, using a nested case-control study design (10,11). Diabetes cases were defined as self-reported diabetes confirmed by a validated supplementary questionnaire (12,13). For cases before 1998, we used the National Diabetes Data Group criteria to define diabetes (14), which included one of the following: one or more classic symptoms (excessive thirst, polyuria, weight loss, hunger, pruritus, or coma) plus a fasting plasma glucose level of ≥7.8 mmol/L (140 mg/dL), a random plasma glucose level of ≥11.1 mmol/L (200 mg/dL), or a plasma glucose level 2 h after an oral glucose tolerance test of ≥11.1 mmol/L (200 mg/dL); at least two elevated plasma glucose levels on different occasions in the absence of symptoms; or treatment with hypoglycemia medication (insulin or oral hypoglycemic agent). We used the American Diabetes Association diagnostic criteria for diabetes diagnosis since 1998 (15). These criteria were the same as those proposed by the National Diabetes Data Group except for the elevated fasting plasma glucose criterion for which the cut point was changed from 7.8 mmol/L (140 mg/dL) to 7.0 mmol/L (126 mg/dL). Only participants with a diabetes diagnosis after the cohort baseline were included as cases. The validity of diabetes self-report was verified in a subsample from the HPFS and the NHS cohorts. A physician blinded to the information reported on the supplementary questionnaire reviewed the medical records according to the diagnostic criteria. Ninety-seven percent of the diabetes cases were confirmed in the HPFS cohort (13). Similarly, 98% of the diabetes cases reported by the supplementary questionnaire were confirmed by medical record review in a subsample of NHS participants (12). Control participants were defined as those free of diabetes at the time of the diagnosis of cases and remained unaffected through follow-up until 2006. We matched the case participants to control participants by age, month and year of blood draw, and fasting status within their respective cohorts. After applying a quality control filter in the NHS and HPFS T2D GWAS (11), duplicate samples, samples with misidentified sex, related samples (siblings or possible first cousins), samples with evidence of contamination, samples with highly variable intensity data, and samples with missing call rates ≥2% were excluded. The population structure was investigated by principal component analysis (16). In this analysis, participants clustered with the HapMap CEU (Utah residents with northern and western European ancestry) samples were genetically inferred to have European ancestry. A total of 3,221 women (1,467 case and 1,754 control participants) and 2,422 men (1,124 case and 1,298 control participants) of genetically inferred European ancestry were included in the current analysis.

Waist and hip circumference were reported to the nearest one-quarter inch in 1986 (NHS), and participants were instructed to measure their waists while standing relaxed at the navel and their hips at the largest circumference, including the buttocks in 1987 (HPFS) (17). Each supplemental questionnaire included a tape measure and detailed instructions, including a diagram. Self-reported anthropometric measures were validated against technician measurements among a subset of participants residing in the Boston area (18). Crude Pearson correlations between self-reported waist circumferences and the average of two technician-measured waist circumferences were 0.95 for men and 0.89 for women. Similarly, correlations for hip measurements were 0.88 for men and 0.84 for women and for WHRs, 0.69 for men and 0.70 for women. Correlations became stronger after correcting for random within-person variability from daily or seasonal fluctuations (18). BMI was calculated as weight in kilograms divided by the square of height in meters. Information about diet and lifestyle factors (e.g., smoking status, alcohol intake), menopausal status and postmenopausal hormone therapy (women only), and medications was derived from the baseline questionnaires (8,9). For men, physical activity was expressed as metabolic equivalents per week by using the reported time spent doing various activities, weighting each activity by its intensity level. For women, physical activity was expressed as hours per week because MET hours were not measured at baseline in the NHS. The Healthy Eating Index, which measures adherence to the U.S. Department of Agriculture Dietary Guidelines for Americans and MyPyramid, was calculated among men and women (19). The validity of the self-reported body weight and physical activity data has been described previously (18,20,21). Self-reported and measured weights were highly correlated at 0.97 for men and 0.97 for women (18).

SNP genotyping and imputation have been described in detail elsewhere (the NHS and HPFS T2D GWA scans) (11). In brief, samples were genotyped and analyzed using the Affymetrix Genome-Wide Human Array 6.0 (Santa Clara, CA) and the Birdseed calling algorithm. All samples used in the current study achieved a call rate of >98%. We used MACH (http://www.sph.umich.edu/csg/abecasis/mach) to impute SNPs on chromosomes 1–22, with National Center for Biotechnology Information build 36 of phase II HapMap CEU data (release 22) as the reference panel.

To estimate the genetic predisposition to central obesity, a genetic score was calculated on the basis of the well-established SNPs in 14 loci (Supplementary Table 1) for the WHR reported by a meta-analysis of GWAS (7). We assumed that each SNP in the panel acts independently in an additive manner, and the genetic score was calculated by using a weighted method. Each SNP was weighted by its relative effect size (β-coefficient) obtained from the reported meta-analysis data (7). Using the same method for the previously reported BMI genetic score (22–24), we first created a weighted score using the following equation: weighted score = β1 × SNP1 + β2 × SNP2 + … + βn × SNPn, where β is the β-coefficient for each individual SNP, and n is number of SNPs. To reflect the number of WHR-increasing alleles, we rescaled the weighted score using the following equation: weighted genetic risk score (GRS) = weighted score × (total number of SNPs / sum of the β-coefficients). Figure 1 shows the association of genetic score with the WHR. In addition, a BMI genetic score based on 32 SNPs was calculated using a similar method to that described previously (22–24).

Comparisons of proportions and means between case and control participants were calculated by χ² and t tests. We used logistic regression to estimate odds ratios (ORs) for T2D risk, adjusting for age and BMI. To examine the accumulative effect of the genetic score, we compared the T2D risk across the quartiles of the genetic score. In multivariate analysis, we further adjusted for smoking (never, past, or current), alcohol intake (0, 0.1–4.9, 5.0–9.9, 10.0–14.9, or ≥15.0 g/day), menopausal hormone therapy use (never, past, or current [women only]), Healthy Eating Index, and physical activity (quintiles). BMI genetic score and WHR were further adjusted for association with WHR genetic score and risk of T2D. Results in women and men were pooled by using inverse variance weights under a fixed model because there was no heterogeneity. A restricted cubic spline regression model was used to test linear relation between the genetic score (as a continuous variable) and risk of T2D (25). This simple method can help to prevent the problems resulting from inappropriate linearity assumptions. We further examined the genetic association with risk of T2D according to joint classification of BMI-GRS and WHR-GRS in which both variables were classified into three categories (tertiles). All reported P values are nominal and two sided, and P ≤ 0.05 was considered statistically significant. The study had 80% power to detect an association with an OR of 1.03 for risk of T2D at a significance level of 0.05. Statistical analyses were performed in SAS version 9.3 software (SAS Institute Inc., Cary, NC).

Table 1 shows the baseline characteristics of participants of two nested case-control studies from the NHS (women) and HPFS (men). Participants with T2D had a significantly higher BMI and lower physical activity level and were more likely to smoke and have a family history of diabetes than participants without T2D. Female participants with T2D consumed less alcohol and were more likely to be postmenopausal than those without diabetes. In addition, the genetic score was not associated with age, BMI, or lifestyle factors, including smoking, alcohol intake, and physical activity (all P > 0.05). The mean genetic scores among men and women were 14.39 ± 2.4, and 14.53 ± 2.3, respectively. The range of genetic scores among men and women was 6.26–23.01 and 5.33–22.32, respectively. The genetic score was significantly associated with the WHR among men (0.005) and marginally related to the WHR among women (P = 0.06) (Fig. 1).

As shown in Table 2, the central obesity genetic score was significantly associated with an increased T2D risk in women (OR 1.03 [95% CI 1.00–1.06] per 1-point genetic score increase) and men (1.03 [1.00–1.07]). Multivariate adjustment for age, family history of diabetes, smoking, menopausal hormone therapy use (women only), physical activity, alcohol intake, and Healthy Eating Index showed a significant association among men (P = 0.02) but a borderline of significance among women (P = 0.08). The pooled OR for T2D was 1.03 (1.01–1.05) per 1-point genetic score increase, adjusting for age and BMI. The ORs for T2D increased across the quartiles of the genetic score (P for trend = 0.014). Compared with those in the lowest quartile of the genetic score, participants in the highest quartile had an OR of 1.22 (1.02–1.42). Multivariate adjustment for age, family history of diabetes, smoking, menopausal hormone therapy use (women only), physical activity, alcohol intake, and Healthy Eating Index did not change the association. Further adjustment of the BMI genetic score did not significantly change the results (P = 0.01), whereas the association was abolished after further adjustment for the WHR (P = 0.12).

We further examined whether the association between the genetic score and T2D risk varied across subgroups stratified by BMI and lifestyle risk factors for T2D (Supplementary Table 2). Although the associations appeared to be more pronounced in participants with a higher BMI and lower physical activity and who consumed modest levels of alcohol and currently smoked, no significant interaction between the genetic score and these risk factors in the combined samples of men and women was found (all P for interaction > 0.16). Results were similar in both sexes when analyses were performed in men and women separately (data not shown).

The central obesity genetic predisposition score showed a linear relationship with increasing T2D risk (P for linearity = 0.006 in the combined samples) (Supplementary Fig. 1).

We found that the GRS for BMI was significantly associated with risk of T2D in both the NHS (P_continuous = 0.01) and the HPFS (P_continuous = 0.02) cohorts (Fig. 2). We further examined the joint effects of BMI-GRS and WHR-GRS on risk of T2D (Fig. 3). Among individuals with the highest tertile of BMI-GRS, the risk of T2D was increased by 23%, 47%, and 46% within subgroups defined by increasing tertiles of WHR-GRS. Among individuals with the lowest tertile BMI-GRS, the OR for T2D was 0%, 0%, and 39% within subgroups defined by increasing tertiles of WHR-GRS.

In two well-established, prospective, nested case-control studies of U.S. women and men, we examined the association between a genetic score comprising 14 independent central obesity–associated variants and risk of T2D. The results indicate that the genetic predisposition to central obesity was significantly associated with an increased T2D risk independent of BMI, dietary, and lifestyle risk factors.

Consistent with our previous analyses (22,26), we estimated a genetic score to evaluate the overall susceptibility to central obesity based on 14 well-established WHR-predisposing variants identified from GWAS. The current study shows robust associations between the central obesity genetic score and risk of T2D in pooled results. Although the genetic association with T2D was weaker in women than in men, there was no significant sex difference. The association between GRS and WHR was also weaker in women than in men. Of note, the findings may partly support the potential causal relationship between central obesity and T2D risk. Because genetic variants are randomly assigned and generally uncorrelated with environmental factors, the observed association between the genetic score and T2D is free of risk for reverse causation and less likely to be affected by confounding (27–29). The current findings provide consistent evidence from two cohorts to show associations between the WHR genetic score and risk of T2D.

Several lines of evidence support the potential causal relationship. The visceral depots of fat most likely contribute to the creation of insulin resistance, with additional effects of elevated fatty acids from central fat depots (3). The higher visceral adipose tissue-to-subcutaneous adipose tissue ratio, a measure of relative body fat distribution, is associated with higher dyslipidemia, insulin resistance, and prevalence of diabetes independent of overall obesity and absolute visceral fat mass (4). Therefore, the waist circumference and WHR have been described as superior measures in predicting diabetes risk in the Diabetes Prevention Program (30).

The potential causal relation between central obesity and diabetes is also supported by evidence from randomized clinical trials. For example, an intervention study showed that twice-weekly progressive resistance training significantly decreases abdominal fat and improves insulin sensitivity and glycemia in older men with T2D (31). A shift of fat distribution from visceral to adipose depots after pioglitazone treatment has been associated with improvements in hepatic and peripheral tissue sensitivity to insulin (32).

In the stratified analysis by lifestyle risk factors for T2D, the associations of WHR genetic score and risk of T2D appeared to be more pronounced in participants who had low physical activity and modest alcohol intake and who currently smoked, although there was no significant interaction between the genetic score and these risk factors. These results suggest that lifestyle factors might modify the genetic association with risk of T2D. High physical activity and smoking cessation may attenuate the genetic susceptibility to T2D. The current findings are in line with a previous study showing that genetic risk for T2D modifies the overall protective effect of physical activity on T2D (33). Furthermore, the joint effects of BMI and WHR genetic scores on T2D suggest that a high BMI genetic score might accentuate the WHR genetic effect on risk of T2D. Of note, the current results highlight the importance in considering the gene-environment interaction when studying the risk factors for diabetes.

The major strengths of this study are the prospective design, high-quality genetic data, and minimal population stratification (11). Although the central obesity genetic score captured the combined information from most of the established genetic variants for the WHR, these variants only explained ∼4% variation of the WHR (7). This may explain the observed moderate effect of the genetic score on T2D risk.

We also acknowledge several limitations. First, although this nested case-control study was conducted in well-established prospective cohorts, some degree of measurement error in self-reported waist and hip circumferences and covariates is inevitable. However, the self-reported waist and hip circumferences were validated with high correlation, and we confirmed the association between genetic score and WHR in this study. Second, residual and unmeasured confounding from other lifestyle behaviors or factors is still possible. Third, although the meta-analysis of GWAS for WHR showed that most individual SNPs had a much stronger genetic association with WHR in women than in men, the study found a weak association of WHR genetic score with WHR and T2D among women. The reason for this discrepancy might be partly due to population variation, which was also observed in the original GWAS meta-analysis and warrants further investigation. Finally, the study was restricted to white participants. Therefore, further investigations in other ethnic groups are needed to validate the findings.

In conclusion, the findings indicate that the genetic predisposition to central adiposity, as estimated by the WHR genetic score, is associated with a higher risk of T2D among women and men from two prospective cohorts. This association is mediated by central obesity. The findings support a potential causal relationship between central obesity and T2D.

Acknowledgments. The authors thank all the participants of the NHS and the HPFS for their continued cooperation.

Funding. This study was supported by grants HL-126024, HL-034594, DK-100383, DK-091718, HL-071981, HL-073168, CA-87969, CA-49449, CA-055075, CA-186107, CA-167552, HL-34594, HL-088521, HL-35464, U01-HG-004399, DK-080140, P30-DK-46200, U01-CA-137088, U54-CA-155626, DK-58845, DK-098311, U01-HG-004728, EY-015473, CA-134958, DK-70756, and DK-46200 from the National Institutes of Health. L.Q. is a recipient of the American Heart Association Scientist Development Award (0730094N).

Duality of Interest. Additional support for genotyping was from Merck Research Laboratories (North Wales, PA). The genotyping of the HPFS and NHS coronary heart disease GWAS was supported by an unrestricted grant from Merck Research Laboratories. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. T.H. and Q.Q. contributed to study concept and design; data acquisition, analysis, and interpretation; statistical analysis; drafting of the manuscript; and critical revision of the manuscript. Y.Z. and S.H.L. contributed to study concept and design and critical revision of the manuscript. J.E.M. contributed to the data acquisition and critical revision of the manuscript. F.B.H. contributed to the study administration, material support, and supervision. L.Q. contributed to the study concept and design; study administration, material support, and supervision; data acquisition, analysis, and interpretation; drafting of the manuscript; and critical revision of the manuscript. L.Q. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.