Abstract P296: Genome Wide Association Study of Weight Change Among Post-Menopausal African Americans: The Women's Health Initiative Study

Genome wide association studies (GWAS) have identified numerous loci associated with BMI, a major risk factor for cardiometabolic disease. Yet, genetic associations with trajectory of weight change have been less studied, particularly in highly vulnerable populations such as racially diverse older women. We conducted a GWAS analysis on longitudinal weight among African American participants in the Women’s Health Initiative SNP Health Association Resource (WHI SHARe) study to better understand the genetic architecture of weight change during the post-menopausal period of the life course. Between 1993 and 2005, we collected longitudinal data on weight in 6,852 African American women between the ages of 50 and 79 at baseline. The average weight at first measure was 81 kg (16) with a range of 46 to 129 kg. The average change in weight per year was 0.09 kg (0.36) ranging from -3.3 to 5.7 kg. The average follow-up duration was 5.3 years (range: 1 to 11). We used time-varying, repeated weight measures across the trial period to perform a growth curve analysis for weight change. Using a mixed model with an unstructured covariance matrix, weight was regressed on year since randomization for both fixed and random effects to derive each individual’s weight change slope. Individuals were genotyped using Affymetrix 6.0 array and ~ 3.2 million SNPs were imputed using MACH (v1.0.16) and a 50(CEU):50(YRI) HapMap 2 reference panel. We performed association analyses with weight change residuals adjusted for clinical covariates and principal components, using an additive genetic model. While no associations reached genome wide significance (p<5E-8), several SNPs reached suggestive significance. Our three most significant signals at 5q22.1, 10q22.1, and 12q15 (all p<8E-7), lie near likely candidate genes involved in lipid breakdown, transport, and/or cholesterol homeostasis (e.g. STARD4, PSAP, IFNG). Interestingly, we observed a high level of phenotypic variance, which we attribute to complex weight fluctuation among post-menopausal women. We are currently conducting sensitivity analyses incorporating information on age since menopause and investigating different methods of longitudinal modeling in an attempt to minimize our trait variability, thereby improving our power to detect genetic effects. Also, we are replicating our top findings in other large longitudinal studies of African American women. While our preliminary analyses did not detect genome-wide significant associations with weight change among the participants of WHI SHARe, many SNPs with near genome wide significance are close to strong biological candidate genes and warrant further exploration. Future studies are needed to further characterize the genetic architecture of weight change at this vulnerable period of the life cycle for women so that effective interventions to prevent weight gain can be implemented.