Abstract P087: Body Mass Index and Change in Obesity Status over 25 Years: The Relationship within Spouse Pairs in the Atherosclerosis Risk in Communities Study

Abstract

Introduction: Environmental causes have been a focus of efforts to address the US obesity epidemic. Spouses share a similar environment and should therefore experience similar changes in body mass index (BMI) and obesity status over time. Existing longitudinal studies of spouses have not examined whether associations of BMI change within pairs are similar by gender.

Methods: We analyzed spouse pairs from the community-based ARIC cohort, recruited from 1986 to 1989 at ages 45-64. We used linear mixed models to assess the association of changes in spousal BMI with changes in individual BMI assessed at 5 clinic visits during 25 years of follow-up. We used survival models accounting for interval censoring to evaluate the association between one spouse becoming obese (BMI>30 kg/m2) and the other becoming obese. Analyses controlled for time-varying demographic, lifestyle, and clinical factors, pre-existing illness, and spousal CVD risk factors and illness. We assessed similarity of the results by gender.

Results: We included 4,003 spouse pairs with information on BMI at more than one visit. Mean BMI in men at baseline was 27.5 kg/m2, increasing to 28.6 in 2011-13. For women, baseline BMI was 27.1 kg/m2, increasing to 28.1 in 2011-13. Baseline BMI correlation within spouses was 0.18 (p <0.001). Among women, mean annual BMI change increased by 0.17 kg/m2 per unit increase in spousal annual BMI change (95% CI: 0.14 - 0.19); this association was lower in men: 0.11 kg/m2 (0.09 - 0.13) (Figure). Non-obese women whose husbands became obese were more likely to become obese themselves compared to women whose husbands stayed non-obese (HR 1.52, 1.09- 2.11). Results for men were similar (HR 1.59, 1.14-2.21).

Conclusions: Our results suggest that changes in BMI and obesity status in spouse pairs are related, with change in men’s BMI having a somewhat stronger influence on their wives’ BMI. This finding was robust to adjustment for individual and spousal characteristics and suggests weight control strategies might benefit from addressing both spouses.