#31-S interaction between childhood socioeconomic position, adult body mass index and 34-YEAR incidence of type 2 diabetes mellitus

Abstract

Abstract PURPOSE: To examine the relationship between childhood socioeconomic position (SEP), adult body mass index (BMI), and the cumulative incidence of Type 2 diabetes in 5,913 adults from Alameda County, CA. METHODS: Childhood SEP was assessed by the respondent's father's occupational category (white- or blue-collar) or years of education. Covariates included sociodemographics, height, blood pressure, depression, behavior, health care access, education, income and occupation. Cox proportional hazard models estimated the risk of incident diabetes in men and women from 1965-1999. Effect modification was assessed by stratified analyses and tests for statistical interaction between childhood SEP and adult BMI (obesity, overweight or normal weight). RESULTS: Low or moderate childhood SEP was associated with a higher incidence of Type 2 diabetes. The effect of childhood SEP was stronger for women, even after adjustment for all covariates [relative hazard: MEN: low childhood SEp = 1.12 (95%CI = 0.7–1.8) /moderate = 1.22 (95%CI = 0.8–1.8) and WOMEN: low childhood SEp = 1.55 (95%CI = 1.0–2.4) /moderate = 1.82 (95%CI = 1.3–2.6)]. Adult BMI modified the impact of childhood SEP on diabetes risk. Obese subjects with low/moderate childhood SEP were at greatest risk. The effect of BMI remained after all model adjustments. The effect of SEP was explained by covariate adjustment for men, but persisted for women [full model relative hazard: low/moderate childhood SEP-obese (RH = 4.38 (95%CI = 2.1–9.3) for men and 5.39 (95%CI = 2.8–10.3) for women), high SEP-obese (RH = 4.27 (95%CI = 1.7–10.7) for men and 3.13 (95%CI = 1.4–6.9) for women) and low/moderate SEP-normal weight (RH = 1.19 (95%CI = 0.7–2.1) for men and 1.46 (95%CI = 1.0–2.2) for women)]. CONCLUSION: Childhood SEP is a robust predictor of incident Type 2 diabetes, although the effect is stronger in women and not completely removed by adjustment for covariates. Multiplicative and additive models suggest a positive interaction between childhood SEP and adult BMI for both men and women.