Background: The association between obesity and dementia risk remains debatable and no studies have assessed this association among diabetic patients. The aim of our study was to investigate the association between body mass index (BMI) and dementia risk among middle and low income diabetic patients.
Methodology/principal findings: The sample included 44,660 diabetic patients (19,618 white and 25,042 African American) 30 to 96 years of age without a history of dementia in the Louisiana State University Hospital-Based Longitudinal Study. During a mean follow-up period of 3.9 years, 388 subjects developed incident dementia. The age- and sex-adjusted hazards ratios (HRs) for incident dementia at different levels of BMI (≤ 25, 25-26.9, 27-29.9, 30-34.9, and ≥ 35 kg/m(2)) were 1.00, 0.53 (95% CI 0.34-0.83), 0.29 (0.18-0.45), 0.37 (0.25-0.56), and 0.31 (0.21-0.48) (P(trend)<0.001) in white diabetic patients, and 1.00, 1.00 (95% CI 0.62-1.63), 0.62 (0.39-0.98), 0.56 (0.36-0.86), and 0.65 (0.43-1.01) (P(trend) = 0.029) in African American diabetic patients. Further adjustment for other confounding factors affected the results only slightly. There was a significant interaction between race and BMI on dementia risk (χ(2) = 5.52, 1df, p<0.025), such that the association was stronger in white patients. In stratified analyses, the multivariate-adjusted inverse association between BMI and risk of dementia was present in subjects aged 55-64 years, 65-74 years, and ≥ 75 years, in men and women, in non-smokers and smokers, and in subjects with different types of health insurance.
Conclusions/significance: Higher baseline BMI was associated with a lower risk of dementia among diabetic patients, and this association was stronger among white than among African American diabetic patients.