Association of Bisphenol A With Diabetes and Other Abnormalities—Reply

In Reply: Dr Wei and Ms Howard and Dr Howard comment on the observed association between BPA and diabetes in our study. Respondents to the NHANES 2003-2004 were asked, “Other than during pregnancy, have you ever been told by a doctor or health professional that you have diabetes or sugar diabetes?” Our analyses included 124 respondents replying yes plus 12 recorded as borderline. In fully adjusted models, geometric mean concentration of urinary BPA in the no group was 2.45 ng/mL (95% confidence interval [CI], 2.26-2.65 ng/mL); in the small borderline group, 5.97 ng/mL (95% CI, 4.01-8.88 ng/mL); and in the yes group, 2.86 ng/mL (95% CI, 2.42-3.37 ng/mL). Excluding the borderline response attenuated outcomes slightly in fully adjusted models: the odds ratio (OR) for responding yes vs no was 1.19 (95% CI, 1.00-1.41; P = .05) per 1-SD increase in BPA concentration. This equated to an OR of 2.38 (95% CI, 1.14-4.98; P for trend = .03) for a yes response in the upper vs lower 25% of BPA concentrations.