Tofu intake and diabetes incidence in Japanese‐American men

Abstract

The Honolulu Heart Program is a long‐term study of cardiovascular disease (CVD) among Japanese American men who were living on Oahu in 1965–68 (T1), when 8006 participated; in 1971–75 (T2), 5877 participated; in 1991–93 (T3), 3741 participated. The objective of this analysis was to evaluate whether soy intake was related to the incidence of Type 2 diabetes mellitus [DM]. All those with a diagnosis of cancer, CVD, stroke, or DM at T1 were eliminated. DM = being told by a physician that you have diabetes, taking an oral agent or insulin, and/or having a blood glucose level > 126 mg/dl or an oral glucose tolerance test > 200. Tofu intake was dichotomized as high [> twice/week] and low (<twice/wk). More had low tofu than high tofu at both T1 [low=2331, high= 523] and T2 [low=1200; high= 503]. Although high intake [n] is similar at T1 and T2, comparisons across time show few with consistent high intake: low/low=1622; low‐high=295; high‐low=248; high‐high=164. There was no significant effect of tofu intake change overall on DM incidence (p=.364). However, tofu intake T2 [Chi square 6.68; p=.0098, CI 1.067, 1.599] and BMI [Chi square 50.4, p=.0001; CI 1.098, 1.179] was associated with DM incidence. The pattern of low‐high tofu intake was associated with DM, whereas low‐low and high‐high was not. This may suggest an effect of increased food or calorie intake rather than tofu per se. Supported by the Illinois Soybean Association, Pacific Health Research Institute, by contract NHLBI‐N01‐HC‐05102, contract NIA‐N01‐AG‐4‐2149 and grants 5 U01 AG019349‐05, R01 AG027060‐01.