Morris R D, Rimm D L, Hartz A J, Kalkhoff R K, Rimm A A
Division of Biostatistics and Clinical Epidemiology, Medical College of Wisconsin, Milwaukee 53226.
Am J Epidemiol. 1989 Jul;130(1):112-21. doi: 10.1093/oxfordjournals.aje.a115302.
Obesity and family history of non-insulin-dependent diabetes mellitus were examined in a cross-sectional study of 32,662 white women in the United States and Canada who were members of the TOPS (Take Off Pounds Sensibly) Club, Inc., with respect to the association between these two factors and the prevalence of diabetes by using questionnaire data originally collected in 1969. A family history index based upon the reported presence of diabetes among siblings, parents, and grandparents was used to grade hereditary factors on a scale of zero to five. The ratio of actual to ideal weight was used to categorize levels of obesity. The prevalence of diabetes increased consistently with increasing levels of either obesity or family history. The two factors appear to present independent risks for diabetes. Among women with no family history of diabetes, the odds ratio for diabetes associated with a weight 75 per cent or more above ideal in comparison with a weight less than 10 per cent above ideal was 7.2. Among women who were less than 10 per cent above ideal weight for their height, a comparison between women with a family history index of 1.5 or greater and women with no family history of the disease showed an odds ratio for diabetes of 6.0. Women with both risk factors (family history index of 1.5 or greater and relative weight of 1.75 or greater) had an overall odds ratio for diabetes of 22.8 when compared with women who had no family history of diabetes and a weight less than 10 per cent above ideal. Further analysis showed the combination of these factors to have a positive synergistic effect. The interaction follows a multiplicative pattern.
