Aims: To relate body mass index (BMI) in middle age to development of diabetes mellitus.
Methods: Participants were 6927 men and 8227 women from the Renfrew/Paisley general population study and 3993 men from the Collaborative occupational study. They were aged 45-64 years and did not have reported diabetes mellitus. Cases who developed diabetes mellitus, identified from acute hospital discharge data and from death certificates in the period from screening in 1970-1976 to 31 March 2004, were related to BMI at screening.
Results: Of Renfrew/Paisley study men 5.4%, 4.8% of women and 5% of Collaborative study men developed diabetes mellitus. Odds ratios for diabetes mellitus were higher in the overweight group (BMI 25 to < 30 kg/m(2)) than in the normal weight group (BMI 18.5 to < 25 kg/m(2)) and highest in the obese group (BMI >or= 30 kg/m(2)). Compared with the normal weight group, age-adjusted odds ratios for overweight and obese Renfrew/Paisley men were 2.73 [95% confidence interval (CI) 2.05, 3.64] and 7.26 (95% CI 5.26, 10.04), respectively. Further subdividing the normal, overweight and obese groups showed increasing odds ratios with increasing BMI, even at the higher normal level. Assuming a causal relation, around 60% of cases of diabetes could have been prevented if everyone had been of normal weight.
Conclusions: Overweight and obesity account for a major proportion of diabetes mellitus, as identified from hospital discharge and death records. With recent increases in the prevalence of overweight, the burden of disease related to diabetes mellitus is likely to increase markedly. Primordial prevention of obesity would be a major strategy for reducing the incidence of diabetes mellitus in populations.