Predictors of type 2 diabetes and diabetes-related hospitalisation in an Australian Aboriginal cohort.

Predictors of diabetes and diabetes-related hospitalisations were examined in 15-88-year-old Aboriginal Australians (256 women, 258 men), surveyed in 1988-1989. Linkage to death records and hospitalisations to 2002 allowed proportional hazards or negative binomial modelling. Forty-five men (18%) and 59 women (24%) developed diabetes. Risk of diabetes was predicted positively by waist girth (hazard ratio (HR) 1.08, 95% CI 1.04, 1.13), smoking (HR 2.05, 95% CI 1.23, 3.39) and eating processed meats>4 times/month (HR 1.58, 95% CI 1.05, 2.40) and negatively by lower alcohol intake (HR 0.69, 95% CI 0.49, 0.99), preferring wine (HR 0.13, 95% CI 0.02, 0.97) and eating bush meats>4 times/month (HR 0.34, 95% CI 0.13, 0.90). Hospitalisation was predicted positively by smoking (Incidence rate ratio (IRR) 3.72, 95% CI 1.70, 8.18) and eating processed meats (IRR 1.03, 95% CI 1.01, 1.06), and negatively by exercise>or=once/week (IRR 0.23, 95% CI 0.08, 0.65), eating bush meats (IRR 0.95, 95% CI 0.91, 0.99) and trimming fat from meats (IRR 0.53, 95% CI 0.30, 0.94). Length of hospital stay was predicted positively by eating processed meats (HR 1.76, 95% CI 1.23, 2.53) and added salt (HR 1.52, 95% CI 1.02, 2.26) and negatively by lower alcohol intake (HR 0.90, 95% CI 0.40, 0.92) and exercise (HR 0.66, 95% CI 0.46, 0.95). Central obesity and adverse lifestyle increase risk for diabetes or related hospitalisation among Aboriginal Australians.