To ascertain whether sleep-disordered breathing (SDB) in the elderly is associated with increased mortality, a prospective cohort study with 4-year follow-up was conducted at a retirement village complex in Sydney, Australia. The subjects were 163 non-demented retirement village residents. Logistic regression was used to assess SDB and co-morbidity as independent predictors of mortality. Respiratory disturbance index (RDI) was measured in the home; those subjects with RDIs > or = 15 were classified as having SDB. Co-morbidity was measured by an index of Burden of Illness based on the medical history obtained at baseline. At 4 years, 27% (4/15) of those subjects with RDIs > or = 15 and 22% (33/148) of those with RDIs < 15 were dead. RDI had an odds ratio (OR) of 1.00 (95% CI: 0.96, 1.04). Burden of Illness had an OR of 1.90 (95% CI: 1.34, 2.71). Adjustment for age and sex did not alter these findings. Significant predictors of mortality from the illness measure were a history of hypertension, Parkinson's disease and other severe illnesses (usually cancer). RDI was not a predictor of mortality in this population of non-demented seniors, where the prevalence of high levels of RDI was low.