Abstract
BACKGROUND: Retinopathy is associated with increased mortality risk in general populations. We evaluated the joint effect of retinopathy and chronic kidney disease (CKD) on mortality in a representative sample of U.S. adults. STUDY DESIGN: Prospective cohort study. SETTING & PARTICIPANTS: 7,640 adults from NHANES (National Health and Nutrition Examination Survey) 1988-1994 with mortality linkage through December 31, 2006. PREDICTORS: CKD, defined as low estimated glomerular filtration rate (<60 mL/min/1.73 m2) or albuminuria (urine protein-creatinine ratio ≥30 mg/g), and retinopathy, defined as the presence of microaneurysms, hemorrhages, exudates, microvascular abnormalities, or other evidence of diabetic retinopathy by fundus photograph. OUTCOMES: All-cause and cardiovascular mortality. MEASUREMENTS: Multivariable-adjusted Cox proportional hazards. RESULTS: Overall, 4.6% of participants had retinopathy and 15% had CKD. Mean age was 56 years, 53% were women, and 81% were non-Hispanic whites. The prevalence of retinopathy in patients with CKD was 11%. We identified 2,634 deaths during 14.5 years' follow-up. In multivariable analyses, compared with individuals with neither CKD nor retinopathy, HRs for all-cause mortality were 1.02 (95% CI, 0.75-1.38), 1.52 (95% CI, 1.35-1.72), and 2.39 (95% CI, 1.77-3.22) for individuals with retinopathy only, those with CKD only, and those with both CKD and retinopathy, respectively. Corresponding HRs for cardiovascular mortality were 0.96 (95% CI, 0.50-1.84), 1.72 (95% CI, 1.47-2.00), and 2.96 (95% CI, 2.11-4.15), respectively. There was a significant synergistic interaction between retinopathy and CKD on all-cause mortality (P=0.04). LIMITATIONS: The presence of retinopathy was evaluated only once. Small sample size of some of the subpopulations studied. CONCLUSIONS: In the presence of CKD, retinopathy is a strong predictor of mortality in this adult population.