Long-term mortality of critically ill patients with diabetes who survive admission to the intensive care unit

Abstract

Objective: Long-term outcomes of critically ill patients with diabetes are unknown. Our objectives were to evaluate the effect of diabetes on both long-term survival rates and the average number of years of life lost for patients admitted to an intensive care unit who survived to hospital discharge. Design and participants: A data linkage study evaluating all adult patients in South Australia between 2004 and 2011 who survived hospitalisation that required admission to a public hospital ICU. Main outcome measures: All patients were evaluated using hospital coding for diabetes, which was cross-referenced with registration with the Australian National Diabetes Services Scheme for a diagnosis of diabetes. This dataset was then linked to the Australian National Death Index. Longitudinal survival was assessed using Cox proportional hazards regression. Life-years lost were calculated using age- and sex-specific life-tables from the Australian Bureau of Statistics. Results: 5450 patients with diabetes and 17 023 patients without diabetes were included. Crude mortality rates were 105.5 per 1000 person-years (95% CI, 101.6-109.6 per 1000 person-years) for patients with diabetes, and 67.6 per 1000 person-years (95% CI, 65.9-69.3 per 1000 person-years) for patients without diabetes. Patients with diabetes were older and had higher illness severity scores on admission to the ICU, were more likely to die after hospital discharge (unadjusted hazard ratio [HR], 1.52 [95% CI, 1.45-1.59]; adjusted HR, 1.16 [95% CI, 1.10-1.21]; P < 0.0001) and suffered a greater number of average life-years lost. Conclusions: Our study indicates that crude mortality for ICU survivors with pre-existing diabetes is considerable after hospital discharge, and the risk of mortality is greater than for survivors without diabetes.