Time-dependent impact of diabetes on the mortality of patients on renal replacement therapy: a population-based study in Germany (2002-2009).

AIMS: To estimate the impact of diabetes on the mortality of patients with incident renal replacement therapy (RRT).
METHODS: We assessed the mortality of 544 incident RRT patients aged ≥ 30 years between 2002 and 2009 (57.9% men, mean age 70.3 years, 49.6% patients with diabetes) by analyzing the data of all dialysis centers covering a German region. We compared the estimated time-dependent hazard ratios of patients with and without diabetes by using the Cox proportional-hazards regression model.
RESULTS: Overall, 319 patients had died (158 diabetic), approximately 50% after 3 years. Up to about 3 years, the mortality rate was lower in diabetic than in nondiabetic patients. Thereafter, the survival curves crossed (interaction diabetes × time, p = 0.002; adjusted hazard ratios for diabetes: baseline, 0.66; year 1, 0.84; year 2, 1.05; year 3, 1.33; year 4, 1.68). The results were similar in men and women; however, the interaction of diabetes and time was significant only in men (p = 0.004). Further significant risk factors of mortality were age, sex, initial central venous catheter, cardiovascular disease, and malignancy.
CONCLUSIONS: In this population-based study, the influence of diabetes was time-dependent, with a lower mortality in diabetic versus non-diabetic patients in the first three years but a higher mortality in these patients after 3 years. Results were similar in men and women.