Risk factors for mortality in diabetic peritoneal dialysis patients

It is well established that the survival rate of diabetic end-stage renal disease patients remains the lowest among all primary diagnoses probably because of higher prevalence of cardiovascular diseases (CVD) associated with diabetes. This study was designed to evaluate the impact of CVD and other risk factors individually or in combination on mortality in diabetic peritoneal dialysis (PD) patients.

In a retrospective study, 213 incident PD patients [118 had diabetes mellitus (DM), 94 were female, mean age 55 ± 13 years] underwent initial assessment of nutritional status, comorbid disease (CMD) survey, residual renal function (RRF), dialysis adequacy and peritoneal transport characteristics at a mean of 9 days (range, 3-24 days) after start of PD and were then followed for 30 ± 24 months (range, 3-115 months). Of 213 patients, 154 patients were reassessed after a mean of 11 months (range, 6-19 months). Nutritional status was assessed by subjective global assessment and other methods. CMD was graded by Davies index and included DM, CVD, liver disease and respiratory disease.

On Kaplan-Meier analysis, patient survival was significantly lower in female DM patients compared to other groups. The 3-year patient survival rate was 46, 70, 82 and 83% for female DM, male DM, male non-DM and female non-DM, respectively (P = 0.003). On Cox proportional hazards multivariate analysis including all patients, old age, presence of CVD or protein-energy wasting (PEW), low serum albumin concentration and low RRF were independent predictors of mortality but not DM per se or female gender. In DM patients, old age, female gender, presence of CVD or PEW and low RRF were independent predictors of mortality while old age was the only risk factor in non-DM patients. After adjustment for age, gender and RRF, DM patients with both CVD and PEW had a risk of mortality that was 3.3 times that of DM patients without CVD and PEW. In DM patients without CVD and PEW, patient survival was not different from that of non-DM patients without CVD and PEW.

DM per se was not a risk factor for mortality in this group of PD patients. Instead, the higher mortality rate in diabetic PD patients, in particular among female patients, was mainly attributable to concurrent morbidity such as CVD and PEW, together with low RRF.