Glycemic Status and Mortality in Chronic Kidney Disease According to Transition Versus Nontransition to Dialysis.

OBJECTIVE:The impact of glycemic control in diabetic patients with chronic kidney disease (CKD) who may or may not transition to dialysis remains uncertain, given recent interest in the conservative management of advanced CKD without dialysis therapy, which may benefit from alternative glycemic control strategies. DESIGN AND METHODS:Among a national cohort of US Veterans, we examined the association of glycemic status, defined by averaged random blood glucose and hemoglobin A1c (HbA1c), with mortality after transitioning to dialysis over 2007-2011 (Transition Cohort) compared with patients in a one-to-one matched cohort of CKD patients with diabetes who did not transition to dialysis (Nontransition Cohort). RESULTS:Among 17,121 patients in the Transition Cohort, averaged random glucose ≥200 mg/dL was associated with higher mortality in expanded case-mix analyses (reference: 100-<120 mg/dL): adjusted hazard ratio (95% confidence interval) 1.26 (1.13-1.40). In the transition cohort, HbA1c 8-<10% and ≥10% were associated with higher mortality (reference: 6-<8%): adjusted hazard ratios (95% confidence interval) 1.21 (1.11-1.33) and 1.43 (1.21-1.69), respectively. Among 8,711 patients in the Nontransition Cohort, averaged random glucose <100 mg/dl and ≥160 mg/dl were associated with higher death risk, whereas HbA1c was not associated with mortality. CONCLUSION:In diabetic CKD patients transitioning to dialysis, higher averaged random glucose and HbA1c were associated with early dialysis mortality, whereas in matched CKD patients who did not transition, both lower and higher glucose levels were associated with higher mortality. These data suggest the need for different glycemic strategies based on whether there are plans to transition to dialysis versus pursue conservative management among diabetic patients with CKD.