Association of Glycemic Status During Progression of Chronic Kidney Disease With Early Dialysis Mortality in Patients With Diabetes

Abstract

Objective: Although early trials suggested that intensive glycemic targets reduce the number of complications with diabetes, contemporary trials indicate no cardiovascular benefit and potentially higher mortality risk. As patients with advanced chronic kidney disease (CKD) transitioning to treatment with dialysis were excluded from these studies, the optimal glycemic level in this population remains uncertain. We hypothesized that glycemic status, defined by hemoglobin A1c (HbA--1c) and random glucose levels, in the pre-end-stage renal disease (ESRD) period is associated with higher 1-year post-ESRD mortality among patients with incident diabetes who have ESRD.

Research design and methods: Among 17,819 U.S. veterans with diabetic CKD transitioning to dialysis from October 2007 to September 2011, we examined the association of mean HbA--1c and random glucose levels averaged over the 1-year pre-ESRD transition period with mortality in the first year after dialysis initiation. All-cause mortality hazard ratios (HRs) were estimated using multivariable survival models. Secondary analyses examined cardiovascular mortality using competing risks methods.

Results: HbA--1c levels ≥8% (≥64 mmol/mol) were associated with higher mortality in the first year after dialysis initiation (reference value 6% to <7% [42-53 mmol/mol]): adjusted HRs [aHRs] 1.19 [95% CI 1.07-1.32] and 1.48 (1.31-1.67) for HbA--1c 8% to <9% [64-75 mmol/mol] and ≥9% [≥75 mmol/mol], respectively). Random glucose levels ≥200 mg/dL were associated with higher mortality (reference value 100 to <125 mg/dL): aHR 1.34 [95% CI 1.20-1.49]). Cumulative incidence curves showed that incrementally higher mean HbA--1c and random glucose levels were associated with increasingly higher cardiovascular mortality.

Conclusions: In patients with diabetes and CKD transitioning to dialysis, higher mean HbA--1c and random glucose levels during the pre-ESRD prelude period were associated with higher 1-year post-ESRD mortality. Clinical trials are warranted to examine whether modulating glycemic status improves survival in this population.

© 2017 by the American Diabetes Association.

Figures

Figure 1

Association between HbA1c levels over the 1-year prelude period and 1-year all-cause mortality rate in the Overall Cohort (A) and Restricted Cohort (B). HbA1c levels of <5%, 5% to <6%, 6% to <7%, 7% to <8%, 8% to <9%, and ≥9%, which are equivalent to <31, 31 to <42, 42 to <53, 53 to <64, 64 to <75, and ≥75 mmol/mol, respectively. The minimally adjusted model includes the patient’s calendar quarter of dialysis initiation. The case-mix–adjusted model includes minimally adjusted covariates plus age, sex, race, ethnicity, cause of ESRD, CCI, diabetes, CHF, and CVD. The expanded case-mix model includes case-mix covariates plus residential region, initial dialysis modality, and BMI. The expanded case-mix plus laboratory model includes expanded case-mix covariates plus serum albumin, hemoglobin, serum bicarbonate, eGFR averaged over the 1-year prelude period, and the last eGFR level measured prior to dialysis initiation.

Figure 2

Cumulative incidence curves of the association between HbA1c over the 1-year prelude period and 1-year cardiovascular (CV) mortality in the Overall Cohort (A) and Restricted Cohort (B). Association between mean random glucose over the 1-year prelude period and 1-year cardiovascular mortality (C). HbA1c levels of <6%, 6% to <7%, 7% to <8%, and ≥8%, which are equivalent to <42, 42 to <53, 53 to <64, and ≥64 mmol/mol, respectively. Random glucose levels of <100, 100 to <150, 150 to <200, and ≥200 mg/dL. Analyses adjusted for case-mix covariates including the patient’s calendar quarter of dialysis initiation, age, sex, race, ethnicity, cause of ESRD, CCI, diabetes, CHF, and CVD.

Figure 3

Subgroup analyses of the association between HbA1c (A and B) and mean random glucose (C and D) over the 1-year prelude period and 1-year all-cause mortality rate adjusted for case-mix covariates. HbA1c levels of <5%, 5% to <6%, 6% to <7% (reference), 7% to <8%, 8% to <9%, and ≥9% (<31, 31 to <42, 42 to <53 [reference], 53 to <64, 64 to <75, and ≥75 mmol/mol, respectively) shown as filled circles in descending order. Glucose levels <100, 100 to <125 (reference), 125 to <150, 150 to <175, 175 to <200, and ≥200 mg/dL shown as filled circles in descending order. The case-mix model includes patient’s calendar quarter of dialysis initiation, age, sex, race, ethnicity, and cause of ESRD, CCI, diabetes, CHF, and CVD. OAD, oral antidiabetes drug.