Glycemic control and the risk of death in 1,484 patients receiving maintenance hemodialysis

Abstract

Background: It is controversial whether tighter glycemic control is associated with better clinical outcomes in people with kidney failure. We aim to determine whether worse glycemic control, measured using serum glucose and hemoglobin A(1c) (HbA(1c)) levels, is independently associated with higher mortality in patients undergoing maintenance hemodialysis.

Study design: Retrospective cohort study.

Setting & participants: 1,484 patients starting maintenance hemodialysis therapy in Alberta, Canada, between 2001 and 2007.

Predictor: Serum glucose and HbA(1c) levels.

Outcome: All-cause mortality.

Measurements: Monthly casual glucose levels from specimens drawn immediately before the first dialysis treatment were averaged over 3 months before and after hemodialysis therapy initiation. Similarly, monthly HbA(1c) values in patients with or at risk of diabetes were averaged.

Results: Overall, median age was 66 years, 41% were women, 75% were white, and 55% had diabetes. All-cause mortality during 8 years (median, 1.5 years) was 43%; it was 49% in those with diabetes. There was no relation between average glucose level and mortality in unadjusted analysis (HR, 1.00 per 18 mg/dL [1 mmol/L]; P = 0.4) or after adjustment for confounders (HR, 0.98 per 18 mg/dL; 95% CI, 0.96-1.01; P = 0.2). Higher HbA(1c) level was not associated with mortality when analyzed in the unadjusted analysis (HR, 1.01 per 1% HbA(1c); P = 0.9) or after adjustment for confounders (HR, 0.98 per 1% HbA1c; 95% CI, 0.88-1.08; P = 0.7). Results were similar when HbA(1c) values were divided into prespecified categories (adjusted P > 0.6 for trend). Markers of malnutrition-inflammation (albumin, hemoglobin, and white blood cell values) or the presence of diabetes did not influence the relation between glycemic control and death (all P for interaction > 0.2).

Limitations: Registry data; casual serum glucose measurements; HbA(1c) values available for only a subset of participants.

Conclusions: Higher casual glucose and HbA(1c) levels were not associated with mortality in maintenance hemodialysis patients with or without diabetes. This may have implications for recommended glycemic targets, quality indicators, and how best to assess glycemic control in this high-risk population.