High-sensitivity Glasgow prognostic score and infection-related mortality in maintenance hemodialysis patients: 10-year outcomes of the Q-Cohort Study.

Abstract

Background: Infection is a major cause of death in patients undergoing maintenance hemodialysis (HD). We examined whether the high-sensitivity Glasgow prognostic score (hsGPS), derived from serum C-reactive protein (CRP) and albumin, is associated with infection-related mortality.

Methods: This retrospective cohort study included 3529 patients undergoing maintenance HD enrolled between December 31, 2006 and December 31, 2007 and followed until December 31, 2016. We used a high-sensitivity Glasgow prognostic score (hsGPS), defined using a lower CRP threshold in a GPS-type manner as follows: score 0, CRP ≤ 0.3 mg/dL and albumin ≥ 3.5 g/dL; score 1, CRP > 0.3 mg/dL or albumin < 3.5 g/dL; and score 2, CRP > 0.3 mg/dL and albumin < 3.5 g/dL. Associations were evaluated using Kaplan-Meier analysis, Cox proportional hazards models, Fine-Gray competing-risk models, and restricted cubic spline analyses.

Results: During a median follow-up of 3199 days, 1748 deaths occurred, including 451 infection-related deaths. In descriptive Kaplan-Meier analysis, infection-related death-free survival was lower in higher hsGPS groups. In the multivariable-adjusted Cox model, hazard ratios for infection-related mortality were 1.79 (95% confidence interval [CI] 1.43-2.24) for hsGPS 1 and 2.38 (95% CI 1.71-3.31) for hsGPS 2, compared with hsGPS 0. Competing-risk analyses showed similar trends, although the associations were attenuated. Restricted cubic spline analyses showed higher risk at lower albumin levels and higher CRP levels.

Conclusions: Higher hsGPS was associated with higher infection-related mortality in patients undergoing maintenance HD. hsGPS may serve as a simple risk stratification tool in this population.