The prognostic value of the C-reactive protein levels in HD patients with death risk from infection

Abstract

Infection is considered the second leading cause of death in dialysis patients with end-stage renal disease (ESRD). However, infection prevalence as primary cause of death still seems to be underreported in the literature. We investigated the role of C-reactive protein (CRP) levels shortly before death as predictor of dying from an infection as primary cause of death in this patient group. Between January 1997 through March 2006, we defined the primary causes of death in 231 of the 481 incident patients in our single-center study, who died during this time and assessed the overall prevalence of infection at different predefined CRP cutpoints (between 2 and 300 mg/l). By means of an adjusted multiple logistic regression model, we calculated the odds ratio of (log) CRP for death in 346 survivors and non-survivors with available CRP levels within 5 days of death. In the 96 non-survivors (i.e. cases) of this group, the association of (log) CRP and causes of death was determined by the multiple linear regression model. Infection as a primary cause of death was initially diagnosed in 42% of the 231 non-survivors by standard parameters and clinically. However, the rate of patients possibly dying from this disease increased accordingly when also including cases without any clinical infection signs but with CRP values higher than a given cutpoint (between 2 and 300 mg/l), e.g. when including all cases with CRP cutpoints higher than 100 mg/l, overall prevalence of infection as cause of death increases to 57% (95% CI = 51-64%). Infection was significantly associated with higher CRP levels compared with cardiac death (p < 0.001), with an odds ratio of log CRP for death of 5.4 (95% CI = 3.8-7.7). Prevalence of infection as primary cause of death in ESRD patients may be even higher than currently stated in the literature. Therefore, to reduce mortality, infections should be further avoided and controlled in the future.