Association of soluble endotoxin receptor CD14 and mortality among patients undergoing hemodialysis

Abstract

Background: CD14 is a key molecule in innate immunity that mediates cell activation and signaling in response to endotoxin and other bacterial wall-derived components. CD14 protein exists in soluble (sCD14) and membrane-bound forms. The correlates of sCD14 in persons undergoing long-term hemodialysis (HD) therapy are not known. We hypothesized that increased sCD14 levels in HD patients are associated with proinflammatory cytokine activation and increased mortality.

Study design: Cohort study.

Setting & participants: 310 long-term HD patients who participated in the Nutritional and Inflammatory Evaluation in Dialysis (NIED) Study, a cohort derived from a pool of more than 3,000 HD outpatients during 5 years in 8 DaVita maintenance dialysis facilities in the South Bay Los Angeles, CA, area.

Predictors: sCD14 levels in serum.

Outcomes: 33-month mortality.

Results: Mean sCD14 level was 7.24 +/- 2.45 microg/mL. Tumor necrosis factor alpha level was the strongest correlate of sCD14 level (r = +0.24; P < 0.001), followed by interleukin 6 level (r = +0.18; P = 0.002), serum ferritin level (r = +0.21; P < 0.001), total iron-binding capacity (r = -0.19; P < 0.001), body mass index (r = -0.15; P = 0.008), vintage (r = +0.14; P = 0.01), low-density lipoprotein cholesterol level (r = +0.13; P = 0.03), and body fat (r = -0.11; P = 0.06). During the 33-month follow-up, 71 (23%) patients died. Multivariable Cox proportional analysis adjusted for case-mix and other nutritional and inflammatory confounders, including serum tumor necrosis factor alpha, C-reactive protein, and interleukin 6 levels, showed that compared with the lowest sCD14 tertile, sCD14 levels in the third tertile (>7.8 microg/mL) were associated with greater death risk (hazard ratio, 1.94; 95% confidence interval, 1.01 to 3.75; P = 0.04).

Limitations: Survivor bias in combined incident/prevalent studies.

Conclusions: Increased sCD14 level is related positively to markers of inflammation and negatively to nutritional status and is an independent predictor of mortality in long-term HD patients. Additional studies are needed to examine the usefulness of sCD14 level in risk stratification and the clinical decision-making process in HD patients.

Figure 1

Scatter plots, regression line and 95% confidence intervals, reflecting the correlation between serum level of soluble endotoxine receptor CD14 and serum ferritin, total binding capacity, log for interleukin-6 (IL-6), and log for tumor necrosis factor alpha (TNF-alpha)

Figure 2

Mean (and SE) of soluble endotoxine receptor CD14 across categories of total body fat and body mass index.

Figure 3

Mortality predictability of soluble endoctoxin receptor CD14 in 310 maintenance hemodialysis patients (April 2004–Jan 2007)

1. Case-mix variables: age, sex, race/ethnicity, diabetes mellitus, log vintage

2. MICS variables: albumin, creatinine, hemoglobin, normalized protein catabolic rate (nPCR), lymphocyte percentage, and body mass index

3. Inflammatory variables: Log C-reactive protein, Log Interleukin-6, Log Tumor necrosis factor-α

Figure 4

Kaplan-Meier proportion of surviving after 2.8 years of observation according to the tertiles of soluble endoctoxin receptor CD14 in 310 HD patients (April 2004–Jan 2007)