Interleukin 6, lipopolysaccharide-binding protein and interleukin 10 in the prediction of risk and etiologic patterns in patients with community-acquired pneumonia: results from the German competence network CAPNETZ

Abstract

Background: The aim of our study was to investigate the predictive value of the biomarkers interleukin 6 (IL-6), interleukin 10 (IL-10) and lipopolysaccharide-binding protein (LBP) compared with clinical CRB and CRB-65 severity scores in patients with community-acquired pneumonia (CAP).

Methods: Samples and data were obtained from patients enrolled into the German CAPNETZ study group. Samples (blood, sputum and urine) were collected within 24 h of first presentation and inclusion in the CAPNETZ study, and CRB and CRB-65 scores were determined for all patients at the time of enrollment. The combined end point representative of a severe course of CAP was defined as mechanical ventilation, intensive care unit treatment and/or death within 30 days. Overall, a total of 1,000 patients were enrolled in the study. A severe course of CAP was observed in 105 (10.5%) patients.

Results: The highest IL-6, IL-10 and LBP concentrations were found in patients with CRB-65 scores of 3-4 or CRB scores of 2-3. IL-6 and LBP levels on enrollment in the study were significantly higher for patients with a severe course of CAP than for those who did not have severe CAP. In receiver operating characteristic analyses, the area under the curve values for of IL-6 (0.689), IL-10 (0.665) and LPB (0.624) in a severe course of CAP were lower than that of CRB-65 (0.764) and similar to that of CRB (0.69). The accuracy of both CRB and CRB-65 was increased significantly by including IL-6 measurements. In addition, higher cytokine concentrations were found in patients with typical bacterial infections compared with patients with atypical or viral infections and those with infection of unknown etiology. LBP showed the highest discriminatory power with respect to the etiology of infection.

Conclusions: IL-6, IL-10 and LBP concentrations were increased in patients with a CRB-65 score of 3-4 and a severe course of CAP. The concentrations of IL-6 and IL-10 reflected the severity of disease in patients with CAP. The predictive power of IL-6, IL-10 and LBP for a severe course of pneumonia was lower than that of CRB-65. Typical bacterial pathogens induced the highest LBP, IL-6 and IL-10 concentrations.

Figure 1

Levels of cytokines in the course of community-acquired pneumonia. (A) Interleukin 6 (IL-6), (B) lipopolysaccharide-binding protein (LBP), and (C) interleukin 10 (IL-10). Boxes: horizontal lines indicate 25th to 75th percentiles, and whiskers indicate median, 2.5th and 97.5th percentiles. CAP, community-acquired pneumonia.

Figure 2

Levels of cytokines and CRB-65 class on patient admission in community-acquired pneumonia. (A) Interleukin 6 (IL-6), (B) lipopolysaccharide-binding protein (LBP), and (C) interleukin 10 (IL-10). Continuous lines denote median values, boxes represent 25th to 75th percentiles and whiskers indicate 2.5th and 97.5th percentiles. CAP, community-acquired pneumonia.

Figure 3

Receiver-operating characteristic curves to predict a severe course of community-acquired pneumonia and high CRB-65 or CRB scores. (A) Receiver operator characteristic (ROC) curves for interleukin 6 (IL-6; area under curve (AUC) 0.689; 95% CI 0.636 to 0.741; P < 0.001), lipopolysaccharide-binding protein (LBP; AUC 0.624; 95% CI 0.567 to 0.681; P < 0.0001), interleukin 10 (IL-10; AUC 0.665; CI 0.611 to 0.718; P < 0.0001), CRB score (AUC 0.701; 95% CI 0.643 to 0.760; P < 0.0001) and CRB-65 score (AUC 0.764; 95% CI 0.716 to 0.812; P < 0.0001), and (B) combination of CRB-65 score plus IL-6 concentration (AUC 0.800; 95% CI 0.759 to 0.842; P < 0.0001) and CRB score plus IL-6 concentration (AUC 0.765; 95% CI 0.720 to 0.810; P < 0.0001), in predicting a severe course of community-acquired pneumonia (CAP).

Figure 4

Levels of cytokines on admission in patients with community-acquired pneumonia, and microbial etiology. Microbial etiology was infection with typical bacteria, atypical bacteria, viruses or mixed organisms, or unknown etiology. The scatter plots represent all data. Median values with interquartile ranges are shown.