Sepsis biomarkers in unselected patients on admission to intensive or high-dependency care

Abstract

Introduction: Although many sepsis biomarkers have shown promise in selected patient groups, only C-reactive protein and procalcitonin (PCT) have entered clinical practice. The aim of this study was to evaluate three promising novel sepsis biomarkers in unselected patients at admission to intensive care. We assessed the performance of pancreatic stone protein (PSP), soluble CD25 (sCD25) and heparin binding protein (HBP) in distinguishing patients with sepsis from those with a non-infective systemic inflammatory response and the ability of these markers to indicate severity of illness.

Methods: Plasma levels of the biomarkers, PCT and selected inflammatory cytokines were measured in samples taken from 219 patients during the first six hours of admission to intensive or high dependency care. Patients with a systemic inflammatory response were categorized as having sepsis or a non-infective aetiology, with or without markers of severity, using standard diagnostic criteria.

Results: Both PSP and sCD25 performed well as biomarkers of sepsis irrespective of severity of illness. For both markers the area under the receiver operating curve (AUC) was greater than 0.9; PSP 0.927 (0.887 to 0.968) and sCD25 0.902 (0.854 to 0.949). Procalcitonin and IL6 also performed well as markers of sepsis whilst in this intensive care unit (ICU) population, HBP did not: PCT 0.840 (0.778 to 0.901), IL6 0.805 (0.739 to 0.870) and HBP 0.607 (0.519 to 0.694). Levels of both PSP and PCT reflected severity of illness and both markers performed well in differentiating patients with severe sepsis from severely ill patients with a non-infective systemic inflammatory response: AUCs 0.955 (0.909 to 1) and 0.837 (0.732 to 0.941) respectively. Although levels of sCD25 did not correlate with severity, the addition of sCD25 to either PCT or PSP in a multivariate model improved the diagnostic accuracy of either marker alone.

Conclusions: PSP and sCD25 perform well as sepsis biomarkers in patients with suspected sepsis at the time of admission to intensive or high dependency care. These markers warrant further assessment of their prognostic value. Whereas previously published data indicate HBP has clinical utility in the emergency department, it did not perform well in an intensive-care population.

Figure 1

Biomarker levels in patients with SIRS at the time of admission to intensive/high dependency care. Boxes show medians and IQRs, whiskers show the 5^th and 95^th percentiles. Dotted horizontal lines show cut-offs defined by ROC analysis for distinguishing sepsis from non-infective SIRS. Brackets show statistically significant differences (independent samples T-test) as follows; * P < 0.001, **P < 0.01, *** P < 0.05).

Figure 2

Receiver Operating Characteristic curves for biomarkers. A) 162 patients with either sepsis or non-infective SIRS. B) 126 patients with dysfunction of one or more organ system (SOFA score ≥ 2 and either sepsis or non-infective SIRS. Data for GMCSF and TNFα are not shown as only a minority of patients had detectable levels of these markers.

Figure 3

Relationship between biomarker levels and severity of illness. Boxes show medians and IQRs, whiskers show 5^th and 95^th percentiles. Brackets show statistically significant differences in pairwise group comparisons (independent samples T-test) as follows; * P < 0.001, **P < 0.01, *** P < 0.05).