Comparison of serial pancreatic stone protein, C-reactive protein and procalcitonin for the diagnosis of infection and sepsis in critically Ill patients: a multicentre prospective study.

Abstract

Background: The serial performance of C-reactive protein (CRP), procalcitonin, and emerging biomarker pancreatic stone protein (PSP) was evaluated for the diagnosis of infection and sepsis in patients admitted to the intensive care unit (ICU).

Methods: All consecutive adult patients with suspected infection or sepsis upon their admission to the ICUs of three multi-speciality hospitals in the UAE were enrolled. CRP, procalcitonin, and PSP levels were measured at admission and repeated within 24-48 h. Patients were categorized into infection vs. non-infection, sepsis vs. non-sepsis groups, and into culture-positive and culture-negative subgroups.

Results: A total of 272 ICU patients were analyzed. All biomarkers could be used to distinguish infection with CRP (AUROC 0.77; 95% confidence intervals [CI] 0.70-0.83) and procalcitonin (AUROC 0.75; 95% CI 0.68-0.81) showing fair performance. Moreover, serial monitoring at 24-48 h improved performance, especially for procalcitonin (p = 0.04). Among patients with infection, PSP levels were higher in culture-positive compared to culture-negative patients, but the difference did not reach statistical significance (median 229 vs. 142 ng/ml, p = 0.05). However, all three biomarkers failed to discriminate sepsis with an AUROC of 0.56 (95% CI 0.49-0.64) for CRP, 0.54 (95% CI 0.46-0.62) for procalcitonin, and 0.58 (95% CI 0.50-0.66) for PSP, respectively. Combining biomarkers improved specificity (93.85%) but with reduced accuracy.

Conclusion: Despite a significant rise in all biomarkers, procalcitonin has overall better performance for diagnosing infections. None of the biomarkers could differentiate sepsis at admission.