Evaluation of plasma calprotectin as a marker for infection in various clinical settings: a prospective observational study

Background Reliably distinguishing infection from sterile inflammation is a major clinical challenge. Uncertainty can lead to unnecessary courses of antibiotics, fueling antimicrobial resistance and adverse effects. Calprotectin, a biomarker released by activated immune cells, may inform decision-making. Methods This prospective, observational, single-centre study recruited patients with suspected infection who provided blood samples on enrolment from the Emergency Department (ED) and Intensive Care Unit (ICU) of a central London university hospital. A separate longitudinal study with five days’ blood sampling was performed in patients undergoing elective major non-cardiac surgery, in whom infection was adjudicated according to the Standardised Endpoints in Perioperative Medicine (StEP) initiative. Diagnostic adjudication was performed blinded to calprotectin. The primary outcome was the ability of calprotectin to diagnose infection. Secondary outcomes included a comparison to C-reactive protein (CRP). Results 427 patients were included, of whom 186 (44%) were female. Of 245 ED patients, 71 (29%) had active cancer and 56 (23%) were on immunosuppressants. The median calprotectin level in the no-infection group was 1.97 mg/L (IQR 1.02–3.39), compared to 2.76 (IQR 1.65–5.08) mg/L in low-probability infection, 2.63 mg/L (IQR 1.83–5.23) in high-probability infection, and 2.64 mg/L (IQR 1.49–4.45) in patients with confirmed infection. Ordinal regression analysis found no meaningful association between calprotectin levels and infection, or bacterial infection. Logistic regression showed an unadjusted AUC of 0.53 (95%-CI 0.45–0.62) for calprotectin and a binary outcome of infection compared to an AUC of 0.63 (95%-CI 0.55–0.71) for CRP. Similar results were seen in a sensitivity analysis excluding patients with cancer or immunosuppression. In 98 ICU patients, neither calprotectin nor CRP showed a meaningful association with an adjudicated diagnosis of infection or ICU death. In the peri-operative cohort, calprotectin levels remained elevated over 5 days, but with no difference between patients developing or not developing infection. Conclusion Calprotectin showed only limited ability to differentiate infection from inflammation across ED, ICU, and elective surgery patients. Excluding patients with cancer or immunosuppression did not alter the overall findings.