The utility of erythrocyte sedimentation rate, C-reactive protein, and procalcitonin in detecting infections in patients with systemic lupus erythematosus: A systematic review

Abstract

Objectives We conducted a systematic review with metanalysis to investigate the utility of erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and procalcitonin (PCT) in diagnosing infections in hospitalized patients with SLE. Methods We searched Medline, Embase, Web of Science, ClinicalTrials.gov , and Cochrane Central Register of Controlled Trials (CENTRAL) with a search strategy developed by a medical librarian. We included retrospective, cross-sectional, case-control, and prospective studies in our analysis. We used the Quality Assessment of Diagnostic Studies (QUADAS-2) to assess for bias and applicability. We obtained mean differences, sensitivities, and specificities in our analysis. Results We included 26 studies in our analysis. Most studies had an unclear or high risk of bias and our results were widely heterogenous. For the diagnosis of infections, the CRP had a pooled sensitivity of 0.75 (95%CI 0.57-0.94) and specificity of 0.72 (0.59-0.85), PCT had a pooled sensitivity of 0.68 (95% CI 0.0.59–0.77) and specificity of 0.75 (0.59–0.90), and for ESR pooled estimates were not calculated but sensitivity ranged from 50 to 69.8 and specificity from 38.5 to 55.6. Modifying cut-offs improved sensitivities and specificities. The ESR, CRP, and PCT mean differences were all greater in infection groups versus non-infection (10.1, 95% CI 3.2–17.0; 46.8, 95% CI 36.5–57.0; 0.53, 95% CI 0.26–0.80; respectively). Discussion Poor sensitivities and specificities were observed for the evaluated biomarkers with substantial heterogeneity in the cut-offs used to determine infection. Although mean biomarker values were increased in the infection group compared with the non-infection, our findings do not support the widespread use of ESR, CRP, or PCT in diagnosing infection in hospitalized patients with SLE due to increased heterogeneity and risk of bias. Further investigation is needed.