Use of C-reactive protein to guide the antibiotic therapy in hospitalized patients: a systematic review and meta-analysis

Abstract

Abstract Background C-reactive protein (CRP) is an inflammatory protein used in clinical practice to identify and monitor inflammatory and infectious processes. Recent data suggest CRP might be useful in guiding antibiotic therapy discontinuation among critical care patients. This meta-analysis analyzed the benefits and risks of CRP-guided protocols to guide antibiotic therapy in hospitalized patients in comparison with standard treatment. Methods Studies were searched in four databases: CENTRAL, Medline, Embase and LILACS. The search was performed until Jan 25th, 2023. The reference lists of the articles retrieved and related review studies were hand-screened to find eligible trials that might have been missed. Primary endpoints included the duration of antibiotic therapy for the index episode of infection. The secondary endpoint was the all-cause hospital mortality and infection relapses. The risk of bias was evaluated using the Cochrane Risk of Bias 2.0 tool. Random effects were used to pool the mean differences and odds ratio of individual studies. The protocol was registered in PROSPERO (CRD42021259977). Results The search strategy retrieved 5209 titles, out of which three studies met the eligibility criteria and were included in this meta-analysis. 727 adult patients were analyzed, of whom 278 were included in the intervention group and 449 were included in the control group. 55,7% of all patients were women. Meta-analysis indicated that experimental groups (CRP-guided) had a lower duration of antibiotic therapy (days) [MMD = -1.82, 95%IC -3.23; -0.40]; with no difference in mortality [OR = 1.19 95%IC 0.67–2.12] or in the occurrence of infection relapse [OR = 3.21 95%IC 0.85–12.05]. Conclusion The use of CRP-guided protocol reduces the total amount of time required for antibiotic therapy when compared to standard protocols of treatment in hospitalized patients with acute bacterial infection. We did not observe statistical differences regarding mortality and infection relapse rates.