Background: C-reactive protein (CRP) is increasingly being used as an inflammatory marker in sepsis. Its main use is for diagnosis, less for prognosis, while it is increasingly used for serial monitoring of response to treatment - with little evidence to support this practice.
Objectives: Assessment of the effectiveness of serial measurements of CRP in the management of patients with blood stream infection (BSI).
Methods: This was a retrospective study of two patient cohorts, one in 2015 (cohort A) and the other in 2018 (cohort B), each included 200 consecutively diagnosed patients with proven BSI. We assessed the following outcome markers: antibiotic management, duration of admission and in-hospital mortality.
Results: Baseline demographic and clinical data of the cohorts were similar. In cohort A, a single CRP test was obtained from 5% of the patients, while in cohort B, 95% of patients had CRP tests with a mean of 5.7 tests/person. Empiric, appropriate antibiotic treatment, increased from 72% in cohort A to 75% in cohort B (NS). Duration of antibiotic treatment did not change (10±8 days). The in-hospital mortality rate decreased from 38.5% in cohort A to 30.5% in cohort B (NS), as did mortality <7 days after diagnosis of BSI, from 17.5% to 14% respectively (NS).
Conclusions: We found statistically insignificant differences between the cohorts. In our view there is currently no solid evidence to support the serial use of CRP tests in the management of patients with BSI.