Early CRP changes track whether antibiotics work in bacteremia patients

When doctors suspect a bloodstream infection, they must choose antibiotics before laboratory results arrive, and choosing poorly can cost lives. A new retrospective study from Tel Aviv Sourasky Medical Center in Israel suggests that a familiar blood test, C-reactive protein (CRP), may offer early clues about whether the empiric antibiotic therapy a patient received was actually appropriate for the bacteria later identified in their blood.

The research, published in BMC Infectious Diseases, analyzed 122 adult patients admitted through the emergency department who had microbiologically confirmed bacteremia. To be included, patients needed an initial CRP level below 31.9 mg/L and a second CRP measurement taken within 24 hours of the first. The researchers defined CRP velocity as the change in CRP between the two measurements divided by the time elapsed, alongside the absolute second value and the raw difference between measurements.

Patients were grouped according to whether their antibiotic therapy matched the antimicrobial susceptibility testing results, as reviewed by infectious disease specialists. Eighty-two patients received appropriate therapy, while 22 received discordant therapy and 18 received no antibiotics between the two CRP measurements; these last two groups were combined into an inappropriate group of 40 for the primary analysis.

The key finding was that CRP dynamics differed markedly between the groups. The second CRP value, the change in CRP, and CRP velocity were all similar between the discordant and no-antibiotic groups, but significantly higher in the patients whose antibiotics were judged appropriate. In other words, patients on effective antibiotics tended to show a sharper early rise in CRP than those whose treatment did not match their infection.

The discriminative performance of these measures was modest but consistent. CRP velocity distinguished appropriate from discordant therapy with an area under the receiver operating characteristic curve of 0.74 (95% CI 0.63–0.85), essentially matching the second CRP value (0.74, 95% CI 0.62–0.85) and the raw CRP change (0.75, 95% CI 0.63–0.86). An AUC of 0.74 indicates discrimination better than chance but far from diagnostic certainty, so these values alone could not guide treatment decisions.

The authors conclude that early CRP dynamics following empiric antibiotic therapy were associated with antibiotic therapy appropriateness and may represent a potential adjunctive biomarker, while emphasizing that validation in larger prospective cohorts is required before any clinical use.

Several limitations deserve emphasis. The study was retrospective, conducted at a single center, and included a relatively small number of patients, particularly in the discordant and no-antibiotic groups. The inclusion criterion of an initial CRP below 31.9 mg/L means the findings may not generalize to patients presenting with already elevated CRP.

Inappropriate antibiotic therapy is a significant health concern, contributing to antibiotic resistance, unfavorable prognosis, and escalated hospitalization costs, which is why the search for early markers of treatment adequacy remains active. For now, the message is cautious: in this single-center retrospective cohort, the trajectory of CRP over the first 24 hours of empiric therapy was associated with whether that therapy ultimately proved appropriate, and larger prospective studies are needed to determine whether this signal is reliable enough to inform real-time clinical decisions.

Subject of Research: Association between early C-reactive protein dynamics and antibiotic therapy appropriateness in bacteremia

Article Title: Early C-reactive protein dynamics are associated with antibiotic therapy appropriateness in patients with bacteremia

Article References: Bernstein, D., Coster, D., Cherny, S., Levinson, T., Shenhar-Tsarfaty, S., Katz-Shalhav, M., Shapira, I., Zeltser, D., Goldiner, I., Berliner, S., Wasserman, A., & Obolski, U. (2026). Early C-reactive protein dynamics are associated with antibiotic therapy appropriateness in patients with bacteremia. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14574-y

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14574-y

Keywords: C-reactive protein, bacteremia, antibiotic therapy, CRP velocity, retrospective cohort study, antimicrobial susceptibility, biomarker, emergency department, antibiotic resistance, BMC Infectious Diseases, Early, C-reactive