IntroductionThe greater predisposition to infections, as well as the possibility of a worse response to treatment, can lead to the excessive use of antimicrobials among cancer patients. C-reactive protein (CRP) has gained prominence as a tool for monitoring therapeutic responses and reducing the duration of antibiotic therapy; however, few studies have analyzed this protein in cancer patient populations. We hypothesize that cancer patients with a good response to antibiotic therapy show a faster decline in serum CRP levels, which would allow us to identify candidates for short-course treatments.ObjectiveTo evaluate the behavior of serum CRP levels among adult cancer patients using antibiotic therapy, and its association with the duration of this treatment, therapeutic response, and clinical recurrence.MethodsThis work consisted of a retrospective study with cancer patients admitted to a university hospital between September 2018 and December 2019. Adults (age ≥ 18 years) who underwent at least one course of antibiotic therapy were included. CRP behavior over the first 7 days of treatment was classified as: i) good response: when the CRP value on the fifth day of therapy reached 50% or less of the peak value detected in the first 48 h of treatment, and ii) poor response: Maintenance, within the same interval, of a CRP value > 50% of the peak value in the first 48 h. The duration of antibiotic therapy was categorized as up to seven full days or more. Outcomes were assessed by events that occurred during the 30 days of hospitalization or until hospital discharge. Primary outcome: Clinical recurrence of the index infection. Secondary outcomes: i) Death from any cause; ii) microbiological recurrence; iii) therapeutic response; iv) colitis associated with Clostridioides difficile; and v) isolation of multi-resistant bacteria, whether in clinical or surveillance samples.ResultsThe final analysis consisted of 212 patients, with a median age (IQ) of 59.2 (48 – 67) years old and a predominance of females (65%), who were hypertensive (35%), smokers (21%), and diabetics (17.8%). There was no difference in clinical recurrence between the two groups (8.1% vs. 12.2%; p = 0.364), with a lower 30-day mortality in the good CRP response group (32.2% vs. 14.5%; p = 0.002). Despite the tendency towards a lower occurrence of other secondary outcomes in the good response group, these differences were not statistically significant. In the poor CRP response group, outcomes like clinical recurrence, mortality, and therapeutic response were significantly worse, regardless of the duration of antibiotic treatment.ConclusionIn this study, cancer patients with a good CRP response during antibiotic therapy presented lower mortality and a higher proportion of satisfactory therapeutic responses. CRP can be a useful tool when combined with other clinical information in optimizing the duration of antimicrobial treatment in a hospitalized cancer population.