The potential of antimicrobials to induce thrombocytopenia in critically ill patients: data from a randomized controlled trial

Abstract

Background: Antimicrobial-induced thrombocytopenia is frequently described in the literature among critically ill patients. Several antimicrobials have been implicated, although experimental evidence to demonstrate causality is limited. We report, using a randomized trial, the potential of antimicrobials to induce thrombocytopenia.

Methods: Randomized trial allocated patients to antimicrobial treatment according to standard- of-care (SOC group) or drug-escalation in case of procalcitonin increases (high-exposure group). Patients were followed until death or day 28. Thrombocytopenia defined as absolute (platelet count ≤ 100 x 109/L) or relative (≥ 20% decrease in platelet count). Analyses were performed in the two randomized groups and as a merged cohort.

Results: Of the 1147 patients with platelet data available, 18% had absolute thrombocytopenia within the first 24 hours after admission to intensive care unit and additional 17% developed this complication during follow-up; 57% developed relative thrombocytopenia during follow-up. Absolute and relative thrombocytopenia day 1-4 was associated with increased mortality (HR: 1.67 [95% CI: 1.30 to 2.14]; 1.71 [95% CI: 1.30 to 2.30], P<0.0001, respectively). Patients in the high-exposure group received more antimicrobials including piperacillin/tazobactam, meropenem and ciprofloxacin compared with the SOC group, whereas cefuroxime was used more frequently in the SOC group (p<0.05). Risk of absolute and relative thrombocytopenia (RR: 0.9 [0.7-1.3], p=0.7439; 1.2 [1.0-1.4], p=0.06; respectively), as well as absolute platelet count (daily difference, high-exposure vs. SOC -1.7 [-3.8-0.5], p=0.14) was comparable between groups. In observational analyses, use of ciprofloxacin and piperacillin/tazobactam predicted risk of relative thrombocytopenia (vs. cefuroxime, RR: 2.08 [1.48-2.92]; 1.44 [1.10-1.89], respectively), however only ciprofloxacin were associated with a reduction in absolute platelet count (p=0.0005).

Conclusion: High exposure to broad-spectrum antimicrobials does not result in a reduction in thrombocytopenia in critically ill patients. However, single use of ciprofloxacin, and less so piperacillin/tazobactam, may contribute to a lower platelet count.

Trial registration: ClinicalTrials.gov NCT00271752 http://clinicaltrials.gov/ct2/show/NCT00271752.

Figure 1. Use of frequent prescribed antimicrobials in the PASS study and occurrence of absolute and relative thrombocytopenia among the two randomized groups.

Unadjusted analysis displaying the use of antimicrobials and occurrence of thrombocytopenia among the two randomized groups displayed as relative rate ratio (RR) during 28 day follow-up. Absolute (one platelet count < 100 x 109/L) or relative (>=20 % decrease in platelet count from ICU admission) thrombocytopenia. RR-Ratio >1.0 indicates that the high-ex. group have a relatively higher risk of occurrence of the asses variable and RR-Ratio <1.0 indicates relatively higher risk of patients in the SOC group of occurrence of the asses variable. RR-ratio=0 indicates no difference between the two groups.

Figure 2. Estimated change in daily platelet count.

Mixed model adjusted for the following time fixed variables: randomisation group, age, gender, BMI, severe sepsis/septic shock at ICU admission, APACHE II score, surgical vs. medical patients. Time-updated use of antimicrobials was included in the model. Ciprofloxacin, Piperacillin/tazobactam (pip/tazo) (used alone or in combinations not including cefuroxime) and none (no antimicrobials) compared to people receiving cefuroxime (used alone or in combinations non including the antibiotic in question).