Background: In patients with traumatic brain injury (TBI), ventilator-associated pneumonia (VAP) is considered a dangerous complication, prompting early aggressive antibiotic treatment and prophylaxis. While this approach increases the selection of multidrug-resistant bacteria (MDR), its clinical benefit has not been demonstrated.
Methods: One-year incidence of VAP in severe TBI patients (ICU stay >48 hours, with either Glasgow Coma Scale ≤8 or receiving intracranial pressure monitoring, or having undergone emergency surgery) and the prevalence of MDR among those who eventually developed it, were compared in two Italian intensive care units (ICUs) adopting different antibiotic approaches. Antibiotic use was guideline-driven and aggressive in the Pisa-based unit (165 eligible patients), and very conservative and coupled with non-pharmacological prevention measures in Cesena (262 patients). Data were also compared with those of 208 Italian ICUs participating in the same infection surveillance program.
Results: Patient case mix and general care were similar in the two units. Overall antibiotic pressure was higher in Pisa (58.9% vs. 26.1% of beds occupied by patients receiving antibiotics, P<0.0001), as was antibiotic prophylaxis in eligible patients (87.3% vs. 7.6%, P<0.0001; Italian ICUs, 69.2%) and empirical therapy in those who developed VAP (60.8% vs. 25.2%, P<0.0001; Italian ICUs, 51.6%). The incidence rate of VAP did not significantly differ (39.8 per 1000 days of mechanical ventilation in Pisa, 49.3 in Cesena, P=0.16), although it occurred earlier in Cesena (23.0% early VAP in Pisa vs. 61.2% in Cesena, P<0.0001). Mortality was higher in Pisa but Cesena transferred more patients to other hospitals, precluding comparison of the two rates. The prevalence of MDR was higher in Pisa (38.2% vs. 9.9%, P<0.0001; Italian ICUs, 30.2%).
Conclusions: Although not conclusive, these results call into question the prevalent aggressive use of antibiotics in TBI patients and urge the scientific community to produce better evidence for clinical recommendations.