RESUMO
BACKGROUND: Clinical approach to ventilator-associated pneumonia (VAP) in the neurocritical care unit (NCCU) varies widely among physicians despite training and validated criteria. METHODS: Prospective observational study of all mechanically ventilated patients with suspected VAP over 18 months in an academic NCCU. Patients meeting VAP criteria by a surveillance program (SurvVAP) were compared to treated patients who did not meet surveillance criteria (ClinVAPonly). We identified appropriate/potentially inappropriate antibiotic treatment and factors associated with excessive antibiotic days (EAD). RESULTS: Of 622 ventilated patients, 83 cases were treated as VAP. Of these, 26 (31.3 %) had VAP by CDC criteria (SurvVAP) (VAP rate = 7.3 cases/1,000 ventilator days). Clinical features significantly more prevalent in SurvVAP cases (vs. ClinVAPonly) were change in sputum character, tachypnea, oxygen desaturation, persistent infiltrate on chest X-ray and higher clinical pulmonary infection score, but not positive sputum culture. Treatment with pneumonia-targeted antibiotics for >8 days was significantly more common in ClinVAPonly versus SurvVAP patients (73.7 vs. 30.8 %, p < 0.001) even after excluding patients with other infections (p = 0.001). Based on current guidelines, the ClinVAPonly group contributed 225 EAD, including 38 vancomycin days, 70 piperacillin-tazobactam days and 85 cephalosporin days with cost figure over four times that of EAD in SurvVAP group. No pre-specified factors were associated with continued VAP treatment beyond 8 days. CONCLUSIONS: Incongruency between clinically and surveillance-defined VAP is common in acute neurological disease although outcomes did not differ between groups. Clinician behaviors rather than clinical factors may contribute to prolonged prescribing.