Issues regarding recommendations on empiric antimicrobial therapy for ventilator-associated pneumonia (VAP) have emerged in specific populations.To develop and validate a score to guide empiric therapy in brain-injured patients with VAP, we prospectively followed a cohort of 379 brain-injured patients in five intensive care units. The score was externally validated in an independent cohort of 252 brain-injured patients and its extrapolation was tested in 221 burn patients.The multivariate analysis for predicting resistance (incidence 16.4%) showed two independent factors: preceding antimicrobial therapy ≥48 h (p<0.001) and VAP onset ≥10 days (p<0.001); the area under the receiver operating characteristic curve (AUC) was 0.822 (95% CI 0.770-0.883) in the learning cohort and 0.805 (95% CI 0.732-0.877) in the validation cohort. The score built from the factors selected in multivariate analysis predicted resistance with a sensitivity of 83%, a specificity of 71%, a positive predictive value of 37% and a negative predictive value of 96% in the validation cohort. The AUC of the multivariate analysis was poor in burn patients (0.671, 95% CI 0.596-0.751).Limited-spectrum empirical antimicrobial therapy has low risk of failure in brain-injured patients presenting with VAP before day 10 and when prior antimicrobial therapy lasts <48 h.
Copyright ©ERS 2016.