Biomarker kinetics in the prediction of VAP diagnosis: results from the BioVAP study

Background: Prediction of diagnosis of ventilator-associated pneumonia (VAP) remains difficult. Our aim was to assess the value of biomarker kinetics in VAP prediction. Methods: We performed a prospective, multicenter, observational study to evaluate predictive accuracy of biomarker kinetics, namely C-reactive protein (CRP), procalcitonin (PCT), mid-region fragment of pro-adrenomedullin (MR-proADM), for VAP management in 211 patients receiving mechanical ventilation for >72 h. For the present analysis, we assessed all (N = 138) mechanically ventilated patients without an infection at admission. The kinetics of each variable, from day 1 to day 6 of mechanical ventilation, was assessed with each variable’s slopes (rate of biomarker change per day), highest level and maximum amplitude of variation (Δ^max). Results: A total of 35 patients (25.4 %) developed a VAP and were compared with 70 non-infected controls (50.7 %). We excluded 33 patients (23.9 %) who developed a non-VAP nosocomial infection. Among the studied biomarkers, CRP and CRP ratio showed the best performance in VAP prediction. The slope of CRP change over time (adjusted odds ratio [aOR] 1.624, confidence interval [CI]_95% [1.206, 2.189], p = 0.001), the highest CRP ratio concentration (aOR 1.202, CI_95% [1.061, 1.363], p = 0.004) and Δ^max CRP (aOR 1.139, CI_95% [1.039, 1.248], p = 0.006), during the first 6 days of mechanical ventilation, were all significantly associated with VAP development. Both PCT and MR-proADM showed a poor predictive performance as well as temperature and white cell count. Conclusions: Our results suggest that in patients under mechanical ventilation, daily CRP monitoring was useful in VAP prediction. Trial registration NCT02078999