Ventilator-associated pneumonia (VAP) is the most common infection in critically ill patients. Initial antibiotic therapy is often broad spectrum, which promotes antibiotic resistance so new techniques are under investigation to obtain early microbiological identification and quantification. This trial compares the performance of a new real-time quantitative molecular-based method with conventional culture in patients with suspected VAP. Patients with suspected VAP who were ventilated for at least 48 h were eligible. An endotracheal aspirate (ETA) and a bronchoalveolar lavage (BAL) were performed at each suspected VAP episode. Both samples were analysed by conventional culture and molecular analysis. For the latter, bacterial DNA was extracted from each sample and real-time PCR were run. In all, 120 patients were finally included; 76% (91) were men; median age was 65 years, and clinical pulmonary infection score was ≥6 for 73.5% (86) of patients. A total of 120 BAL and 103 ETA could be processed and culture results above the agreed threshold were obtained for 75.0% (90/120) of BAL and 60.2% (62/103) of ETA. The main isolated bacteria were Staphylococcus aureus, Pseudomonas aeruginosa and Haemophilus influenzae. Performance was 89.2% (83.2%-93.6%) sensitivity and 97.1% (96.1%-97.9%) specificity for BAL samples and 71.8% (61.0%-81.0%) sensitivity and 96.6% (95.4%-97.5%) specificity for ETA samples when the molecular biology method was compared with conventional culture method (chosen as reference standard). This new molecular method can provide reliable quantitative microbiological data and is highly specific with good sensitivity for common pathogens involved in VAP.

Major concern for intubated patients is ventilator-associated pneumonia (VAP). Early detection of VAP and its causative microorganism(s) is a key challenge for clinicians. Diagnosis is based on clinical, radiological, and microbiological elements, the latter being provided 24-48h after sampling. According to practices, clinicians can sample endotracheal aspirates (ETAs) so as to check for patient colonization or perform ETA in case of VAP suspicion. In this proof-of-concept study, we report the evaluation of a semiautomated molecular method to rapidly quantify Staphylococcus aureus, one of the most involved microorganisms in VAP, directly from raw ETA samples. After evaluation using artificial ETA samples, our method was applied on 40 clinical ETA samples. All S. aureus-positive samples were successfully detected and quantified. Our method can provide an efficient sample preparation protocol for all raw ETA samples, combined with an accurate quantification of the bacterial load, in less than 3h 30min.

No abstract