ClusterVAP: study protocol for multicentre proteomic endotyping of ventilator-associated pneumonia

Area of Science:

• Critical Care Medicine
• Infectious Diseases
• Proteomics

Background:

• Ventilator-associated pneumonia (VAP) is a common intensive care unit infection with high mortality.
• Current diagnostic methods for VAP are often non-specific, leading to misdiagnosis and inappropriate antibiotic use.
• Identifying distinct patient subgroups (pneumoclusters) could refine VAP diagnosis and treatment.

Purpose of the Study:

• To define patient subgroups (pneumoclusters) in VAP using proteomic profiling.
• To characterize these pneumoclusters based on clinical, microbiological, and radiological data.
• To compare 30-day outcomes across identified clusters to understand clinical trajectories.

Main Methods:

• Exploratory, observational, prospective, multicentre study (ClusterVAP).
• Inclusion of mechanically ventilated patients with suspected lower respiratory tract infection.
• Proteomic profiling of bronchoalveolar lavage (BAL) supernatants using liquid chromatography tandem mass spectrometry.
• Unsupervised consensus clustering to define pneumoclusters.
• Analysis of clinical, microbiological, radiological data, and 30-day outcomes.

Main Results:

• Unsupervised consensus clustering successfully defined distinct pneumoclusters.
• Characterization of pneumoclusters revealed variations in clinical presentation, microbiology, and radiology.
• Comparative analysis of 30-day outcomes (mortality, ventilator-free days, etc.) showed significant differences across clusters.
• Candidate protein biomarkers were identified for pragmatic cluster assignment.

Conclusions:

• Proteomic profiling and clustering can identify distinct pathophysiological subgroups of VAP.
• This approach may enable more precise VAP diagnosis, differentiating true infections from mimics.
• Targeted therapy based on pneumoclusters holds potential for improved patient outcomes and antimicrobial stewardship.