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

## TL;DR

This study aims to improve the diagnosis of ventilator-associated pneumonia by identifying subgroups of patients with similar biological profiles to guide better treatment and reduce unnecessary antibiotics.

## Contribution

The study introduces a novel approach to classify VAP using proteomic profiling and unsupervised clustering to define distinct pneumoclusters.

## Key findings

- Proteomic profiling will be used to identify distinct pneumoclusters in VAP patients.
- Candidate protein biomarkers will be derived to facilitate pragmatic cluster assignment.
- 30-day clinical outcomes will be compared across clusters to describe different clinical trajectories.

## Abstract

Ventilator-associated pneumonia (VAP) is the most frequent healthcare-associated infection in intensive care units and is associated with high morbidity and mortality. Current diagnostic criteria lack specificity, leading to misclassification and unnecessary antibiotic use. Identifying patient subgroups with a common pathophysiological basis (pneumoclusters) may distinguish true VAP of varying aetiology and severity from non-infectious mimics, enabling more targeted therapy and improved antimicrobial stewardship.

ClusterVAP is an exploratory, observational, prospective, multicentre cross-sectional study conducted in intensive care units across Sweden, France, Portugal, Denmark and the UK. Mechanically ventilated patients aged 18 years or older with newly developed clinical signs of lower respiratory tract infection will undergo bronchoalveolar lavage (BAL) or mini BAL sampling on clinical indication. Proteomic profiling using liquid chromatography tandem mass spectrometry will be performed on BAL supernatants. Unsupervised consensus clustering will define pneumoclusters, which will be characterised using clinical, microbiological and radiological data. 30-day outcomes, including mortality, ventilator-free days, antibiotic-free days, intensive care unit-free days and hospital-free days, will be compared across clusters to describe clinical trajectories. Candidate protein biomarkers for pragmatic cluster assignment will be derived using differential expression analysis.

Ethical approval will be obtained at all participating sites. Deferred consent will be used where permitted, with subsequent patient or proxy consent according to local regulations. Results will be disseminated through peer-reviewed publications and scientific conferences.

NCT07245888.