More has been written about ventilator-associated pneumonia (VAP) than any other nosocomial infection, with more PubMed records than the next two most prevalent nosocomial infections combined. However, despite this monumental research effort and plethora of devices, strategies and programs designed to prevent VAP, the condition itself remains elusive. We lack universally agreed diagnostic criteria 1, remain uncertain as to the optimal method of obtaining specimens for microbiological culture and indeed even whether microbiological testing is required 1 . At the heart of the problem of VAP lies the huge diversity of pathologies which can cause radiographically detectable pulmonary infiltration 2 and the high frequency of bacterial colonisation of the proximal respiratory tract of ventilated patients 3 . Distinguishing what is almost certainly benign pulmonary infiltrate with upper respiratory colonisation from genuine infection of the alveolar space remains a challenge we have yet to fully surmount.