Comment

Can better oral care reduce non-ventilator hospital-acquired pneumonia?

Catia Cilloniz, Michael Klompas

Lancet Infect Dis Available online 8 June 2026

https://doi.org/10.1016/S1473-3099(26)00282-3Get rights and content

Non-ventilator hospital-acquired pneumonia (NV-HAP) is the most frequent and deadly health-care-associated infection.1, 2 The Hospital Acquired Pneumonia Prevention (HAPPEN) trial, presented by Nicole M White and colleagues in The Lancet Infectious Diseases, suggests that enhanced oral care could lower NV-HAP rates.³ The HAPPEN study was a 1-year, stepped-wedge, cluster-randomised trial conducted in nine inpatient units across three hospitals in Australia; each unit was randomly assigned to a variable period of usual care followed by a period of enhanced oral care, including patient and nurse education, free oral care products, and regular audit and feedback. Research nurses supported implementation. NV-HAP occurred in 46 (1·0%) of 4523 patients during control periods versus 32 (0·7%) of 4347 patients during intervention periods (cumulative hazard ratio 0·40, 95% CI 0·19-0·82).

This is the largest and most rigorous NV-HAP prevention study to date and provides the best evidence thus far that better oral care can lower NV-HAP rates. Strengths include randomisation, multicentre enrolment, large sample size, and rich data on secondary outcomes. The intervention was developed through consumer engagement and led to remarkable increase in oral care compliance, rising from 15·9% to 61·9%. Large-scale randomised trials are rare in infection control despite the frequency and morbidity of nosocomial infections; the authors are to be congratulated for pulling off this remarkable study.

But while the HAPPEN study's findings are promising, their interpretation is complicated by the stepped-wedge design, the intensity of the intervention, and the limitations of NV-HAP definitions. First, the stepped-wedge design compelled each unit to begin with usual care before starting the oral care intervention. The study began in June, 2024, corresponding to winter in Australia, so the plurality of control months were winter months whereas most intervention months occurred after winter. This is concerning because NV-HAP rates are highest in the winter.4, 5 The authors did not adjust for seasonality or respiratory viral activity so there is some risk that the observed decrease in NV-HAP rates was due to the passing of winter rather than enhanced oral care.

Second, oral hygiene was only one component of a multifaceted behavioural and organisational initiative that included dedicated research nurses (20 h per week per cluster), ward-based education, patient engagement, and audit and feedback.³ As such, the reduction in NV-HAP might reflect increased clinical vigilance, improved general care processes, and changes in documentation in addition to reducing oral bioburden. This makes it difficult to attribute changes specifically to oral care versus more systematic changes in care. The intensity of the intervention may also limit generalisability.

Third, the European Centre for Disease Prevention and Control (ECDC) surveillance criteria used to define NV-HAP risk ascertainment bias.⁶ The ECDC criteria require many subjective judgements such as radiographic interpretations and adjudicating “purulent sputum” and “changes in breath sounds”. In an open-label stepped-wedge design, where clinical staff were aware of intervention status and actively engaged in implementation activities, and surveillance staff recognised that units were increasingly likely to be in intervention status as the study progressed, changes in diagnostic intensity, documentation, and surveillance judgements could have lowered observed NV-HAP rates.⁷ Similarly, the dramatic increase in oral care might reflect improved recording in addition to enhanced performance.

One way to determine the extent to which lower NV-HAP rates reflect better care compared with changes in diagnostic practices or surveillance determinations is to look at alternative measures of infection and more concrete outcomes such as length of stay or mortality. The authors found no statistically significant difference in combined lower and upper respiratory tract infections between the intervention and control groups. Mortality rates were also similar between groups, and the intervention was not associated with an increase in adverse events.³ The authors suggest the neutral impact on lower and upper respiratory tract infections might be because oral care does not affect viral transmission pathways, but this explanation is difficult to understand as lower respiratory tract infections include aspiration pneumonia, lung abscess, and empyema, processes more likely due to bacterial microaspiration than viral inoculation.⁸ Very small numbers of microbiologically confirmed cases further complicated interpretation.³

In summary, the HAPPEN trial provides encouraging evidence that better oral care could prevent NV-HAP and represents an important advance in a field where high-quality randomised trials remain uncommon. The authors are to be commended for conducting this large pragmatic trial with clinically important outcomes rather than surrogate indicators of care quality. At the same time, considerable uncertainty remains due to a possible seasonal effect, the mixing of biological and systems-level interventions, and possible ascertainment bias. Further studies with objective endpoints, more microbiological power, and scalable implementation models are needed.