Editorial 

Prone Positioning in Viral Bronchiolitis—Back to the Back?

Joseph G. Kohne, Susan R. Conway, Steven L. Shein

JAMA 2026;336;(4):290-292. doi:10.1001/jama.2026.11742

Acute viral bronchiolitis is the most common diagnosis in US children’s hospitals, and respiratory failure is the costliest diagnosis across all pediatric acute care.¹ Despite the prevalence of this condition and its impact on children, families, and health systems, evidence-based treatments for acute bronchiolitis in hospitalized children are few and guidelines largely recommend against most interventions.² In this issue of JAMA, Baudin and colleagues for the PROPOSITIS Investigators evaluated a promising but understudied option: prone positioning.³ In a well-designed randomized clinical trial (RCT) conducted from 2021 to 2023, the authors enrolled 451 infants younger than 6 months with viral bronchiolitis who were receiving high-flow nasal cannula (HFNC) support. The primary end point of the study was escalation of care to noninvasive or invasive positive pressure ventilation (notably, only 3 children in the cohort were intubated). The primary intention-to-treat analysis did not reach statistical significance—with an odds ratio for escalation with prone positioning of 0.66 (95% CI, 0.40-1.07)—which was surprising given that their pilot physiological study suggested that prone positioning either improves effort of breathing or has negligible impact.⁴

Mechanistically, prone positioning is a plausible way to improve lung mechanics with limited safety risk, as evidenced in PROPOSITIS with similar comfort measures, feeding tolerance, and serious adverse event rates between groups. Prone positioning decreases compression of the densely perfused dorsal lung fields by anterior thoracic structures, which improves ventilation-perfusion (V̇/Q̇) matching, and improved homogeneity of ventilation decreases lung stress and strain.^5,6 Additionally, prone positioning improves abdominothoracic synchrony, decreases abdominal pressure on the lung fields, and may contribute to improved secretion clearance.⁷ Prone positioning is effective in restrictive lung diseases—most notably in adults with severe acute respiratory distress syndrome (ARDS), in whom early initiation of prone positioning is associated with a mortality benefit, and also in those with COVID-19, in whom awake prone positioning is associated with a decreased risk of intubation.^8,9Smaller trials of prone positioning among adults and children with primarily obstructive pathophysiology suggest benefit as well, as prone positioning can open small airways and has been associated with decreased airway resistance and improved oxygenation.^7,10 Bronchiolitis is classically an obstructive condition, but many children diagnosed with bronchiolitis are observed to have restrictive pathology, so it stands to reason that prone positioning has the potential to prove beneficial in this broad clinical syndrome.¹¹

At first glance, PROPOSITIS shows that prone positioning does not improve outcomes in children with severe bronchiolitis. But a deeper look suggests that the impact of prone positioning was heterogeneous in the study cohort. Among children who were able to tolerate 8 hours of prone positioning, there were fewer escalations (9% vs 20%; odds ratio, 0.35 [95% CI, 0.18-0.67]). This suggestion of a subset of responders is unfortunately offset by a 30% escalation rate among prone positioning group participants who were turned supine by their clinical team before the study goal prone positioning duration of 24 hours, which included “some infants [who] did not tolerate prone positioning at all” and a group (n = 49) who spent between 2 hours and 18 hours in prone positioning and had a nearly 50% rate of escalation to positive-pressure ventilation. Similar findings suggesting the presence of a heterogeneous treatment effect among other medications commonly prescribed to children with critical bronchiolitis have led to recognition that, as clinically diagnosed, bronchiolitis represents a broader clinical syndrome rather than a singular disease with unifying physiology. As with ARDS and sepsis, experts have begun calling for phenotype-driven treatments as the future of bronchiolitis research.¹²

Prone positioning was not universally beneficial in this trial, nor is it among any form of respiratory failure treated with prone positioning. Among adults with ARDS, response to prone positioning varies with ARDS severity, biological phenotype, and pattern of lung pathology identified on computed tomography.¹³ Even Baudin and colleagues noted in their preceding small physiological study that prone positioning improved esophageal pressure–time products in only 8 of 14 children with bronchiolitis (57%) receiving continuous positive airway pressure, with response being associated with higher baseline respiratory effort but no other patient-level factors.⁴ Nor is improvement in physiology always linked to improved outcomes, as children with ARDS have improved oxygenation with prone positioning but not improved mortality or ventilator-free days.¹⁴ It is possible that age, virus, body habitus, respiratory mechanics, inflammatory phenotype, or countless other patient-level factors may drive response to prone positioning in critical bronchiolitis. However, a qualitative comparison of the children who did and did not tolerate prone positioning (eTables 2 and 3 in the article’s Supplement 3) does not reveal an obvious difference, although there were insufficient data to investigate all putative biomarkers of responsiveness (eg, inflammatory status, vital sign changes). Identifying phenotypic factors that predict response to and tolerance of prone positioning may be needed to apply it effectively.

Additionally, it is possible that clinician comfort—and experience—influenced this study’s results. Prone positioning has not been reported as a common intervention in critical bronchiolitis. The median duration of prone positioning in this study was just above the minimum protocolized dose (24 hours), and there was minimal crossover to prone positioning among those randomized to supine positioning, both of which suggest that clinical teams at the study centers were using prone positioning very little outside of the study mandates. The idea that clinicians, even within a trial, are heavily influenced by prior experience and comfort comes from this same research group. As a leading force in conducting RCTs in severe bronchiolitis, they have now completed 3 large trials: TRAMONTANE (completed in 2015), TRAMONTANE 2 (2017), and PROPOSITIS (2025).^15,16 All 3 studies had nearly identical inclusion criteria and illness severity (as measured by median modified Wood Clinical Asthma Score at baseline). Yet the rate of escalation among children receiving HFNC support decreased across these trials, from 51% to 39% to 25%. And only in the most recent trial did clinicians elect not to intensify support for some children who met failure criteria—more than 30 participants in PROPOSITIS had treatment failure but continued to receive HFNC support. These declining treatment failure rates may be related to a shifting epidemiology, but they are likely driven, at least in part, by clinicians’ growing comfort with HFNC. How a clinician’s experience may impact outcomes of prone positioning is unknown, as even seemingly objective measures like the modified Wood Clinical Asthma Score leave room for interpretation (eg, mild vs moderate accessory muscle use), and choices to escalate can be clinician specific. Nevertheless, it is very possible that some PROPOSITIS participants might have benefited from prone positioning if they had been given more time in the prone position.

The PROPOSITIS Investigators have reported the largest RCT to date evaluating the clinical impact of prone positioning in patients with bronchiolitis, and one of the largest RCTs in severe bronchiolitis overall. The trial was well designed and executed: while clinicians and parents could not be blinded to treatment allocation, those adjudicating treatment failure were blinded; including only participants younger than 6 months reduces confounding by toddlers, whom some would diagnose with asthma; a nearly 95% consent rate suggests wide acceptance by families (although the approach rate was only about 25%); and even the post hoc modifications to their published per-protocol analytic methods suggest a collaborative peer review process. Importantly, some features of the study cohort may differ from those of other clinicians’ and centers’ practices, which may impact generalizability. More than one-third of the cohort had no reported abnormalities on chest radiograph; 30% received antibiotics; less than 3% received bronchodilators or corticosteroids; and there were notably long intervals between the median durations of respiratory support (3 days), intensive/intermediate care unit length of stay (4 days), and hospital length of stay (7 days).

We now have 1 more strategy to consider for preventing escalation of respiratory support in young infants, provided a child tolerates it. The findings presented point to a need for further investigation, although study design should consider the incomplete tolerance of prone positioning and the low and shifting failure rates among infants treated with HFNC. Given the changing background of the disease process, the variability within the disease itself, and the heterogeneity in treatment response, identifying ways to enrich and refine the study populations and study outcomes beyond the level of respiratory support required will be key to identifying effective interventions for patients with critical bronchiolitis. Until then, while prone positioning likely improves clinical outcomes in many children with severe bronchiolitis, clinicians are back to not knowing exactly when and how to use it safely and effectively.