Awake Prone Positioning in Adults With COVID-19: An Individual Participant Data Meta-Analysis

Jian Luo; Ivan Pavlov; Elsa Tavernier; Yonatan Perez; Aileen Kharat; Bairbre McNicholas; Oriol Roca; David L Vines; Miguel Ibarra-Estrada; Waleed Alhazzani; Kimberley Lewis; Steven Q Simpson; Garrett Rampon; Ling Liu; Qin Sun; Haibo Qiu; Yi Yang; Giuseppe Lapadula; Edward Tang Qian; Cheryl L Gatto; Todd W Rice; Ken Kuljit S Parhar; Jason Weatherald; Allan J Walkey; Nicholas A Bosch; Mai-Anh Nay; Thierry Boulain; Guillaume Fossat; Tim RE Harris; C Louise Thwaites; Nguyen Thanh Phong; Paolo Bonfanti; Sajad Yarahmadi; Seyed Mohammadreza Hashemian; Devachandran Jayakumar; Stephanie Parks Taylor; Stacy A Johnson; Claude Guerin; John G Laffey; Stephan Ehrmann; Jie Li

doi:10.1001/jamainternmed.2025.0011

. 2025 Mar 10;185(5):572–581. doi: 10.1001/jamainternmed.2025.0011

Awake Prone Positioning in Adults With COVID-19

An Individual Participant Data Meta-Analysis

Jian Luo ¹, Ivan Pavlov ², Elsa Tavernier ³, Yonatan Perez ⁴, Aileen Kharat ⁵, Bairbre McNicholas ^6,⁷, Oriol Roca ^8,^9,¹⁰, David L Vines ¹¹, Miguel Ibarra-Estrada ¹², Waleed Alhazzani ^13,¹⁴, Kimberley Lewis ^15,¹⁶, Steven Q Simpson ¹⁷, Garrett Rampon ¹⁸, Ling Liu ¹⁹, Qin Sun ¹⁹, Haibo Qiu ¹⁹, Yi Yang ¹⁹, Giuseppe Lapadula ²⁰, Edward Tang Qian ²¹, Cheryl L Gatto ²¹, Todd W Rice ²¹, Ken Kuljit S Parhar ²², Jason Weatherald ²³, Allan J Walkey ²⁴, Nicholas A Bosch ²⁵, Mai-Anh Nay ²⁶, Thierry Boulain ²⁶, Guillaume Fossat ²⁶, Tim RE Harris ^27,²⁸, C Louise Thwaites ²⁹, Nguyen Thanh Phong ³⁰, Paolo Bonfanti ²⁰, Sajad Yarahmadi ³¹, Seyed Mohammadreza Hashemian ³², Devachandran Jayakumar ³³, Stephanie Parks Taylor ³⁴, Stacy A Johnson ³⁵, Claude Guerin ³⁶, John G Laffey ^6,⁷, Stephan Ehrmann ^37,³⁸, Jie Li ^11,^✉, for the Awake Prone Positioning Meta-Analysis Group

¹Respiratory Medicine Unit and Oxford National Institute for Health and Care Research (NIHR) Biomedical Research Centre, NDM Experimental Medicine, University of Oxford, Oxford, United Kingdom

²Department of Emergency Medicine, Hôpital de Verdun, Montréal, Quebec, Canada

³Methods in Patients-Centered Outcomes and Health Research, INSERM UMR 1246, Nantes, France

⁴Service de Médecine Intensive–Réanimation, Hôpital de Hautepierre, Hôpitaux Universitaires de Strasbourg, Strasbourg, France

⁵Department of Pulmonology, Hôpitaux Universitaires de Genève, Geneva, Switzerland

⁶Anaesthesia and Intensive Care, School of Medicine, University of Galway, Galway, Ireland

⁷Department of Anaesthesia and Intensive Care Medicine, Galway University Hospitals, Galway, Ireland

⁸Servei de Medicina Intensiva, Parc Taulí Hospital Universitari, Institut de Recerca Part Taulí (I3PT-CERCA), Sabadell, Spain

⁹Departament de Medicina, Universitat Autònoma de Barcelona, Bellaterra, Spain

¹⁰Ciber Enfermedades respiratorias (Ciberes), Instituto de Salud Carlos III, Madrid, Spain

¹¹Department of Cardiopulmonary Sciences, Division of Respiratory Care, Rush University, Chicago, Illinois

¹²Unidad de Terapia Intensiva, Hospital Civil FrayAntonio Alcalde, Universidad de Guadalajara, Guadalajara, Jalisco, México

¹³Critical Care and Internal Medicine Department, College of Medicine, Imam Abdulrahman Bin Faisal University, Saudi Arabia

¹⁴Scientific Research Center, Ministry of Defense Health Services, Riyadh, Saudi Arabia

¹⁵Department of Medicine, McMaster University, Hamilton, Ontario, Canada

¹⁶Department of Health Research Methods, Evidence, and Impact, McMaster University, Hamilton, Ontario, Canada

¹⁷Division of Pulmonary, Critical Care, and Sleep Medicine, University of Kansas, Kansas City

¹⁸Department of Pulmonary, Critical Care, and Sleep Medicine, University of Missouri-Kansas City, Kansas City, Missouri

¹⁹Jiangsu Provincial Key Laboratory of Critical Care Medicine, Department of Critical Care Medicine, Zhongda Hospital, School of Medicine, Southeast University, Nanjing, China

²⁰IRCCS Fondazione San Gerardo, University of Milano Bicocca, Italy

²¹Vanderbilt University Medical Center, Nashville, Tennessee

²²Department of Critical Care Medicine, University of Calgary, Calgary, Alberta, Canada

²³Department of Medicine, Division of Pulmonary Medicine, University of Alberta, Edmonton, Alberta, Canada

²⁴Department of Medicine, University of Massachusetts Chan Medical School, Worcester

²⁵The Pulmonary Center, Boston University Chobanian & Avedisian School of Medicine, Boston, Massachusetts

²⁶Intensive Care Unit, Centre Hospitalier Universitaire d’Orléans, Orléans, France

²⁷Corporate Department of Emergency Medicine, Hamad Medical Corporation, Doha, Qatar

²⁸Queen Mary University London, United Kingdom

²⁹Oxford University Clinical Research Unit, Ho Chi Minh City, Vietnam

³⁰Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam

³¹Social Determinants of Health Research Center, School of Nursing and Midwifery, Lorestan University of Medical Sciences, Khorramabad, Iran

³²Chronic Respiratory Disease Research Center (CRDRC), National Research Institute of Tuberculosis and Lung Diseases (NRITLD), Clinical Tuberculosis and Epidemiology Research Center, Shahid Beheshti University of Medical Sciences (SBMU), Tehran, Iran

³³Department of Critical Care, Dr Kamakshi Memorial Hospital, Chennai, India

³⁴Department of Internal Medicine, University of Michigan, Ann Arbor

³⁵Division of General Internal Medicine, University of Utah School of Medicine, Salt Lake City

³⁶University of Lyon, Faculty of Medicine Lyon-Est and IMRB INSERM 955 Créteil, Lyon, France

³⁷Médecine Intensive Réanimation, CIC INSERM 1415, CRICS-TriggerSEP F-CRIN research network, Centre Hospitalier Régional Universitaire de Tours, Tours, France

³⁸Centre d’étude des pathologies respiratoires, INSERM U1100, Université de Tours, Tours, France

Accepted for Publication: January 3, 2025.

Published Online: March 10, 2025. doi:10.1001/jamainternmed.2025.0011

^✉

Corresponding Author: Jie Li, PhD, Department of Cardiopulmonary Sciences, Division of Respiratory Care, Rush University, 600 S Paulina St, AAC 765, Chicago, IL 60612 (jie_li@rush.edu).

Author Contributions: Dr Luo had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Luo, Pavlov, Tavernier, and Perez contributed equally to this work.

Concept and design: Luo, Pavlov, Tavernier, Perez, McNicholas, Vines, Alhazzani, Lewis, Simpson, Qiu, Parhar, Thwaites, Guerin, Laffey, Ehrmann, Li.

Acquisition, analysis, or interpretation of data: Luo, Pavlov, Tavernier, Perez, Kharat, McNicholas, Roca, Vines, Ibarra-Estrada, Alhazzani, Lewis, Simpson, Rampon, Liu, Sun, Yang, Lapadula, Qian, Gatto, Rice, Parhar, Weatherald, Walkey, Bosch, Nay, Boulain, Fossat, Harris, Thwaites, Phong, Bonfanti, Yarahmadi, Jayakumar, Taylor, Johnson, Laffey, Ehrmann, Li.

Drafting of the manuscript: Luo, Pavlov, Lewis, Parhar, Yarahmadi, Taylor, Li.

Critical review of the manuscript for important intellectual content: Luo, Pavlov, Tavernier, Perez, Kharat, McNicholas, Roca, Vines, Ibarra-Estrada, Alhazzani, Lewis, Simpson, Rampon, Liu, Sun, Qiu, Yang, Lapadula, Qian, Gatto, Rice, Parhar, Weatherald, Walkey, Bosch, Nay, Boulain, Fossat, Harris, Thwaites, Phong, Bonfanti, Yarahmadi, Jayakumar, Taylor, Johnson, Guerin, Laffey, Ehrmann, Li.

Statistical analysis: Luo, Pavlov, Tavernier, McNicholas, Liu.

Obtained funding: Lewis, Qiu, Weatherald, Thwaites.

Administrative, technical, or material support: Simpson, Sun, Qiu, Yang, Gatto, Rice, Thwaites, Jayakumar, Johnson, Laffey, Ehrmann, Li.

Supervision: Tavernier, McNicholas, Ibarra-Estrada, Alhazzani, Lewis, Gatto, Walkey, Nay, Thwaites, Phong, Guerin, Laffey, Ehrmann, Li.

Conflict of Interest Disclosures: Dr Li reported receiving research funding from the American Association for Respiratory Care, COPD Foundation, MEKICS Co. Ltd, Vincent Ltd, and Rice Foundation; she reported speaker fees from the American Association for Respiratory Care, Aerogen Ltd, Vincent Ltd, and Fisher & Paykel Healthcare Ltd outside the submitted work; she reported service as the section editor of Respiratory Care. Dr Pavlov reported receiving a grant from Open AI Inc; he reported receiving a speaker fee and travel funding from Fisher & Paykel Healthcare Ltd outside the submitted work. Dr Perez reported receiving research support to his institution and travel expenses reimbursement from Fisher & Paykel Healthcare Ltd and research support to his institution by Open AI. Dr Vines reported receiving research funding from the Rice Foundation and Tri-anim Health Services, as well as book royalties from Elsevier outside the submitted work. Dr Harris reported receiving funding from the Medical Research Council, Qatar outside the submitted work. Dr Roca reported receiving speaker fees from Hamilton Medical AG, Fisher & Paykel, and Aerogen Ltd outside the submitted work; he reported holding stock from Tesai Care SL. Dr Weatherald reported receiving grants from Janssen, 35Pharma, Merck, Astra Zeneca, and Sanofi outside the submitted work. Dr Qian reported receiving a speaker fee from Karl Storz Endoscopy outside the submitted work. Dr Rice reported receiving grant funding from the National Institutes of Health National Center for Advancing Translational Sciences; he reported consulting work from Cumberland Pharmaceuticals, Inc and Cytovale, Inc outside the submitted work; and he reported holding stock from Cumberland Pharmaceuticals, Inc. Dr Nay reported receiving research support to her institution and speaker’s fees from Fisher & Paykel Healthcare, and reported travel support from Pfizer outside the submitted work. Dr Bonfanti reported board membership with Viiv, Gilead, MSD, and AstraZeneca. Dr Lewis reported receiving grant funding from the Canadian Institutes of Health Research. Dr Nguyen Thanh reported receiving grant funding from Wellcome Trust outside the submitted work. Dr Lapadula reported receiving speaker fees from Insmed and Pfizer, travel expenses from ViiV Healthcare, and board membership with Insmed, ViiV Healthcare, Gilead srl outside the submitted work. Dr Laffey reported receiving consulting fees from Celenkos outside the submitted work. Dr Ehrmann reported receiving consulting fees from Aerogen Ltd and research support to his institution and travel expenses reimbursement from Fisher & Paykel Healthcare and Aerogen Ltd; he reported receiving research support to his institution by Open AI outside the submitted work. No other conflicts were reported.

Group Information: A complete list of the members of the Awake Prone Positioning Meta-Analysis Group appears in Supplement 2.

Data Sharing Statement: See Supplement 3.

Additional Contributions: We thank Dr José M. Ordóñez-Mena, DScHum, at University of Oxford for advice on imputation of missing data. He received no additional compensation for his contributions.

^✉

Corresponding author.

PMCID: PMC11894540 PMID: 40063016

Key Points

Question

Is awake prone positioning (APP) associated with better odds of survival without intubation compared with supine positioning in patients with COVID-19 and acute hypoxemic respiratory failure?

Findings

In this meta-analysis using individual participant data of 3019 patients from 14 randomized clinical trials, APP was found to improve survival without intubation and reduce the risk of both intubation and hospital mortality. A prolonged duration of APP, specifically 10 or more hours per day, was associated with better outcomes.

Meaning

These results suggest that among patients with COVID-19 and acute hypoxemic respiratory failure, APP is associated with improved clinical outcomes.

This meta-analysis examines awake prone positioning to treat patients with COVID-19 and acute hypoxemic respiratory failure and its association with increased odds of survival without intubation.

Abstract

Importance

The impact of awake prone positioning (APP) on clinical outcomes in patients with COVID-19 and acute hypoxemic respiratory failure (AHRF) remains uncertain.

Objective

To assess the association of APP with improved clinical outcomes among patients with COVID-19 and AHRF, and to identify potential effect modifiers.

Data Sources

PubMed, Embase, the Cochrane Library, and ClinicalTrials.gov were searched through August 1, 2024.

Study Selection

Randomized clinical trials (RCTs) examining APP in adults with COVID-19 and AHRF that reported intubation rate or mortality were included.

Data Extraction and Synthesis

Individual participant data (IPD) were extracted according to PRISMA-IPD guidelines. For binary outcomes, logistic regression was used and odds ratio (OR) and 95% CIs were reported, while for continuous outcomes, linear regression was used and mean difference (MD) and 95% CIs were reported.

Main Outcomes and Measures

The primary outcome was survival without intubation. Secondary outcomes included intubation, mortality, death without intubation, death after intubation, escalation of respiratory support, intensive care unit (ICU) admission, time from enrollment to intubation and death, duration of invasive mechanical ventilation, and hospital and ICU lengths of stay.

Results

A total of 14 RCTs involving 3019 patients were included; 1542 patients in the APP group (mean [SD] age, 59.3 [14.1] years; 1048 male [68.0%]) and 1477 in the control group (mean [SD] age, 59.9 [14.1] years; 979 male [66.3%]). APP improved survival without intubation (OR, 1.42; 95% CI, 1.20-1.68), and it reduced the risk of intubation (OR, 0.70; 95% CI, 0.59-0.84) and hospital mortality (OR, 0.77; 95% CI, 0.63-0.95). APP also extended the time from enrollment to intubation (MD, 0.93 days; 95% CI, 0.43 to 1.42 days). In exploratory subgroup analyses, improved survival without intubation was observed in patients younger than age 68 years, as well as in patients with a body mass index of 26 to 30, early implementation of APP (ie, less than 1 day from hospitalization), a pulse saturation to inhaled oxygen fraction ratio of 155 to 232, respiratory rate of 20 to 26 breaths per minute (bpm), and those receiving advanced respiratory support at enrollment. However, none of the subgroups had significant interaction with APP treatment. APP duration 10 or more hours/d within the first 3 days was associated with increased survival without intubation (OR, 1.85; 95% CI, 1.37-2.49).

Conclusions and Relevance

This IPD meta-analysis found that in adults with COVID-19 and AHRF, APP was associated with increased survival without intubation and with reduced risks of intubation and mortality, including death after intubation. Prolonged APP duration (10 or more hours/d) was associated with better outcomes.

Introduction

Awake prone positioning (APP), a noninvasive and low-cost treatment, has been widely used in patients with acute hypoxemic respiratory failure (AHRF) since the COVID-19 pandemic.¹ Physiological studies have shown that, in addition to improving oxygenation, APP can significantly reduce the work of breathing and enhance ventilation homogeneity,^2,3,4 potentially lowering the risk of lung injury.^5,6 In contrast, a 2023 physiological study showed that APP was associated with more intense inspiratory effort when compared with supine position due to positional increases in airway resistance and prolonged expiratory time.⁷ These heterogeneous physiological responses to APP may account for the variability in outcomes observed across randomized clinical trials (RCTs). Indeed, despite the publication of numerous RCTs^{8,9,10,11,12,13,14,15,16,17,18,19,20} and meta-analyses,^21,22,23,24 several key questions remain unanswered. First, the impact of APP on mortality remains unclear. While earlier meta-analyses^21,22 did not find a mortality benefit, a 2024 RCT²⁵ has demonstrated that prolonged APP can reduce the risk of death. Second, while patients receiving advanced respiratory support, such as high-flow nasal cannula (HFNC), continuous positive airway pressure (CPAP), or noninvasive ventilation (NIV), have been reported to benefit most from APP,^21,22 its effectiveness in patients receiving conventional oxygen therapy remains uncertain.^26,27 Third, it is unclear which patient populations, beyond those on advanced respiratory support, are most likely to benefit from APP, and what factors contribute to treatment success. To address these uncertainties, we conducted an individual participant data meta-analysis (IPD-MA) of RCTs comparing APP with supine positioning in patients with COVID-19 and AHRF.

Method

Literature Search

We prospectively registered the protocol for this IPD-MA in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42022343625) and conducted the review in accordance with the Preferred Reporting Items for Systematic Review and Meta-analysis of Individual Participant Data (PRISMA-IPD).²⁸ Two reviewers (A.K. and Y.P.) independently and in duplicate, conducted a systematic electronic search of PubMed, Embase, the Cochrane Library, and ClinicalTrials.gov from January 1, 2021, to August 1, 2024, based on our previous work.²¹ We searched for studies published in English only. (Detailed search strategies are available in the eAppendix in Supplement 1.) We included published and unpublished RCTs that compared APP with standard care in adult, nonintubated patients with COVID-19 and AHRF, if they reported our primary outcome of interest. Studies that randomized patients undergoing invasive mechanical ventilation or extracorporeal membrane oxygenation were excluded. Quasi-RCTs were included only in secondary sensitivity analyses. A group of investigators (J. Luo, J. Li, and I.P.) resolved discrepancies concerning study eligibility by discussion and consensus. All included studies obtained approval from the local research ethics boards and required obtaining informed consent from participants.

Outcomes

The primary outcome was survival without intubation. Secondary outcomes included intubation, death during intensive care unit (ICU) and hospital stay, death without intubation, death after intubation, escalation of respiratory support (defined as a sequential increase in oxygen or respiratory support levels [room air < conventional oxygen therapy < HFNC < CPAP/NIV < invasive mechanical ventilation]), ICU admission, time from enrollment to intubation, time from enrollment to death, duration of invasive mechanical ventilation, and lengths of stay in hospital and ICU.

Data Collection and Integration

The corresponding authors of eligible trials were invited to participate and share deidentified IPD. A data usage agreement was established before distributing a standardized data collection form, which included demographics, comorbidities, type of respiratory support, baseline oxygenation and vital signs, duration of APP, and outcomes. To ensure uniformity in data analysis, the primary outcome was censored to days 28 to 30 if data were available; otherwise, it was reported as originally collected. APP duration was reported as the daily average time (in hours) spent on APP (daily average time on APP [D0 through D3]). This metric was calculated as the total APP duration divided by the number of days APP was actually applied from day 0 to day 3 (data beyond day 3 were largely incomplete).

Assessment of Data Integrity and Risk of Bias

Data integrity and completeness were verified against the original study publications, when available, and any invalid, inconsistent, or missing data were sought for correction by communication with the respective study investigators. Two independent groups of investigators (S.E. and B.M.; J. Luo and J. Li) assessed the risk of bias for each included study using the revised tools for randomized trials (RoB 2) and cluster-randomized trials (RoB 2 CRT).²⁹ For each study, reviewers assigned one of the following risk-of-bias categories (low risk, some concerns, and high risk) based on an assessment of the following domains: (1) randomization process, (2) intended intervention (effect of assignment/adhering to intervention), (3) missing outcome data, (4) outcome measurement, and (5) selection of the reported result. Any discrepancies were resolved by additional 2 reviewers (A.K. and Y.P.).

Imputation of Missing Data

Patterns of missing data were tested using a Little test for missing completely at random (MCAR) at both study and intervention levels. Missing data were imputed using multivariate imputation by chained equations (MICE) under a fully conditional specification approach, accounting for the clustering of participants within studies.³⁰ Two models were attempted for multiple imputation, including Bayesian linear regression (referred to as “norm”) and predictive mean matching. The consistency of imputed data was further assessed by intraclass correlation with higher intraclass correlation indicating better representation. Imputation of missing data was conducted only for confounding variables, not for outcome variables. Complete-case, which removed participants with missing data, was used for primary analyses, while imputed data was used only for sensitivity analyses.

Statistical Analysis

To estimate the overall treatment association of APP with differences in this IPD-MA, a 1-stage approach was conducted using a generalized linear mixed model with stratified trial intercept and trial-specific centering of treatment variable to account for both clustering of participants within trials and heterogeneity across trials.^31,32 For binary outcomes, logistic regression was used and odds ratios (ORs) with 95% CIs were reported, while for continuous outcomes, linear regression was used and the mean difference (MD) and 95% CIs were reported. For time-to-event outcomes, a Cox proportional hazards model was used and reported as hazard ratio (HR). For the identification of patient-level treatment effect modifiers, treatment-covariate interaction was estimated between intervention and variables including age, body mass index (BMI; calculated as weight in kilograms divided by height in meters squared), time from hospitalization to enrollment, ratio of pulse oxygen saturation to fraction of inhaled oxygen (SpO₂:FiO₂ratio), respiratory rate, the unit of care at enrollment (ICU vs non-ICU), and type of respiratory support at enrollment (advanced respiratory support [HFNC, NIV, or CPAP] vs conventional oxygen therapy). Exploratory subgroup analyses were performed based on tertiles of continuous variables or categories of binary variables using the same regression model within each categorized analysis, and significance level was adjusted for multiple testing using the Bonferroni method.³³

To assess the independent association of APP duration with binary outcomes, univariate and multivariable logistic regression were performed in the APP group, using a full model with risk factors chosen a priori and confirmed by stepwise factor selection based on Akaike Information Criteria value. Cutoff values of APP duration were defined as the inflection time point where the trend in the percentage of outcome events changed, depicted graphically by the number and percentage of patients who had an event in 2-hour bins.

For conventional meta-analyses of aggregate data, treatment effects were estimated using the Mantel-Haenszel method for fixed-effects models and the inverse variance method for random-effect models, respectively. Two sets of conventional meta-analyses were conducted separately for the studies providing IPD and those with aggregate data.

Heterogeneity was evaluated by the Cochran Q test and the degree of heterogeneity was quantified by I²: insignificant heterogeneity (0% to 30%), moderate heterogeneity (30% to 60%), substantial heterogeneity (50% to 90%), and considerable heterogeneity (75% to 100%).³⁴

All analyses were performed on intention-to-treat population and conducted separately on complete-case data and imputed data from RCTs, both without and with quasi-RCTs. Results of complete-case IPD from RCTs were reported primarily, while results of IPD from imputation or quasi-RCTs and results of conventional meta-analyses were used as sensitivity analyses. All tests were 2-sided with a significance level of P < .05. All data were analyzed using R software version 4.2.2 (R Project for Statistical Computing).

Results

Study Selection and Quality Assessment

Our search identified a total of 3139 records, and 26 studies (5333 patients) were eligible for inclusion, out of which 5 studies (796 patients) were not published, resulting in 21 studies^{8,9,10,11,12,13,14,15,16,17,18,19,20,25,35,36,37,38,39,40,41} with 4537 patients (2315 APP vs 2222 control) included in a conventional meta-analysis (eFigure 1 in Supplement 1). After contacting corresponding authors to share individual participant data, 6 (1017 patients [22.4%]) were not included because IPD data could not be made available, and the other 15 (3520 patients [77.6%]) agreed to participate. We obtained IPD from 14 RCTs,^{8,9,10,11,12,13,14,15,16,17,20,25,35,36} comprising 3019 patients (1542 APP vs 1477 control). In addition, we included data from 1 quasi-RCT,¹⁹ resulting in 3520 patients (1800 APP vs 1720 control) for sensitivity analyses. Details of studies included in IPD-MA are summarized in eTable 1 in Supplement 1. From 1719 patients (approximately 60% of total patients) undergoing HFNC or NIV, data on flow was reported in 1241 patients (72.2%). Using the new definition of ARDS,⁴² ie, flow 30 or more L/min and SpO₂:FiO₂ ratio of 315 or lower, 1229 patients (99.0%) had ARDS based on the assumption of bilateral opacities. There were no significant imbalances across demographic variables between the 2 groups. In the APP group, the mean (SD) age was 59.3 (14.1) years, and 1048 of 1542 participants (68.0%) were male; in the control group, the mean (SD) age was 59.9 (14.1) years, with 979 of 1477 male participants (66.3%) (Table 1).

Table 1. Demographic Characteristics in RCTs.

Characteristics	APP		Control
Characteristics	No.	Event, No. (%)	No.	Event, No. (%)
Age, mean (SD), y	1542	59.3 (14.1)	1476	59.9 (14.1)
Sex
Female	1542	494 (32.0)	1477	498 (33.7)
Male	1542	1048 (68.0)	1477	979 (66.3)
BMI, mean (SD)	1500	28.7 (5.6)	1443	28.4 (5.6)
Comorbidities
Hypertension	936	391 (41.8)	892	372 (41.7)
COPD	947	52 (5.5)	890	42 (4.7)
CKD	1218	82 (6.7)	1181	78 (6.6)
Chronic liver disease	1208	17 (1.4)	1164	21 (1.8)
Diabetes	1542	466 (30.2)	1477	472 (32.0)
Time from hospitalization to enrollment, mean (SD), h	1339	1.6 (2.1)	1285	1.5 (2.1)
Use of steroids	1326	1143 (86.2)	1287	1106 (85.9)
Unit of care at enrollment (ICU)
ICU	1542	879 (57.0)	1477	857 (58.0)
Non-ICU	1542	663 (43.0)	1477	620 (42.0)
Type of respiratory support at enrollment
Advanced respiratory support	1542	872 (56.5)	1476	847 (57.4)
Conventional oxygen therapy	1542	670 (43.5)	1476	629 (42.6)
SpO₂:FiO₂ ratio, mean (SD)	1509	215.2 (104.7)	1451	212.2 (103.6)
Respiratory rate, mean (SD), bpm	1129	24.1 (6.2)	1083	24.2 (6.4)

Open in a new tab

Abbreviations: APP, awake prone positioning; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); bpm, breaths per minute; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; ICU, intensive care unit; SpO₂:FiO₂, ratio of pulse oxygen saturation to fraction of inhaled oxygen; RCT, randomized clinical trial.

No data were missing for the primary outcome. At the intervention level, all treatment effect modifiers were consistent with MCAR (Little test, P > .05). However, at the study level, systematic missingness occurred, with only BMI, time from hospitalization to enrollment, and respiratory rate identified as not missing at random (Little test, P < .001). Although masking participants and health care staff were not feasible due to the nature of the intervention, 17 trials^{8,9,10,11,12,13,14,16,18,20,25,35,36,37,39,40,41} were judged as low risk of bias, while 2 cluster-randomized trials^15,17 were considered high risk of bias (eFigures 2 and 3 in Supplement 1).

Overall Outcomes From Treatment With APP

Compared with supine positioning, APP was associated with increased survival without intubation (OR, 1.42; 95% CI, 1.20-1.68) and with reduced odds of intubation (OR, 0.70; 95% CI, 0.59-0.84) and hospital mortality (OR, 0.77; 95% CI, 0.63-0.95) (Table 2). While the overall patients (intention-to-treat population) in the APP group had a lower risk of death after intubation than the control group (OR, 0.76; 95% CI, 0.61-0.96), mortality was similar between APP and control groups within the selected subgroup of intubated patients (OR, 1.07; 95% CI, 0.79-1.46). APP extended the time to intubation compared with supine positioning (MD, 0.93 days; 95% CI, 0.43 to 1.42 days). No significant differences were observed between APP and control groups in terms of ICU mortality, need for escalation of respiratory support, need for ICU admission, duration of invasive mechanical ventilation, ICU length of stay, hospital length of stay, or adverse events.

Table 2. Primary and Secondary Outcomes in RCTs.

Outcomes	No. of study	Event, No. or mean/total		OR or MD (95% CI)		I², % (95% CI)
Outcomes	No. of study	APP	Control	IPD-MA	Conventional MA	I², % (95% CI)
Primary outcome
Survival without intubation	14	1149/1542	1004/1477	OR: 1.42 (1.20-1.68)	OR: 1.43 (1.20-1.69)	0 (0- 58)
Secondary outcomes
Intubation	14	336/1542	413/1477	OR: 0.70 (0.59-0.84)	OR: 0.70 (0.59-0.83)	0 (0-58)
Mortality	14	215/1542	255/1477	OR: 0.77 (0.63-0.95)	OR: 0.78 (0.64-0.96)	0 (0-58)
Death without intubation	14	57/1542	61/1477	OR: 0.88 (0.60-1.28)	OR: 0.81 (0.50-1.33)	0 (0-71)
Death after intubation	14	158/1542	194/1477	OR: 0.76 (0.61-0.96)	OR: 0.77 (0.61-0.97)	0 (0-58)
Death in nonintubated patients	14	57/1206	61/1064	OR: 0.72 (0.46-1.13)	OR: 0.68 (0.40-1.18)	8 (0-73)
Death in intubated patients	12	158/336	194/413	OR: 1.07 (0.79-1.46)	OR: 1.08 (0.79-1.47)	0 (0-60)
ICU mortality	9	86/807	89/781	OR: 0.91 (0.66-1.25)	OR: 0.90 (0.66-1.24)	0 (0-71)
Need for escalation of respiratory support	13	412/1501	464/1435	OR: 0.83 (0.59-1.18)	OR: 0.83 (0.59-1.19)	53 (11-75)
Need for ICU admission	10	112/518	95/468	OR: 0.92 (0.60-1.40)	OR: 0.88 (0.57-1.36)	34 (0-71)
Time to intubation, mean, d	8	4.5/308	3.6/375	MD: 0.93 (0.43 to 1.42)	MD: 0.96 (0.30 to 1.63)	0 (0-79)
Time to death, mean, d	10	13.0/211	13.0/248	MD: −0.25 (−2.25 to 1.75)	MD: −0.18 (−2.25 to 1.88)	45 (0-78)
Duration of IMV, mean, d	5	11.9/233	11.5/273	MD: −0.78 (−3.11 to 1.55)	MD: −0.72 (−3.16 to 1.73)	57 (0-86)
ICU LOS, mean, d	8	12.2/696	13.2/705	MD: −1.97 (−4.37 to 0.44)	MD −2.01 (−4.38 to 0.35)	81 (62-91)
Hospital LOS, mean, d	12	14.1/1434	14.6/1380	MD: −0.81 (−2.07 to 0.45)	MD −0.81 (−2.04 to 0.42)	62 (27-80)
Adverse events
Skin breakdown	10	26/1191	27/1143	OR: 0.84 (0.48-1.47)	OR: 0.84 (0.48-1.47)	0 (0-79)
Vomiting	9	31/1195	32/1154	OR: 0.78 (0.46-1.34)	OR: 0.78 (0.45-1.34)	0 (0-90)
Central arterial line dislodgement	9	38/1311	30/1228	OR: 1.14 (0.69-1.90)	OR: 1.08 (0.54-2.15)	0 (0-85)
Back pain	6	18/367	9/349	OR: 1.59 (0.64-3.93)	OR: 1.52 (0.62-3.72)	0 (0-90)
Bloating sensation	5	21/357	12/332	OR: 1.31 (0.52-3.35)	OR: 1.31 (0.51-3.35)	NA^a
Discomfort	8	76/567	58/524	OR: 3.47 (0.64-18.86)	OR: 2.64 (0.52-13.40)	80 (52-91)
Cardiac arrest	9	13/1164	11/1130	OR: 0.82 (0.33-2.02)	OR: 1.03 (0.26-4.09)	23 (0-92)

Open in a new tab

Abbreviations: APP, awake prone positioning; ICU, intensive care unit; IMV, invasive mechanical ventilation; IPD-MA, individual participant data meta-analyses; LOS, length of stay; MA, meta-analyses; MD, mean difference; NA, not available; OR, odds ratio; RCT, randomized clinical trial.

^{^a}

Four out of 5 studies reported no event on bloating sensation, resulting in only 1 study contributing to the final meta-analysis result; therefore, I² was not available.

In the time-to-event analysis, APP was associated with reduced risk of the composite outcome of mortality or intubation (HR, 0.75; 95% CI, 0.65-0.86) (Figure 1), intubation (HR, 0.74; 95% CI, 0.64-0.86) and mortality (HR, 0.81; 95% CI, 0.67-0.98) within 30 days (eFigures 4 and 5 in the Supplement 1).

Figure 1. — APP indicates awake prone positioning; HR, hazard ratio. Shading indicates 95% CIs.

Similar results were found in the sensitivity analyses, which excluded cluster-randomized trials (eTable 3 in Supplement 1) and included the quasi-RCT (eTable 4 and eFigures 6 through 8 in Supplement 1) and conventional meta-analyses (Table 2; eTables 5 and 6 in Supplement 1), except for mortality in the analyses involving the quasi-RCT.

Exploratory Subgroup and Interaction Analyses

For survival without intubation, the P values for interaction were not significant (Figure 2). The treatment effect of APP on survival without intubation was estimated as follows: patients aged 55 to 68 years had an OR of 1.57 (95% CI, 1.19-2.07), patients under 55 years old had an OR of 1.67 (95% CI, 1.18-2.35), patients with a BMI between 26 and 30 had an OR of 1.57 (95% CI, 1.18-2.08), patients who received APP 0.8 to 1 day during hospitalization had an OR of 1.60 (95% CI, 1.12-2.28), patients who received APP 0.8 days of hospitalization or less had an OR of 1.50 (95% CI, 1.11-2.04), patients with a SpO₂:FiO₂ ratio between 155 and 232 had an OR of 1.85 (95% CI, 1.25-2.72), patients with respiratory rates between 20 and 26 breaths per minute (bpm) had an OR of 1.53 (95% CI, 1.15-2.03), and patients receiving advanced respiratory support at enrollment had an OR of 1.46 (95% CI, 1.21-1.77) (Figure 2). These findings were consistent across both imputed data (eFigures 9 and 10 in the Supplement 1) and data from the quasi-RCT (eFigures 11 through 13 in the Supplement 1).

Similar results were also found for intubation (eFigure 14 in Supplement 1) and confirmed by imputed data and data from the quasi-RCT (eFigures 15 through 19 in Supplement 1), but significant treatment effect modification by age was found in the latter (P for interaction = .047). No significant treatment effect modifiers were identified for mortality (eFigures 20 through 25 in Supplement 1), death without intubation (eFigures 26 through 31 in Supplement 1), or death after intubation (eFigures 32 through 37 in Supplement 1).

Association of APP Duration and Outcomes

In 11 RCTs^{8,9,11,12,13,16,17,20,25,35,36} with data on APP duration, more than 93% (1174 of 1258 [93.3%]) of patients started APP on day 0 or day 1 from enrollment. After adjustment for age, BMI, time from hospitalization to enrollment, SpO₂:FiO₂ratio, respiratory rate, unit of care at enrollment, and type of respiratory support at enrollment, multivariable logistic regression showed that APP duration was an independent factor associated with survival without intubation. There was a 3% increase in the odds of survival without intubation for every hour increase in APP duration (OR, 1.03; 95% CI, 1.00-1.07) (eTable 7 in Supplement 1). We identified a significant inflection cutoff at 10 hours/d within the first 3 days (Figure 3). Patients with 10 hours or more of APP per day had a significantly lower baseline SpO₂:FiO₂ ratio compared with those with less than 10 hours/d, with similar baseline respiratory rates (eFigure 38 in Supplement 1). Patients with 10 or more hours/d of APP had higher odds of survival without intubation (OR, 1.85; 95% CI, 1.37-2.49), especially those who were consecutively proned for the first 3 days (OR, 2.59; 95% CI, 1.30-5.14) (eTable 8 in Supplement 1). These findings were confirmed by sensitivity analyses (eTables 7, 9 and 10, eFigures 39 through 41 in Supplement 1). Although APP duration was not independently associated with higher odds of intubation (OR, 0.97; 95% CI, 0.94-1.00) (eTables 11 and 12 in Supplement 1) or mortality (OR, 0.98; 95% CI, 0.95-1.02) (eTables 13 and 14 in Supplement 1), significant differences were observed when comparing patients with an APP duration of 10 or more hours/d to those with less than 10 hours/d during the first 3 days. Patients with 10 or more hours/d had reduced odds of intubation (OR, 0.69; 95% CI, 0.50-0.94) (eTables 8 and 10, eFigures 42 through 45 in Supplement 1) and mortality (OR, 0.49; 95% CI, 0.34-0.70) (eTables 8 and 10, eFigures 46 through 49 in Supplement 1), especially those who were consecutively proned for the 3 first days (intubation: OR, 0.54; 95% CI, 0.30-0.95; mortality: OR, 0.38; 95% CI, 0.15-0.97) (eTables 8 and 10 in Supplement 1).

Figure 3. — The number of patients who survived without intubation (yes) and those who did not (no) and the percentage of patients who survived without intubation in the total number of patients are depicted at each APP duration (daily average time on APP during day 0 to day 3) interval of 2 hours/d by stacked bars (left y-axis) and connected points (right y-axis), respectively.

Discussion

This study presents the most comprehensive analysis to date of APP for COVID-19 and AHRF. We have confirmed that APP reduces intubation and mortality, and identified key factors that are associated with its success, offering actionable insights for clinical practice.

Consistent with previous meta-analyses of aggregated data, our findings confirm that APP significantly reduces the risk of intubation compared with the supine position.^21,22,23,24 Unlike prior meta-analyses of RCTs,^21,22,23 which found no significant impact on hospital mortality, our study demonstrates that APP reduces mortality in patients with COVID-19 and AHRF. This outcome is likely attributed to larger sample sizes and particularly to analyses conducted at the individual participant level. However, in our sensitivity analysis that included a quasi-RCT,¹⁹ we did not observe a significant difference in mortality. This discrepancy is presumably due to the high number of patients with “do-not-intubate” orders in the quasi-RCT,¹⁹ because a significant reduction in the postintubation mortality by APP was still observed. More insights are expected from the ongoing CORONA trial (NCT04402879), which will further explore this issue. Among patients who failed treatment and ultimately required intubation, the APP group experienced longer durations between enrollment and intubation, yet their mortality did not significantly differ from those intubated in the supine position group. While assessing outcomes only in intubated patients can introduce bias compared with a strict intention-to-treat analysis of all randomized patients, these findings help alleviate clinicians’ concerns that APP may delay intubation and thereby increase postintubation mortality.^43,44,45

In addition to confirming that APP benefits most patients receiving advanced respiratory support,²¹ our exploratory subgroup analyses revealed new insights. Specifically, patients under 68 years old, with a BMI of 26 to 30, early implementation of APP (within 1 day), a SpO₂:FiO₂ ratio of 155 to 232 at enrollment, a respiratory rate of 20 to 26 bpm at enrollment, and those treated in an ICU were significantly associated with improved survival without intubation. Although interaction analyses did not show significant treatment effect modification by any of these factors potentially due to limited sample size,⁴⁶ the treatment effect in these subgroups remained significant after adjusting for multiple comparisons using the Bonferroni method.³³ These findings may help identify the patients most likely to benefit from APP, but should not preclude implementation of APP in patients who do not exhibit these beneficial characteristics since no significant increase of severe adverse events were observed with APP.

APP duration is an independent factor associated with treatment success; efforts should be made to improve patient compliance and comfort in the prone position in order to extend APP duration.⁴⁷ Although this finding aligns with previous studies,^8,48,49 our study analyzed data at the individual participant level after adjustment for other confounding factors, highlighting the importance of sustained prone positioning for optimal clinical outcomes, with a preferred daily duration of at least 10 hours in the first 3 days. Furthermore, the use of IPD allowed us to address potential biases in patient severity between those with longer and shorter APP duration. Given that the duration of APP was not randomized in the original RCTs, there was a concern that the observed benefit of longer APP might be driven by what may be described as a “healthy worker bias,” with healthier patients tolerating longer durations of APP. This assumption is not borne out in our IPD meta-analysis, as patients with longer APP durations (10 or more hours/d) had worse baseline oxygenation, reinforcing the conclusion that longer APP is independently associated with better clinical outcomes. In addition, approximately 44% of patients were intubated or dead within the first 3 days of enrollment, which necessarily resulted in shorter exposure to APP and could have introduced immortal time bias into our analysis of the dose-response relationship of APP on the outcomes. Reassuringly, the benefit of longer duration of APP was even stronger in a subgroup analysis restricted to patients who did not encounter the outcomes during the first 3 days, and had therefore an equal window of opportunity for exposure to APP.

Strengths and Limitations

Our study has several strengths. First, to our knowledge this study features the largest sample size to date, utilizing IPD from all eligible RCTs, representing the highest level of evidence. The findings from the comprehensive analyses are confirmed by sensitivity analyses including the most updated conventional meta-analysis of 21 RCTs, making them conclusive. Second, our findings are also practical for guiding clinical practice, including identifying the patient population that benefits most from APP and the optimal APP duration. Third, these results offer a valuable reference for future studies involving patients without COVID-19.

However, our study also has some limitations. First, despite extensive efforts, we could not obtain data from all RCTs for various reasons, among them regulatory complexities concerning sharing data collected across multiple sites. Second, not all studies collected all outcomes, resulting in several outcomes having data from only a few RCTs with limited sample sizes, which could result in insignificant differences in findings for those outcomes. Third, the impact of pandemic timing on the benefits of APP has not been fully addressed. Subgroup analyses based on pandemic timing indicated a stronger treatment effect of APP during the first and second waves compared to the control group. However, these findings should be interpreted with caution. Pandemic timing was analyzed as a study-level covariate, limiting the granularity of the analysis, and the sample size for patients enrolled during the third and fourth wave subgroups was relatively small (less than 20% of the total patients in each treatment arm), which reduces statistical power and precision and increases the potential for bias. Additionally, heterogeneity in co-interventions, virus variants, and clinical practices across pandemic periods may have influenced outcomes. For example, earlier trials with longer APP durations may have shown greater benefit, while later trials with shorter APP durations may have shown diminished benefits due to changes in cotreatments, virus variants, or overall care practices. These factors complicate efforts to attribute differences in outcomes solely to APP duration. Finally, while our models accounted for clustering within trials, they did not explicitly adjust for the time period of enrollment as a random effect, which could have introduced residual confounding.

Conclusions

This IPD meta-analysis establishes the efficacy of APP in improving survival outcomes and reducing intubation for adult patients with COVID-19–induced AHRF, with potential benefits in specific patient subgroups and with sustained application. APP was also associated with significant reductions in mortality, including death after intubation. These findings support the broader implementation of APP in clinical practice.

Supplement 1.

eAppendix. Search strategy

eFigure 1. Study Flow Diagram

eTable 1. Details of Each Included Trial for IPD-MA

eTable 2. Demographic Characteristics in Sensitivity Analysis for IPD-MA

eFigure 2. Risk of Bias Assessment of RCTs

eFigure 3. Risk of Bias Assessment of Cluster-Randomized Trials

eFigure 4. Comparison of Time-to-Intubation in RCTs

eFigure 5. Comparison of Time-to-Death in RCTs

eTable 3. Primary and Secondary Outcomes in RCTs Excluding Cluster-Randomized Trials

eTable 4. Primary and Secondary Outcomes in Sensitivity Analysis

eFigure 6. Comparison of Time-to-Composite Outcome of Mortality or Intubation in Sensitivity Analysis

eFigure 7. Comparison of Time-to-Intubation in Sensitivity Analysis

eFigure 8. Comparison of Time-to-Death in Sensitivity Analysis

eTable 5. Results of Conventional Meta-analyses From RCTs

eTable 6. Results of Conventional Meta-analyses From Sensitivity Analyses

eFigure 9. Interaction and Subgroup Analyses of Primary Outcome in RCTs After Imputation (Norm)

eFigure 10. Interaction and Subgroup Analyses of Primary Outcome in RCTs After Imputation (pmm)

eFigure 11. Interaction and Subgroup Analyses of Primary Outcome in Sensitivity Analysis

eFigure 12. Interaction and Subgroup Analyses of Primary Outcome in Sensitivity Analysis After Imputation (Norm)

eFigure 13. Interaction and Subgroup Analyses of Primary Outcome in Sensitivity Analysis After Imputation (pmm)

eFigure 14. Interaction and Subgroup Analyses of Intubation in RCTs

eFigure 15. Interaction and Subgroup Analyses of Intubation in RCTs After Imputation (Norm)

eFigure 16. Interaction and Subgroup Analyses of Intubation in RCTs After Imputation (pmm)

eFigure 17. Interaction and Subgroup Analyses of Intubation in Sensitivity Analysis

eFigure 18. Interaction and Subgroup Analyses of Intubation in Sensitivity Analysis After Imputation (Norm)

eFigure 19. Interaction and Subgroup Analyses of Intubation in Sensitivity Analysis After Imputation (pmm)

eFigure 20. Interaction and Subgroup Analyses of Mortality in RCTs

eFigure 21. Interaction and Subgroup Analyses of Mortality in RCTs After Imputation (Norm)

eFigure 22. Interaction and Subgroup Analyses of Mortality in RCTs After Imputation (pmm)

eFigure 23. Interaction and Subgroup Analyses of Mortality in Sensitivity Analysis

eFigure 24. Interaction and Subgroup Analyses of Mortality in Sensitivity Analysis After Imputation (Norm)

eFigure 25. Interaction and Subgroup Analyses of Mortality in Sensitivity Analysis After Imputation (pmm)

eFigure 26. Interaction and Subgroup Analyses of Death Without Intubation in RCTs

eFigure 27. Interaction and Subgroup Analyses of Death Without Intubation in RCTs After Imputation (Norm)

eFigure 28. Interaction and Subgroup Analyses of Death Without Intubation in RCTs After Imputation (pmm)

eFigure 29. Interaction and Subgroup Analyses of Death Without Intubation in Sensitivity Analysis

eFigure 30. Interaction and Subgroup Analyses of Death Without Intubation in Sensitivity Analysis After Imputation (Norm)

eFigure 31. Interaction and Subgroup Analyses of Death Without Intubation in Sensitivity Analysis After Imputation (pmm)

eFigure 32. Interaction and Subgroup Analyses of Death After Intubation in RCTs

eFigure 33. Interaction and Subgroup Analyses of Death After Intubation in RCTs After Imputation (Norm)

eFigure 34. Interaction and Subgroup Analyses of Death After Intubation in RCTs After Imputation (pmm)

eFigure 35. Interaction and Subgroup Analyses of Death After Intubation in Sensitivity Analysis

eFigure 36. Interaction and Subgroup Analyses of Death After Intubation in Sensitivity Analysis After Imputation (Norm)

eFigure 37. Interaction and Subgroup Analyses of Death After Intubation in Sensitivity Analysis After Imputation (pmm)

eFigure 38. Baseline SpO₂ FiO₂ Ratio and Respiratory Rate Between Patients With Longer and Shorter APP

eTable 7. Independent Factors Associated With APP Success for Primary Outcome in RCTs

eTable 8. Comparison of Outcomes Between Daily APP ≥10 Hours and <10 Hours in RCTs

eFigure 39. Dose Effect of APP Duration on APP Success for Primary Outcome in RCTs After Imputation (pmm)

eTable 9. Independent Factors Associated With APP Success for Primary Outcome in Sensitivity Analysis

eTable 10. Comparison of Outcomes Between Daily APP ≥10 Hours and <10 Hours in Sensitivity Analysis

eFigure 40. Dose Effect of APP Duration on APP Success for Primary Outcome in Sensitivity Analysis

eFigure 41. Dose Effect of APP Duration on APP Success for Primary Outcome in Sensitivity Analysis After Imputation (pmm)

eTable 11. Independent Factors Associated With APP Success for Intubation in RCTs

eFigure 42. Dose Effect of APP Duration on APP Success for Intubation in RCTs

eFigure 43. Dose Effect of APP Duration on APP Success for Intubation in RCTs After Imputation (pmm)

eTable 12. Independent Factors Associated With APP Success for Intubation in Sensitivity Analysis

eFigure 44. Dose Effect of APP Duration on APP Success for Intubation in Sensitivity Analysis

eFigure 45. Dose Effect of APP Duration on APP Success for Intubation in Sensitivity Analysis After Imputation (pmm)

eTable 13. Independent Factors Associated With APP Success for Mortality in RCTs

eFigure 46. Dose Effect of APP Duration on APP Success for Mortality in RCTs

eFigure 47. Dose Effect of APP Duration on APP Success for Mortality in RCTs After Imputation (pmm)

eTable 14. Independent Factors Associated With APP Success for Mortality in Sensitivity Analysis

eFigure 48. Dose Effect of APP Duration on APP Success for Mortality in Sensitivity Analysis

eFigure 49. Dose Effect of APP Duration on APP Success for Mortality in Sensitivity Analysis After Imputation (pmm)

jamainternmed-e250011-s001.pdf^{(8.1MB, pdf)}

Supplement 2.

Nonauthor Collaborators

jamainternmed-e250011-s002.pdf^{(161.2KB, pdf)}

Supplement 3.

Data Sharing Statement

jamainternmed-e250011-s003.pdf^{(12.6KB, pdf)}

References

1.McNicholas BA, Ibarra-Estrada M, Perez Y, et al. Awake prone positioning in acute hypoxaemic respiratory failure. Eur Respir Rev. 2023;32(168):220245. doi: 10.1183/16000617.0245-2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
2.Lehingue S, Allardet-Servent J, Ferdani A, et al. Physiologic effects of the awake prone position combined with high-flow nasal oxygen on gas exchange and work of breathing in patients with severe COVID-19 pneumonia: a randomized crossover trial. Crit Care Explor. 2022;4(12):e0805. doi: 10.1097/CCE.0000000000000805 [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Chiumello D, Chiodaroli E, Coppola S, et al. Awake prone position reduces work of breathing in patients with COVID-19 ARDS supported by CPAP. Ann Intensive Care. 2021;11(1):179. doi: 10.1186/s13613-021-00967-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Menga LS, Delle Cese L, Rosà T, et al. Respective effects of helmet pressure support, continuous positive airway pressure, and nasal high-flow in hypoxemic respiratory failure: a randomized crossover clinical trial. Am J Respir Crit Care Med. 2023;207(10):1310-1323. doi: 10.1164/rccm.202204-0629OC [DOI] [PMC free article] [PubMed] [Google Scholar]
5.Li J, Ibarra-Estrada M, Guérin C. Kittredge Lecture: Prone positioning for patients with COVID-19-induced acute hypoxemic respiratory failure: flipping the script. Respir Care. 2023;68(10):1449-1464. doi: 10.4187/respcare.11227 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Al Duhailib Z, Parhar KKS, Solverson K, Alhazzani W, Weatherald J. Awake prone position in patients with acute hypoxic respiratory failure: a narrative review. Respir Med Res. 2023;84:101037. doi: 10.1016/j.resmer.2023.101037 [DOI] [PubMed] [Google Scholar]
7.Grieco DL, Delle Cese L, Menga LS, et al. Physiological effects of awake prone position in acute hypoxemic respiratory failure. Crit Care. 2023;27(1):315. doi: 10.1186/s13054-023-04600-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
8.Ehrmann S, Li J, Ibarra-Estrada M, et al. ; Awake Prone Positioning Meta-Trial Group . Awake prone positioning for COVID-19 acute hypoxaemic respiratory failure: a randomised, controlled, multinational, open-label meta-trial. Lancet Respir Med. 2021;9(12):1387-1395. doi: 10.1016/S2213-2600(21)00356-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Alhazzani W, Parhar KKS, Weatherald J, et al. ; COVI-PRONE Trial Investigators and the Saudi Critical Care Trials Group . Effect of awake prone positioning on endotracheal intubation in patients with COVID-19 and acute respiratory failure: a randomized clinical trial. JAMA. 2022;327(21):2104-2113. doi: 10.1001/jama.2022.7993 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Rampon G, Jia S, Agrawal R, et al. Smartphone-guided self-prone positioning vs usual care in nonintubated hospital ward patients with COVID-19: a pragmatic randomized clinical trial. Chest. 2022;162(4):782-791. doi: 10.1016/j.chest.2022.05.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Nay MA, Hindre R, Perrin C, et al. Prone position versus usual care in hypoxemic COVID-19 patients in medical wards: a randomised controlled trial. Crit Care. 2023;27(1):240. doi: 10.1186/s13054-023-04529-z [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Fezzi M, Antolini L, Soria A, et al. Early prone positioning does not improve the outcome of patients with mild pneumonia due to SARS-CoV-2: results from an open-label randomised controlled trial—the EPCoT study. ERJ Open Res. 2023;9(4):00181-02023. doi: 10.1183/23120541.00181-2023 [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Harris TRE, Bhutta ZA, Qureshi I, et al. A randomised clinical trial of awake prone positioning in COVID-19 suspects with acute hypoxemic respiratory failure. Contemp Clin Trials Commun. 2024;39:101295. doi: 10.1016/j.conctc.2024.101295 [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Jayakumar D, Ramachandran Dnb P, Rabindrarajan Dnb E, Vijayaraghavan Md BKT, Ramakrishnan Ab N, Venkataraman Ab R. Standard care versus awake prone position in adult nonintubated patients with acute hypoxemic respiratory failure secondary to COVID-19 infection—a multicenter feasibility randomized controlled trial. J Intensive Care Med. 2021;36(8):918-924. doi: 10.1177/08850666211014480 [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Taylor SP, Bundy H, Smith WM, Skavroneck S, Taylor B, Kowalkowski MA. Awake prone positioning strategy for nonintubated hypoxic patients with COVID-19: a pilot trial with embedded implementation evaluation. Ann Am Thorac Soc. 2021;18(8):1360-1368. doi: 10.1513/AnnalsATS.202009-1164OC [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Johnson SA, Horton DJ, Fuller MJ, et al. Patient-directed prone positioning in awake patients with COVID-19 requiring hospitalization (PAPR). Ann Am Thorac Soc. 2021;18(8):1424-1426. doi: 10.1513/AnnalsATS.202011-1466RL [DOI] [PMC free article] [PubMed] [Google Scholar]
17.Kharat A, Dupuis-Lozeron E, Cantero C, et al. Self-proning in COVID-19 patients on low-flow oxygen therapy: a cluster randomised controlled trial. ERJ Open Res. 2021;7(1):00692-02020. doi: 10.1183/23120541.00692-2020 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Fralick M, Colacci M, Munshi L, et al. ; COVID Prone Study Investigators . Prone positioning of patients with moderate hypoxaemia due to covid-19: multicentre pragmatic randomised trial (COVID-PRONE). BMJ. 2022;376:e068585. doi: 10.1136/bmj-2021-068585 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Qian ET, Gatto CL, Amusina O, et al. ; Vanderbilt Learning Healthcare System Platform Investigators . Assessment of awake prone positioning in hospitalized adults with COVID-19: a nonrandomized controlled trial. JAMA Intern Med. 2022;182(6):612-621. doi: 10.1001/jamainternmed.2022.1070 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Phong NT, Duc DH, Hai HB, et al. ; OUCRU COVID RESEACH GROUP . Awake prone positioning effectiveness in moderate to severe COVID-19 a randomized controlled trial. Version 1; peer review: awaiting peer review. Wellcome Open Res. 2024;9:543. doi: 10.12688/wellcomeopenres.22792.1 [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Li J, Luo J, Pavlov I, et al. ; Awake Prone Positioning Meta-Analysis Group . Awake prone positioning for non-intubated patients with COVID-19-related acute hypoxaemic respiratory failure: a systematic review and meta-analysis. Lancet Respir Med. 2022;10(6):573-583. doi: 10.1016/S2213-2600(22)00043-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Weatherald J, Parhar KKS, Al Duhailib Z, et al. Efficacy of awake prone positioning in patients with covid-19 related hypoxemic respiratory failure: systematic review and meta-analysis of randomized trials. BMJ. 2022;379:e071966. doi: 10.1136/bmj-2022-071966 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Cao W, He N, Luo Y, Zhang Z. Awake prone positioning for non-intubated patients with COVID-19-related acute hypoxic respiratory failure: a systematic review based on eight high-quality randomized controlled trials. BMC Infect Dis. 2023;23(1):415. doi: 10.1186/s12879-023-08393-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Beran A, Mhanna M, Srour O, et al. Effect of prone positioning on clinical outcomes of non-intubated subjects with COVID-19. Respir Care. 2022;67(4):471-479. doi: 10.4187/respcare.09362 [DOI] [PubMed] [Google Scholar]
25.Liu L, Sun Q, Zhao H, et al. ; Chi-ARDS Net (Chinese ARDS Research Network) . Prolonged vs shorter awake prone positioning for COVID-19 patients with acute respiratory failure: a multicenter, randomised controlled trial. Intensive Care Med. 2024;50(8):1298-1309. doi: 10.1007/s00134-024-07545-x [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Pavlov I, Li J, Kharat A, et al. Awake prone positioning in acute hypoxaemic respiratory failure: an international expert guidance. J Crit Care. 2023;78:154401. doi: 10.1016/j.jcrc.2023.154401 [DOI] [PubMed] [Google Scholar]
27.Myatra SN, Alhazzani W, Belley-Cote E, et al. Awake proning in patients with COVID-19-related hypoxemic acute respiratory failure: a rapid practice guideline. Acta Anaesthesiol Scand. 2023;67(5):569-575. doi: 10.1111/aas.14205 [DOI] [PubMed] [Google Scholar]
28.Stewart LA, Clarke M, Rovers M, et al. ; PRISMA-IPD Development Group . Preferred Reporting Items for Systematic Review and Meta-Analyses of individual participant data: the PRISMA-IPD Statement. JAMA. 2015;313(16):1657-1665. doi: 10.1001/jama.2015.3656 [DOI] [PubMed] [Google Scholar]
29.Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898. doi: 10.1136/bmj.l4898 [DOI] [PubMed] [Google Scholar]
30.Van Buuren S, Brand JPL, Groothuis-Oudshoorn CGM, et al. Fully conditional specification in multivariate imputation. J Stat Comput Simul. 2006;76(12):1049-1064. doi: 10.1080/10629360600810434 [DOI] [Google Scholar]
31.Debray TP, Moons KG, van Valkenhoef G, et al. ; GetReal Methods Review Group . Get real in individual participant data (IPD) meta-analysis: a review of the methodology. Res Synth Methods. 2015;6(4):293-309. doi: 10.1002/jrsm.1160 [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Riley RD, Tierney JF, Stewart LA. Individual Participant Data Meta-Analysis: A Handbook for Healthcare Research. John Wiley & Sons Ltd; 2021. doi: 10.1002/9781119333784 [DOI] [Google Scholar]
33.Bristol DR. P value adjustments for subgroup analyses. J Biopharm Stat. 1997;7(2):313-321. doi: 10.1080/10543409708835189 [DOI] [PubMed] [Google Scholar]
34.Higgins J, Thomas J, Chandler J, et al. Cochrane Handbook for Systematic Reviews of Interventions, version 6.3, 2022. Cochrane Training. Accessed August 22, 2024. https://training.cochrane.org/handbook/current
35.Hashemian SM, Jamaati H, Malekmohammad M, Tabarsi P, Khoundabi B, Shafigh N. Efficacy of early prone positioning combined with noninvasive ventilation in COVID-19. Tanaffos. 2021;20(2):82-85. [PMC free article] [PubMed] [Google Scholar]
36.Yarahmadi S, Ebrahimzadeh F, Mohamadipour F, Cheraghian T, Eskini M. Effect of prone position on clinical outcomes of nonintubated patients with COVID-19: a randomised clinical trial. Collegian. 2023;30(3):449-456. doi: 10.1016/j.colegn.2022.12.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Rosén J, von Oelreich E, Fors D, et al. ; PROFLO Study Group . Awake prone positioning in patients with hypoxemic respiratory failure due to COVID-19: the PROFLO multicenter randomized clinical trial. Crit Care. 2021;25(1):209. doi: 10.1186/s13054-021-03602-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Gad S. Awake prone positioning versus non invasive ventilation for COVID-19 patients with acute hypoxemic respiratory failure. Egypt J Anaesth. 2021;37:85-90. doi: 10.1080/11101849.2021.1889944 [DOI] [Google Scholar]
39.Gopalakrishnan M, Khichar S, Saurabh S, et al. Effectiveness of early awake self proning strategy in non-intubated patients with COVID-19 hypoxemia: an open-labelled randomized clinical trial from Jodhpur, India. Monaldi Arch Chest Dis. 2022;93(4). doi: 10.4081/monaldi.2022.2431 [DOI] [PubMed] [Google Scholar]
40.Nasrallah BZN, Mahmoud MS, ElGendy HMA, Youssri Mahmoud NM, ElGendy MAEA. Patients self-proning with high-flow nasal cannula improves oxygenation in mild ARDS patients: a randomized clinical trial. Anaesth Pain Intensive Care. 2023;27(3):351-355. doi: 10.35975/apic.v27i3.2079 [DOI] [Google Scholar]
41.Javed H, Qayyum F, Aslam Khan MN, et al. Effect Of eight hours per day of intermittent self prone positioning for seven days on the severity of Covid-19 pneumonia/ acute respiratory distress syndrome. J Ayub Med Coll Abbottabad. 2023;35(1):68-75. doi: 10.55519/JAMC-01-11069 [DOI] [PubMed] [Google Scholar]
42.Matthay MA, Arabi Y, Arroliga AC, et al. A new global definition of acute respiratory distress syndrome. Am J Respir Crit Care Med. 2024;209(1):37-47. doi: 10.1164/rccm.202303-0558WS [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Camous L, Pommier JD, Martino F, Tressieres B, Demoule A, Valette M. Very late intubation in COVID-19 patients: a forgotten prognosis factor? Crit Care. 2022;26(1):89. doi: 10.1186/s13054-022-03966-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
44.Manrique S, Claverias L, Magret M, et al. Timing of intubation and ICU mortality in COVID-19 patients: a retrospective analysis of 4198 critically ill patients during the first and second waves. BMC Anesthesiol. 2023;23(1):140. doi: 10.1186/s12871-023-02081-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Kang BJ, Koh Y, Lim CM, et al. Failure of high-flow nasal cannula therapy may delay intubation and increase mortality. Intensive Care Med. 2015;41(4):623-632. doi: 10.1007/s00134-015-3693-5 [DOI] [PubMed] [Google Scholar]
46.Leon AC, Heo M. Sample sizes required to detect interactions between two binary fixed-effects in a mixed-effects linear regression model. Comput Stat Data Anal. 2009;53(3):603-608. doi: 10.1016/j.csda.2008.06.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Busico M, Laiz MM, Urrutia JG, et al. Strategies to achieve adherence to prone positioning in awake COVID-19 patients with high-flow nasal oxygen—a case series. Can J Respir Ther. 2022;58:151-154. doi: 10.29390/cjrt-2022-035 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Esperatti M, Busico M, Fuentes NA, et al. ; Argentine Collaborative Group on High Flow and Prone Positioning . Impact of exposure time in awake prone positioning on clinical outcomes of patients with COVID-19-related acute respiratory failure treated with high-flow nasal oxygen: a multicenter cohort study. Crit Care. 2022;26(1):16. doi: 10.1186/s13054-021-03881-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
49.Ibarra-Estrada M, Li J, Pavlov I, et al. Factors for success of awake prone positioning in patients with COVID-19-induced acute hypoxemic respiratory failure: analysis of a randomized controlled trial. Crit Care. 2022;26(1):84. doi: 10.1186/s13054-022-03950-0 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials