Awake prone positioning reduces mortality, intubation, and hospital stay in acute hypoxemic respiratory failure: a systematic review and meta-analysis of 6,164 patients

Abstract

Background

Acute hypoxemic respiratory failure (AHRF) is a major cause of morbidity and mortality and often requires advanced respiratory support. Awake prone positioning (APP) has emerged as a simple, low-cost intervention to improve oxygenation in non-intubated patients; however, its clinical effectiveness and safety remain uncertain.

Aim

This systematic review and meta-analysis aimed to evaluate the effectiveness and safety of awake prone positioning in non-intubated adult patients with acute hypoxemic respiratory failure.

Methods

A systematic search of PubMed, Scopus, and Web of Science was conducted from database inception to February 2026. Randomized controlled trials and observational comparative studies evaluating APP versus usual care were included. Primary outcomes were mortality, intubation, and length of hospital stay. Secondary outcomes included ICU stay, invasive mechanical ventilation, ICU admission, escalation of respiratory support, time to invasive ventilation, and adverse events. Risk of bias was assessed using ROB 2 for randomized trials and the Newcastle–Ottawa Scale for observational studies. Meta-analysis was performed using a random-effects model.

Results

Twenty-four studies involving 6,164 patients were included. APP significantly reduced mortality (OR = 0.60, 95% CI 0.42–0.86, p = 0.005), intubation (OR = 0.69, 95% CI 0.60–0.79, p < 0.00001), length of hospital stay (MD = − 0.70 days, 95% CI − 1.07 to − 0.32, p = 0.0003), ICU stay (MD = − 2.84 days, 95% CI − 5.44 to − 0.24, p = 0.03), and invasive mechanical ventilation (OR = 0.42, 95% CI 0.31–0.58, p < 0.00001). No significant differences were observed in ICU admission, escalation of respiratory support, or adverse events.

Conclusion

Awake prone positioning was associated with improved clinical outcomes, including reduced mortality, intubation, and hospital stay, without an apparent increase in adverse events. However, these findings should be interpreted cautiously given the observed heterogeneity and potential publication bias. Further high-quality randomized trials are needed to confirm these results.

Introduction

Acute hypoxemic respiratory failure (AHRF) is a life-threatening condition characterized by impaired oxygen exchange and inadequate arterial oxygenation, often resulting from severe pulmonary infections, acute respiratory distress syndrome (ARDS), or inflammatory lung injury [1]. It represents a major cause of morbidity and mortality worldwide and frequently requires advanced respiratory support, including supplemental oxygen, non-invasive ventilation, or invasive mechanical ventilation [2]. The progression from mild hypoxemia to respiratory failure requiring intubation is associated with increased complications, prolonged hospitalization, and higher mortality rates [3]. Therefore, early identification and implementation of effective non-invasive interventions to improve oxygenation and prevent clinical deterioration remain critical priorities in the management of these patients [4].

Prone positioning has long been recognized as an effective strategy to improve oxygenation in intubated patients with ARDS [5]. Physiologically, prone positioning enhances ventilation-perfusion matching, promotes more uniform distribution of tidal volume, reduces dorsal lung compression, and improves alveolar recruitment [6]. These mechanisms collectively improve oxygenation and reduce ventilator-induced lung injury [7]. Large randomized controlled trials have demonstrated that prone positioning significantly reduces mortality in mechanically ventilated patients with moderate to severe ARDS, leading to its widespread adoption as a standard therapeutic intervention in this population [8–10].

During the coronavirus disease 2019 (COVID-19) pandemic, the unprecedented number of patients presenting with severe hypoxemia placed enormous strain on healthcare systems worldwide [11]. Many patients developed AHRF requiring high levels of oxygen support, and intensive care unit (ICU) capacity and mechanical ventilators became limited resources [12]. In this context, awake prone positioning (APP), which involves placing non-intubated, spontaneously breathing patients in the prone position, emerged as a simple, low-cost, and feasible intervention to improve oxygenation and potentially reduce the need for invasive mechanical ventilation [13]. Unlike traditional prone positioning in intubated patients, APP can be implemented early in the disease course and does not require sedation, specialized equipment, or intensive monitoring [14].

Several physiological studies have demonstrated that awake prone positioning can improve oxygenation by redistributing pulmonary perfusion, enhancing dorsal lung recruitment, and reducing ventilation-perfusion mismatch [15, 16]. Early observational studies during the COVID-19 pandemic suggested that APP may reduce intubation rates and improve clinical outcomes [17]. These findings led to widespread adoption of awake prone positioning in clinical practice, particularly in patients receiving high-flow nasal oxygen or non-invasive ventilation [18]. The intervention was especially attractive due to its safety, ease of implementation, and potential to preserve ICU resources [18].

Despite its widespread use, evidence regarding the effectiveness of awake prone positioning remains inconsistent. While some randomized controlled trials and observational studies have reported significant reductions in intubation rates and improved oxygenation, others have found no significant benefit in mortality, ICU admission, or length of hospital stay. Variability in study designs, patient populations, duration of prone positioning, adherence to the intervention, and clinical settings may contribute to these conflicting results. Furthermore, concerns remain regarding potential adverse events, patient tolerance, and the optimal duration and timing of prone positioning. So, the present systematic review and meta-analysis aimed to evaluate the effectiveness and safety of awake prone positioning in non-intubated adult patients with acute hypoxemic respiratory failure. Specifically, this study examined its impact on key clinical outcomes, including mortality, intubation, and length of hospital stay, as well as secondary outcomes such as ICU stay, invasive mechanical ventilation, escalation of respiratory support, and adverse events.

Methods

Study design and reporting standards

This systematic review and meta-analysis was conducted to evaluate the effectiveness and safety of APP in non-intubated adult patients with acute hypoxemic respiratory failure. The review was designed and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines to ensure transparency, reproducibility, and methodological rigor [19]. All stages of the review process, including literature search, study selection, data extraction, quality assessment, and statistical synthesis, were guided by the methodological recommendations outlined in the Cochrane Handbook for Systematic Reviews of Interventions [20]. Given the clinical importance of awake prone positioning during the COVID-19 pandemic and the variability in study designs evaluating this intervention, both randomized controlled trials and observational comparative studies were included.

Search strategy

A comprehensive and systematic literature search was performed to identify studies evaluating the clinical outcomes of awake prone positioning in patients with acute hypoxemic respiratory failure. The following electronic databases were searched: PubMed, Scopus, and Web of Science. The search covered all available records from database inception until February 2026. The search strategy incorporated a combination of Medical Subject Headings (MeSH) and free-text keywords related to prone positioning, respiratory failure, and COVID-19. Key search terms included “awake prone positioning,” “prone position,” “self-proning,” “non-intubated,” “acute hypoxemic respiratory failure,” “COVID-19,” “SARS-CoV-2,” “oxygen therapy,” “high-flow nasal cannula,” and “non-invasive ventilation.” Boolean operators (AND, OR) were used to appropriately combine search terms, and search strategies were adapted for each database to maximize sensitivity. Reference lists of relevant articles and previously published systematic reviews were also screened to identify additional eligible studies. Only full-text articles published in English were included.

Literature screening and study selection

All retrieved records were imported into reference management software (Endnote) [21], and duplicate entries were removed prior to screening. The screening process was conducted in two stages. Initially, titles and abstracts were reviewed to identify potentially relevant studies. Articles that appeared to meet the eligibility criteria were subsequently retrieved in full text and assessed in detail. The selection process was conducted systematically to ensure consistency and minimize the risk of selection bias. Any uncertainties regarding study eligibility were resolved through careful reassessment of the full text and discussion until agreement was reached. The complete study selection process, including identification, screening, eligibility assessment, and final inclusion, is presented in the PRISMA flow diagram.

Eligibility criteria

Studies were considered eligible if they evaluated adult patients diagnosed with acute hypoxemic respiratory failure who were managed without invasive mechanical ventilation at the time of enrollment and received awake prone positioning as part of their treatment. Eligible studies were required to include a comparator group receiving usual care or standard positioning without prone positioning to allow for direct comparison of outcomes. Both randomized controlled trials and observational comparative studies were included, provided they reported quantitative clinical outcome data. To ensure clinical relevance of the findings, studies were required to report at least one of the predefined primary or secondary outcomes of interest, including mortality, intubation, length of hospital stay, or other clinically meaningful endpoints such as ICU admission, escalation of respiratory support, or adverse events. Studies involving intubated patients at baseline, pediatric populations, or non-comparative designs were excluded. Additionally, case reports, case series without control groups, conference abstracts, review articles, editorials, and studies lacking sufficient data for analysis were excluded. Only peer-reviewed full-text articles published in English were considered eligible to maintain consistency and ensure adequate reporting quality.

PICO framework

The research question was structured according to the PICO framework. The population (P) included adult patients with acute hypoxemic respiratory failure managed without invasive mechanical ventilation. The intervention (I) consisted of awake prone positioning applied during spontaneous breathing. The comparator (C) was usual care or standard positioning without prone positioning. The primary outcomes (O) were mortality, endotracheal intubation, and length of hospital stay. Secondary outcomes included duration of ICU stay, need for invasive mechanical ventilation, ICU admission, escalation of respiratory support, time to initiation of invasive ventilation, and adverse events, including skin breakdown, vomiting, intravascular line dislodgement, and cardiac arrest.

Data extraction

Data extraction was conducted using a standardized data collection form developed specifically for this review to ensure consistency and completeness. Extracted study characteristics included study ID (first author and publication year), country of origin, study design, total sample size, participant demographics including mean or median age and proportion of female participants, clinical setting, follow-up duration, and main findings. Outcome data were extracted separately for intervention and control groups. These included mortality, intubation rates, length of hospital stay, duration of ICU stay, invasive mechanical ventilation requirement, ICU admission, escalation of respiratory support, time to invasive ventilation, and adverse events. When studies reported multiple follow-up time points, the most clinically relevant time point within the hospitalization period was selected. All extracted data were carefully reviewed to ensure accuracy and consistency before inclusion in the quantitative synthesis.

Outcomes

The primary outcomes of this meta-analysis were mortality, intubation, and length of hospital stay. These outcomes were selected because they represent critical clinical endpoints that directly reflect disease severity, progression, and patient recovery. Mortality is the most definitive measure of clinical effectiveness, while intubation represents progression to more severe respiratory failure requiring invasive intervention. Length of hospital stay was included as an indicator of disease burden and healthcare resource utilization. Secondary outcomes included duration of ICU stay, invasive mechanical ventilation requirement, ICU admission, escalation of respiratory support, time to initiation of invasive ventilation, and adverse events such as skin breakdown, vomiting, intravascular line dislodgement, and cardiac arrest.

Quality assessment and risk of bias

The methodological quality and risk of bias of the included studies were assessed using validated tools appropriate for each study design. For randomized controlled trials, the Cochrane Risk of Bias tool version 2 (ROB 2) was used [22]. This tool evaluates potential sources of bias across five key domains: the randomization process, deviations from intended interventions, missing outcome data, measurement of outcomes, and selective reporting of results. Each domain was assessed individually, and studies were classified as having low risk of bias, some concerns, or high risk of bias based on predefined criteria [22]. For observational studies, methodological quality was assessed using the Newcastle–Ottawa Scale (NOS), which is widely used to evaluate the quality of non-randomized studies included in systematic reviews and meta-analyses [23]. The NOS assesses studies across three main domains: selection of study participants, comparability of study groups, and adequacy of outcome assessment and follow-up. Studies were awarded stars based on predefined criteria within each domain, with higher scores indicating better methodological quality [23].

Statistical analysis and heterogeneity assessment

Quantitative synthesis was performed using Review Manager (RevMan) software. Dichotomous outcomes were analyzed using odds ratios with 95% confidence intervals, while continuous outcomes were analyzed using mean differences with corresponding confidence intervals. Forest plots were generated to visually present the pooled effect estimates.

A random-effects model was used for all analyses to account for potential variability between studies, including differences in patient populations, clinical settings, and intervention protocols. Statistical heterogeneity was assessed using the I² statistic, with values greater than 50% indicating substantial heterogeneity.

Publication bias was evaluated using funnel plots when sufficient studies were available. Sensitivity analyses were performed to assess the robustness of the pooled estimates and to determine whether individual studies had a disproportionate influence on the overall results.

To assess the potential impact of study design on the pooled estimates, a sensitivity analysis was performed excluding observational studies. This approach was used to evaluate the robustness of the findings derived primarily from randomized controlled trials.

A leave-one-out sensitivity analysis was performed to assess the influence of individual studies on the pooled estimates and to evaluate the robustness of the results in the presence of heterogeneity.

A sensitivity analysis restricted to randomized controlled trials was performed to assess the robustness of the findings and to evaluate the potential impact of including observational studies.

Results

Search strategy and study selection

The initial database search yielded 4184 records. After removal of duplicates, 2170 unique studies remained for title and abstract screening. Of these, 2114 articles were excluded for clear irrelevance, leaving 56 full-text reports assessed for eligibility. Following full-text evaluation according to the predefined inclusion criteria, 24 studies were ultimately included in the qualitative and quantitative synthesis, comprising 2 cohort studies and 22 RCTs studies. The complete study selection process is illustrated in the PRISMA flow diagram. (Fig. 1).

Fig. 1.

PRISMA flowchart of included studies

Characteristics of the included studies

A total of 24 studies involving 6164 patients were included in this meta-analysis, with 3,164 patients in the APP group and 3000 patients in the control group. Most studies were randomized controlled trials (n = 22), while two were observational cohort studies. The studies were conducted across multiple countries, including Canada, the United States, France, Egypt, India, China, Iran, Argentina, Tunisia, Switzerland, Sweden, Pakistan, Qatar, and Mexico, reflecting a diverse and internationally representative patient population. Sample sizes varied widely, ranging from small single-center trials to large multicenter studies. The mean age of participants generally ranged from approximately 41 to 69 years, and the proportion of female participants varied across studies. Most studies were conducted in hospital settings, including intensive care units, isolation wards, and tertiary care hospitals. Follow-up duration ranged from short-term assessments of 24 h to longer periods of up to 90 days, with most studies reporting outcomes at approximately 28 to 30 days. Overall, the included studies evaluated the effectiveness of awake prone positioning in patients with acute hypoxemic respiratory failure. While several studies reported reductions in intubation rates and improved respiratory outcomes, findings related to mortality and hospital length of stay were inconsistent across studies (Table 1).

Table 1.

Baseline characteristics of included studies

Study ID	Country	Design	Sample size		Age		Sex (Female)		Setting	Follow-up	Main findings
			APP	control	APP	control	APP	control
Alhazzani 2022 [24]	Canada	RCT	205	195	56.8 (12.5)	58.3 (13.2)	56 (27)	61 (31)	Multicenter hospitals	30 days	Awake prone positioning did not significantly reduce intubation or mortality compared with usual care in nonincubated COVID-19 patients.
Ehrmann 2021 [25]	France, Spain, USA, Canada	RCT	564	557	61·5 (13·3)	60·7 (14·0)	184 (33%)	191 (34%)	Multicenter meta-trial	28 days	Awake prone positioning significantly reduced treatment failure and intubation compared with standard care (RR 0.86; HR for intubation 0.75), with no mortality difference.
Fralick 2022 [26]	Canada	RCT	126	122	59.5 (45–68)	54 (44–62)	44 (35)	45 (37)	Multicenter trial	Hospital stay	Awake prone positioning did not reduce the composite outcome of death, mechanical ventilation, or respiratory failure compared with standard care (14% vs. 14%, OR 0.92).
Gad 2021 [27]	Egypt	RCT	15	15	49.0(38–62)	46.0(33–51)	6	7	critical care isolation unit	3 days.	Awake prone positioning significantly improved oxygenation (SaO₂ and PaO₂) and helped reduce intubation rates, though NIV showed superior CO₂ reduction and no difference in mortality or length of stay.
Gopalakrishnan 2023 [10]	India	RCT	257	245	58.2 (12.2)	61.3 (13.0)	55 (21.4)	67 (27.3)	Tertiary hospital	30 days	Awake self-proning did not reduce mortality or mechanical ventilation requirement compared with standard care.
Harris 2024 [28]	Qatar	RCT	31	30	42.4 ± 10.9	41.2 ± 9.5	29 (93.5)	25 (83.3)	Two-center hospitals	30 days	Awake prone positioning did not significantly reduce escalation of oxygen support or improve respiratory outcomes compared with control.
Harrois 2025 [14]	France	RCT	224	221	59 (12)	60 (11)	57 (25)	59 (27)	Multicenter trial	28 days	Awake prone positioning showed a high probability of reducing intubation or death compared with standard care (posterior probability 93.8%, OR 0.74).
Hashemian 2021 [29]	Iran	RCT	45	30	NR	NR	29(64.4)	23(73.3)	Single-center ICU	ICU stay	Awake prone positioning combined with NIV improved oxygenation compared with NIV alone in COVID-19 patients.
Ibarra-Estrada 2022 [30]	Mexico, Canada, USA	RCT	216	214	58.6 ± 15.8	58.2 ± 15.8	84 (38.9)	88 (41.1)	Multicenter trial	28 days	Awake prone positioning significantly reduced intubation rate compared with standard care (30% vs. 43%, RR 0.70).
Ismail 2023 [31]	Tunisia	Cohort	192	81	55 (28–60)	56 (28–75)	NR	NR	Single-center ICU	ICU stay	Awake prone positioning significantly reduced invasive ventilation use and mortality compared with control patients.
Javed 2023 [32]	Pakistan	RCT	36	36	63.79 ± 15.26	63.79 ± 15.26	23	24	Isolation wards	90 days	Awake prone positioning improved early respiratory physiology but did not improve long-term survival compared with control.
Jayakumar 2021 [33]	India	RCT	30	30	54.8 + 11.1	57.3 + 12.1	5 (16.7%)	5 (16.7%)	Tertiary hospitals	Hospital stay	Awake prone positioning was feasible and safe but did not significantly improve clinical outcomes or reduce intubation compared with standard care.
Johnson 2021 [34]	USA	RCT	15	15	52 (40–65)	62 (49–75)	8 (53.3)	8 (53.3)	Single center	72 h	Patient-directed awake prone positioning showed poor adherence and did not improve oxygenation or clinical outcomes compared with usual care.
Kaur 2021 [35]	USA	RCT	92	33	61.1 ± 12.3	64.9 ± 10.4	56 (61)	23 (67)	Multicenter trial	28 days	Early awake prone positioning reduced 28-day mortality compared with late initiation, with no significant difference in intubation rate.
Kharat 2021 [36]	Switzerland	RCT	10	17	54 ± 14	60 ± 11	4	6	Single center	24 h	Awake prone positioning reduced oxygen requirements clinically but without statistically significant differences compared with usual care.
Liu 2024 [37]	China	RCT	205	204	67.6 (10.4)	68.9 (9.6)	59 (29)	68 (33)	Multicenter trial	28 days	Prolonged awake prone positioning significantly reduced intubation and mortality compared with shorter-duration prone positioning.
Nasrallah 2023 [9]	Egypt	RCT	45	45	62.24 ± 15.15	58.47 ± 10.32	13 (28.9)	19 (42.2)	University hospitals	ICU stay	Awake prone positioning with HFNC significantly improved oxygenation and reduced intubation rate and ICU stay compared with HFNC alone.
Nay 2023 [38]	France	RCT	135	132	58.4 (12.1)	59.2 (11.0)	96 (71)	93 (71)	Multicenter wards	28 days	Awake prone positioning did not significantly reduce the composite outcome of NIV, intubation, or death compared with usual care.
Olmos 2025 [39]	Argentina	Cohort	197	302	57 (47–66)	59 (46–71)	56 (28)	77 (25)	Multicenter trial	Hospital stay	Bundle-of-care strategy increased awake prone positioning duration and was associated with significantly reduced intubation risk compared with standard care.
Qian 2022 [40]	USA	RCT	258	243	61.6 (15.4)	60.3 (15.2)	112 (43.4)	105 (43.2)	Academic centers	28 days	Awake prone positioning was not associated with improved clinical outcomes and showed no reduction in mechanical ventilation or mortality compared with usual care.
Rampon 2022 [41]	USA	RCT	159	143	52 (39–62)	54 (43–63)	63 (39.6)	54 (40.3)	Multicenter trial	Hospital stay	Smartphone-guided awake prone positioning did not significantly reduce respiratory deterioration or ICU transfer compared with usual care.
Rosén 2021 [42]	Sweden	RCT	39	36	65 [55–70]	66 [53–74]	7	13	Multicenter trial	30 days	Awake prone positioning did not reduce intubation rate compared with standard care (33% vs. 33%, HR 1.01).
Taylor 2020 [43]	USA	RCT	27	13	56 (45–66)	60 (54–63)	10 (37)	3 (23)	Teaching hospital	48 h	Awake prone positioning improved oxygenation trends but feasibility was limited by poor adherence and implementation barriers.
Yarahmadi 2023 [8]	Iran	RCT	41	41	53 ± 7.1	53 ± 7.1	23 (56.10)	23 (56.10)	Single center	Hospital stay	Awake prone positioning improved oxygenation and respiratory parameters but did not significantly reduce intubation or improve survival compared with supine positioning.

APP awake prone positioning, RCT randomized controlled trial, NIV non-invasive ventilation, HFNC high-flow nasal cannula, ICU intensive care unit, RR risk ratio, HR hazard ratio, OR odds ratio, NR not reported, SaO₂ arterial oxygen saturation, PaO₂ partial pressure of arterial oxygen, CO₂ carbon dioxide, USA United States of America

Quality assessment and risk of bias

The methodological quality of the included studies varied according to study design. Among the randomized controlled trials assessed using the ROB 2 tool, eighteen studies were judged to be at low risk of bias across all domains, while four studies were rated as having some concerns, most related to deviations from intended interventions or outcome measurement. (Fig. 2).

Fig. 2.

Risk of bias of of included RCTs using ROB2 tool

The two cohort studies were evaluated using the NOS and all were rated as good quality, indicating adequate selection of participants, comparability of cohorts, and appropriate outcome assessment. (Table 2).

Table 2.

Quality assessment of cohort studies using NOS scale

Study ID	Selection (max 4)	Comparability (max 2)	Outcome (max 3)	Total (max 9)
Ismail 2023	☆☆☆☆	☆☆	☆☆☆	☆☆☆☆☆☆☆☆☆
Olmos 2025	☆ ☆ ☆ ☆	☆	☆ ☆☆	☆ ☆ ☆ ☆ ☆ ☆ ☆ ☆

Data analysis

Mortality

Mortality was reported in 23 studies comprising a total of 6088 participants. The pooled analysis using a random-effects model demonstrated that awake prone positioning was associated with a statistically significant reduction in mortality compared with usual care (OR = 0.60, 95% CI 0.42 to 0.86, p = 0.005), favoring the APP group (Fig. 3). This represents a clinically meaningful reduction in the risk of death among patients managed with awake prone positioning. Moderate heterogeneity was observed among the included studies (I² = 72%), suggesting some variability in treatment effect across different study populations. Visual inspection of the funnel plot demonstrated substantial asymmetry, indicating a moderate likelihood of publication bias. (Fig. 4).

Fig. 3.

Odds ratio of mortality between APP and usual care

Fig. 4.

Funnel plot (publication bias) of mortality between APP and usual care

Intubation

Intubation was reported in 19 studies including 5,283 participants. The pooled analysis showed that awake prone positioning was associated with a statistically significant reduction in intubation compared with usual care (OR = 0.69, 95% CI 0.60 to 0.79, p < 0.00001), favoring the APP group (Fig. 5). This finding suggests that awake prone positioning reduces the progression to invasive airway support. No heterogeneity was present (I² = 6%). Funnel plot analysis showed clear asymmetry, suggesting moderate risk of publication bias (Fig. 6).

Fig. 5.

Odds ratio of Intubation between APP and usual care

Fig. 6.

Funnel plot (publication bias) of Intubation between APP and usual care

Length of hospital stay

Length of hospital stay was reported in 13 studies comprising 3,726 participants. The pooled analysis demonstrated that awake prone positioning was associated with a statistically significant reduction in hospital length of stay compared with usual care (MD = − 0.70 days, 95% CI − 1.07 to − 0.32, p = 0.0003), favoring the APP group (Fig. 7). This indicates that awake prone positioning may facilitate earlier recovery and discharge. High heterogeneity was observed (I² = 81%). Funnel plot analysis showed clear asymmetry, suggesting moderate risk of publication bias (Fig. 8).

Fig. 7.

Mean difference of Duration of hospital stay (days) between APP and usual care

Fig. 8.

Funnel plot (publication bias) of Duration of hospital stay (days) between APP and usual care

Duration of ICU stay

Duration of ICU stay was reported in 8 studies including 1,486 participants. The pooled analysis showed a statistically significant reduction in ICU stay duration in favor of awake prone positioning (MD = − 2.84 days, 95% CI − 5.44 to − 0.24, p = 0.03), favoring the APP group. High heterogeneity was observed among the included studies (I² = 94%). (Fig. 9).

Fig. 9.

Mean difference of Duration of ICU stay (days) between APP and usual care

Invasive mechanical ventilation

The requirement for invasive mechanical ventilation was reported in 7 studies comprising 1,532 participants. The pooled analysis demonstrated a statistically significant reduction in invasive mechanical ventilation in the APP group compared with usual care (OR = 0.42, 95% CI 0.31 to 0.58, p < 0.00001), favoring awake prone positioning. High heterogeneity was observed (I² = 87%). (Fig. 10).

Fig. 10.

Odds ratio of Invasive mechanical ventilation between APP and usual care

ICU admission

ICU admission was reported in 6 studies including 1,184 participants. The pooled analysis showed no statistically significant difference between awake prone positioning and usual care (OR = 0.92, 95% CI 0.64 to 1.33, p = 0.67). No heterogeneity was observed (I² = 0%). (Fig. 11).

Fig. 11.

Odds ratio of Admission to ICU between APP and usual care

Skin breakdown

Skin breakdown was reported in 2 studies comprising 1,551 participants. The pooled analysis demonstrated no statistically significant difference between the APP and control groups (OR = 0.69, 95% CI 0.29 to 1.65, p = 0.41). No Heterogeneity was observed (I² = 0%). (Fig. 12).

Fig. 12.

Odds ratio of Skin breakdown between APP and usual care

Vomiting

Vomiting was reported in 3 studies including 1,626 participants. The pooled analysis showed no statistically significant difference between the two groups (OR = 0.74, 95% CI 0.41 to 1.31, p = 0.30). No heterogeneity was observed (I² = 0%). (Fig. 13).

Fig. 13.

Odds ratio of vomiting between APP and usual care

Intravascular line dislodgement

Intravascular line dislodgement was reported in 3 studies including 1,626 participants. The pooled analysis demonstrated no statistically significant difference between awake prone positioning and usual care (OR = 1.29, 95% CI 0.80 to 2.09, p = 0.30). No heterogeneity was observed (I² = 0%). (Fig. 14).

Fig. 14.

Odds ratio of Intravascular lines dislodgement between APP and usual care

Cardiac arrest

Cardiac arrest was reported in 3 studies including 1,626 participants. The pooled analysis showed no statistically significant difference between the APP and control groups (OR = 2.61, 95% CI 0.49 to 13.73, p = 0.26). No heterogeneity was observed (I² = 0%). (Fig. 15).

Fig. 15.

Odds ratio of Cardiac arrest at any time between APP and usual care

Time to initiation of invasive mechanical ventilation

Time to initiate invasive mechanical ventilation was reported in 2 studies including 1,151 participants. The pooled analysis demonstrated no statistically significant difference between the APP and usual care groups (MD = 4.66 h, 95% CI − 14.58 to 23.91, p = 0.63). High heterogeneity was observed (I² = 95%). (Fig. 16).

Fig. 16.

Mean difference of Time to initiate invasive ventilation (hs) between APP and usual care

Escalation of respiratory support

Escalation of respiratory support was reported in 3 studies comprising 151 participants. The pooled analysis showed no statistically significant difference between awake prone positioning and usual care (OR = 1.86, 95% CI 0.63 to 5.49, p = 0.26). Moderate heterogeneity was observed (I² = 56%). (Fig. 17).

Fig. 17.

Odds ratio of escalation between APP and usual care

Sensitivity analysis excluding the two observational cohort studies demonstrated consistent effect estimates for the primary outcomes, including mortality and intubation, with no meaningful change in direction or statistical significance. This suggests that the inclusion of observational studies did not significantly influence the overall results.

Sensitivity analyses demonstrated that the pooled effect estimates for mortality and intubation remained directionally consistent after exclusion of studies with higher risk of bias, suggesting that no individual study had a disproportionate influence on the overall results. Although the magnitude of effect varied slightly across analyses, the overall conclusions remained unchanged.

Discussion

This systematic review and meta-analysis suggests that awake prone positioning significantly improves key clinical outcomes in non-intubated adult patients with acute hypoxemic respiratory failure. The pooled analysis showed that awake prone positioning was associated with a significant reduction in mortality, intubation, invasive mechanical ventilation, and both hospital and ICU length of stay compared with usual care, indicating its effectiveness in preventing disease progression and improving patient recovery. These findings suggest that awake prone positioning enhances respiratory function sufficiently to reduce the need for invasive interventions and shorten the duration of hospitalization and critical care. Importantly, awake prone positioning did not increase the risk of adverse events, including skin breakdown, vomiting, intravascular line dislodgement, or cardiac arrest, confirming its safety and tolerability. However, no significant differences were observed in ICU admission, escalation of respiratory support, or time to invasive ventilation, which may reflect variations in patient characteristics and clinical management rather than lack of therapeutic benefit.

Our findings are broadly consistent with the recent meta-analysis by Karanth et al., [44] which included 22 RCTs comprising 3,615 patients and demonstrated that awake prone positioning (APP) significantly reduced both intubation (RR 0.80, 95% CI 0.72–0.90) and mortality (RR 0.86, 95% CI 0.74–0.99). Similarly, our pooled analysis showed a significant reduction in intubation (OR 0.69, 95% CI 0.60–0.79) and mortality (OR 0.60, 95% CI 0.42–0.86), reinforcing the protective effect of APP on clinically meaningful outcomes. Notably, Karanth et al. reported no heterogeneity for intubation (I² = 0%) and mortality (I² = 0%), whereas our analysis demonstrated moderate heterogeneity for mortality (I² = 72%) but minimal heterogeneity for intubation (I² = 6%), possibly reflecting differences in included populations, settings, and analytic approach. Furthermore, Karanth et al. identified a significant cultural effect, with stronger intubation reduction in high Power Distance Index (PDI ≥ 80) countries (RR 0.67) and no significant benefit in lower-PDI settings, suggesting that adherence and implementation fidelity may influence outcomes [44].

Our findings both align with and expand upon the meta-analysis by Cheema et al., [45] which included 11 randomized controlled trials and demonstrated that awake prone positioning significantly reduced the risk of intubation (RR 0.84, 95% CI 0.74–0.95) but did not significantly reduce mortality (RR 0.93, 95% CI 0.77–1.11). In contrast, our analysis not only confirmed a significant reduction in intubation (OR 0.69, 95% CI 0.60–0.79) but also demonstrated a statistically significant reduction in mortality (OR 0.60, 95% CI 0.42–0.86), suggesting a stronger overall survival benefit. Additionally, while Cheema et al. reported no significant differences in hospital length of stay, ICU stay, or escalation of respiratory support, our pooled analysis showed significant reductions in both hospital length of stay (MD − 0.70 days) and ICU stay duration (MD − 2.84 days), as well as a significant reduction in the requirement for invasive mechanical ventilation (OR 0.42). These differences likely reflect the inclusion of a larger number of studies and more recent trials in our analysis, which provided greater statistical power and more comprehensive outcome assessment [45].

Our findings also expand upon the meta-analysis by Wang et al., [46] which included 10 randomized controlled trials with 2294 patients and reported that awake prone positioning significantly reduced the risk of intubation (RR 0.84, 95% CI 0.74–0.95) but did not demonstrate a significant reduction in mortality (RR 0.93, 95% CI 0.77–1.11). These results are partially consistent with our analysis, which similarly confirmed a significant reduction in intubation risk, although with a stronger effect size (OR 0.69, 95% CI 0.60–0.79). However, unlike Wang et al., our study demonstrated a statistically significant reduction in mortality (OR 0.60, 95% CI 0.42–0.86), suggesting that the survival benefits of awake prone positioning may be more substantial than previously recognized. Furthermore, while Wang et al. did not observe significant differences in mortality or other clinical recovery outcomes, our analysis identified additional benefits, including significant reductions in hospital length of stay, ICU stay duration, and the need for invasive mechanical ventilation. Both analyses consistently found no increase in adverse events associated with awake prone positioning, supporting its safety profile. The differences in mortality and recovery outcomes between the two analyses may be explained by the inclusion of a larger number of studies, more recent high-quality randomized trials, and a substantially greater sample size in our meta-analysis, which enhanced the statistical power to detect clinically meaningful differences [46].

Our findings are also consistent with the meta-analysis by Graziani et al., [47] which included 34 studies involving 6808 patients and demonstrated that awake prone positioning significantly reduced mortality (OR 0.60, 95% CI 0.46–0.79) and the composite outcome of death or intubation (OR 0.73, 95% CI 0.60–0.89). These results closely align with our analysis, which likewise showed a significant reduction in mortality (OR 0.60, 95% CI 0.42–0.86), reinforcing the growing evidence that awake prone positioning may confer a meaningful survival benefit. However, unlike Graziani et al., [47] who did not observe a statistically significant reduction in intubation overall (OR 0.85, 95% CI 0.56–1.27), our study demonstrated a significant reduction in intubation risk (OR 0.69, 95% CI 0.60–0.79), suggesting a more consistent protective effect against respiratory deterioration. Additionally, while their analysis focused primarily on mortality and intubation outcomes, our study further demonstrated significant reductions in hospital length of stay, ICU stay duration, and invasive mechanical ventilation requirements, providing broader evidence of clinical and healthcare system benefits [47].

Awake prone positioning was not applied uniformly across the included studies, with considerable variability in the duration, timing of initiation, and type of concurrent respiratory support. These factors are likely to influence the effectiveness of the intervention, as prolonged or early application may enhance physiological benefits, while differences in respiratory support modalities (e.g., high-flow nasal cannula or non-invasive ventilation) may modify treatment response. The lack of standardized protocols across studies may therefore contribute to the observed variability in outcomes and should be considered when interpreting the pooled results.

The findings of this meta-analysis have important clinical implications for the management of non-intubated patients with acute hypoxemic respiratory failure. Awake prone positioning represents a simple, low-cost, and non-invasive intervention that can be easily implemented in a wide range of clinical settings, including general wards, high-dependency units, and intensive care units [44]. Its ability to significantly reduce mortality, intubation, and the need for invasive mechanical ventilation highlights its potential to prevent clinical deterioration and improve patient outcomes. Furthermore, the observed reduction in hospital and ICU length of stay suggests that awake prone positioning may help alleviate the burden on healthcare systems by reducing resource utilization and improving patient flow [48]. Importantly, the absence of increased adverse events supports its safety and feasibility, making it suitable for routine clinical practice when appropriately monitored. Awake prone positioning may be particularly valuable in resource-limited settings or during periods of increased healthcare demand, where reducing the need for invasive mechanical ventilation and ICU admission is critical [49]. These findings support the early and routine use of awake prone positioning as part of standard supportive care for patients with acute hypoxemic respiratory failure.

Although awake prone positioning was associated with a statistically significant reduction in hospital length of stay, the magnitude of this effect (− 0.70 days) is relatively modest and may have limited clinical relevance at the individual patient level. However, even small reductions in length of stay may translate into meaningful benefits at the healthcare system level, particularly during periods of high resource utilization.

The high heterogeneity observed in outcomes such as ICU and hospital length of stay may be explained by several factors. Differences in baseline patient severity across studies, variability in adherence to awake prone positioning protocols, and inconsistency in the duration and timing of the intervention are likely contributors. Additionally, variations in healthcare settings and discharge practices may have influenced length-of-stay outcomes. Despite this heterogeneity, sensitivity analysis confirmed the robustness of the overall findings, suggesting that the observed effects are consistent across different study conditions. The substantial heterogeneity observed in outcomes such as ICU stay, invasive mechanical ventilation, and hospital length of stay is likely multifactorial. Differences in baseline patient severity, variations in clinical setting (ICU versus general wards), and heterogeneity in respiratory support strategies may have contributed to the observed variability. In addition, adherence to awake prone positioning protocols and the duration and timing of the intervention were inconsistently reported and likely varied significantly across studies, further contributing to heterogeneity. These factors highlight the complexity of interpreting pooled estimates in this context. The substantial heterogeneity observed in some outcomes, particularly hospital and ICU length of stay, may be explained by differences in patient populations, disease severity, respiratory support strategies, prone positioning duration, adherence to APP protocols, and variations in healthcare systems across studies.”

Funnel plot asymmetry observed in several outcomes suggests the presence of potential publication bias. This may reflect the preferential publication of studies demonstrating positive effects of awake prone positioning, which could lead to an overestimation of the true treatment effect. Although formal statistical tests such as Egger’s regression or trim-and-fill analysis can be used to further assess publication bias, their reliability may be limited in the presence of heterogeneity and a relatively small number of studies for certain outcomes. Therefore, the possibility of publication bias should be considered when interpreting the results.

Notably, the effects of awake prone positioning were not consistent across all outcomes. While significant improvements were observed in mortality and intubation, no significant differences were found in ICU admission, escalation of respiratory support, or time to invasive ventilation. These discordant findings suggest that the impact of awake prone positioning on disease progression may be more complex and not uniformly reflected across all clinical endpoints. Therefore, caution is warranted in attributing a direct mechanistic effect, and the observed benefits may be influenced by multiple interacting clinical factors.

Strengths and limitations

This meta-analysis has several important strengths. It included a large sample size of 6,164 patients from 24 studies, most of which were randomized controlled trials with overall low risk of bias, enhancing the reliability and validity of the findings. The analysis evaluated multiple clinically relevant outcomes, including mortality, intubation, invasive mechanical ventilation, and hospital and ICU length of stay, providing a comprehensive assessment of both effectiveness and safety. Additionally, the inclusion of recent large multicenter trials improves the relevance and generalizability of the results to current clinical practice.

However, several limitations should be acknowledged. Significant heterogeneity was observed for some outcomes, likely due to differences in patient populations, clinical settings, and prone positioning protocols. Variations in the duration, timing, and adherence to awake prone positioning across studies may have influenced treatment effects. Most studies were conducted during the COVID-19 pandemic, which may limit generalizability to other causes of acute hypoxemic respiratory failure. Finally, evidence of possible publication bias was observed, which may have influenced the pooled estimates. Subgroup analyses were not feasible due to inconsistent and incomplete reporting of key variables across the included studies. Important factors such as underlying etiology (COVID-19 vs. non-COVID-19), type of respiratory support (high-flow nasal cannula, non-invasive ventilation, or conventional oxygen), duration and timing of awake prone positioning, and differences in clinical setting (ICU vs. ward) were not uniformly reported, limiting the ability to perform reliable subgroup analyses. This restricts a more detailed exploration of potential sources of heterogeneity and effect modifiers.

Most included studies were conducted in COVID-19 populations, which may limit the generalizability of the findings to patients with non-COVID acute hypoxemic respiratory failure.

Although subgroup analyses based on factors such as duration of awake prone positioning, severity of hypoxemia, respiratory support modality, and ICU versus non-ICU setting would have been clinically valuable, these analyses were not feasible because several relevant variables were inconsistently reported or insufficiently detailed across the included studies. Additionally, substantial variability in intervention protocols and outcome reporting limited the reliability of subgroup-based pooled analyses.

Although funnel plots were visually assessed for primary outcomes, formal statistical tests for publication bias were not consistently performed because several outcomes included a limited number of studies, reducing the reliability of such analyses. The observed asymmetry in funnel plots for some primary outcomes suggests the possibility of publication bias, which may have resulted in overestimation of the beneficial effects of awake prone positioning. Additionally, studies reporting positive outcomes may have been more likely to be published during the COVID-19 pandemic, further contributing to potential reporting bias.

Conclusion

This systematic review and meta-analysis suggests that awake prone positioning may be associated with improved clinical outcomes in non-intubated adult patients with acute hypoxemic respiratory failure; however, the findings should be interpreted cautiously due to heterogeneity, variability in intervention protocols, and possible publication bias. Awake prone positioning was associated with significant reductions in mortality, intubation, invasive mechanical ventilation, and both hospital and ICU length of stay, without increasing adverse events. These findings support awake prone positioning as a safe, effective, and practical intervention that can reduce disease progression and improve recovery. Given its simplicity, low cost, and favorable safety profile, awake prone positioning may be considered as part of supportive care in appropriately selected patients. Future large, well-designed randomized trials are needed to determine the optimal timing, duration, and patient selection to maximize its clinical benefit.

Abbreviations

APP

Awake prone positioning

AHRF

Acute hypoxemic respiratory failure

ARDS

Acute respiratory distress syndrome

CI

Confidence interval

COVID-19

Coronavirus disease 2019

HFNC

High-flow nasal cannula

ICU

Intensive care unit

MD

Mean difference

NOS

Newcastle–Ottawa Scale

OR

Odds ratio

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

RCT

Randomized controlled trial

ROB 2

Risk of Bias tool version 2

RR

Risk ratio

Authors’ contributions

M.A., M.H.E.D.M., and H.A.A contributed to the conceptualization and methodology of the study. M.A. MHEDM, HAA, performed data curation and drafted the original manuscript. M.H.E.D.M. contributed to formal analysis and software, while H.B., O.Y.A., and D.R.F.A.-J. were involved in data curation, investigation, and validation. I.M.A. contributed to formal analysis and visualization. All authors contributed to writing—review and editing of the manuscript. All authors approve the final version of the manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

All data analyzed during this study are included in the published articles cited in this review. Extracted data and analysis files are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.