In COVID-19 acute hypoxemia, awake prone positioning vs. usual care did not reduce intubation at 30 d

Abstract

Question:

In adults with COVID-19 and acute hypoxemia who are not intubated, does awake prone positioning vs. usual care without awake prone positioning reduce intubation rates at 30 days?

Design:

Randomized controlled trial (COVI-PRONE [Awake Prone Position in Hypoxemic Patients with Coronavirus Disease 19]).

Blinding:

Treatment allocation concealed; unblinded.^*

Setting:

21 hospitals in Canada, Kuwait, Saudi Arabia, and the USA.

Patients:

400 patients aged ≥18 years (mean age, 58 y; 71% men) who were treated in an intensive care or monitored acute care unit, had suspected or confirmed COVID-19, needed ≥40% oxygen or noninvasive positive pressure ventilation, and were not intubated. Key exclusions: contraindication to, or risk for complications from, prone positioning; pretrial self-prone positioning; or need for immediate intubation.

Interventions:

Protocolized awake prone positioning for a target 8 to 10 h/d, with 2 to 3 breaks of 1 to 2 hours each if needed (n = 205); or usual care without awake prone positioning (n = 195).

Funding:

Canadian Institute of Health Research; King Abdullah International Medical Research Center in Saudi Arabia; McMaster University.

Results:

Awake prone positioning vs. usual care in adults with COVID-19 and acute hypoxemia who were not intubated (intention-to-treat analysis)

Outcomes	Event rates		RRR (95% CI)†
	Awake prone positioning	Usual care
Endotracheal intubation at 30 d ‡	34%	41%	13% (−7 to 29)
Mortality at 60 d	22%	24%	6% (−33 to 35)
	Mean values		Mean difference (CI)
Ventilation-free§ days at 30 d	21	19	2.04 (−0.50 to 4.59)
Hospital-free days at 60 d	34	31	3.52 (−1.05 to 8.08)
Intensive care unit-free days at 60 d	39	35	4.07 (−0.67 to 8.81)
Adverse events occurred in 10% of patients in the awake prone positioning group vs. 0% in the usual care group; no serious adverse events occurred in either group. In prespecified subgroup analysis, awake prone positioning vs. usual care reduced endotracheal intubation at 30 d in patients with Spo2:Fio2 ratio >150 at baseline (n = 162; 15% vs. 32%; hazard ratio, 0.44 [95% CI, 0.23 to 0.87]) but not in those with Spo2:Fio2 ratio ≤150 (n = 238; Fio2 ratio ≤150 (n = 238; 47% vs. 46%), P for interaction = 0.03; treatment effects did not differ for other subgroups based on age, sex, or respiratory support at baseline.

Abbreviations defined in Glossary. Primary outcome indicated by boldface.

^†RRR and CI calculated using usual care event rates and risk ratio (endotracheal intubation) or hazard ratio (mortality) reported in article.

^‡400 patients were needed to detect an absolute 13.5% reduction in endotracheal intubations with awake prone positioning from 45% with usual care (80% power, 2-sided α = 0.05).

^§Invasive mechanical ventilation or noninvasive ventilation.

Commentary:

Prone positioning is part of routine management in intubated patients without COVID-19 who have acute respiratory distress syndrome (ARDS) because of its mortality benefit (1) and is recommended for patients with COVID-19 and refractory hypoxemia receiving mechanical ventilation (2). For refractory hypoxemia in patients not receiving mechanical ventilation, awake prone positioning is not recommended to avoid intubation (2).

Although Alhazzani and colleagues found that awake self-prone positioning did not reduce intubations, COVI-PRONE offers important insights. First, median prone duration achieved was 4.8 h/d, which was shorter than in trials of intubated patients with ARDS (3) but longer than in other trials of awake prone positioning in COVID-19. An extended duration of prone positioning may be needed for clinical benefit. Second, having adequate power to detect clinical benefit or harm is difficult in studies of critically ill patients. This is exacerbated in COVID-19 management trials that must estimate potential effect sizes and sample sizes with limited previous research; the inflated sample size in COVI-PRONE may still have lacked power. A subgroup analysis in COVI-PRONE found prone positioning reduced intubation at 30 days in patients with Spo₂:Fio₂ ratio >150, suggesting a possible role for the therapy in those who are less severely ill at hospital presentation.

Another recent randomized controlled trial of awake prone positioning in nonintubated patients with COVID-19 in North America was stopped early for futility (4), suggesting COVI-PRONE results are not unique to the population studied and are generalizable to patients in the USA. Future studies evaluating awake prone positioning in COVID-19 may need to consider illness severity at presentation and duration of prone positioning. Based on COVI-PRONE, evidence is insufficient to routinely offer awake self-prone positioning to all nonventilated patients to prevent intubation, although those with higher Spo₂:Fio₂ levels may benefit.