22.06.2020
Feasibility and physiological effects of prone positioning in non-intubated patients with acute respiratory failure due to COVID-19 (PRON-COVID): a prospective cohort study
Therapeutic PulmonologyAnesthesia-intensive care
Coppo A et al
Lancet Respir Med
Main result
- Inclusion of 56 patients with 79% male, mean age = 57.4 years and mean BMI = 27.5
- Prone positioning feasible (maintained for at least 3 hours) in 83.9% of patients
- Oxygenation improved significantly from the supine to the prone position: PaO2/FiO2 ratio 180.5 mm Hg in the supine position compared to 285.5 mm Hg in the prone position (p<0.0001).
- Resupination (transition from prone to supine position) allowed the maintenance of good oxygenation in 50% of patients, without significant improvement (PaO2/FiO2 ratio 192.9 mm Hg 1 h after resupination; p=0.29).
- Patients who maintained increased oxygenation had increased levels of inflammatory markers: CRP=12.7 mg/L in responders vs. 8.4 mg/L in non-responders, platelets=241.1 vs. 319.8, and a shorter time from hospital admission to the supine position (2.7 days in responders vs. 4.6 days in non-responders).
- 28% of the 46 patients were intubated, with 30% responders versus 26% non-responders (p=0.74). Five patients died during follow-up due to underlying disease unrelated to the study procedure.
Takeaways
Prone positioning was feasible in 84% of patients and significantly more effective in rapidly improving blood oxygenation in oxygen-requiring awake patients with COVID-19 pneumonia. The effect was maintained after procubitus in half of the patients. Further studies are needed to determine the potential benefit of this technique in ultimately improving overall respiratory and clinical outcomes.
Strength of evidence Weak
- Monocentric retrospective study on small numbers
- Non-consecutive inclusions and broad and subjective exclusion criteria that discuss possible selection bias
- Electronic data collection without quality control mentioned
- Questionable robustness of a mixed model on this small population, no consideration of the multiplicity of statistical tests or precision on the management of missing data
Objectives
To study the feasibility and effect on gas exchange of prone positioning in awake non-intubated patients with COVID-19 pneumopathy
Method
- Prospective monocentric cohort feasibility study conducted in a hospital in Northern Italy
- Non-consecutive inclusion of patients over a 15-day period if less than 75 years of age, confirmed diagnosis of COVID-19 oxygen-requiring pneumonia without intubation, excluding patients with chronic respiratory or cardiac disease.
- Intervention: after receiving basic demographic and clinical characteristics, especially respiratory, basic, positioning of the patient in prone position. Collect clinical data at 10 min, then maintain the procubitus for at least 3 hours if possible before returning to the supine position. Re-collect clinical data one hour after returning to this position. If it was not possible to maintain the prone position for 3 hours, then it was considered not feasible and the reason was collected. Data were collected from the electronic record.
- Primary endpoint = change in oxygenation (arterial partial pressure of oxygen [PaO2]/FiO2) between baseline and 10 min after supine = pulmonary recruitment index.
- Secondary endpoints = safety and feasibility of the ventral decubitus, effect of the procubitus on partial CO2 pressure, on dyspnea, on predictors of response to the procubitus knowing that the good responders were those with an increased PaO2/FiO2 ratio.
- Statistics: The headcount cacul is based on a reasonable and documented hypothesis giving the result of 40 subjects to be included, with a correction factor of 20% to take into account the infeasibility of the technique. Use of numerous tests of non-parametric two-by-two comparisons, then a mixed model with random effect. Finally, performing sensitivity analyses by changing the threshold (percentage) for the increase in the PaO2/FiO2 ratio to consider a subject as a responder.
bibliovid.org and its content are bibliovid property.