Noninvasive respiratory support for postoperative patients in Intensive Care Unit: a systematic review and network meta‑analysis ; Support respiratoire non invasif post opératoire en unité de soins intensifs : revue systématique et méta analyse

Background: postoperative respiratory complications are a major cause of morbidity and mortality. Discrepancies were found in recent randomized controlled trials (RCTs) regarding the place of noninvasive respiratory support strategies for postoperative patients in Intensive Care Unit (ICU). Methods: in this systematic review and network meta-analysis, we searched MEDLINE, Cochrane Center Register of Controlled Trials and Web of Science for RCTs evaluating either noninvasive positive pressure ventilation (NIPPV), high-flow nasal oxygenation (HFNO), conventional oxygen therapy (COT), or NIPPV+HFNO for postoperative patients in ICU. Primary outcome was 28-day mortality. Secondary outcome was 7-day reintubation. We generated pooled relative risks (RR) and numbers needed to treat (NNT). We assessed risk of bias using the Cochrane 2.0 tool. The study was registered on PROSPERO (CRD XXXX). Findings: in 23 RCTs including 9,018 patients, NIPPV (RR 0·61 [95%confidence interval (CI) 0·47–0·81], NNT=26, high confidence) was the only strategy that significantly reduced 28-day mortality compared to COT. Compared to HFNO, neither NIPPV (RR 0·79 [95%CI 0·52–1·22]) nor NIPPV+HFNO (RR 0·87 [95%CI 0·34–2·21]) reduced 28-day mortality. NIPPV was significantly associated with lower 7-day reintubation compared to COT (RR 0·52 [95%CI 0·37–0·74], NNT=14, moderate confidence) and to HFNO (RR 0·58 [95%CI 0·34–0·99], NNT=16, low confidence). Interpretation: the results suggest that compared to COT, NIPPV reduces mortality and reintubation, in postoperative patients admitted to ICU. The NNT to avoid one death with NIPPV was of 26. NIPPV might be superior to HFNO in preventing reintubation in postoperative patients admitted to Intensive Care Unit. Future studies should further evaluate NIPPV+HFNO. ; Contexte : les complications respiratoires postopératoires sont une cause majeure de morbidité et de mortalité. Des divergences ont été observées dans les essais cliniques randomisés récents concernant l'utilisation des stratégies de support ...