Timing of tracheostomy and patient outcomes in critically ill patients requiring extracorporeal membrane oxygenation: a single-center retrospective observational study

BACKGROUND: Extracorporeal membrane oxygenation (ECMO) is an integral method of life support in critically ill patients with severe cardiopulmonary failure; however, such patients generally require prolonged mechanical ventilation and exhibit high mortality rates. Tracheostomy is commonly performed in patients on mechanical ventilation, and its early implementation has potential advantages for favorable patient outcomes. This study aimed to investigate the association between tracheostomy timing and patient outcomes, including mortality, in patients requiring ECMO. METHODS: We conducted a single-center retrospective observational study of consecutively admitted patients who were supported by ECMO and underwent tracheostomy during intensive care unit (ICU) admission at a tertiary care center from April 2014 until December 2021. The primary outcome was hospital mortality. Using the quartiles of tracheostomy timing, the patients were classified into four groups for comparison. The association between the quartiles of tracheostomy timing and mortality was explored using multivariable logistic regression models. RESULTS: Of the 293 patients treated with ECMO, 98 eligible patients were divided into quartiles 1 (≤ 15 days), quartile 2:16–19 days, quartile 3:20–26 days, and 4 (> 26 days). All patients underwent surgical tracheostomy and 35 patients underwent tracheostomy during ECMO. The complications of tracheostomy were comparable between the groups, whereas the duration of ECMO and ICU length of stay increased significantly as the quartiles of tracheostomy timing increased. Patients in quartile 1 had the lowest hospital mortality rate (19.2%), whereas those in quartile 4 had the highest mortality rate (50.0%). Multivariate logistic regression analysis showed a significant association between the increment of the quartiles of tracheostomy timing and hospital mortality (adjusted odds ratio for quartile increment:1.55, 95% confidence interval 1.03–2.35, p for trend = 0.037). CONCLUSIONS: The timing of tracheostomy in patients requiring ECMO was significantly associated with patient outcomes in a time-dependent manner. Further investigation is warranted to determine the optimal timing of tracheostomy in terms of mortality. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s40560-022-00649-w.