Long-term survival of critically ill patients treated with prolonged mechanical ventilation: a systematic review and meta-analysis

Background: Prolonged dependence on mechanical ventilation after critical illness is an emerging public health challenge; however, long-term outcomes are incompletely understood. We aimed to systematically analyse long-term survival of critically ill patients treated with prolonged mechanical ventilation.

Methods: We searched PubMed, CINAHL, and the Cochrane Library between 1988 and Nov 6, 2013, with no language restrictions, for studies on prolonged mechanical ventilation. We included studies of adult populations treated with mechanical ventilation for more than 14 days, who were admitted to a ventilator weaning unit, or who had a tracheostomy for acute respiratory failure. We abstracted data with a standardised collection template and assessed study quality (ie, risk of bias) using a customised Newcastle-Ottawa Scale. We did a stratified analysis based on study setting (eg, acute vs post-acute care hospitals), and used a random-effects model to calculate pooled statistics (proportions with 95% CIs) for all outcomes. We did sensitivity analyses based on study quality (ie, high-quality studies only) and country of origin (USA vs non-USA and USA vs UK). The primary outcome was mortality at 1 year. Secondary outcomes were in-hospital mortality, discharge destination among survivors, successful liberation from mechanical ventilation while in hospital, and mortality at timepoints longer than 1 year.

Findings: Of 6326 studies identified, 402 underwent full manuscript review, and 124 studies from 16 countries met the inclusion criteria. 39 studies reported mortality at 1 year, which was 59% (95% CI 56-62). Among the 29 high-quality studies, the pooled mortality at 1 year was 62% (95% CI 57-67). Pooled mortality at hospital discharge was 29% (95% CI 26-32). However, only 19% (16-24) were discharged to home and only 50% (47-53) were successfully liberated from mechanical ventilation. For studies in post-acute care hospitals, outcomes were worse in the USA than internationally (mortality at 1 year was 73% [95% CI 67-78] in the USA vs 47% [29-65] in non-USA countries; in-hospital mortality was 31% [26-37] vs 18% [14-24]; and liberation from ventilation was 47% [42-51] vs 63% [59-68]; p<0·0001 for all).

Interpretation: Although a high proportion of patients survived to hospital discharge, fewer than half of patients survived beyond 1 year. Future studies should focus on optimum patient selection for prolonged mechanical ventilation and integration of long-term outcome information into clinical decision making.

Funding: Cooper University Health Care and Cooper Medical School of Rowan University.