The optimal timing of tracheostomy in nonneurologically injured mechanically ventilated critically ill adult patients is uncertain. We conducted a systematic review of randomized controlled trials to evaluate the effect of early versus late tracheostomy or prolonged intubation in this population.
We searched MEDLINE, Embase, CENTRAL, CINAHL, and Web of science databases for randomized controlled trials comparing early tracheostomy (<10 d of intubation) with late tracheostomy or prolonged intubation in adults.
We selected trials comparing early tracheostomy (defined as being performed less than 10 d after intubation) with late tracheostomy (performed on or after the 10th day of intubation) or prolonged intubation and no tracheostomy in nonneurologically injured patients. The primary outcome was overall mortality. Secondary outcomes included ventilator-associated pneumonia, duration of mechanical ventilation, ICU, and hospital length of stay.
Two reviewers screened citations, extracted data, assessed the risk of bias, and classification of Grading of Recommendations, Assessment, Development, and Evaluation independently.
Our search strategy yielded 8,275 citations, from which nine trials (n = 2,457) were included. We did not observe an effect on the overall mortality of early tracheostomy compared with late tracheostomy or prolonged intubation (risk ratio, 0.91, 95% CI, 0.82–1.01; I2 = 18%). Our results were consistent in all subgroup analyses. No differences were observed in ICU and hospital length of stay, duration of mechanical ventilation, incidence of ventilator-acquired pneumonia, and complications. Our trial sequential analysis showed that our primary analysis on mortality was likely underpowered.
In our systematic review, we observed that early tracheostomy, as compared with late tracheostomy or prolonged intubation, was not associated with a reduction in overall mortality. However, we cannot exclude a clinically relevant reduction in mortality considering the level of certainty of the evidence. A well-designed trial is needed to answer this important clinical question.