COMPARISON OF EARLY VERSUS LATE TRACHEOSTOMY IN PATIENTS WITH PROLONGED MECHANICAL VENTILATION ON MORTALITY AND COMPLICATIONS : A SYSTEMATIC REVIEW OF RANDOMIZED CONTROLLED TRIALS AND PRIMARY STUDIES

Abstract

Introduction: Prolonged invasive mechanical ventilation is a major determinant of morbidity, mortality, and resource consumption in the ICU. Tracheostomy is performed to facilitate ventilator weaning, reduce sedation requirements, and decrease laryngotracheal injury; however, its optimal timing remains debated. This study compares the effects of early tracheostomy (ET) and late tracheostomy (LT) on mortality and complications.

Methods: Systematic review conducted in accordance with PRISMA 2020. Included designs: RCTs, cohort, case-control, cross-sectional, and observational studies; systematic reviews were excluded from the characteristics table but used as comparators. Risk of bias: RoB 2 and NOS. Certainty of evidence: GRADE. Structured narrative and semi-quantitative synthesis.

Results: Twenty-three primary studies (>33,000 patients; 12 RCTs, 11 observational) were included. ET did not reduce 30-day mortality (TracMan: 30.8% vs 31.5%; ARR 0.7%; 95% CI −5.4 to 6.7). ET consistently improved process outcomes: shorter sedation duration (MD −2.78 days; 95% CI −3.68 to −1.88), increased ventilator-free days (+2.12 days; 95% CI 0.94–3.30), shorter duration of mechanical ventilation (−2.77 days; 95% CI −5.10 to −0.44), and shorter ICU length of stay (−6.36 days; 95% CI −9.84 to −2.88). VAP incidence decreased in high-risk subpopulations (OR 0.60–0.86). Procedural complications did not differ (5.5% vs 7.8%). The proportion of patients who underwent tracheostomy was markedly higher with ET (91.9% vs 44.9%; OR 16.1).

Discussion: The discordance between randomized and observational evidence is explained by immortal time bias and confounding by indication. Failure of clinical prediction of the need for prolonged ventilation is the principal determinant of why ET fails to demonstrate a mortality benefit. Bayesian analysis indicates a 99% posterior probability of benefit for short-term mortality, suggesting that "no benefit" conclusions based on frequentist testing may be overly conservative. In neurological and severe trauma populations, the benefit–risk balance shifts in favour of ET.

Conclusion: ET confers no mortality benefit in the general ICU population, but consistently improves sedation duration, ventilator-free days, duration of ventilation, and ICU length of stay without an increase in complications. An ET strategy should be applied selectively in patients with a high probability of requiring ventilation >14 days—particularly severe brain injury, severe stroke, high cervical spinal cord injury, and severe trauma—rather than as a universal policy.