Early tracheostomy versus Prolonged Endotracheal Intubation in Severe Head Injury

 

Bouderka et al. J Trauma: Injury, Infection & Critical Care 2004;57(2): 251-254. doi: 10.1097/01.TA.0000087646.68382.9A

Clinical Question

• Does early tracheostomy reduce duration of mechanical ventilation and mortality in comparison with prolonged intubation in patients with head injury?

Design

• Randomised Controlled Trial
• Non-blinded
• With estimated reduction in duration of ventilation from intervention of 25%, alpha of 0.05, 34 patients would give the study 80% power

Setting

• Single ICU, Morocco

Population

• Isolated severe head injury, with admission GCS <9, and GCS <8 without sedation on day 5
• 62 patients randomised

Intervention

• Early tracheostomy (Day 5/6) performed by critical care physician

Control

• Prolonged endotracheal intubation

Outcome

• Comparing early tracheostomy vs. prolonged endotracheal intubation
  • Primary outcome: length of mechanical ventilation (days)
    • 14.5 ±7.3 vs. 17.5 ±10.6, P=0.02
  • Secondary Outcomes
    • Mortality: 38.7% vs. 22.5%, P=0.27
    • Nosocomial pneumonia: 58% vs. 61.3%, P=0.79

 

Authors’ Conclusion

• Early tracheostomy decreased the length of mechanical ventilation but not mortality

Strengths

• Appropriate inclusion criteria with homogenous population
• Randomised
• Appropriate power calculation

Weaknesses

• 150 patients with severe head injury had been admitted to the ICU but only 62 randomised. Only reported exclusion criteria for 6 patients.
• Non-blinded. (Impractical to blind treating clinicians but could have blinded staff analysing the data)
• The statistics for the primary outcome are questionable. With only a small number of patients, a small difference found between treatment groups and a large range for the duration of ventilation, a P value of 0.02 is questionable. The author of the study was contacted for further information so that we could verify the statistics. Unfortunately this was not possible as the original data has now been lost.