Web site designed and maintained by
Chris Cairns
© SICS EBM Group 2004
|
|
|
|
Early tracheostomy 2
Lead author's name and fax: MA Bouderka: mabouderka@yahoo.fr
Three-part Clinical Question: Does early tracheostomy in severe head injury alter outcomes of death, pneumonia, ICU stay and duration of mechanical ventilation?
Search Terms: This is one of the five studies that contributed to the meta-analysis of Griffiths et al.
The Study: Non-blinded concealed randomised controlled trial with intention-to-treat analysis.
The Study Patients: Head injury patients expected to require prolonged ventilation. These patients had isolated head injury with an admission GCS <8; a GCS <8 off sedation on day 5 and cerebral contusion on CT scan. These are relatively objective criteria for entry into the study. There was no discussion of ventilation, weaning nor sedation protocols. Pneumonia was defined according to the CDC criteria. It is unclear if the diagnosis of pneumonia was independent of the investigators (ie prone to detection bias). The technique used to perform the tracheostomy was not given, although it was undertaken by a critical care physician.
Control group (N = 31; 31 analysed): Control group patients continued with trans-laryngeal intubation and did not proceed to tracheostomy at any stage.
Experimental group (N = 31; 31 analysed): These patients were randomised to receive tracheostomy on day 5 or 6.
The Evidence: (control is the prolonged trans-laryngeal intubation group)
Comments and EBM questions:
1) Do the methods allow accurate testing of the hypothesis?
Yes. The study was designed to look at ventilator time in severe head injury. It compares relatively early tracheostomy versus prolonged trans-laryngeal intubation (ie no tracheostomy). This is of course different to the early vs late tracheostomy debate, which is perhaps the issue of the meta-analysis. Ventilatory, weaning and extubation criteria were not discussed. Sedation of patients was not discussed. Few baseline criteria were given to demonstrate similarity between the groups, although age, sex and SAPS were similar. Perhaps patients here have a predominant need for airway protection and the authors discussed this. It is worth noting that an artificial airway was present for longer in the tracheostomy group (26.3 days (+/- 13.7) vs 19.4 days (+/- 10.4) P=0.03). At 5-6 days, the tracheostomy placement was not especially early.
2) Do the statistical tests correctly test the results to allow differentiation of statistically significant results?
Yes. The study was powered to a reduction in ventilator time by 25%. By this calculation, 34 patients should have been recruited. It is unclear why 62 patients were recruited in the end; there was no mention of an interim analysis. It appears that 150 patients with severe head injury were admitted during the study period, but exclusion criteria were only given for 6 patients. Statistical tests were otherwise straightforward. There was genuine randomisation. The error around the mean (presumably) was not defined (eg SD, SEM etc). There appears to be intention to treat analysis since all patients received their allotted treatment. Students t-test was used to analyse ventilation time, without demonstration of normality of the data.
3) Are conclusions valid in light of the results? Yes. The conclusions are valid in view of the stated aim of the study: tracheostomy reduced ventilator days but did not alter the incidence of death nor pneumonia. It could be debated as to how relevant this study is to the meta-analysis, although clearly that is not a weakness of this study. In terms of complications, there were a small number in each arm, without apparent differences; there was not any bronchoscopic follow up. The major potential confounding factor is perhaps continued use of sedation in those patients that continued with trans-laryngeal intubation.
4) Did results get omitted, and why?
Yes. 82 eligible patients were unaccounted for. ICU stay was a stated secondary outcome, but the results were not presented.
5) Did they suggest areas of further research?
No. The discussion section focused on a limited review of the topic.
6) Did they make any recommendations based on the results and were they appropriate? Not specifically.
7) Is the study relevant to my clinical practice?
Largely. If you look after patients with severe head injury and perform tracheostomy. Tracheostomy at 5-6 days may reduce the number of ventilated days but does not affect the other outcomes measured.
8) What level of evidence does this study represent?1-
(RCT with high risk of bias. The selected sample had a large number of drop-outs.)
9) What grade of recommendation can I make on this result alone? C
10) What grade of recommendation can I make when this study is considered along with other available evidence? This is the only prospective randomised study of its type in this patient group. There are mixed results in other patient groups (early vs late tracheostomy) and these have been subject to meta-analysis.
11) Should I change my practice because of these results?
Perhaps. This paper does not really provide convincing evidence to perform tracheostomy over prolonged trans-laryngeal intubation in patients with severe head injury. One could debate how useful a reduction in ventilator days from 17 to 14 is. The study does not answer the question about tracheostomy timing.
12) Should I audit my current practice because of these results?
Yes. It may be reasonable to audit the timing and outcome of patients with severe head injury who proceed to tracheostomy.
Appraised by: Richard Price; SpR, Dept. Anaesthetics, Gartnavel Hospital, Dumbarton Road, Glasgow, G12 0YN. Monday, 27 March 2006
©SICS EBM 2006
| |||||||||||||||||||||||||||||||||||||||||||||||||||
