University of Michigan
Department of Pediatrics
Evidence-Based Pediatrics Web Site

Question

• Will the use of peri-extubation steroids decrease the rate of re-intubation in critically ill intubated children?

Clinical Bottom Lines

1. Peri-extubation steroids show a trend toward reduced rates of re-intubation in neonates, but no clear statistically significant benefit.(RR=0.1, 95% CI .01-1.68)¹
2. The use of peri-extubation steroids in children reduces the incidence of stridor.(RR=0.53, 95% CI of 0.28-0.97, NNT=21). However, there is only a trend that this treatment reduces re-intubation rates. (RR=0.49, 95% CI 0.01-19.65)
3. There is no indication for the use of peri-extubation steroids in children older than 15 or adults to reduce re-intubations or stridor (RR=0.95, 95 % CI 0.52-1.72., and RR=0.86, 95% CI 0.57-1.3)
   

Summary of Key Evidence

1. Cochrane review meta-analysis of seven randomized controlled trials comparing steroids to placebo in neonates, children and adults.¹
2. Three age groups were examined separately, decided upon prior to start of data collecting, as authors felt there was significant anatomical and physiological differences between these groups that they warranted separate review.
3. Heterogeneity in the studies explained by differences in age, severity of illness and duration of steroid use prior to extubation.
4. They excluded studies of patients with known tracheitis, laryngitis, laryngotracheobronchitis, external or surgical trauma to the larynx or subglottis, and studies which failed to clearly document reasons for re-intubating patients.
5. Of the 251 studies identified, only seven met the criteria to be included in the analysis.
6. All studies except one examined steroid use both before and after extubation, while one article investigated treating with steroids in patients' with stridor following extubation. There was a wide variation in peri-extubation steroid protocols, some administering only one dose 30 min prior to extubation and others using multiple doses starting 12 hours prior and continuing for up to 36 hours.
7. Study was valid in that it had an exhaustive search for articles, the articles chosen were of high methodologic quality and weighted on both the quality and size of the studies. There was reproducibility in the assessments of the studies.
8. Meta-analysis limited by the small number of available studies and the small sample size within the individual studies.
   

Additional Comments

• Although review states primary data was rate of re-intubation they also explored incidence of stridor as a secondary outcome. This too is relevant to patient care as up to 37% of extubated patients will have stridor. This lengthens ICU stays and increases the need for vasoconstrictive medications.
• Studies separated by age groups prior to onset of analysis as authors felt the differences in the anatomy and physiology of the airway was sufficient to warrant separate reviews. This was based on the fact that the narrowest portion of the airway in neonates and children is at the level of the cricoid cartilage, as opposed to the vocal cords in adults. This is why cuffed tubes are used in adults but not in children. There is also the difference in absolute size of the airway. As it is significantly smaller in neonates and children, any small changes secondary to swelling can lead to large changes in the total diameter. This in turn leads to increased risks of stridor and airway obstruction necessitating re-intubation.
• Airway edema is a common complication of intubations, and for this reason, patients with anatomical abnormalities compromising their airway may actually benefit more from this intervention, as small amounts of edema may have even greater effects on the airway.
• In previous studies that did not specify obstruction as the reason for re-intubation, there is evidence that neonatal patients do benefit from peri-extubation steroids. This may be a reflection of the benefits of steroids on pulmonary compliance rather than on airway edema.
• There was very limited discussion of the side effects or potential risks of steroid use in this population of critically ill patients. Only one article addressed this as a component of the study and found no significant increase in ill effects compared to placebo.

Citation

CAT Author: Heather Van Mater, MD

CAT Appraisers: John G. Frohna, MD, MPH

Date appraised: November 12, 2003