Best evidence in critical care medicine. Steroids to prevent post-extubation airway obstruction in adult critically ill patients.

Structured abstract Background: The efficacy of corticosteroids for reducing the incidence of post-extubation laryngeal edema remains controversial. Objective: To evaluate whether methylprednisolone (MP) started 12 hr prior to planned tracheal extubation, can prevent post-extubation laryngeal edema. Design: Prospective, randomized, blinded, placebocontrolled trial performed in 15 intensive care units (ICU) across France. Methods: Inclusion criteria were adult (age >18 yr) patients who were scheduled for planned tracheal extubation after having received mechanical ventilation for > 36 hr. Exclusion criteria included pregnancy, history of post-extubation upper-airway obstruction, primary ICU admission diagnosis of throat disease or surgery, tracheostomy, and those who had been chronically treated with non-steroidal anti-inflammatory drugs or corticosteroids. Patients were randomly allocated to receive either corticosteroid or placebo. Allocation was stratified by site, and was concealed in sequentially numbered sealed opaque envelopes. Analyses were performed on a per-protocol and intention-to-treat basis. Intervention: Patients were allocated to either intravenous MP 20 mg initiated 12 hr prior to planned tracheal extubation, and continued every four hours thereafter with the final dose given immediately prior to extubation (total dose 80 mg) or identical volume of intravenous placebo (normal saline). Primary endpoint: The primary endpoint was occurrence of laryngeal edema within 24 hr of planned tracheal extubation. Laryngeal edema was clinically defined as the development of upper-airway obstruction after extubation, and further classified as either minor or major. Minor laryngeal edema was defined as evidence of stridor associated with respiratory distress requiring medical intervention. Major laryngeal edema was defined as severe distress prompting tracheal re-intubation, and confirmed with direct visualization by laryngoscopy. Results: 761 patients were randomized, but 63 could not be assessed due mostly to self-extubation (n = 16) or cancelled extubation (n = 44) between the time of randomization and planned tracheal extubation. In total, 698 patients were included in the final analysis (n = 355, in the MP group; n = 343, in the placebo group). The median duration of intubation prior to randomization was six days (intra-quartile range 4–11). The groups were similar at baseline apart from a higher rate of nasotracheal intubation in those allocated to MP. Methylprednisolone significantly reduced the incidence of post-extubation laryngeal edema (3% vs 22%, P < 0.0001), the global incidence of tracheal re-intubations (4% vs 8%, P = 0.02), and the proportion of re-intubations secondary to laryngeal edema (8% vs 54%, P = 0.005). Methylprednisolone did not affect the severity of laryngeal edema or the delay of occurrence relative to extubation. Similar findings were evident by intention-to-treat analysis of assessed patients [3% (11/358) vs 22% (76/345), P < 0.0001]. By multivariable analysis, several factors were found associated with increased risk of laryngeal edema, including female sex, shorter height-toendotracheal tube diameter ratio, ICU admission for trauma, and tracheal intubation for a shorter duration. No significant adverse effects were reported. Conclusions: Methylprednisolone 20 mg iv given every four hours, commencing 12 hr before a planned tracheal extubation, substantially reduced the incidence of post-extubation laryngeal edema and the rate of re-intubation. Such treatment should be considered in adult patients before a planned extubation that follows a period of tracheal intubation of more than 36 hr. Source of funding: The authors report that the sponsor of the study had no role in study design, data Best Evidence in Critical Care Medicine