Background: adult patients with asthma commonly present to the ED with exacerbations, and many are appropriate for discharge from the ED after initial treatment with a bronchodilator and a corticosteroid. Most patients are treated with 5 days of prednisone on discharge per the GINA guidelines; however, even short courses of corticosteroids can be associated with side effects. Prior investigation in children suggests that a short course of a longer-acting steroid, dexamethasone, can be just as effective as preventing symptom relapse. Does treatment with dexamethasone as compared to prednisone change the incidence of symptom relapse in adults with acute asthma exacerbation?   

Adverse effects of steroids 

PICO: does 1-2 doses of oral dexamethasone decrease symptom relapse over standard therapy of 5 days of oral prednisone in adults who are discharged from the ED after presenting with asthma exacerbation?   

Study #1: Kravitz et al. 2011.  

This was a prospective, double-blinded randomized control trial conducted at two EDs in Philadelphia, PA. Participants included adults aged 18-45 with a diagnosis of asthma established > 6 months ago and a peak expiratory flow rate of < 80% than predicted. The upper age limit of 45 was chosen in an attempt to exclude patients with COPD. The study specifically excluded patients who were admitted to the hospital due to asthma exacerbation. The objective of the study was to compare the number of patients who would return to normal function within 3 days and the frequency of relapse after acute asthma exacerbation between patients receiving 2 days of oral dexamethasone and patients receiving 5 days of oral prednisone. The ideal sample size was pre-determined; with a goal of 80% power and a two-tailed α = 0.05, 88 patients were needed in each group to detect a 15% difference in return to function within 3 days, assuming an 80% return to function within 3 days. 285 patients were enrolled and randomized, with 28 of those patients ultimately excluded because of their admission for their asthma exacerbation. 257 patients were ultimately randomized and included in the study, with 129 in the dexamethasone group and 128 in the prednisone group. All patients who were included in the study had their peak flow measured before any treatment, underwent initial treatment with 5 mg nebulized albuterol and 2.5 mg nebulized ipratropium bromide, and then had their peak flow measured again. Participants could receive additional nebulized medication at the discretion of the treating physician. Patients in the prednisone group received 5 medication packets labeled 1-5, with each packet containing 60 mg of oral prednisone. Patients in the dexamethasone group received 5 medication packets also labeled 1-5; however, packets 1-2 contained 16 mg of oral dexamethasone, and packets 3-5 contained a placebo pill. These packets were identical and prepared by the pharmacy, so neither the treating physician nor the study staff were able to discern which patient got which treatment. The first dose of medication from the packets was administered in the emergency department. The patients were then discharged home. Patients were contacted at 2 weeks after their initial presentation. Questions asked included how many days passed before they were able to return to their normal daily activities, whether there had been a relapse (repeated ED or primary care visit or admission to the hospital), and how many times per day albuterol was used per day in the week after the ED visit. 25/129 patients (19.3%) were lost to follow-up in the dexamethasone group, and 32/128 (25%) were lost to follow-up in the prednisone group. Thus, 104 patients were analyzed in the dexamethasone group, and 96 patients were analyzed in the prednisone group. These analyzed patients were ultimately similar in their baseline characteristics, including in peak flow measurements, age, sex, asthma severity score, discharge peak flow, and prior ED visits and admissions for asthma. There was no information included about the 28 patients who were randomized but excluded due to their admission, and there was no information included about the patients who were lost to follow-up. 72 patients (80%) in the prednisone group returned to normal function within 3 days compared to 91 patients (90%) in the dexamethasone group. It should be noted that data for return to function within 3 days was missing for 6 patients in the prednisone group and 3 patients in the dexamethasone group. This is a 10% difference, with a 95% confidence interval defined as 0%-20%, meaning that this result is statistically significant. In the prednisone group, 1 patient was admitted to the hospital, 6 had repeat ED visits, and 5 had a primary care visit. In the dexamethasone group, 3 patients were admitted to the hospital, 5 had repeat ED visits, and 3 had a primary care visit. It is unknown whether there was overlap among these patients (i.e., did the same patient who had a repeated ED visit also have a primary care visit?). There was no calculation of pooled relapse rate as previously defined. There was a 2% difference in hospital admissions favoring the prednisone group, which was not statistically significant (95% CI -6% to 2%). There was a 1% difference in repeat ED visits favoring the dexamethasone group, which was not statistically significant (95% CI -5% to 8%). There was a 2% difference in repeat ED visits favoring the dexamethasone group, which was not statistically significant (95% CI -3% to 8%). This indicates a signal of benefit towards the dexamethasone group in terms of return to function within 3 days, but it should be noted that data was missing and that the sample size was small, which could have influenced the statistical significance of the results. 

Study #2: Rehrer et al. 2016. 

This was a prospective, triple-blinded randomized controlled trial conducted a single urban ED in Oakland, CA. Participants included patients aged 18-55 with a preexisting diagnosis of asthma who presented to the ED with an asthma exacerbation requiring > 1 treatment of nebulized albuterol and were ultimately discharged from the ED. The upper age limit of 55 was chosen in an attempt to exclude patients with COPD.  The study specifically excluded patients who were admitted to the hospital due to asthma exacerbation. The respiratory therapist identified patients eligible for enrollment after their first treatment of nebulized albuterol. The objective of the study was to compare the number of patients who would relapse within 14 days (defined as an unscheduled visit to a healthcare provider for persistent or worsening symptoms) between patients who received one dose of oral dexamethasone versus patients who received 5 days of oral prednisone. This was specifically a noninferiority study. The ideal sample size was pre-determined; with a goal 80% power and an α = 0.05, 520 patients would be needed to detect an 8% difference in relapse rates, assuming a 16% relapse rate based on prior data. There was a plan to potentially adjust the sample size after an interim analysis. Interim analysis was conducted at 200 patients, and relapse rate was calculated at 11%. The sample size was adjusted using an 11% relapse rate to 375. 465 subjects were enrolled and randomized, with 227 in the dexamethasone group and 238 in the prednisone group. Randomization was conducted by the pharmacy. 10 patients in the dexamethasone group and 6 patients in the prednisone group were ultimately admitted and therefore excluded from analysis. Thus, 217 patients were discharged in the dexamethasone group, and 232 were discharged in the prednisone group. 44/217 patients (20%) were lost to follow-up in the dexamethasone group, making the final size for analysis 173 patients. 29/232 patients (12.5%) were lost to follow-up in the prednisone group, making the final size for analysis 203 patients. This is a total of 376 patients, which just met the goal of 375 patients after the interim analysis for goal sample size. These analyzed patients had some differences in their baseline characteristics – for example, 31.5% of patients in the prednisone group were current smokers as compared to 23.7% of patients in the dexamethasone group, and 26.6% of patients in the prednisone group were prescribed inhaled steroids at discharge as compared to 17.3% of patients in the dexamethasone group. However, other characteristics, such as age, sex, ethnicity, vital signs at presentation, and initial and discharge peak flow were similar. Notably, patients in each group received the same number of nebulized albuterol treatments (3) and nebulized ipratropium treatments (1) prior to discharge. There was no information about the 16 patients who were ultimately excluded due to their admission, and there was no information included about the patients who were lost to follow-up. Patients in the prednisone group received one 60-mg dose of oral prednisone in the ED and were discharged with a bottle containing 4 tabs of prednisone 60 mg. Patients in the dexamethasone group received one 12-mg dose of oral dexamethasone in the ED and were discharged with a bottle containing 4 tabs of placebo. Capsules and bottles were identical. Patients were contacted at two weeks, and a survey was administered in either English or Spanish. The main question was whether the participant had consulted a physician within the last 14 days of visiting the ED and, if so, whether it was a routine follow-up or an emergency visit for worsening or persistent symptoms. Other questions asked included about adverse effects, current treatment, and current symptoms. 9.8% of patients in the prednisone group had an ED visit within the past 14 days as compared to 12.1% in the dexamethasone group. This is a 2.3% difference, with a 95% confidence interval defined as -4.1%-8.6%, meaning that this result is not statistically significant. 2.9% of patients in the prednisone group were admitted as compared to 3.4% in the dexamethasone group. This is a 0.5% difference, with a 95% confidence interval defined as -4.1%-3.1%, meaning that this result is not statistically significant. 64.5% of patients in the prednisone group felt better with treatment as compared to 59.5% in the dexamethasone group. This is a -5.0% difference, with a 95% confidence interval defined as -4.8%-14.8%, meaning that his result is not statistically significant. Adverse events including sleep disturbance, abdominal pain, vomiting, and mood disturbance were also tracked among the groups, with significantly fewer patients in the dexamethasone group experiencing abdominal pain. However, it is important to note that the experiencing any adverse event was not statistically significant, with 28.6% of the prednisone group and 24.3% of the dexamethasone group reporting any adverse outcome for a difference of -4.3% (95% CI -4.6%-13%). Adherence to study medications was also tracked, with 91.3% of patients in the prednisone group and 93.1% in the dexamethasone group completing the medications given. There was no data included about whether patients were using inhaled steroids as prescribed on discharge. This indicates that with regard to the primary outcome of relapse, dexamethasone is noninferior to prednisone; however, there is a suggestion that patients may feel better with the prednisone and that patients may experience decreased incidence of certain adverse events.  

Conclusions: