Inhaled Budesonide in Addition to Oral Corticosteroids to Prevent Asthma Relapse Following Discharge From the Emergency Department:  A Randomized Controlled Trial FREE

Asthma affects approximately 7% of adults in North America,^1- 4 and patients frequently present to the emergency department (ED) with acute asthma. They often exhibit features that place them at increased risk for complications and even death.^5- 6 Improved strategies to treat acute asthma and reduce relapses after ED visits are needed to improve patient quality of life⁶ and reduce societal costs associated with this disease.^3,7

Our understanding of asthma has improved over the past decade^8- 10 and there is now general agreement on the primary role of inflammation in the pathophysiology of asthma. Control of inflammation has become the cornerstone of treatment during acute asthma.^11- 13 For patients with acute asthma discharged from the ED, the anti-inflammatory agents used most frequently are systemic corticosteroids (CS). There are 7 randomized, placebo-controlled trials^14- 20 examining outpatient treatment of asthma with CS that have been summarized in systematic reviews.^21- 22 These reviews indicate that CS reduce relapses by 65%; as few as 5 patients require treatment to prevent 1 relapse to additional care.

The role of inhaled CS (ICS) in acute asthma is less well-defined. Research suggests that while more than 70% of patients are discharged from North American EDs taking oral CS,²³ only 10% of patients are discharged taking ICS. This proportion is lower for those patients who were not receiving them before presentation.²³ This practice is consistent with the National Asthma Education and Prevention Program^11- 12 and other consensus guidelines,¹³ which do not make firm recommendations regarding starting ICS in the ED. To date, no randomized controlled trials have been published that examine the use of ICS in addition to oral CS for patients discharged from the ED after an exacerbation of asthma.

This study examined the effect of adding an ICS (budesonide) to a fixed dose of oral prednisone for patients discharged from the ED with acute asthma. The main objective of this study was to determine if adding ICS would reduce relapses requiring additional care. The secondary outcomes included monitoring changes in quality of life, pulmonary function, and the use of β₂-agonists over the follow-up period.

The study was conducted at the Sudbury General Hospital in Sudbury, Ontario. This community teaching ED provides 24-hour coverage in a community of 164,000 (1996 Canadian Census Statistics). Between November 15, 1995, and September 15, 1997, patients treated in the ED for acute asthma were eligible for enrollment.

Physicians in the ED referred patients whom they believed were likely to be discharged home after completion of treatment for acute asthma. The diagnosis of asthma was made on clinical grounds (a previous history of asthma, response to β₂-agonist therapy, and worsening symptoms necessitating treatment in the ED). Patients aged 16 to 60 years with peak expiratory flow rate (PEFR) of less than 80% predicted for sex, age, and height before treatment were eligible.

Those receiving oral or inhaled CS during the week before ED presentation were excluded. Patients were also excluded if they presented for prescription renewals, had complicated comorbid medical illnesses (renal or cardiovascular diseases, human immunodeficiency virus), were pregnant or breast-feeding, were geographically inaccessible for follow-up, had cognitive impairment, or were unable to communicate in French or English.

The study was approved by the hospital's research ethics committee, and informed consent was obtained from each patient. Patients were assessed for eligibility in the ED by a trained, on-call research assistant. For patients' safety, all participants were provided with a pocket card describing how to access research assistants and a patient information sheet outlining their involvement in the study.

Patients were assigned to treatment, using a computer-generated randomization method, in blocks of 10 to ensure equal numbers and elimination of seasonal biases (Figure 1). Study personnel and physicians remained blinded to the allocation of treatment until completion of data analysis. The randomization code was generated by the nonsite statistician (R.A.M.) and the drug packages were prepared by the hospital pharmacy. Study drugs were kept in packages of similar appearance in a research office. The treatment code was revealed on 2 occasions at the request of physicians treating study patients (1 acute deterioration; 1 nonasthma hospitalization).

The dosage, method of administration, and frequency of medication in the ED was determined by referring ED physicians. However, a recommended treatment protocol was made available to all physicians and consisted of nebulized β₂-agonist (salbutamol) every 20 minutes for the first hour,²⁴ a dose of oral prednisone (50 mg),^21- 22 supplemental oxygen, and, when indicated, nebulized ipratropium bromide.²⁵ The decision to discharge patients after ED therapy was made at the discretion of the treating physician. All physicians were provided with pulmonary function results (forced expiratory volume in 1 second [FEV₁], PEFR, and forced vital capacity); spirometry was performed by the research assistant using a hand-held spirometer.

All enrolled patients were discharged taking a nontapering course of oral prednisone (50 mg/d) for 7 days.^26- 28 The prednisone was provided in a blister pack dispensing format with clear written instructions. Patients were assigned in a double-blind fashion to inhaled budesonide (Pulmicort, Astra Pharma Inc, Mississauga, Ontario) or placebo Turbuhaler (Astra Pharma Inc) for 3 weeks. Budesonide is a potent ICS,²⁹ and Turbuhaler is a breath-activated dry powder inhaler that does not require a coordinated technique. Patients were given instructions for the use of Turbuhaler and asked to inhale forcefully and deeply through the mouth.³⁰

Turbuhaler devices were identical in appearance, and the drug delivery was indistinguishable by taste or smell. The dose of budesonide administered was 2 inhalations twice per day (400 µg per activation), for a total daily dose of 1600 µg. In addition, patients used short-acting inhaled β₂-agonists regularly and were provided with instruction on their use. If β₂-agonist supplies were insufficient, patients were supplied with metered-dose inhalers designed to provide 100 µg of salbutamol per inhalation. Other agents such as theophylline, long-acting β₂-agonists, ipratropium bromide, and antibiotics were prescribed at the discretion of the treating physician or continued if the patient had already been using them. Patients were asked to stop taking nedocromil sodium and sodium cromoglycate for the duration of the study. In addition, all patients were asked to refrain from using other ICS in conjunction with the study drugs, unless prescribed after a relapse.

All patients were asked to report relapses to the on-call research assistant via telephone contact or presentation to the ED. Patients were contacted by telephone 3 days following the ED visit to determine their respiratory status and to encourage compliance with the in-person reassessments at 7 and 21 days; they were recontacted by telephone at 14 days. Patients were followed up until the primary study end point (relapse) or for 21 days, whichever occurred first. Computerized hospital records were searched for the study period at the only 2 acute care hospitals within the region to validate ED visits and hospitalizations. Others have used relapse in outpatient asthma trials^14- 20; it was defined as an "unscheduled visit for worsening asthma symptoms" for this trial. This outcome was assessed by the study investigators after study completion, but before study unblinding.

Quality of life was measured using the Asthma Quality of Life Questionnaire (AQLQ).³¹ This questionnaire has been used previously in acute^32- 33 and chronic^31,34- 35 asthma and has been shown to be reliable, valid, sensible to patients, and responsive to clinically important changes in patient function. The 32-item questionnaire examines 4 domains (physical, symptom, environment, and emotion) of asthma using a 7-point Likert response scale for each question. It provides standardized overall and domain-specific AQLQ values between 1 and 7, with 1 indicating severe disease and 7 indicating minimal perturbation in total or domain scores.

Additional secondary outcomes also included β₂-agonist inhalations in the previous 24 hours. Three symptoms related to the severity of asthma (wheezing, breathlessness, and cough) and a global assessment of asthma, both using a 7-point Likert scale, were also determined at 7, 14, and 21 days. An adverse effect profile was determined for each patient from a list of symptoms at all follow-up sessions. Compliance was measured by self-report for both pills and the use of Turbuhaler. Compliance with Turbuhaler was also measured by counting the remaining activations from returned devices.³⁶ Most secondary outcomes were measured using in-person (at entry, day 7, and day 21) or telephone interviews (day 14). Pulmonary function results were recorded at entry, ED discharge, day 7, and day 21 by the research assistants.

Continuous variables are reported as means and SDs for normally distributed data, median and interquartile ranges (25%-75%) for skewed results. Primary analyses were based on an intention-to-treat principle. Time to relapse was analyzed using a Kaplan-Meier survival curve for time free of relapse and the log-rank test. An adjusted analysis controlling for imbalance on important baseline covariates was performed using logistic regression. All results were considered to indicate significance at P<.05. The AQLQ results were reported as separate domain scores and as a total score; all are continuous variables. A mean difference of 0.5 per domain or total score was used as the a priori significance level for all of the AQLQ analyses.^32,35 The secondary analyses used t tests for continuous, normally distributed results, and χ² for categorical variables.

The sample size was calculated a priori using the following assumptions: α=.05; relapse rate at 21 days equaled 25% to 30%. We estimated a reduction to 10% to 12% would be sufficient to warrant ED treatment with this agent, and thus a sample size of 180 patients (85-90 per group) provided an 80% to 90% power to identify this difference. We enrolled patients until 180 follow-up sessions were successfully completed.

Over the study period, 1006 patients aged 16 to 60 years were seen in the ED with acute asthma; 743 patients (74%) were excluded from the trial. The most common reasons for exclusion were the following: 278 patients (37%) were receiving ICS and 59 (8%) were receiving oral CS at the time of presentation, 116 (16%) presented with medical comorbidity, and 83 (11%) were not prescribed prednisone. Other reasons for the remaining 207 patients were: 46 (6%) were hospitalized, 40 (5%) were geographically inaccessible, 47 (6%) had been enrolled before, 29 (4%) received his/her first diagnosis of asthma, 18 (2%) left before treatment, 10 (1%) had pretreatment PEFR of more than 80% predicted, and 17 (2%) presented to the ED for prescription refills only. Of the remaining 263 patients, 51 (19%) refused enrollment, 21 (8%) were missed by treating physicians, and 191 (73%) agreed to participate in the trial (Figure 1).

Of the 191 participants originally randomized, 3 were subsequently excluded; 2 patients required hospitalization prior to ED discharge and 1 was inappropriately enrolled and removed from the study in stable condition after 1 day. Of the remaining 188 patients, 94 received budesonide and 94 received placebo. Three patients taking budesonide dropped out of the study for nontreatment reasons and 5 were lost to follow-up (2 taking budesonide; 3 placebo); however, these patients were kept in the database.

The overall study results are based on the 188 appropriately enrolled patients. The baseline characteristics of the patients in both groups were similar for most outcomes (Table 1, Table 2, and Table 3). However, there were clinically important differences between the groups based on sex, age, and severity (percent of predicted PEFR).

The medications used regularly at the time of presentation did not differ between the groups. The number of treatments of salbutamol in 24 hours prior to presentation was high in both groups. Few patients in this study were using asthma aids (eg, spacer devices, action plans, peak flowmeters, daily diaries) to treat their asthma (Table 3). Quality of life was rated as severely impaired in both groups; however, based on percent predicted PEFR, the asthma was more severe in the budesonide group (45% vs 54%).

The medications administered in the ED were similar in both groups. All received nebulized β₂-agonists, while many received nebulized ipratropium bromide and 50 mg of oral prednisone prior to discharge. By discharge, PEFR had increased over baseline values by 121 L/min or 22% predicted in the budesonide group compared with 117 L/min or 21% predicted in the placebo group (P=.73).

Follow-up. Overall, 14 patients (7.4%; 95% confidence interval [CI], 4%-12%) relapsed within the first day after discharge. By 7 to 10 days, 26 patients (13.8%; 95% CI, 9%-20%) had experienced a relapse and 3 patients (2%) had dropped out of the study. In-person follow-up was obtained for 138 of the remaining patients (87%); 2 others (1%) provided information via the telephone.

By 14 days, 30 patients (16.0%; 95% CI, 11%-22%) had experienced a relapse, and telephone follow-up was obtained for 132 (85%) of the remaining patients. At 21 days, 35 patients (18.6%; 95% CI, 13%-25%) had experienced a relapse. In-person follow-up was obtained for 117 (78%) of the remaining patients; 33 patients (22%) completed telephone follow-up. Computerized hospital record review at 21 days did not identify any study patients who had presented to the ED or were admitted to the hospital who were not already known. In summary, dropouts and losses to follow-up were uncommon (8 patients [4%]). All other 180 patients (96%) provided relapse data at 21 days.

Relapse Rates. The budesonide group experienced fewer relapses (12 patients [12.8%]; 95% CI, 7%-21%) than the placebo group (23 patients [24.5%]; 95% CI, 16%-34%) by 21 days (P=.049; Table 4). This represents a 48% relapse reduction and suggests as few as 9 patients would require treatment with budesonide to prevent 1 relapse.

The trends in relapse favored the budesonide group, and the difference between groups was statistically significant at 21 days (Figure 2). A consistent positive effect for the budesonide treatment (P=.04) was demonstrated using logistic regression when controlling for sex, age, and severity. No relapses resulted in death; 2 patients (2%) taking budesonide and 1 patient (1%) taking placebo required admission to a hospital at relapse.

Quality of Life.Table 4 shows that the AQLQ score in the budesonide group was higher (better quality) than that for the placebo group (P=.001). In addition, improvements in all domains of the AQLQ were greater in the budesonide group; activities (P=.002), symptom (P=.002), emotional (P=.001), and environmental (P=.01) improvements were all more than 0.5 better in the budesonide group. The trends mirror the results of the relapse data with budesonide patients separating from the placebo patients by the first follow-up (data not shown).

β₂-Agonist Inhaler Use. Patients were asked to report the total number of β₂-agonist inhalations used in the 24-hour period prior to follow-up contact. The budesonide patients were using fewer salbutamol inhalations in 24 hours compared with placebo patients (P=.01) at 21 days. While the values were similar at 7 days, by 14 days budesonide patients were using 1.4 (95% CI,−0.8 to 2.9) fewer inhalations and by 21 days, 1.7 (95% CI, 0.4-3.1) fewer inhalations in the 24-hour period.

Pulmonary Function Data. Mean and percent of predicted PEFR spirometry findings revealed no differences between the groups at the time of the final visit (P=.39 and .11, respectively). The improvement in pulmonary functions over time was not different between the groups (Table 4).

Symptoms. By the study's conclusion, the budesonide group had fewer symptoms of cough (P=.004), breathlessness (P=.001), wheezing (P=.001), and nighttime wakening (P=.001) compared with placebo (Table 4).

Global Assessment. Patients in the budesonide group assessed their asthma as more improved than those in the placebo group at 21-day follow-up (6.2 vs 5.2; P=.001).

Adverse Effects. In general, the study drugs were well tolerated (Table 4). Adverse effects were greater in the placebo group for both hoarseness and sore throat (P=.02). However, the incidence of adverse effects thought to be associated with ICS use (eg, insomnia, acne, fluid retention) were equal between the treatment groups.

Compliance. Compliance with therapy was measured throughout the study; 80% or more of the patients' behavior was designated compliant. Self-reported compliance with the use of oral prednisone was high within the first week of care in both groups (94% for budesonide vs 96% for placebo; P=.73). Self-reported compliance with Turbuhaler was similar between the groups at 7 (100% for both budesonide and placebo) and 21 (92% for budesonide vs 93% for placebo; P=.95) days.

To our knowledge, this is the first study to examine whether adding an ICS enhances the effectiveness of oral CS in patients with acute asthma discharged from the ED. Our results demonstrate that the addition of high-dose budesonide (1600 µg/d) reduces the rate of relapse, improves quality of life, and reduces β₂-agonist use after 21 days. The benefit of the ICS begins to emerge early, as outcomes diverge (Figure 2) before the completion of the oral prednisone regimen. It appears that the full benefit of ICS is reached by 10 days, after which relapses are rare.

However, previous reports from trials of ICS in the treatment of unstable asthma and asthma treated in the ED have yielded conflicting results. Salmeron and colleagues³⁷ reported on a double-blind, randomized trial of high-dose beclomethasone dipropionate (1500 µg/d) vs placebo in patients with unstable chronic asthma. They found that ICS maintained optimal pulmonary function induced by a 2-week course of oral CS and led to fewer asthma attacks than those experienced in the placebo group. Their results corroborate the combined approach we used in this study.

Another trial examined patients discharged from the ED and found no benefit in pulmonary function for the addition of flunisolide to prednisone therapy.³⁸ Since the complete findings of this study are yet to be published, comparisons are difficult at this point. However, these results may indicate that the effect of therapy is variable and dependent on either drug or other patient factors.

Three other studies have examined the use of high-dose ICS compared with oral CS in outpatient asthma. Investigators in primary care compared high-dose fluticasone propionate (2 mg/d) with a short course of prednisone in patients with acute asthma whom physicians felt required oral CS.³⁹ There were no differences in PEFR or presence of persistent symptoms. A second trial examined the effect of high-dose budesonide (1600 µg/d) vs prednisone in patients following ED discharge and found no difference in FEV₁.⁴⁰ Finally, another trial compared high-dose budesonide (2400 µg/d) vs prednisone in patients following ED discharge, and found no difference in subsequent relapse.⁴¹ However, 2 of these studies enrolled patients with mild asthma,^39,41 and none followed up patients to 3 weeks as was done in our study.^39- 41 Furthermore, several studies concentrated on pulmonary function as outcomes.^39- 40 Finally, 1 study is published only as an abstract⁴¹ and 2 are limited by small sample sizes.^40- 41

While these trials conclude treatment equivalence, and make no mention of the combined approach, the available evidence is insufficient to support this strategy. While some patients may benefit from receiving ICS alone, physicians should remain cautious about adopting this policy. Furthermore, our results suggest the combination of oral and inhaled CS provides the safest and most effective outpatient treatment regimen.

Several aspects of the current study deserve further comment. First, this was the first acute asthma trial to collect disease-specific quality-of-life outcomes. Previously, the AQLQ instrument has been shown to be valid, reliable, sensible, and relevant^31- 35; moreover, it is more responsive than other measures, thereby allowing efficiencies in research design.^32,42 The AQLQ provides a comprehensive evaluation of outcomes using a total score or 4 domain scores; all AQLQ domains demonstrated clinically (0.5 per domain or total score) and statistically significant changes in favor of the budesonide group.

Second, there was uniform acceptance of the protocol by physicians and few refusals by patients. The follow-up was comprehensive and complete, and the sample of patients should be generalizable to other ED settings. Finally, the broad range of secondary outcomes corroborated the findings of the primary analysis and add strength to the results.

There are several potential weaknesses of the study that require discussion. First, relapse was used as the primary study end point. While it could be argued that this outcome is subject to interpatient variability, it has been used extensively in the past and is clinically important.^14- 21 In addition, due to psychometric properties, the use of relapse may reduce the chance of identifying a difference between groups, thus resulting in a conservative estimation of effect.³² Second, despite randomization, there were some small differences between the groups at the start of the study. These differences resulted in somewhat more severely ill patients in the budesonide group. However, we adjusted the analyses to account for these differences and the results did not change significantly.

Finally, while treatment for 3 weeks is sensible to patients and physicians, we have not investigated the benefits of extending treatment beyond this period. Relapses may occur long after ED discharge,^6,14,43 and this period warrants more attention in future research. However, our follow-up period is one of the longest reported in the acute asthma ED literature,²¹ and certainly was sufficient to ensure patient safety and return to a stable state.

Notwithstanding these concerns, the results of this trial have several implications. First, we believe the study design and methods produce results that are valid and relevant to the ED physician. While use of a clinical diagnosis of asthma may have included some nonasthma patients, we feel this mirrors clinical practice and makes these results more generalizable. In addition, we used outcomes that are clinically important to both patients and physicians.

Second, these results suggest that physicians should change their approach to asthma treatment and consider adding inhaled to oral CS at discharge from the ED. This is an effective method of reducing relapses because for every 9 patients treated with this regimen, 1 potential relapse can be prevented. Adoption of combined CS prescribing by ED physicians has the potential to improve both short- and long-term outcomes.⁴⁴ Based on historical features, many patients in this study could be considered moderate-to-severe asthma sufferers. By prescribing ICS, ED physicians are providing patients with evidence-based therapy that they had not been receiving. Moreover, other evidence indicates patients who are not prescribed ICS at ED discharge are unlikely to have them started when contacted at follow-up.²³ By adopting a policy of initiating ICS therapy at the time of discharge, ED physicians have the opportunity to affect long-term outcomes of patients with asthma.

The results of this research indicate budesonide delivered by Turbuhaler is effective in this setting. However, further research is needed to define the optimal ICS regimen following ED discharge. Furthermore, a dose of 1600 µg/d is higher than many ED physicians may be comfortable prescribing. While this dose is supported by some empirical evidence,^{39- 40,45- 48} it is associated with few adverse effects, and is well-tolerated by patients, additional studies are required to determine the dose, duration, and delivery method for ICS, which will optimize outcomes in the discharge period. Finally, studies are required to determine whether oral CS can be safely reduced or eliminated in certain populations with the use of ICS.^39- 41