|
|
| |||||||||||||||||||||||||||||||
Background Although recent evidence has strongly supported the use of glucocorticoid therapy in children hospitalized with croup, the benefit of this therapy in children with less severe croup has not been documented. This randomized, double-blind trial compared a nebulized glucocorticoid, budesonide, with placebo in outpatients with mild-to-moderate croup.
Methods Children three months to five years of age were eligible for the study if their croup scores fell in the mild-to-moderate range (scores of 2 to 7 out of a possible 17). The patients were randomly assigned to receive either 2 mg (4 ml) of nebulized budesonide (27 children) or 4 ml of nebulized normal saline (27 children); they were then assessed hourly for up to four hours by investigators who were unaware of the assigned treatments.
Results The median croup score at entry into the study was 4 in both groups. At the final study assessment, the median score was significantly lower in the budesonide group than in the placebo group (1 vs. 3, P = 0.005). The patients in the budesonide group were discharged from the emergency department significantly earlier than those in the placebo group (P = 0.002). One week after enrollment, 21 patients assigned to placebo had received dexamethasone, as compared with 15 patients assigned to budesonide (P = 0.10), and 7 patients assigned to placebo had been admitted to the hospital, as compared with 1 patient assigned to budesonide (P = 0.05).
Conclusions We conclude that nebulized budesonide leads to a prompt and important clinical improvement in children with mild-to-moderate croup who come to the emergency department.
Previous studies have not addressed whether the benefits of glucocorticoid therapy extend to children with milder disease, many of whom are evaluated in emergency departments. The argument against the routine use of glucocorticoids in outpatients has been that the majority of children with croup have self-limited illnesses16,17. It is not known, however, whether treatment at an early stage of the illness would reduce the severity of the clinical symptoms, prevent hospitalization or prolonged visits to the emergency department, and thereby both improve health outcomes and reduce costs.
We designed a randomized, placebo-controlled trial to determine whether nebulized budesonide leads to a clinically important improvement in respiratory symptoms within four hours for children with mild-to-moderate croup who come to the emergency department. Budesonide is a synthetic glucocorticoid with relatively strong topical antiinflammatory effects and low systemic activity as compared with beclomethasone18. Nebulized budesonide was selected for the study because it can be administered without the discomfort of intramuscular injection and because it begins to act as early as one hour after administration19.
Methods
Selection of Patients
Patients coming to the emergency department at the Children's Hospital of Eastern Ontario between October 1, 1992, and October 15, 1993, were eligible for the study if they met the following criteria: an age of three months to five years; a syndrome consisting of hoarseness, inspiratory stridor, and barking cough; and a croup score of 2 or higher after breathing humidified oxygen for at least 15 minutes. Three research assistants trained by the investigators were on call to enroll patients in the study between 9 a.m. and midnight every day except holidays. Patients were excluded if they had been given a diagnosis of epiglottitis or chronic upper or lower airway disease (not including asthma), if corticosteroids had been administered within the preceding two weeks, or if they had severe croup. We defined children as having severe croup if their croup scores were 8 or higher or if they required treatment with racemic epinephrine immediately on arrival, as determined by the treating physician. The study was approved by the research ethics committee at the Children's Hospital of Eastern Ontario. Informed, written consent was obtained from all parents before the enrollment of patients.
Outcome Measures
All the patients underwent base-line clinical assessment by a research assistant, consisting of determination of the croup score,20 respiratory rate, and heart rate while the patient breathed humidified oxygen or was inside a plastic enclosure (a croup tent). The croup score measured the degree of stridor (on a scale on which 0 denotes no stridor, 1 stridor audible with the stethoscope with the patient at rest, and 2 stridor audible without the stethoscope with the patient at rest), the severity of retraction of the intercostal and subcostal regions (0 denotes none, 1 mild, 2 moderate, and 3 severe), the entry of air into the lungs (0 denotes normal, 1 decreased, and 2 severely decreased), cyanosis (0 denotes none, 4 cyanosis with agitation, and 5 cyanosis at rest), and level of consciousness (0 denotes normal, and 5 altered).
In addition, the research assistants, parents, and treating physicians were asked at the end of the study period to rate independently whether the patient's condition had improved, remained the same, or worsened. This global assessment of change was then evaluated on a 15-point Likert scale ranging from -7 ("a very great deal worse") to +7 ("a very great deal better"), with 0 representing no change. The parents and the treating physicians were unaware of the patients' croup scores.
Study Design
Eligible patients were randomly assigned to receive a single dose containing either 2 mg (4 ml) of budesonide solution (Pulmicort, Astra, Pharma, Mississauga, Ont.) or 4 ml of 0.9 percent saline solution, administered by an updraft nebulizer with a continuous flow of oxygen at 5 to 6 liters per minute. Because budesonide is slightly opaque, the pharmacy provided both budesonide and normal saline in opaque brown syringes to ensure blinding. The research assistants then placed the study drug directly into an opaque nebulizer reservoir. Once nebulized, the study drugs were indistinguishable by sight and smell when tested before and during the study.
Randomization was performed in blocks of 10 by the pharmacy department, with a random-number table. Block randomization was used to ensure that there was an equal number of patients in the budesonide and placebo groups during the season when a respiratory virus was prevalent. The randomization list was kept concealed from the research assistants, parents, and emergency physicians and from the child's regular physician until the end of the trial. Thus, all treatment decisions during the week after enrollment were made by people unaware of the patient's treatment assignment.
The treating physicians followed the patients throughout the study period and were free to discharge them at any point if they deemed the patients clinically ready for discharge. The patients were given humidified oxygen through a large plastic tube (a mist stick) during the study period according to the protocol. To prevent the clinical trial from interfering with usual clinical practice, the treating physicians were free to use other interventions, such as racemic epinephrine or dexamethasone, or to place a patient in a croup tent for the delivery of humidified oxygen. These interventions were defined as not being part of the study protocol. Once a patient received racemic epinephrine, measurements for that patient were not included in the hourly assessments. While the patients remained in the emergency department, they were assessed by the research assistant every hour for four hours, until the croup score was 1 or less, or until the treating physician considered the patient well enough to be sent home. The research assistant telephoned the family after one week to inquire about any further visits to doctors, further treatments, or hospitalizations.
Estimate of Sample Size
A two-point improvement in the croup score or a return of the score to 1 or less was considered to be clinically important and to constitute a response. We estimated that 40 percent of the patients assigned to placebo would have clinically important improvements, as compared with 70 percent of those assigned to budesonide. With a two-sided alpha level of 0.05 and 80 percent power, a sample containing 48 patients per group would be required.
During the trial, however, the clinicians in the emergency department came to consider the administration of dexamethasone to be the standard of care for children with croup that remained symptomatic after therapy with a mist stick. Because this change in clinical practice impaired our ability to recruit eligible patients and to evaluate the independent effect of budesonide, the study was stopped before the specified sample was reached and before the data were analyzed.
Statistical Analysis
The data were analyzed with SPSS/PC+ V4.01 software (SPSS, Chicago). Continuous data were analyzed with an independent two-tailed t-test when the data were parametric. Nonparametric data were analyzed with a Mann-Whitney U test. Categorical data were analyzed with the chi-square statistic or Fisher's exact test. Yates' continuity correction was used for two-by-two tables. Pearson's correlation coefficients were used to measure the association between the global assessment of change (on a 15-point Likert scale) and changes in the croup score, with a one-sided test for statistical significance. The survival analysis was performed with Kwikstat V3.01 (TexaSoft, Cedar Hill, Tex.), and differences between groups were tested with the Mantel-Haenszel test. Agreement between observers was measured by the kappa statistic with quadratic weights with use of PC-Agree software (McMaster University, Hamilton, Ont.).
Results
Characteristics of the Patients
During the study period, 390 patients had a diagnosis of croup on discharge from the emergency department. Of these patients, 163 presented at times when the study team was not on call, and in 24 additional cases the emergency department failed to contact the study team. A further 146 patients did not meet the eligibility criteria for the study because they had croup that was too mild (88 patients), did not meet the age requirement (31 patients), did not meet our definition of croup (16 patients), had croup that was too severe (5 patients), had epiglottitis (3 patients), had previous upper airway disease (2 patients), or had recently been treated with steroids (1 patient). The parents of the remaining 57 patients were approached about participation in the study, and 54 agreed on behalf of their children.
Twenty-seven patients were randomly assigned to the budesonide group and 27 to the placebo group. Table 1 shows that most base-line demographic and clinical variables were evenly distributed between the two groups, except for temperature and the duration of cough, differences that would have favored the placebo group.
|
The research assistants and physician investigators rated the croup scores of 12 patients concurrently to measure interobserver reliability. The weighted kappa statistic (±SE) was 0.95 ±0.02, indicating excellent agreement.
Outcome Variables
As compared with the patients assigned to placebo, the patients treated with budesonide had significantly lower croup scores at the final study assessment (median, 3 vs. 1; P = 0.005) (Table 2). There were no significant differences between the groups in heart rate or respiratory rate at the final study assessment (Table 2). When the a priori definition of response was used, at four hours 19 of the budesonide group had a response (70 percent), as compared with 10 of the placebo group (37 percent) (P = 0.03).
|
|
Other Interventions
Patients received other interventions at the discretion of their treating physicians, who remained unaware of the treatment assignments. Two patients in the placebo group received racemic epinephrine, as compared with none of the patients in the budesonide group (P = 0.49). There was no difference between the groups in the proportion of patients for whom a croup tent was ordered (P = 1.0). Six patients in the budesonide group (22 percent) received dexamethasone during the study period, as compared with eight of those in the placebo group (30 percent) (P = 0.75). The mean (±SD) interval before the administration of dexamethasone did not differ significantly between groups (budesonide, 148 ±35 minutes; placebo, 122 ±34; P = 0.19).
In addition to the patients who received dexamethasone in the emergency department during the study period, 9 patients assigned to budesonide (33 percent) and 11 patients assigned to placebo (41 percent) received dexamethasone before discharge from the emergency department (P = 0.34).
Follow-up
A survival analysis of the time until patients were sent home from the emergency department or the holding unit revealed a significantly earlier discharge for patients treated with budesonide than for those receiving placebo (P = 0.002) (Figure 1). By the end of the four-hour study period, 12 patients in the budesonide group (44 percent) were sent home, as compared with 6 patients in the placebo group (22 percent). Six patients assigned to placebo (22 percent) were admitted to the hospital for a median stay of two days (range, one to five), as compared with one patient assigned to budesonide (4 percent), who remained in the hospital two days.
|
Possible Adverse Effects of Treatment
No adverse events were noted in the budesonide group. No patient in that group had clinical deterioration, either in the emergency department or after discharge. One patient in the placebo group had a burning sensation on the face.
Discussion
This randomized, controlled trial has demonstrated an important clinical benefit of budesonide treatment in patients with mild-to-moderate croup. The improvement associated with budesonide was reflected in lower croup scores, improved global ratings by independent observers, earlier discharge from the emergency department, and a lower rate of hospital admissions and other interventions by physicians who were unaware of the treatment assignment.
A useful index of therapeutic benefit is the number of patients who would need to be treated by an intervention for one patient to have the outcome event of interest21. On the basis of the data from this trial, for example, five patients with croup scores of 2 to 7 would need to be treated with nebulized budesonide for one hospitalization to be prevented. For one patient to have a clinically important response under our definition, three patients would need to be treated with budesonide. These results indicate the potential for a substantial reduction in health care expenditures for croup, because the cost of a single dose of nebulized budesonide is dwarfed by the cost of even a two-day admission to the hospital.
The demonstration that budesonide has a therapeutic effect within four hours in children with mild-to-moderate croup is consistent with observations from other studies and studies in animals22,23. Among patients hospitalized for croup who were assessed with a clinical scoring system similar to the one we used, Husby and colleagues observed a significant improvement two hours after administering 2 mg of nebulized budesonide15.
When a clinical score is used as an outcome measure, it should be valid, reliable, and responsive to important changes in clinical status24. These methodologic properties of the croup score have received little attention in the past, although the score has been used as the primary outcome measure in most clinical trials in children with this condition. As far as we are aware, the only available validation study showed that the croup score correlated with changes in the diameter of the tracheal lumen before and after treatment with racemic epinephrine in 12 children whose mean croup score before treatment was 425. Our study has generated data on both validity and responsiveness by showing that the croup score correlates with the global judgments of three groups of independent observers who have different perspectives on the patient.
One other aspect of the study methods deserves comment. Although dexamethasone treatment had the potential to interfere with assessment of the independent effect of budesonide, by one week after enrollment 78 percent of the placebo group had received dexamethasone, as compared with 56 percent of the budesonide group. Only those in the placebo group were treated with racemic epinephrine. The overall effect of both interventions would have been to reduce the likelihood of detecting the observed difference between the patients in the budesonide group and those in the placebo group.
Past admonitions against the routine use of glucocorticoids in patients with croup have stressed the potential for infrequent but serious gastrointestinal hemorrhage and other adverse effects after dexamethasone therapy16,26. To some observers, these risks seemed unnecessarily large for children with mild or moderate croup in whom recovery was virtually assured, albeit occasionally after a brief hospitalization. Because budesonide does not carry a risk of either gastrointestinal hemorrhage or painful injection, the demonstration of its efficacy in mild-to-moderate croup shifts the risk-benefit ratio in favor of treatment. Further study will help determine whether giving budesonide alone at an early phase of the illness has a sustained effect on the course of croup and whether children with croup are most responsive to glucocorticoid therapy at an early point in the illness, before excessive subglottic edema has developed.
Glucocorticoids should be used with caution in patients with preexisting immunodeficiency, recent exposure to varicella, or possible tuberculosis. When used in previously healthy children with croup, however, glucocorticoids given orally or parenterally have had few important adverse effects27,28. The systemic toxic effects of inhaled steroids should be even more limited than those associated with the oral or parenteral forms. Although our study had 100 percent follow-up at one week, we would emphasize that our sample was not large enough for rare side effects to be identified29.
We conclude that nebulized budesonide is an effective initial therapy for children with mild-to-moderate croup who come to the emergency department and remain symptomatic after mist therapy. With its rapid action and prolonged effect, budesonide may be a valuable agent for inhibiting the subglottic inflammatory edema that contributes to stridor in children with croup.
Supported by a grant (04440) from the Ontario Ministry of Health.
We are indebted to Pat Harman and Joanne Momy for their tireless work as research assistants, to Colline Blanchard for randomization and drug preparation, to Dr. Marilyn Li and all the pediatricians and nurses in the Emergency Department for help in recruiting patients, to Dr. Martin Osmond and Grace Dueck for their critical review of an earlier version of this manuscript, and to Robert Stenstrom for assistance with the early development of this project.
Source Information
From the Department of Pediatrics, University of Ottawa, Ottawa, Ont. (T.P.K., L.K.W., T.S.); Scarborough Grace Hospital, Scarborough, Ont. (M.E.F.); and the Department of Pediatrics, Johns Hopkins University, Baltimore (P.C.R.).
Address reprint requests to Dr. Rowe at the Department of Pediatrics, Johns Hopkins Hospital, 600 N. Wolfe St./Brady 212, Baltimore, MD 21287.
References
| |||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
