|
|
Background Treatment of acute Kawasaki disease with intravenous immune globulin and aspirin reduces the risk of coronary-artery abnormalities and systemic inflammation, but despite intravenous immune globulin therapy, coronary-artery abnormalities develop in some children. Studies have suggested that primary corticosteroid therapy might be beneficial and that adverse events are infrequent with short-term use.
Methods We conducted a multicenter, randomized, double-blind, placebo-controlled trial to determine whether the addition of intravenous methylprednisolone to conventional primary therapy for Kawasaki disease reduces the risk of coronary-artery abnormalities. Patients with 10 or fewer days of fever were randomly assigned to receive intravenous methylprednisolone, 30 mg per kilogram of body weight (101 patients), or placebo (98 patients). All patients then received conventional therapy with intravenous immune globulin, 2 g per kilogram, as well as aspirin, 80 to 100 mg per kilogram per day until they were afebrile for 48 hours and 3 to 5 mg per kilogram per day thereafter.
Results At week 1 and week 5 after randomization, patients in the two study groups had similar coronary dimensions, expressed as z scores adjusted for body-surface area, absolute dimensions, and changes in dimensions. As compared with patients receiving placebo, patients receiving intravenous methylprednisolone had a somewhat shorter initial period of hospitalization (P=0.05) and, at week 1, a lower erythrocyte sedimentation rate (P=0.02) and a tendency toward a lower C-reactive protein level (P=0.07). However, the two groups had similar numbers of days spent in the hospital, numbers of days of fever, rates of retreatment with intravenous immune globulin, and numbers of adverse events.
Conclusions Our data do not provide support for the addition of a single pulsed dose of intravenous methylprednisolone to conventional intravenous immune globulin therapy for the routine primary treatment of children with Kawasaki disease. (ClinicalTrials.gov number, NCT00132080
[ClinicalTrials.gov]
.)
The role of corticosteroids in the primary treatment of Kawasaki disease has been the subject of retrospective case series and open trials, with insufficient evidence to make recommendations concerning their use. Although one early study showed a detrimental effect of corticosteroid therapy in patients with Kawasaki disease,6 the results of other studies have suggested that corticosteroids may be beneficial in preventing coronary-artery aneurysms.7,8,9,10,11,12 The effect of primary treatment with pulsed corticosteroids on coronary dimensions has not been tested in a double-blind, placebo-controlled trial.
To study the efficacy and safety of pulsed corticosteroid therapy, added to conventional treatment with intravenous immune globulin plus aspirin, in the primary treatment of acute Kawasaki disease, we conducted a multicenter, randomized, double-blind, placebo-controlled trial within the Pediatric Heart Network.13 Patients were randomly assigned to receive either a pulsed dose of intravenous methylprednisolone or placebo; intravenous immune globulin and aspirin were administered to both groups.
Methods
Patients
Patients were recruited from December 2002 through December 2004 from eight centers in North America. Eligible patients were between days 4 and 10 of illness, with day 1 defined as the first day of fever. At least one of the following was also required for eligibility: the patient met four or more principal clinical criteria1; the patient had a coronary-artery z score1 of 2.5 or more for the proximal right coronary artery or the left anterior descending coronary artery, as measured by two-dimensional echocardiography, and met two principal clinical criteria (for patients younger than 6 months) or three principal clinical criteria (for patients 6 months of age or older); or the patient had a coronary aneurysm as defined according to criteria of the Japanese Ministry of Health14 and met at least one principal clinical criterion. Exclusion criteria were previous treatment with intravenous immune globulin; treatment with corticosteroids, other than inhaled forms, in the previous 2 weeks; the presence of a disease known to mimic Kawasaki disease1; previous diagnosis of Kawasaki disease; contraindication to corticosteroid use; and inability to take aspirin. Written informed consent was obtained from parents or legal guardians; assent from patients was also obtained when appropriate, according to the guidelines of local institutional review boards, which approved the study protocol.
Procedures
Patients were randomly assigned to receive either intravenous methylprednisolone (30 mg per kilogram of body weight over 2 to 3 hours) or placebo infusion, within strata according to age (<1 year or 
1 year) and sex, with the use of dynamic balancing at each center. After the study drug was infused, all children received diphenhydramine (Benadryl), 1 mg per kilogram, followed by intravenous immune globulin, 2 g per kilogram over 10 hours. They also received aspirin, 80 to 100 mg per kilogram per day, until they were afebrile for 48 hours; then they received aspirin, 3 to 5 mg per kilogram daily, until study completion. Children who had a temperature of 38.3°C or higher 36 hours or more after completion of the initial treatment with intravenous immune globulin, without another probable source of fever, were retreated with 2 g of intravenous immune globulin per kilogram. A third treatment with intravenous immune globulin, 2 g per kilogram, was administered to patients with recrudescent or persistent fever 36 hours or more after intravenous immune globulin retreatment, without another probable source of fever. Patients with continued fever after the third dose were treated at the discretion of center physicians.
Echocardiograms and laboratory data were obtained at baseline and at means (±SD) of 7.8±1.8 days (median, 8.0) and 36.5±4.3 days (median, 36.0) after randomization. Using two-dimensional echocardiography, we measured the internal lumen diameters of the left main coronary artery, the proximal and distal left anterior descending coronary arteries, and the circumflex, posterior descending, and proximal and distal right coronary arteries. In addition, coronary arteries were classified on the basis of the presence or absence of aneurysms according to criteria of the Japanese Ministry of Health.14 The diagnosis of pericardial effusion required more than 1 mm of fluid. At a core laboratory, all echocardiograms were interpreted in a fashion that was blind to patient identity and illness day.
Temperatures were measured immediately before each aspirin dose was administered, but the site (e.g., rectal) of temperature measurement was not standardized. Children were hospitalized until they had been afebrile for more than 12 hours. Parents recorded the temperatures of the patients daily after discharge.
Adverse events were classified according to severity, expectedness, and attributability (a possible or probable relation to factors in the study). Classification was adjudicated by a Pediatric Health Network subcommittee to ensure consistency across centers.
Statistical Analysis
The primary outcome variable was the larger of the z scores for the right coronary artery and the left anterior descending coronary artery at week 5 after randomization. For the five echocardiograms (2.6%) at week 5 on which only one primary segment was visualized, the maximum z score was based on the dimension of the single segment. We calculated that we needed to enroll 194 patients, including 10% in excess to account for one interim analysis and the possibility of missing data, for the study to have a statistical power of 85% to detect a mean difference of 0.50±1.10 in the maximum z score, with a two-sided significance level of 5%. One interim analysis was reviewed by an independent data and safety monitoring board. All reported P values are two-sided and are not adjusted for multiple testing, unless otherwise specified. P values of less than 0.05 were considered to indicate statistical significance.
We compared the distributions of data between the two study groups as follows: for continuous variables, using a t-test if the data were normally distributed and Wilcoxon's rank-sum test otherwise; for time from admission to initial hospital discharge, using a log-rank test; and for categorical variables, using a Fisher's exact test unless otherwise specified. We compared the numbers of adverse events and episodes of retreatment in each study group using Poisson regression. We transformed coronary-artery dimensions to z scores (standard-deviation units) on the basis of body-surface area.1 We performed four prespecified subgroup analyses, as well as a post hoc analysis according to the presence or absence of retreatment with intravenous immune globulin, using a test for interaction between the subgroup factor and study group.
We performed secondary analyses of all outcome variables, excluding data for six patients who were discovered after enrollment to have met an exclusion criterion, for two patients who did not receive corticosteroid therapy despite randomization to the intravenous methylprednisolone group, and for eight patients who were enrolled because they had coronary abnormalities but who would not have met the classic criteria for Kawasaki disease.1 The inferences reached were similar to those derived from analysis of the entire data set.
Results
During the 2-year study period, 589 children were treated for Kawasaki disease (Figure 1). Of these, 276 were ineligible for the trial; 185 met at least one exclusion criterion, most often being ill for more than 10 days (102 patients), and the remaining 91 did not meet the inclusion criteria. Of the 313 eligible children, 199 (64%) had parental consent for enrollment in the study. Of these, 101 were randomly assigned to receive intravenous methylprednisolone, 30 mg per kilogram, and 98 were randomly assigned to receive placebo. Patients in the two study groups had similar baseline characteristics (Table 1).
|
|
|
The time to first hospital discharge was marginally shorter in the intravenous methylprednisolone group than in the placebo group, but the two groups had similar total numbers of days in the hospital (including readmissions) and days of fever both after randomization and after the onset of illness (Table 3). Similarly, the groups did not differ significantly in the percentage of patients who were retreated at least once with intravenous immune globulin or in the total number of episodes of retreatment with intravenous immune globulin (Table 3).
|
|
|
To explore whether the effect of intravenous methylprednisolone on coronary outcomes differed according to the severity of illness, we performed a post hoc comparison of the efficacy of the drug in the 27 patients with persistent fever who required retreatment with intravenous immune globulin (12 in the intravenous methylprednisolone group and 15 in the placebo group) with the efficacy in the 172 children who did not require retreatment (those who had a response to intravenous immune globulin). The efficacy of intravenous methylprednisolone at week 1 and week 5 differed between these two subgroups with regard to mean maximum z scores (P=0.03 and P=0.006, respectively) and the percentages of patients with coronary-artery abnormalities (P=0.02 and P<0.001, respectively). Among the patients who were retreated with intravenous immune globulin, the mean maximum z score tended to be lower in the intravenous methylprednisolone group than in the placebo group at week 1 (1.35±1.21 and 2.73±2.40, respectively; P=0.07) and was significantly lower at week 5 (0.77±0.86 vs. 2.66±2.37, P=0.01). Correspondingly, the percentages of patients with coronary-artery abnormalities in the intravenous methylprednisolone group and the placebo group were 25% (3 of 12 patients) and 67% (10 of 15 patients), respectively, at week 1 (P=0.05) and 0% (0 of 11 patients) and 60% (9 of 15 patients), respectively, at week 5 (P=0.002). Images for one patient at week 5 were insufficient for the classification of the presence or absence of coronary abnormalities.
The patients retreated with intravenous immune globulin in the intravenous methylprednisolone group had a shorter time to first hospital discharge than did those in the placebo group (median, 3 days and 4 days, respectively; P=0.05 by the log-rank test), but the two subgroups had similar numbers of days of fever and days in the hospital, as well as similar laboratory results. Five of the children retreated with intravenous immune globulin (two in the intravenous methylprednisolone group and three in the placebo group) received rescue intravenous methylprednisolone because of persistent or recrudescent fever after two episodes of retreatment with intravenous immune globulin.
Discussion
We found that primary therapy with pulsed intravenous methylprednisolone, administered as a single dose of 30 mg per kilogram before conventional therapy with intravenous immune globulin (2 g per kilogram), did not improve coronary-artery outcomes at week 1 or week 5 after study enrollment. Pulsed intravenous methylprednisolone somewhat shortened the duration of the initial period of hospitalization and accelerated the recovery of some laboratory markers of the acute-phase response, but the total number of days of fever and of hospitalization did not differ significantly between study groups. The addition of intravenous methylprednisolone to conventional therapy was not associated with fewer adverse side effects. In post hoc subgroup analyses of children with persistent fever who received retreatment with intravenous immune globulin, coronary outcomes were better in the intravenous methylprednisolone group than in the placebo group. Thus, children at highest risk for resistance to intravenous immune globulin and for coronary abnormalities may benefit from corticosteroid therapy. However, a single pulsed dose of intravenous methylprednisolone in addition to conventional therapy is not indicated for routine primary treatment of all children with Kawasaki disease.
Previous prospective studies of the use of corticosteroids in primary treatment of children with Kawasaki disease have been inconclusive with regard to the effect on coronary-artery abnormalities. The authors of a meta-analysis concluded that, if combined with aspirin-containing regimens as initial therapy, corticosteroids significantly reduced the incidence of coronary-artery aneurysms.11 Of the eight studies included,6,10,15,16,17,18,19,20 only one had blind interpretation of echocardiography15; two were prospective,15,16 and one included intravenous immune globulin administration according to current guidelines.15 The conclusions from the meta-analysis are therefore limited by the quality and design of the studies.21,22 Recently, Inoue et al.12 performed a multicenter, prospective randomized trial in which intravenous immune globulin and aspirin were administered with or without the addition of intravenous prednisolone until defervescence, followed by the daily administration of oral prednisolone until C-reactive protein levels normalized. Patients in the corticosteroid group had a lower prevalence of coronary dilation (as defined according to criteria of the Japanese Ministry of Health) during the first month of illness than did the conventional group. Beyond 1 month, this difference was no longer significant. Assignment to corticosteroids or placebo and interpretation of echocardiograms, performed at local centers, were not blind. Our study, a larger-scale clinical trial, involved the optimal, currently recommended regimen of intravenous immune globulin in all patients, was double-blind and placebo-controlled, and included blind interpretation of echocardiograms.
Although intravenous methylprednisolone did not affect coronary-artery outcomes among patients whose fevers responded to intravenous immune globulin, it appeared to be beneficial in patients who required retreatment with intravenous immune globulin. The post hoc nature of these analyses and practical limitations in prospectively identifying patients who do not have a response to intravenous immune globulin are important caveats to this conclusion. However, in several recent publications, investigators have constructed risk scores for the Japanese population in order to predict resistance to intravenous immune globulin from baseline data.23,24,25,26 More aggressive primary treatment with corticosteroids might benefit children who are determined at baseline to be at high risk for such resistance.
Additional study limitations should be noted. We studied a single dose of intravenous methylprednisolone only, since we believed that this regimen was the safest addition to standard treatment with intravenous immune globulin in a relatively low-risk population.27,28,29,30 The results of our trial do not preclude the efficacy of other corticosteroid regimens for primary treatment or that of corticosteroid rescue therapy for children with persistent fever or aneurysms after conventional primary treatment.9,30,31,32,33 The anatomical site of temperature measurement was not standardized, which decreased the accuracy of fever assessment. Data were not collected beyond week 5 after randomization. However, new coronary aneurysms rarely develop after the first month of illness, and approximately half of coronary aneurysms regress, through myointimal proliferation, to a normal internal lumen diameter.1 Thus, any differences between the two groups at week 5 would only have diminished over time. Finally, the study was underpowered for subgroup analyses and for detection of between-group differences in the numbers of adverse events.
In summary, our data do not provide support for the addition of a single dose of pulsed intravenous methylprednisolone to conventional therapy in the routine primary treatment of Kawasaki disease. Although pulsed corticosteroid therapy was associated with more rapid resolution of serum inflammatory markers and a somewhat shorter initial length of stay in the hospital than placebo, it did not improve coronary-artery outcomes or reduce the numbers of adverse events, days in the hospital, or days of fever. Since post hoc subgroup analysis suggested that primary therapy with intravenous methylprednisolone might benefit children with persistent fever after treatment with intravenous immune globulin, future prospective studies should explore the usefulness of corticosteroid or other immunomodulatory therapies in children at highest risk for resistance to intravenous immune globulin.
Supported by grants from the National Institutes of Health (U01 HL068285 and RR 02172, to Dr. Newburger, Ms. Baker, and Dr. Sundel; U01 HL068270, to Drs. Sleeper and Colan, Mr. Mitchell, and Ms. Klein; U01 HL068288, to Dr. McCrindle; U01 HL068292, to Dr. Minich; U01 HL068290, to Dr. Gersony; U01 HL068279, to Dr. Vetter; U01 HL068281, to Dr. Atz; and U01 HL068269, to Dr. Li) and from the Higgins Family Cardiology Research Fund (to Dr. Colan).
No potential conflict of interest relevant to this article was reported.
Source Information
From Children's Hospital and Harvard Medical School, Boston (J.W.N., A.L.B., S.D.C., R.P.S.); New England Research Institutes, Watertown, MA (L.A.S., P.D.M., G.L.K.); the University of Toronto, Hospital for Sick Children, Toronto (B.W.M.); Primary Children's Medical Center, Salt Lake City (L.L.M.); Columbia University Medical Center, New York (W.G.); Children's Hospital of Philadelphia, Philadelphia (V.L.V.); Medical University of South Carolina, Charleston (A.M.A.); Duke University Medical Center, Durham, NC (J.S.L.); and Children's Hospital of Los Angeles and University of Southern California, Los Angeles (M.T.).
Address reprint requests to Dr. Newburger at the Department of Cardiology, Children's Hospital, 300 Longwood Ave., Boston, MA 02115, or at jane.newburger{at}cardio.chboston.org.
References
The following persons participated in the enrollment of patients, data collection, or study coordination: Investigators: Children's Hospital Boston — D.R. Fulton, A.L. Woodward, F. Dedeoglu, B. Binstadt, S. Kim, R. Fuhlbrigge, E. McGrath; Children's Hospital of Los Angeles — W. Mason, V. Guerrero; Children's Hospital of Philadelphia — A. Hogart; S. Paridon, J. Rychik, M. Harkins; Columbia University Medical Center — L. Imundo, D. Levy, D. Hsu, S. Mital; Duke University Medical Center — P. Anderson, A.M. Nawrocki, D. Harrington; Hospital for Sick Children — T. Bradley, R. Yeung, E. Radojewski; Medical University of South Carolina — G. Shirali, M. Scheurer, A. Jones; Primary Children's Medical Center — R.V. Williams, D.U. Frank, L.Y. Tani, L.M. Lambert, A.L. Smart: Core Laboratories: C-Reactive Protein Core Laboratory — S.D. Douglas, D.E. Campbell, J.M. McMann, P. Wuthnow, N. Gilper, M. Wilson; Echocardiographic Core Laboratory — D. Cabral, E. Marcus; Data Coordinating Center: New England Research Institutes — P. Nash, D. Gallagher; Study Sponsor: National Heart, Lung, and Blood Institute — G.D. Pearson, V. Pemberton, M. Stylianou, J. Massicot-Fisher, T. Hoke, L. Mahony (network chair); Data and Safety Monitoring Board: J.D. Kugler (chair), D.J. Driscoll, S.A. Hunsberger, C.L. Webb, L.S. Wissow, K.D. Davis, J.S. Tweddell.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Treatment of Kawasaki Disease
Inoue Y., Kobayashi T., Morikawa A., Taddio A., Rosé C. D., Greil G. F., Manning W. J., Newburger J. W., Sleeper L. A., Burns J. C.
Extract |
Full Text |
PDF
N Engl J Med 2007;
356:2746-2748, Jun 28, 2007.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
