 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ELECTRONIC ARTICLE |
,
,,||
* Clinical Trials Research Center, IWK Health Centre, Dalhousie University, Halifax, Canada
Department of Pediatrics, Dalhousie University, Halifax, Nova Scotia, Canada
Department of Community Health and Epidemiology, Dalhousie University, Halifax, Nova Scotia, Canada
|| Department of Microbiology and Immunology, Dalhousie University, Halifax, Nova Scotia, Canada
¶ Centre Hospitalier de l’Université Laval, Laval, Quebec, Canada
# Department of Mathematics and Statistics, Dalhousie University, Halifax, Nova Scotia, Canada
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSCONCLUSIONSREFERENCES |
|---|
Methods. Children who were 6 months to 16 years of age and had cough illness that was suspected to be or was culture confirmed as pertussis were randomized to azithromycin (10 mg/kg on day 1 and 5 mg/kg on days 2–5 as a single dose) or erythromycin estolate (40 mg/kg/day in 3 divided doses for 10 days) with stratification by center. The primary outcome measure was bacteriologic cure of infection as determined by cultures of nasopharyngeal aspirates. Culture-positive participants had a second aspirate collected at the end of therapy (days 5–7 for azithromycin, days 10–12 for erythromycin) and 1 week after therapy. Bacteriologic cure was defined as negative cultures at the end of therapy. Bacteriologic relapse was defined as a positive culture 1 week after completion of therapy and after a negative end-of-therapy culture. Secondary outcomes were pertussis diagnosed by serology and polymerase chain reaction (PCR), treatment-associated adverse events, compliance, and presence of clinical symptoms at the end of the treatment course. Serology was performed using standard enzyme-linked immunosorbent assay methods. A participant was considered to have pertussis when the PCR was positive or a 4-fold increase in pertussis toxin antibody between baseline and follow-up visits was observed. PCR was performed using a 1046-bp ClaI DNA fragment from B pertussis. Adverse events (nausea, vomiting, diarrhea, any gastrointestinal complaint, or other) were determined by a parent-completed diary that was reviewed with study personnel during study visits. Compliance was measured by review of the parent medication diary during study visits and observation of medication containers by the pharmacist at study completion. Symptoms were determined by history collected by study personnel at enrollment and subsequently from the diary. The design of the study was an equivalence trial, aimed at demonstrating that the bacteriologic failure rates with the 2 therapies did not differ by >8%. For the safety analysis, all participants who received at least 1 dose of study drug were included. In the per-protocol efficacy analysis, all culture-positive participants with end-of-treatment cultures were considered.
Results. A total of 477 children were enrolled and randomly assigned to either azithromycin (n = 239) or erythromycin (n = 238). Of these children, 114 (24%) grew B pertussis from nasopharyngeal specimens (azithromycin group: 58 of 239 [24%]; erythromycin group: 56 of 238 [23%]); these children composed the efficacy cohort for the per-protocol and intention-to-treat analyses. Serology and PCR added 52 children to the number considered to have pertussis for a total of 35% (166 of 477) of all children who presented with cough illness. In the safety analysis (antibiotic side effects, compliance) and comparison of cough symptoms after treatment, all randomized children are reported in their assigned treatment group. At end of therapy, bacterial eradication was demonstrated in all 53 patients in the azithromycin group and all 53 patients in the erythromycin group with follow-up cultures available (eradication 100%; 95% confidence interval [CI]: 93.3–100). No bacterial recurrence was demonstrated in children with 1 week posttreatment nasopharyngeal cultures available (51 and 53 participants in the azithromycin and erythromycin arms, respectively [0%, 95% CI: 0–7.0; and 0%, 95% CI: 0–6.7]). No serious adverse events attributable to study drug were observed. Gastrointestinal adverse events were reported less frequently in azithromycin (18.8%; 45 of 239) than in erythromycin estolate (41.2%; 98 of 238) recipients (90% CI on difference: –29.0% to –15.7%) as a result of less nausea (2.9% vs 8.4%; 95% CI: –8.9% to –2.0%), less vomiting (5.0% vs 13.0%; 95% CI: –4.9% to –1.4%), and less diarrhea (7.1% vs 11.8%; 95% CI: –9.0% to –0.3%). Children who were randomized to azithromycin were much more likely to have complied with antimicrobial therapy over the treatment period. In the azithromycin group, 90% of children took 100% of prescribed doses, whereas only 55% of children in the erythromycin group took 100% of prescribed doses.
Conclusions. In this large, multicenter, randomized trial, we found that azithromycin is as effective as erythromycin estolate for the treatment of pertussis in children. Gastrointestinal adverse events were much more common with erythromycin treatment than azithromycin. Compliance with therapy was markedly better with azithromycin than with erythromycin in this study.
Key Words: Bordetella pertussis • macrolides • respiratory tract infection
Abbreviations: PCR, polymerase chain reaction • CI, confidence interval
Although universal immunization against Bordetella pertussis (whooping cough) infection has resulted in dramatic reductions in the incidence of pertussis,1 outbreaks continue to occur in countries with excellent vaccine coverage2–5 and are associated with morbidity and mortality.6–8 Treatment given during the catarrhal phase of pertussis may ameliorate symptom severity9 but has no effect on paroxysms, emesis, or apnea associated with cough illness if given during the paroxysmal or convalescent phases. The rationale for antibiotic treatment of suspected or proven pertussis is to prevent transmission of infection by eradicating the bacteria from the nasopharynx.
Erythromycin is recommended as the drug of choice for the prophylaxis and treatment of pertussis,10 although it is poorly tolerated because of gastrointestinal side effects in up to 30% of patients,11,12 which may lead to noncompliance with therapy.13 Newer macrolide antimicrobials such as azithromycin and clarithromycin are better tolerated and offer the advantage of less frequent dosing and shorter duration of therapy, factors associated with improved antibiotic compliance.14 We compared the safety and efficacy of azithromycin with erythromycin estolate in the treatment of pertussis in a large, randomized trial that enrolled children from 11 sites across Canada and 1 in the United States from 1995 to 1998.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
Participants
Eligible children were aged 6 months to 16 years and had either culture-proven B pertussis infection or a cough illness suspected by a physician to be pertussis that met the definition for a suspect case. A suspect case was a child with at least 1 of the following: 1) paroxysmal cough of any duration; 2) cough with inspiratory whoop; 3) cough ending in apnea, vomiting, or gagging with no other known cause; or 4) cough of any type in a child in contact with a culture-proven case of pertussis.
Children were not eligible when they had known allergy to any macrolide antimicrobial; immunodeficiency; had hepatic, renal, cardiovascular, or hematologic disease; had underlying lung disease with chronic symptoms; had gastrointestinal absorption disorder; had concomitant use of theophylline, digitalis, phenytoin, cyclosporin, carbamazepine, warfarin, triazolam, terfenadine, astemizole, ergot alkaloids, or zidovudine; or were already receiving macrolide, chloramphenicol, tetracycline, trimethoprim-sulfamethoxazole, or clindamycin.
Children were recruited through the primary care practices of physicians in 1 American and 11 Canadian urban centers. All participants or their parents gave written informed consent. The study was approved by the local research ethics boards in all participating centers.
Interventions
Children who were assigned to azithromycin received 10 mg/kg (maximum: 500 mg) by mouth on the first day of treatment and 5 mg/kg (maximum daily dose: 250 mg) once daily on the second to fifth days of treatment. In the erythromycin group, 3 doses of erythromycin estolate (40 mg/kg/day; maximum: 1 g) were given by mouth for 10 days.15 Parents/guardians were instructed not to administer any medication other than acetaminophen or ibuprofen until after the last study visit had been completed.
Randomization
Participants were randomly assigned to 2 parallel groups, the erythromycin and azithromycin groups, with stratification by center. A computer-generated randomization list was prepared by the statistician and provided to each center’s pharmacy. The interventions were assigned by the pharmacist, and the allocation sequence was concealed from other study personnel until after the child was enrolled. Because of the differences in duration of therapy, dosing, and ease of product recognition, group assignment was not blinded after randomization.
Outcomes
The primary outcome measure was bacteriologic cure of infection as determined by cultures of nasopharyngeal aspirates for B pertussis. Culture-positive participants had a second aspirate collected at the end of therapy (days 5–7 for azithromycin, days 10–12 for erythromycin) and 1 week after therapy. Collection of nasopharyngeal aspirates and inoculation of specimens for B pertussis culture were performed according to standard methods.16–18 Bacteriologic cure was defined as negative cultures at the end of therapy. Bacteriologic relapse was defined as a positive culture 1 week after completion of therapy and after a negative end-of-therapy culture.
Secondary outcome measures were pertussis diagnosed by serology and polymerase chain reaction (PCR), treatment-associated adverse events, compliance, and presence of clinical symptoms at the end of the treatment course. Serology was performed using standard enzyme-linked immunosorbent assay methods.19 A participant was considered to have pertussis when the PCR was positive or a 4-fold increase in pertussis toxin antibody between baseline (visit 1) and follow-up visits (visit 4) was observed. PCR using a 1046-bp ClaI DNA fragment from B pertussis was performed.20 Adverse events (nausea, vomiting, diarrhea, any gastrointestinal complaint, or other) were determined by a parent-completed diary that was reviewed with study personnel during study visits. Compliance was measured by review of the parent medication diary during study visits and observation of medication containers by the pharmacist at study completion. Symptoms were determined by history collected by study personnel at enrollment and subsequently from the diary.
Statistical Methods
The design of the study was an equivalence trial, aimed at demonstrating that the bacteriologic failure rates with the 2 therapies did not differ by >8%. For the safety analysis, all participants who received at least 1 dose of study drug were included. For the efficacy analysis, we performed intention-to-treat analysis and per-protocol analyses in culture-positive participants. In the per-protocol analysis, all culture-positive participants with end-of-treatment cultures were considered. In the intention-to-treat analysis, all culture-positive participants who received at least 1 dose of study drug were considered. We also report the prevalence of pre- and posttreatment cough-related symptoms in those who had laboratory-confirmed pertussis (culture and/or serology and/or PCR) in both treatment groups.
Baseline characteristics of treatment groups were compared using Fisher exact and t tests. Rates of bacteriologic failure were estimated by treatment group, together with exact binomial confidence intervals (CIs). Following Baughman et al,21 the hypothesis of equivalence is rejected when a 90% 2-sided CI for the observed difference in rates is not contained in the equivalence interval (–8% to 8%). Unless otherwise specified, the intervals reported are the simple asymptotic intervals.22 Adverse events rates were estimated and compared in the same manner. The software program SAS Version 8.2 (SAS Institute Inc, Cary, NC) was used for these analyses.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
|
Efficacy
At the end of therapy, bacterial eradication was demonstrated in all patients in the azithromycin and erythromycin groups for whom cultures were available (53 of 53 vs 53 of 53; eradication 100%; 95% CI: 93.3–100). No bacterial recurrence was demonstrated in the 51 patients in the azithromycin group (0%; 95% CI: 0–7.0) or the 53 patients in the erythromycin group (0%; 95% CI: 0–6.7) with 1 week posttreatment cultures available.
In the intention-to-treat analysis, the children who did not have cultures at the end of therapy (n = 5 azithromycin, n = 3 erythromycin) or 1 week after treatment (n = 6 azithromycin, n = 5 erythromycin) were assumed to be treatment failures. Participants with protocol deviations or violations were also assumed to have failed therapy. In this comparison, azithromycin was theoretically less effective than erythromycin in eradicating B pertussis at the end of treatment (93.2% vs 94.6% eradication; 95% CI: 83.5–98.1 vs 85.1–98.9; 90% CI on difference: –8.73 to 5.89) and may have had a higher recurrence rate at 1 week posttreatment (10.2% vs 5.4%; 95% CI: 3.9–20.8 vs 1.1–14.9; 90% CI on difference: 3.34–12.96). In the intention-to-treat cohort, exact 90% CI on the eradication rate was –16.2 to 12.1 and on the recurrence rate was –9.0 to 19.7.
At study completion, cough-related symptoms persisted in many children with laboratory-confirmed pertussis (culture and/or serology and/or PCR) and in those with undiagnosed cough illness (Fig 2). However, the prevalence of cough-related symptoms had decreased in all children, whether analyzed according to their treatment group or laboratory diagnosis. Equivalence testing of the prevalence of each symptom at study completion for the azithromycin and erythromycin groups yielded the following results: paroxysmal cough (67.5% vs 66.3%; CI: –10.8% to 13.2%), vomiting (30% vs 24.4%; CI: –5.8% to 16.9%), and apnea/cyanosis (28.8% vs 25.6%; CI: –8.2% to 14.5%); prevalence of whoop was decreased (41.3% vs 48.8%; CI: –20.2% to 5.1%).
|
|
Children who were randomized to azithromycin were much more likely to have complied with antimicrobial therapy during the treatment period (Fig 4). In the azithromycin group, 90% of children took 100% of prescribed doses compared with 55% of the erythromycin group.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
Nasopharyngeal cultures were also obtained 1 week after completion of therapy to assess disease relapse, and no relapses occurred in the 91% (104 of 114) of children who had this outcome measure ascertained. In the intention-to-treat analyses, however, in which the 2 participants who had protocol violations were assumed to have failed therapy in a "worst-case scenario," 89% of azithromycin recipients and 94.6% of erythromycin recipients would have remained culture negative. The CIs around these point estimates indicate that as many as 15% of children on erythromycin could have positive cultures 1 week after therapy and as many as 21% of children on azithromycin. We suggest that these are extraordinarily high estimates that are not consistent with clinical experience with macrolides and indicate that the intention-to-treat analysis is overly conservative in this instance. In case series, erythromycin failures are reported to occur in 10% (n = 1 of 10)24 to 11% (n = 2 of 18)25 of patients and in randomized controlled trials in 1% (n = 2 of 168),11 4.3% (n = 1 of 23),26 and 6% (n = 1 of 17)9 of patients on erythromycin. Relapse did not occur in patients who were treated with azithromycin in 2 case series25,27 but has been reported in 1 case report.28 Relapse of infection with clarithromycin is equally uncommon.25,26 The only randomized controlled trials of pertussis treatment other than this report are of 7 compared with 14 days of erythromycin11 and clarithromycin compared with erythromycin.26
Compliance with therapy was markedly better with azithromycin than with erythromycin in this study, with 95% of children taking >90% of doses compared with only 60% of children on erythromycin. The gastrointestinal side effects of erythromycin are widely known26,29,30 and thought to be attributable to the agonist effect of the drug on intestinal motilin receptors. We hypothesize that inability to comply with therapy may actually account for some erythromycin failures in the treatment and prophylaxis of pertussis. However, it is noteworthy that no treatment failures were observed on a 10-day course of erythromycin, even with poor compliance. We have previously shown that a 7-day course of erythromycin is as effective as a 14-day course.11 In the absence of evidence of a clear treatment benefit, it seems unreasonable to recommend erythromycin as the drug of choice,10 because up to 30% of patients will have significant gastrointestinal side effects. A course of azithromycin is considerably more expensive than erythromycin, however, and this must be considered when an antimicrobial choice is made.
Only 
24% of participants who were enrolled in this study ultimately had culture-proven B pertussis infection, although this was suspected by their primary care clinician as the cause of their infection; serologic and PCR diagnosis increased this percentage to 35%. It is likely that the most common cause of infection in these children, 65% of those who presented with cough illness, was viral. Chlamydia pneumoniae and Mycoplasma pneumoniae31 may also present with cough illness but are susceptible to macrolides. Alternatively, infection may have initiated airway damage leading to cough, and antimicrobial therapy would be unlikely to alter these symptoms. The cause of cough illness is difficult if not impossible to diagnose clinically, and this is even more difficult in the earlier nonspecific catarrhal stage of pertussis. This leads to late treatment of children with pertussis and accounts for the long period of enrollment in this multicenter trial. Accurate and early diagnosis would be possible only with easy access to rapid diagnostic tests. When such tests are available, it is possible that early treatment of pertussis will reduce transmission and potentially interrupt development of prolonged cough illness.
|
ACKNOWLEDGMENTS |
|---|
PICNIC Investigators are Joanne Langley, MD, Scott Halperin, MD, and Bruce Smith, PhD, Dalhousie University (Halifax and Bathurst); François Boucher, MD, Centre Hospitalier de l’Université Laval (Quebec City); Elaine Mills, MD, and Earl Rubin, MD, Montreal Children’s Hospital (Montreal); Elaine Wang, MD, and Stanley Read, MD, The Hospital for Sick Children (Toronto); Barbara Law, MD, University of Manitoba (Winnipeg); Ben Tan, MD, University of Saskatchewan (Saskatoon); Wendy Vaudry, MD, University of Alberta (Edmonton); H. Dele Davies, MD, University of Calgary (Calgary; now the Chair, Pediatrics and Human Development, Michigan State University); Ole Hammerberg, MD, University of Western Ontario (London); and Michael Pichichero, MD, University of Rochester (Rochester).
We thank Kate MacIntosh for assistance in study coordination and Steve Ioannou and Danielle Carr of Pfizer Canada Inc for assistance in study management.
|
FOOTNOTES |
|---|
Reprint requests to (J.M.L.) Clinical Trials Research Center, IWK Health Centre, 5850 University Ave, Halifax, NS B3J 3G9, Canada. E-mail: joanne.langley{at}dal.ca
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Singh and K. Lingappan Whooping cough: the current scene. Chest, November 1, 2006; 130(5): 1547 - 1553. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Bonacorsi, C. Farnoux, P. Bidet, V. Caro, S. Aizenfisz, M. Benhayoun, Y. Aujard, N. Guiso, and E. Bingen Treatment failure of nosocomial pertussis infection in a very-low-birth-weight neonate. J. Clin. Microbiol., October 1, 2006; 44(10): 3830 - 3832. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. S. Braman Postinfectious Cough: ACCP Evidence-Based Clinical Practice Guidelines Chest, January 1, 2006; 129(1_suppl): 138S - 146S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. S. Dworkin Adults Are Whooping, but Are Internists Listening? Ann Intern Med, May 17, 2005; 142(10): 832 - 835. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Mattoo and J. D. Cherry Molecular Pathogenesis, Epidemiology, and Clinical Manifestations of Respiratory Infections Due to Bordetella pertussis and Other Bordetella Subspecies Clin. Microbiol. Rev., April 1, 2005; 18(2): 326 - 382. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Journal Watch Arch. Dis. Child., January 1, 2005; 90(1): 106 - 107. [Full Text] [PDF]
|
|
|
|
 |
|
 Pertussis: Still with Us, but Easier to Treat Journal Watch Pediatrics and Adolescent Medicine, August 30, 2004; 2004(830): 2 - 2. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2004 American Academy of Pediatrics. All rights reserved. |
||
|
||
