Objective and Methods. Although 14 days of erythromycin is recommended for the treatment ofBordetella pertussis infection, there have been no prospective controlled studies to support the contention that this long course of therapy is required to eradicate the microorganism from the nasopharynx or to prevent bacteriological relapse. We randomly allocated children and adults with culture-positive community-acquired pertussis to either 7 or 14 days of erythromycin estolate treatment (40 mg/kg/d; maximum dose 1 g/d). Nasopharyngeal aspirate cultures were obtained by study nurses during home visits before and at the end of treatment, and 1 week after the completion of treatment. B pertussis-specific antibodies were measured before treatment and 1 month later. Information about clinical symptoms, adverse reactions, and compliance were collected at each scheduled contact.
Results and Conclusions. A total of 168 participants were eligible for analysis (74 treated for 7 days and 94 treated for 14 days). Bacteriological persistence (positive end of therapy culture) occurred once in each group, and bacteriological relapse (positive culture 1 week after completion of treatment) occurred in one participant treated for 7 days. The overall failure rate (persistence plus relapse) of 2.70% in the 7-day group was not different than the rate of 1.06% in the 14-day group. The study had a power of 99.99% at the 5% level to detect a difference in failure rates of 10% and a power of 80% to detect a difference of 5%. We conclude that 7 days of erythromycin estolate is as effective as 14 days for the eradication ofB pertussis.
Pertussis (whooping cough) is an acute respiratory infection caused by Bordetella pertussis and characterized by a paroxysmal cough which often ends in an inspiratory whoop and/or vomiting. Through the widespread use of pertussis vaccine in North America, the incidence of pertussis has decreased by more than 90% in Canada and the United States1,2; however, outbreaks of pertussis continue to occur, with a resurgence of cases in recent years.3-11 Although antimicrobial therapy has little effect on the course of the illness once the individual has developed a paroxysmal cough,12-15 treatment with erythromycin is recommended to eradicate the bacteria from the nasopharynx, thereby limiting transmission of the infection.16 Throughout the last three decades, recommendations for the duration of therapy has ranged from 5 to 14 days.14-20 After a series of editorials, reviews, and letters which asserted that 14 days of erythromycin estolate was required to prevent bacteriological relapse of pertussis,21-24 the American Academy of Pediatrics recommended 14 days of erythromycin for the treatment of pertussis.25 In Canada, 10 days of treatment has been recommended.26 All recommendations on the duration of treatment have been based on reviews of observational studies. A prospective examination of the optimal or minimal effective duration of therapy for pertussis has not been done. We describe the first randomized, controlled, clinical trial designed to determine whether 14 days of erythromycin estolate is superior to 7 days for the treatment of pertussis.
Children and their household contacts with culture-positive pertussis were eligible for participation in the study. Potential participants were identified by a daily review of positive nasopharyngeal aspirate cultures (NPA) in the clinical microbiology laboratory of the IWK-Grace Health Centre. After obtaining permission from the family physician to contact the family, participants were recruited by study personnel. Exclusions for enrollment were allergy to erythromycin, preexisting liver disease, or pregnancy. Written informed consent was obtained from the participants or their parents. The protocol was approved by the Research Ethics Board of the IWK-Grace Health Centre.
Randomization, Study Drug, and Study Procedures
Participants were randomized to receive 7 or 14 days of treatment using a table of random numbers. Erythromycin estolate (Eli Lilly, Canada) was given at a dose of 40 mg/kg/d in three divided doses with a maximum of 1000 mg/d. For participants who had already begun erythromycin before study enrollment, sufficient medication was provided to complete the course allocated by the randomization process. An unblinded study design was chosen because blinding would have required administration of a placebo for the second week in participants randomized to the 7-day group and collection of an additional NPA from all participants to ensure an end-of-therapy culture. The added trauma attributable to an unnecessary NPA and the increased complexity of design were not considered justified because the outcome measures were bacteriological and not likely to be susceptible to bias. Technologists in the microbiology laboratory were unaware of treatment group allocation.
Clinical data and study specimens were obtained through four home visits by nurses employed as study personnel. At the first visit, a questionnaire was completed by study personnel detailing the presence and duration of symptoms, medication was supplied, and blood was taken for baseline serology. Follow-up visits were scheduled for the end-of-therapy (day 7 or 14, respectively) and for one week postcompletion of therapy (day 14 or 21, respectively) at which time the symptom questionnaire was completed and NPAs were repeated. Compliance and antibiotic-associated adverse reactions were also assessed by questionnaire and by measurement of medication remaining in the bottles at the end-of-therapy visit. The number of doses taken was expressed as a percentage of the number prescribed (21 doses in the 7-day treatment group and 42 doses in the 14-day group). A convalescent serum specimen was obtained and the questionnaire was completed at the final visit 28 days after study entry.
Specimen Collection and Culture and Serological Methods
Cultures for B pertussis were obtained by nasopharyngeal aspirate.27 Secretions were rinsed from the catheter with .8 mL phosphate-buffered saline containing 1% casaminoacids (Difco Laboratories, Detroit, MI).28 Rinsed secretions were used to inoculate two Regan-Lowe plates, one with and one without cephalexin.29 Suspicious colonies were confirmed as B pertussis using specific agglutinating antisera; B parapertussis specific antisera was used as a negative control (Wellcome Diagnostics, UK).
Blood was collected by finger prick or venipuncture and serum separated and stored below −70°C. IgG and IgA antibodies against pertussis toxin (PT), filamentous hemagglutinin (FHA), the 69-kilodalton protein (69K; also known as pertactin), and fimbriae 2 and 3 (FIM) were measured by enzyme immunoassay.30 Antibody titers were expressed as reciprocal dilutions using single point, parallel line calculations.27
Treatment outcomes were based on rates of bacteriological failure. Bacteriological cure was defined as negative cultures at both the end-of-therapy and one week posttherapy specimen collections. Bacteriological persistence was defined as a positive culture at the end-of-therapy. Bacteriological relapse was defined as a positive culture one week postcompletion of therapy after a negative end-of-therapy culture. Any bacteriological failure was defined as bacteriological persistence or relapse.
The sample size of the study was estimated to provide a power of 80% to detect a difference in failure rates of 10% at theP < .05 level with an estimated range of failure rates in the 14-day group of 5%. With the true failure rate of a 14-day course of 1%, the study had a 99.99% power at the 5% level to detect a difference in rates of 10% and more than 80% power to detect a difference of 5%. Study groups were assessed for comparability at study entry and after treatment. Proportions were compared by Fisher's exact test or by χ2. Exact-power 95% confidence intervals (CI) were computed to estimate the difference in failure rates in the two groups. Ordinal outcomes were assessed by Wilcoxon tests and logistic regression. Continuous outcomes were compared byt tests. Durations of symptoms were compared by the log-rank test. Serological tests were compared after logarithmic conversion. Pretreatment, posttreatment, and the difference between pretreatment and posttreatment mean log-titers were compared by t tests.
Baseline Population Characteristics
A total of 193 culture-positive individuals were enrolled into the study; 87 were randomly allocated to the 7-day treatment group and 106 were allocated to the 14-day group (Fig 1). The unequal allocation was a result of the use of a random number table without blocking. There were 7 (8.0%) withdrawals from the 7-day group and 10 (9.4%) withdrawals from the 14-day group (P = .8, Fisher exact test). The reason for withdrawal was unwillingness to comply with any follow-up specimen collection (NPA cultures or serum). In one of these cases, side effects of medication were cited as the reason for withdrawal and the family would not permit further specimen collection.
An additional 8 participants were excluded from the analysis because of protocol violations (6 in the 7-day group and 2 in the 14-day group;P = .14). Protocol violations in the 14-day group were a change in the erythromycin formulation by the family physician from estolate to ethyl succinate in one case and a 50% decrease in dosage by the family physician in a second case. The protocol violations in the 7-day group were restarting erythromycin for otitis media after completion of the study medication (one case), administration of 9 days of treatment instead of 7 days (one case), and inadvertent dispensing of 14 days of treatment to a participant allocated to 7 days (four cases).
After withdrawals and exclusions were taken into account, 74 participants were eligible for analysis in the 7-day group and 94 participants were eligible for analysis in the 14-day group. There were no differences in the baseline characteristics of the two groups at study entry (Table 1). Antibody titers were also measured at the time of study entry. There were no differences in baseline IgG antibodies against PT, 69K, or FIM, or in IgA antibodies against PT, FHA, or FIM (data not shown). The group allocated to receive 14 days of therapy had higher IgG antiFHA antibody titers (399.0 vs 227.8; P = .01) and IgA anti69K antibody titers (15.8 vs 9.8; P = .02) than the group allocated to 7 days of treatment.
End-of-therapy cultures were available on 69 (93.2%) of the 74 participants treated for 7 days and 84 (89.4%) of the participants treated for 14 days (P = .43). The median day of specimen collection was the last day of treatment (standard deviation, .10 days; range, one day before completion to 4 days postcompletion). There was one (1.45%) participant with bacteriological persistence in the 7-day group and one (1.19%) with bacteriological persistence in the 14-day group (P = .98); the difference in the rates was .26% (95% CI, −7.7 to 10.4) (Table 2).
One week postcompletion of therapy cultures were available from 72 (97.3%) of the 74 participants in the 7-day group and 83 (83.3%) of the 94 participants in the 14-day group (P = .04). The median day of specimen collection was 7 days posttreatment (standard deviation, .19 days; range, 4 to 25 days). There was one (1.39%) bacteriological relapse in the 7-day group and no relapses in the 14-day group (P = .77). The difference between the relapse rates was 1.39% (95% CI, −5.8 to 10.8).
The total failure rate (persistence or relapse) was 2 (2.7%) of the 74 participants in the 7-day group and 1 (1.06%) of the 94 participants in the 14-day group (P = .74). The difference between the rates was 1.64% (95% CI, −6.0 to 11.7). If only participants for whom both follow-up cultures were available were analyzed, (67 [90.5%] of the 74 participants in the 7-day group and 73 [77.7%] of the 94 participants in the 14-day group), overall failure was documented in 2 (3.0%) of the 7-day group and 1 (1.4%) in the 14-day group (P = .82). The difference in the rates was 1.6% (95% CI, −8.0 to 12.6%).
Medication Adverse Events and Compliance
A mean of 92.4% of doses were taken by participants in the 7-day group and 94.5% in the 14-day group (P = .37). Adverse reactions associated with antimicrobial therapy were common (Fig 2). An adverse reaction of any type was reported by 33.8% of 7-day recipients and 44.7% of 14-day participants (P = .20). Gastrointestinal symptoms were the most common adverse reactions reported (28.0% and 37.2%, respectively; P = .25). Although all adverse reactions were more common in the 14-day group, none of the differences reached statistical significance.
The three treatment failures were explored in further detail. Case 1 was a 4-month-old girl in the 7-day group who had previously received one dose of pertussis vaccine; all 21 doses of the medication were reported as taken. Case 2 was an 8-month-old boy in the 7-day group with bacteriological relapse. He was previously immunized with two doses of pertussis vaccine and had not missed any of his 21 prescribed doses of erythromycin. The single failure in the 14-day group (Case 3) occurred in a 5-year, 10-month old girl who had received five doses of pertussis vaccine; this girl reported missing only 2 of 42 prescribed erythromycin doses.
Clinical Course and Antibody Response to Infection
The presence of clinical symptoms and the duration of symptoms were similar whether participants were treated for 7 or 14 days (Table3). The proportion of participants reporting cough (95.9% vs 98.9%), paroxysmal cough (90.4% vs 92.6%), whoop (71.2% vs 73.4%), and apnea (13.7% vs 12.8%) were all similar in the 7-day and 14-day treatment groups. The duration of symptoms was more variable but none of the differences reached statistical significance, nor were there any apparent trends. Congestion and cough lasted longer in the 14-day treatment group (62.6 and 64.3 days, respectively) compared with the 7-day group (35.5 and 40.6 days, respectively) whereas cough that was worse at night and whoop lasted longer in the 7-day group (53.0 and 39.5 days compared with 45.6 and 28.5 days in the 14-day group).
There was no apparent effect of the duration of therapy on the antibody response to infection (Fig 3). The most vigorous antibody response was detected against PT. Geometric mean antiPT IgG antibodies rose from 389.7 to 1584.2 in the 7-day treatment group and from 522.0 to 1575.5 in the 14-day treatment group (P > .1). AntiFHA IgG antibody titers rose from 227.8 to 396.2 in the 7-day group and from 399.0 to 660.4 in the 14-day group; both the acute and convalescent titers were higher in the 14-day group (P < .05) although there was no difference in the fold antibody rise between acute and convalescent titers (1.7 and 1.6, respectively). Similar degrees of antibody rise were demonstrated for antifimbriae IgG antibodies and anti69K IgG antibodies; no differences were detected between the two treatment groups.
IgA response was less dramatic than the IgG response although significant rises in titer were seen for both treatment groups for all antigens except antiPT in the 7-day treatment group. There were no differences in geometric mean IgA antibody titers between the two treatment groups. The antibody response of the three treatment failures were similar to the treatment successes (data not shown).
A minimum of 14 days of treatment of pertussis has been recommended by some20-25,31 to ensure bacteriological eradication and to prevent relapse, although shorter courses have been recommended by others.14,17,26 This study is the first prospective, randomized, clinical trial assessing the duration of erythromycin therapy for pertussis and demonstrates that 7 days of erythromycin estolate is as effective as 14 days for the treatment of pertussis.
B pertussis is susceptible to a number of antimicrobials in vitro;20,32-34 however, results do not correlate well with clinical efficacy.35 Although erythromycin, chloramphenicol, tetracycline, and kanamycin were equally effective in eradicating B pertussis from the nasopharynx in a mean of 3 to 4 days,13 its superior safety profile has made erythromycin the treatment of choice. Because of its superior bioavailability, erythromycin estolate has been recommended as the preferred formulation23-26 although the ethyl-succinate was also effective in the only comparative study reported.36 Although a single study supported the use of ampicillin in the treatment of pertussis,37 most studies have found it to be no better than placebo.13,14,32Duration of therapy has been variable with successful treatment reported after a minimum of 5 days16,17,19,37 or 7 days.13,14,18 Recently, a minimum of 14 days of therapy has been recommended based on reanalyses of treatment failures reported in the literature and the assertion that no failures had occurred in individuals treated with the longer courses.20-25 Although failures have been reported with 14 days of treatment38 and one occurred in the present study, 14 days of therapy continued to be recommended by the American Academy of Pediatrics.
The results of this study indicate that a 7-day course of erythromycin treatment is as effective as a 14-day course for the treatment of pertussis. Bacteriological persistence occurred in one case in each group whereas a single relapse was documented in a child treated for 7 days. The study was of sufficient power that if a difference exists between the two treatment regimens it is too small to be clinically important. We are unable to provide an explanation for the three bacteriological failures that occurred with erythromycin treatment. Compliance with the medication was not a factor. Other possibilities include differences in intestinal absorption of the drug36and erythromycin resistance of the B pertussisstrain.39
Although there was a trend toward more adverse reactions with the longer course of treatment, none of the differences reached statistical significance. Surprisingly, there were no differences in compliance detected between the two treatment groups because duration of therapy is inversely associated with compliance.40 However, these compliance data were derived from a clinical trial and one must be careful to extrapolate this result to general practice in which family support structures such as weekly home visits by a nurse and the availability of 24-hour nurse or investigator contact do not exist. Even if shorter course therapy does not improve compliance, it will decrease medication costs and inconvenience to patients.
In untreated infection, B pertussis is cleared from the nasopharynx in a mean of 2 to 3 weeks.13 Therefore, late initiation of treatment could obscure potential differences between the courses of treatment if the nasopharyngeal cultures would have spontaneously become negative. However, this is unlikely to be an explanation for our findings because pertussis is well recognized by physicians in our community4 and the duration of symptoms before study entry was similar in both groups and to that reported in other studies.41
The results of this study suggest that the current recommendations for 14 days of treatment for pertussis should be reconsidered and that 7 days of treatment may be sufficient. However, even with a 7-day course of treatment, adverse reactions occurred in more than one-third of participants. The availability of newer macrolides which are associated with fewer gastrointestinal side effects may provide a further benefit if they are shown to be effective in the treatment of B pertussis infection.
This study was supported by grant 6603–1245–54 from the National Health Research and Development Program, Health Canada. Erythromycin estolate was a generous donation of Eli Lilly Canada Inc., Scarborough, Ontario, Canada.
We thank Ms Petra Rykers for data management, the rest of the staff of the Clinical Trials Research Center, and collaborating family physicians and pediatricians for the performance of the study, and the staff of the Microbiology Division at the IWK-Grace Health Centre.
- Received September 5, 1996.
- Accepted November 5, 1996.
Reprint requests to (S.A.H.) IWK-Grace Health Centre, 5850 University Avenue, Halifax, Nova Scotia, B3J 3G9 Canada.
- NPA =
- nasopharyngeal aspirate culture •
- PT =
- pertussis toxin •
- FHA =
- filamentous hemagglutinin •
- 69K =
- 69 kilodalton protein •
- FIM =
- fimbriae 2 and 3 •
- CI =
- confidence interval
- Centers for Disease Control
- ↵Centers for Disease Control. Pertussis surveillance—United States, 1989–1991. MMWR. 1992;41(no SS-8):11–19
- Farizo KM,
- Cochi SL,
- Zell ER,
- Brink EW,
- Wassilak SW,
- Patriarca PA
- Statistics Canada
- Ewanowich CA,
- Chui LWL,
- Peranchych MG,
- Peppler MS,
- Marusyk RG,
- Albritton WL
- Brem J,
- Kabiling TR
- Islur J,
- Anglin CS,
- Middleton PJ
- Nelson JD
- Altemeier WA III.,
- Ayoub EM
- Baraff LJ,
- Wilkins J,
- Wehrle PF
- Bass JW
- ↵American Academy of Pediatrics, Committee on Infectious Diseases. Report of the Committee on Infectious Diseases—Red Book. Elk Grove, IL: American Academy of Pediatrics; 1994
- National Advisory Committee on Immunization, Advisory Committee on Epidemiology, Canadian Paediatric Society
- Halperin SA,
- Bortolussi R,
- Wort AJ
- Regan J,
- Lowe F
- Adasek PJ,
- Meyer MN,
- Ray CG
- Bannatyne RM,
- Cheung R
- Kurzynski TA,
- Boehm DM,
- Rott-Petri JA,
- Schell RF,
- Allison PE
- ↵Lewis S, Erickson B, Cage G, et al. Erythromycin-resistant Bordetella pertussis—Yuma County, Arizona, May-October 1994. MMWR. 43;44:807–810
- ↵Wang E, Haynes RB. Compliance with antimicrobial therapy in children. Antimicrobial Therapy in Infants and Children. New York, NY: Marcel Dekker; 1988:104–114
- Copyright © 1997 American Academy of Pediatrics