|
|
Background Single-dose azithromycin is effective in the treatment of severe cholera in children, but its effectiveness in adults has not been evaluated.
Methods We conducted a double-blind, randomized trial comparing the equivalence of azithromycin and ciprofloxacin (each given in a single 1-g dose of two 500-mg tablets) among 195 men with severe cholera caused by Vibrio cholerae O1 or O139. Patients were hospitalized for five days. A stool culture was performed daily. Primary outcome measures were clinical success (the cessation of watery stools within 48 hours after drug administration) and bacteriologic success (the inability to isolate V. cholerae after 48 hours).
Results Therapy was clinically successful in 71 of 97 patients receiving azithromycin (73 percent) and in 26 of 98 patients receiving ciprofloxacin (27 percent) (P<0.001) and bacteriologically successful in 76 of 97 patients receiving azithromycin (78 percent) and in 10 of 98 patients receiving ciprofloxacin (10 percent) (P<0.001). Patients who were treated with azithromycin had a shorter duration of diarrhea than did patients treated with ciprofloxacin (median, 30 vs. 78 hours); a lower frequency of vomiting (43 percent vs. 67 percent); fewer stools (median, 36 vs. 52); and a lower stool volume (median, 114 vs. 322 ml per kilogram of body weight). The median minimal inhibitory concentration of ciprofloxacin for the 177 isolates of V. cholerae O1 was 0.25 µg per milliliter, which was 11 to 83 times as high as that in previous studies at this site.
Conclusions Single-dose azithromycin was effective in the treatment of severe cholera in adults. The lack of efficacy of ciprofloxacin may result from its diminished activity against V. cholerae O1 strains currently circulating in Bangladesh. (ClinicalTrials.gov number, NCT00229944
[ClinicalTrials.gov]
.)
The choice of antimicrobial drug for the treatment of cholera is guided by the patterns of resistance of the infecting organism. The development of resistance in V. cholerae O1 and O139 has diminished the use of a number of drugs that have historically been used in the treatment of cholera.3,4 Tetracycline and its derivatives have been the mainstay of cholera therapy for 40 years2 and are effective when used in a single dose.5,6,7 There are situations, however, such as cholera infections in children or pregnant women, in which the use of the tetracycline class of drugs is relatively contraindicated because of the risk of toxicity. Resistance to tetracycline and its derivatives has also been reported.8,9
Erythromycin has often been used as an alternative to tetracycline for the treatment of cholera.10 Unlike the tetracyclines, erythromycin has not been shown to be effective in a single dose, and currently recommended courses of treatment require 12 doses over a period of three days.10 As compared with erythromycin, azithromycin, a macrolide antibiotic derived from erythromycin, has better in vitro activity against V. cholerae,11 has a longer half-life,12 and has less gastrointestinal toxicity.13 It has also been shown to be effective when used in a single dose for the treatment of cholera in children.11 In this study, we evaluated the equivalence in adults of single-dose azithromycin and single-dose ciprofloxacin, a regimen that we have previously demonstrated to be effective in the treatment of severe cholera in adults.14
Methods
Study Patients and Protocol
Patients were enrolled at the Dhaka Diarrhoea Treatment Centre of the ICDDR, B–Centre for Health and Population Research in Bangladesh between December 11, 2002, and May 2, 2004. Eligible patients were men between the ages of 18 and 60 years who had had watery diarrhea for no more than 24 hours, severe dehydration (according to World Health Organization criteria10), a high purging rate (stool volume, 
20 ml per kilogram of body weight during a four-hour observation period after initial rehydration), and V. cholerae O1 or O139 isolated from a culture of stool or a rectal-swab sample. Patients were excluded from the study if they had a concomitant illness or had received an antimicrobial agent effective in the treatment of cholera. Women were excluded from the study because of difficulty in separating urine from stool and because social norms make it difficult for them to stay away from their homes.
Consecutive patients meeting initial eligibility requirements who presented for care at the Dhaka Diarrhoea Treatment Centre from 6 a.m. to 6 p.m. daily were identified. Such patients were rehydrated during a period of two to four hours and screened for V. cholerae O1 or O139 infection with the use of dark-field microscopy of a stool specimen. Patients with a presumptive diagnosis of cholera who had a volume of diarrhea of at least 20 ml per kilogram during the next four hours were eligible for participation in the study.
Eligible patients who provided written informed consent to participate in the study were randomly assigned, in a double-blind fashion, to receive orally either two 500-mg tablets of azithromycin and a placebo formulation of ciprofloxacin or two 500-mg tablets of ciprofloxacin and a placebo formulation of azithromycin. (Pfizer provided the azithromycin and azithromycin placebo tablets, and Square Pharmaceuticals provided the ciprofloxacin and ciprofloxacin placebo tablets but had no other role in study design, data accrual or interpretation, or the writing or approval of the manuscript.) Patients were admitted to the study ward of the treatment center for five days. (Study days were enumerated from the time of the administration of the first dose of study drug.) Every six hours, vital signs, the presence of watery stool (defined as stool that could be poured), and fluid balance (volume of stool, urine, vomitus, and oral and intravenous fluid intake) were assessed. Each day after admission, interval histories were obtained and physical examinations were performed. Hydration was maintained with a rice-based oral rehydration preparation, and intravenous fluids were administered only if the oral intake of fluid was insufficient to maintain hydration. Patients were asked to return for a follow-up visit 7 to 10 days after discharge to assess clinical and bacteriologic status.
A stool specimen for the isolation of V. cholerae O1 or O139, salmonella, shigella and Campylobacter jejuni was obtained before administration of the study drug, on study day 3, and at the follow-up visit. A rectal-swab sample for isolation of V. cholerae O1 or O139 was obtained before drug administration, daily while patients were in the treatment center, and at the follow-up visit if a stool sample could not be obtained. Microscopical examination of stool for semiquantitative measurement of leukocytes and erythrocytes and for detection of enteric parasitic infection was performed before the initiation of the study drugs. A blood sample was obtained after rehydration but before the study drug was administered for a complete blood count and a determination of serum electrolyte and creatinine levels. V. cholerae isolates were tested for susceptibility to azithromycin and ciprofloxacin by both the disk-diffusion test and the E test (AB Biodisk). Susceptibility to tetracycline, trimethoprim–sulfamethoxazole, chloramphenicol, erythromycin, and furazolidone was done with the disk-diffusion method only. The study protocol was approved by the ethics review committee of the ICDDR, B–Centre for Health and Population Research and the institutional review board of the New England Medical Center, where Dr. Bennish was affiliated at the time of the study.
Study Outcomes
The primary study outcomes were the clinical success of therapy (defined as the cessation of watery stools within 48 hours after the administration of the study drug) and bacteriologic success of therapy (defined as the inability to isolate V. cholerae O1 or O139 from stool or rectal-swab samples within 48 hours after the administration of the study drug). Secondary outcome measures were the rate of clinical or bacteriologic relapse (defined, respectively, as the reappearance of watery stool or the isolation of V. cholerae from the stool at any time during the hospital stay after being absent for 24 hours), the duration of diarrhea (defined as the time to the end of the last 6-hour period in which there was watery stool), the total volume of watery stool, the frequency and volume of vomitus, the duration of excretion of V. cholerae, and the volume of oral and intravenous fluids that was administered. All secondary outcome measures were determined from the time of administration of antimicrobial therapy.
Statistical Analysis
Sample size was calculated for an equivalence study. A regimen of 1 g of oral ciprofloxacin had a rate of clinical success of 94 percent and a rate of bacteriologic success of 95 percent in a previous study.14 With a type I error of 5 percent and a type II error of 20 percent, 91 patients were required in each treatment group to show that the 95 percent confidence interval for the difference in the rates of clinical and bacteriologic success between the treatment groups did not exceed 10 percent.15
The binomial method was used to calculate differences in medians between study groups and the confidence intervals for those differences; Newcombe's method was used to determine confidence intervals for differences in proportions.16 We constructed a Kaplan–Meier survival curve to show differences in the duration of diarrhea and of excretion of V. cholerae between the treatment groups and tested the significance of the differences in duration with the use of the log-rank test. All tests of significance were two-tailed. To determine factors that were independently associated with the clinical success or failure of therapy, we performed a backward, stepwise, multiple logistic-regression analysis, including variables in the initial model that could be associated biologically with the failure of therapy and that occurred in at least 10 percent of patients.
The analysis of the primary outcomes was conducted on an intention-to-treat basis, with all patients eligible for study who were assigned to treatment included in the analysis. Treatment was considered to have failed clinically and bacteriologically in patients in whom the primary end points could not be assessed because the patients did not complete the study. Analyses of secondary outcomes were conducted only among patients who completed the study.
Results
Enrollment and Characteristics of Patients
A total of 198 of the 325 patients who were screened met the eligibility criteria and consented to take part in the trial, with 99 patients being assigned to each of the treatment groups (Figure 1). One patient in the ciprofloxacin group and two in the azithromycin group did not have V. cholerae O1 or O139 isolates and were excluded from the analysis. Four patients in the ciprofloxacin group and nine in the azithromycin group did not complete the study, and treatment in these patients was considered to have failed (Figure 1).
|
|
Azithromycin was clinically successful in 71 of 97 patients (73 percent), and ciprofloxacin was clinically successful in 26 of 98 patients (27 percent) (absolute difference, 47 percent; 95 percent confidence interval, 33 to 58 percent; P<0.001) (Table 2). Bacteriologic success was achieved in 76 patients treated with azithromycin (78 percent) and in only 10 patients treated with ciprofloxacin (10 percent) (absolute difference, 68 percent; 95 percent confidence interval, 56 to 77 percent; P<0.001) (Table 2).
|
Secondary Outcomes
Patients who were treated with azithromycin had a significantly shorter duration of diarrhea, fewer stools, a lower volume of stool, and a lower frequency of vomiting than did patients who received ciprofloxacin. They also required less intravenous and oral fluids and excreted V. cholerae for a shorter period (Table 2 and Figure 2).
|
Of the patients who completed the study, 67 of the 88 who received azithromycin (76 percent) and 69 of 94 who received ciprofloxacin (73 percent) returned for a follow-up visit. No patient reported having diarrhea after discharge from the treatment center, and none had a stool or rectal-swab sample that yielded V. cholerae O1 or O139 on culture.
Adverse Events
Three patients in the ciprofloxacin group and one in the azithromycin group reported having abdominal pain. There were no serious adverse events attributable to either drug during the hospital stay or at follow-up.
Antimicrobial Susceptibility
All 168 isolates of V. cholerae O1 obtained from patients in the study were susceptible to tetracycline, doxycycline, erythromycin, azithromycin, and ciprofloxacin and were resistant to furazolidone and trimethoprim–sulfamethoxazole when tested by the disk-diffusion method. The 14 isolates of V. cholerae O139 were susceptible to all seven drugs. All V. cholerae O1 and O139 isolates were susceptible to azithromycin and ciprofloxacin when tested by the E test with the use of standard threshold levels for susceptibility.17,18 However, median minimal inhibitory concentrations (MICs) of ciprofloxacin for V. cholerae O1 isolates were 11 times as high as those for V. cholerae O139 isolates (0.250 µg per milliliter vs. 0.0230 µg per milliliter), with a difference of 0.227 (95 percent confidence interval, 0.174 to 0.234; P=0.001) and 11 to 83 times as high as those for V. cholerae O1 isolates in previous studies of single-dose ciprofloxacin therapy for cholera.14,19
Outcomes of Therapy
Therapy failed in 98 of the 195 patients in the study (50 percent). Factors predictive of treatment failure were the receipt of ciprofloxacin as treatment, the number of stools before admission, and stool volume during the observation period (Table 3).
|
Our study has three important findings: a single dose of azithromycin is effective in the treatment of cholera; a single dose of ciprofloxacin is clinically and bacteriologically ineffective in the treatment of cholera caused by strains of V. cholerae O1 that have diminished in vitro susceptibility to ciprofloxacin, such as those found in patients in this study; and the current thresholds of antimicrobial susceptibility to ciprofloxacin may be inappropriate for V. cholerae O1. In this study, the duration of watery diarrhea with azithromycin (median, 30 hours; mean, 39 hours) was similar to that previously reported for a single 300-mg dose of doxycycline (median, 32 hours),5 a single 1-g dose (mean, 35 hours) or 2-g dose of tetracycline (mean, 43 hours),6 and a single 1-g dose of ciprofloxacin in the treatment of fully susceptible strains of V. cholerae O1 or O139 (median, 30 hours; mean, 31 hours),14 and it was shorter than that for a single 400-mg dose of furazolidone (mean, 74 hours).20 The frequency and volume of stool, the need for intravenous fluids, the frequency and volume of vomiting, and bacteriologic success rates also compared favorably with those of other single-dose treatment regimens.5,6,7,14,20
In contrast, despite previous reports14,19 of the efficacy of ciprofloxacin for the treatment of cholera, measures of clinical outcome in this study of the 87 patients with V. cholerae O1 infection treated with ciprofloxacin were similar to those of patients with cholera in previous studies who received placebo or no antimicrobial treatment.6,20 The most likely reason for the lack of efficacy of ciprofloxacin was the diminished susceptibility of V. cholerae O1 isolates. The median MIC of V. cholerae isolates in this study was 83 times as great as that for V. cholerae O1 isolates in a 1996 study of single-dose ciprofloxacin among 66 adults (which showed a 94 percent rate of clinical success and a 95 percent rate of bacteriologic success14), 11 times as great as that recently reported in a study of single-dose ciprofloxacin therapy for V. cholerae O1 infection among 78 children (which showed a 60 percent rate of clinical success and a 42 percent rate of bacteriologic success19); and 11 times as great as that for V. cholerae O139 isolates in this study (which showed a 100 percent rate of clinical success and a 57 percent rate of bacteriologic success). This failure of therapy occurred despite the ability of ciprofloxacin to achieve concentrations in stool that are considerably higher than the MIC of the V. cholerae O1 isolates in this study14 — a finding that emphasizes the limitations in our understanding of the pharmacodynamics of antimicrobial therapy.21
All V. cholerae O1 isolates in this study were susceptible to ciprofloxacin according to the MIC, as determined by the E test (with susceptibility defined as a MIC 
1 µg per milliliter) or the disk-diffusion method (zone of inhibition, 21 mm).22 In our study, the lack of clinical response of infections caused by V. cholerae O1 with diminished susceptibility is similar to the findings of studies of salmonella with diminished susceptibility to the fluoroquinolones23 and, to a lesser degree, of Neisseria gonorrhoeae.24 The changing susceptibility of N. gonorrhoeae resulted in a decision by the Centers for Disease Control and Prevention to adopt new dilution breakpoints for the interpretation of resistance of N. gonorrhoeae to the fluoroquinolones.24
Our study had a number of limitations, including the fact that the trial was conducted only among men; however, there is no reason to think that women would have had a different response. There were differences between the two treatment groups on admission and during the observation period. It is unlikely, however, that these differences accounted for the very large differences in outcome between the two groups. In a multiple logistic-regression analysis, the drug treatment that patients received was the most important predictor of clinical outcome.
What are the practical implications of this study? First, single-dose azithromycin has now been established as an effective drug for the treatment of cholera caused by susceptible strains of V. cholerae in both adults and children. Inexpensive generic formulations of the drug are now available in many countries. (The cost of 1 g of azithromycin in Bangladesh is 95 cents, similar to the 90-cent cost of 12 250-mg doses of erythromycin.) Unfortunately, since azithromycin has been used for this indication, strains of V. cholerae O1 that are resistant to both erythromycin and azithromycin have been identified.25
Second, for the clinician treating patients with cholera or for sentinel surveillance programs, a determination of whether ciprofloxacin will be effective for the treatment of cholera will be problematic without adequate support from a microbiologic laboratory. In most areas where cholera is endemic or epidemic, if susceptibility testing is done at all, the simpler disk-diffusion method is likely to be used.26 On the basis of currently defined thresholds for resistance, the use of the disk-diffusion method will not identify strains of V. cholerae with diminished susceptibility that may be clinically resistant to single-dose, and possibly multiple-dose, ciprofloxacin therapy.
Third, this study emphasizes the critical problem posed by increasing antimicrobial resistance. Since V. cholerae O1, in Bangladesh and elsewhere, has developed substantial antimicrobial resistance to trimethoprim–sulfamethoxazole, furazolidone, and tetracycline, and resistance is emerging to ciprofloxacin and azithromycin, options for effective antimicrobial treatment are meager. There is a great need for the development of new, affordable, effective antimicrobial compounds.27
Supported by a grant from Pfizer (UO1 AI45508-01 and K24 AIHDO01671, to Dr. Bennish), grants from the National Institute of Allergy and Infectious Diseases, and a grant (62925, to Dr. Bennish) from the Wellcome Trust.
Dr. Bennish reports having spoken at a conference sponsored by Pfizer in January 2005, with the speaker's fee donated to charity. No other potential conflict of interest relevant to this article was reported.
We are indebted to Dr. Charles Knirsch of Pfizer for his support of the study, to Hosne Ara Shova for maintaining the drug inventory and dispensing study medication, to Humayun Kabir for assistance with data entry, and to the nurses and staff of the Clinical Research Ward at the ICDDR, B–Centre for Health and Population Research for providing care and support to study patients and their attendants.
Source Information
From the International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh (D.S., M.M.K., W.A.K., S.A., M.A.S.); the School of Family Medicine and Public Health, University of KwaZulu-Natal, Durban, South Africa (M.L.B.); and the Nuffield Department of Clinical Medicine, University of Oxford, Oxford, United Kingdom (M.L.B.).
Address reprint requests to Dr. Saha at the Clinical Sciences Division, ICDDR, B–Centre for Health and Population Research, G.P.O. Box 128, Dhaka 1000, Bangladesh, or at dsaha{at}icddrb.org.
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. |
Previous
