LAU, CHUEN-YEN MD, MS*; QURESHI, AZHAR K. MD, DrPH†
CHLAMYDIA TRACHOMATIS (CT) infection is the most common bacterial cause of sexually transmitted disease (STD) in the United States. It affects approximately 4 million Americans annually at an estimated cost of $2.4 billion per year. The disease affects all age groups, with greatest risk for individuals younger than age 25 years. 1,2 Chlamydial infection is common not only in the United States; it is a serious international health issue. It is estimated that at any point in 1995, 81 million people worldwide were infected with CT. Infection is more prevalent in females across all regions. 3
Genital CT infection has significant health ramifications. Initial clinical manifestations vary from no symptoms (80% of females and 50% of males) to pain, bleeding, and vaginal or urethral discharge. Untreated infections can progress to pelvic inflammatory disease, Fitz-Hugh Curtis perihepatitis, tubal damage such as infertility and ectopic pregnancy, chronic pain from adhesions, neonatal transmission, conjunctivitis, epididymoorchitis, and arthritis. 4 There is also evidence of an association between genital CT infection and cervical squamous cell carcinoma, as well as risk of HIV-1 transmission, in women. 5,6
The Centers for Disease Control and Prevention currently recommends either doxycycline (100 mg orally twice a day for 7 days) or azithromycin (1 g orally once) for treatment of chlamydial urethritis and cervicitis. 7 Thus far, randomized trials have not shown a significant difference in efficacy between the two regimens. 8 However, investigators in many of these trials utilized small patient samples, and despite high attrition rates they did not perform intention-to-treat analysis. A review in 1996 also showed no significant difference between regimens. 9 However, this review did not include all trials comparing doxycycline with azithromycin, did not account for study size or quality, and did not exclude patients with nongonococcal urethritis due to pathogens other than CT. Furthermore, reports of new trials comparing the two regimens have been published since 1996. Per The British Medical Journal, “more comparisons of azithromycin and doxycycline are needed to rule out a clinically important difference between them.”8 In addition to antibiotic efficacy, safety must be considered in determining appropriate therapy. The reported frequency of adverse events is similar for both medications. 10,11
In the current study, we used formal statistical methods (meta-analysis) to determine whether azithromycin or doxycycline is more efficacious and safe in the treatment of CT urethritis and cervicitis.
Published English-language studies were identified via an exhaustive computer search with use of the following databases: MEDLINE and Pre-MEDLINE (1975 to August 2001), HealthSTAR (1975 to August 2001), Journals at Ovid Full Text (August 2001), Evidence Based Medicine (EBM) Reviews: Best Evidence (September 1991 to January/February 2001), EBM Reviews: Cochrane Database of Systematic Reviews (second quarter of 2001) and EBM Reviews: Database of Abstracts and Reviews of Effectiveness (second quarter of 2001). Medical subject headings CT and doxycycline or CT and azithromycin were used. Bibliographies of computer-identified articles were manually searched for additional trials for inclusion.
Both authors independently evaluated all trials comparing doxycycline with azithromycin for treatment of CT cervicitis in females or urethritis in males. Criteria for inclusion in the analysis were (1) randomized trial design; (2) medication regimens of oral doxycycline (100 mg twice daily for 7 days) and oral azithromycin (1 g once); (3) males aged >15 years and nonpregnant females aged >15 years; and (4) evaluation of microbial cure at follow-up. Microbial cure was defined as CT negativity in biological assay (culture or enzyme immunoassay). One study involved the use of DNA amplification tests (polymerase chain reaction) for evaluation of cure. Studies were excluded if they did not fulfill all of the aforementioned criteria.
For follow-up, the last available time point was selected. The most commonly reported follow-up times were 2 weeks and 4 weeks. Seven study reports presented data from follow-up at 3 to 5 weeks; 4 studies, from follow-up at 2 weeks; and 2 studies, cumulative data from undifferentiated follow-up times (recorded at the 4-week time-point).
Data were extracted from selected studies. Microbial cure at the last available follow-up and number of patients who had adverse effects were tabulated. Adverse events were separated by type (gastrointestinal, neurological, dermatological, and other). When occurrence of adverse events among patients with chlamydial infection was not distinguished from that of patients with nongonococcal urethritis, results for patients with nongonococcal urethritis were used as proxies for those for patients with CT infection. Attrition rates based on last available follow-up were calculated with the intention-to-treat analytical approach.
For each trial, the efficacy difference (ED; the difference in cure rates between azithromycin and doxycycline) of treatment success for azithromycin versus doxycycline was computed. The numerator, treatment success, was defined as the number of subjects assigned to a particular antibiotic group who complied with the treatment regimen and were microbiologically cured at follow-up. The denominator for computing treatment success comprised all subjects assigned to an antibiotic group who began the medication. To pool dichotomous outcomes (microbial cure versus no microbial cure), an overall weighted average of the ED from each trial was calculated by assigning each a weight derived from the standard error of the ED. 12 Thus, each trial's contribution to the pooled estimate of ED is directly proportional to the amount of information provided. A similar approach was used to calculate the pooled risk difference (RD) of adverse events for the two antibiotics. Subjects lost to follow-up were excluded from the analysis.
Each pooled RD and ED was tested for statistical significance, and 95% CIs were computed. In computing the pooled estimates, it was assumed that ED and RD were uniform across trials and that differences in results of individual trials could be attributed to chance. Hence, a fixed-effects model was used. 13 A chi-square test of homogeneity was used to test the assumption of uniform ED and RD. 14
A stratified meta-analysis was performed to assess possible bias in the results. Data were stratified by various subgroups: type of diagnostic assay (culture/nonculture), sex (male/female), attrition rate (≤10%/>10%), follow-up time (2 weeks/>2 weeks), publication date (pre-1995/post-1995), study design (open/blind) and study sponsorship (Pfizer/non-Pfizer [Pfizer Corporation holds the patent for azithromycin]).
Begg's and Egger's tests were used to check for publication bias (selective publication of trials independent of scientific merit) in pooled estimates. 15 To increase statistical power, a P value of 0.10 was used as the criterion for determining statistical significance of the tests for publication bias.
Thirteen trials fulfilling the inclusion criteria were identified. One was excluded because the data for the doxycycline treatment group were not segregated from those for several other antibiotic treatment groups. 16 Of the 12 included studies, 5 were masked and 7 were open-label. Adverse event data were provided from nine trials. Publication dates ranged from 1990 to 1999, and eight studies were conducted before 1995. Two studies included female patients only; six studies, male patients only; and another four, both. For evaluation of microbial cure, nine studies used culture, two used enzyme immunoabsorbent assay, and one used DNA amplification tests.
A total of 1543 patients (726 males and 817 females) were evaluated for microbial cure. In the azithromycin group, microbial cure occurred in 853 of 884 patients (96.5%). In the doxycycline group, microbial cure occurred in 645 of 659 (97.9%). Adverse events were reported by 319 (25.0%) of 1274 azithromycin-treated patients and 205 (22.9%) of 897 doxycycline-treated patients. Table 1 contains specific information on study design, follow-up, sample size, sex of patients, microbial cure rate, and attrition. 10,11,17–26
Figure 1 shows the pooled ED for microbial cure of azithromycin versus doxycycline. The pooled ED is 0.008 (95% CI, −0.007–0.022). This result is not statistically significant (Z = 1.05;P = 0.296). The test for homogeneity shows that results for individual trials are consistent with the overall pooled ED (chi-square = 10.48;df = 11;P = 0.488).
The RD for an adverse event with administration of azithromycin versus doxycycline is shown in Figure 2. The pooled RD is 0.009 (95% CI, −.019–0.037). This result is not statistically significant (Z = 0.62;P = 0.533). The test for homogeneity shows that the results are consistent across trials (chi-square = 6.63;df = 8;P = 0.577).
Table 2 shows that the most frequently reported adverse events were gastrointestinal in nature (87.3%). Gastrointestinal adverse events include diarrhea, abdominal pain, nausea, vomiting, dyspepsia, constipation, flatulence, and other (unspecified) symptoms. Other nongastrointestinal symptoms were neurological (fatigue, malaise, sweating, dizziness, headache, and other [unspecified] symptoms); dermatological (skin rash and drug eruption); and miscellaneous (genitourinary, fever, and unspecified). Reporting of adverse events was nonuniform across studies, precluding further analysis by type of event.
Neither Begg's nor Egger's test showed evidence of publication bias. For Egger's test, the bias coefficient was 0.53 (95% CI, −0.70–1.66), which is not statistically significant (t = 1.03;P = 0.327). The Begg's funnel plot did not reveal any trials outside of the pseudo-95% CI (Z = 0.55;P = 0.583).
We stratified the data by various subgroups to assess possible bias in the results. Results showed no evidence of bias by type of diagnostic assay (culture/nonculture), follow-up time (2 weeks/>2 weeks), attrition rate (≤10%/>10%), sex (male/female), publication date (pre-1995/post-1995), study design (open/blind), or study sponsorship (Pfizer/non-Pfizer).
In this systematic review, we found that doxycycline and azithromycin are equally efficacious and safe for treatment of genital CT infection. We found sufficient data to generate statistically reliable conclusions about antibiotic choice. All meta-analyses are subject to potential bias because of systematic and random errors. 27 However, our results show no evidence of systematic bias by assay type, follow-up time, attrition rate, sex, site of infection, publication date, study design, or sponsorship. In addition, there is no statistical evidence of publication bias. Estimates of pooled effects are homogenous, indicating that the magnitude of random error is not greater than expected.
We recognize that the estimates of efficacy presented herein were obtained from studies conducted under optimal conditions. However, doxycycline is administered in a multidose regimen, and its efficacy may be compromised in practice by partial compliance. Given the long life cycle of CT, our choice of last available follow-up may not be optimal. The validity of chlamydial culture testing at <3 weeks after completion of therapy has not been established. At shorter follow-up times (≤2 weeks), false-negative results can occur because of small numbers of chlamydial organisms. 7 Our results do not show evidence of bias due to length of follow-up.
Microbial cure is probably a more appropriate measure of effectiveness than clinical cure because the presence of multiple infections in some patients confounds evaluation of clinical response to antibiotic therapy. Furthermore, 50% to 80% of genital CT infections are asymptomatic. 4 Thus, clinical cure rates were not considered in this analysis.
Doxycycline and azithromycin are not associated with statistically different rates of adverse events. Approximately 23% to 25% of patients report adverse events with each antibiotic. However, most complaints are minor and gastrointestinal in nature. Serious complications with either antibiotic are extremely rare. 28 We did not find a report of a randomized trial that directly compared the intensity and duration of adverse events associated with azithromycin and doxycycline. For analysis, we assumed that adverse events in males treated for chlamydial urethritis are comparable to those in males treated with the identical regimen for nongonococcal urethritis of any etiology. Given that urethritis is a local infection and pharmacokinetics are unlikely to vary with different infectious etiologies of urethritis, this is a reasonable assumption.
Compliance of patients is a critical consideration in treatment decisions. Administration of doxycycline involves a twice-daily 7-day regimen, whereas azithromycin is given as a one-time dose. Because patients are more likely to take azithromycin as prescribed and there is no difference in cure rates, one might conclude that azithromycin is the treatment of choice. Trials with less frequent doxycycline regimens have shown high efficacy rates. Several have used 200 mg on the first day with 6–9 subsequent doses of 100 mg daily. In these trials, cure rates are 97%–100% at the 2-week follow-up, with an overall cure rate of 98%. 29–35 Moreover, the 18- to 22-hour half-life of doxycycline suggests that adequate serum concentrations will be maintained with a once-daily protocol. 28 Although observational studies suggest that partial compliance with doxycycline is sufficient to effect cure, dose–response trials are needed to evaluate these findings. 36–38
The marked cost differential between single-dose and multidose therapy makes the medication choice a central issue in the current environment of limited resources. Hence, there is controversy about the regimen of choice for uncomplicated genital CT infection. 39 The per-gram wholesale cost of azithromycin powder is approximately $10 (U.S.); the cost of generic doxycycline is <$1.50 for a 7-day course. Two major cost-effectiveness studies comparing the antibiotics concluded that azithromycin is more cost-effective. However, one assumed a doxycycline efficacy of <93% and an azithromycin efficacy of 96%; the other did not take antibiotic price into account and considered partial compliance with doxycycline equivalent to treatment failure. 40,41 Since these assumptions are unlikely to be accurate, a comprehensive cost-effectiveness analysis should be performed.
This meta-analysis showed that doxycycline and azithromycin are equally efficacious in treatment of genital CT infection. No difference in adverse event rates was found. A comprehensive cost-effectiveness evaluation may affect antibiotic choice for genital CT infection. Dose–response trials may also be needed to evaluate less complicated doxycycline regimens. However, our results strongly suggest that there is no need for new head-to-head trials to establish the efficacy and safety of azithromycin and doxycycline in the treatment of genital CT infection.
1. US Preventative Services Task Force. Guide to Clinical Preventative Services, 2nd ed. Rockville, Maryland: Williams & Wilkins, 1996.
2. Braverman PK. Incident Chlamydia trachomatis infections among inner-city adolescent females. Clin Pediatr 2000; 39: 378.
3. Gerbase AC, Rowley JT, Heymann DHL, Berkley SFB, Piot P. Global prevalence and incidence estimates of selected curable STD's. Sex Transm Infect 1998; 74 (suppl 1): 2–6.
4. Clinical Effectiveness Group. National guideline for management of Chlamydia trachomatis genital tract infection. Sex Transm Infect 1999; 75 (suppl 1): 4–8.
5. Antilla T, Saikuu P, Koskela P, et al. Serotypes of Chlamydia trachomatis and risk for development of cervical squamous cell carcinoma. JAMA 2001; 285: 47–51.
6. Laga M, Manoka A, Kivuvu M, et al. Non-ulcerative sexually transmitted diseases as risk factors for HIV-1 transmission in women: results from a cohort study. AIDS 1993; 7: 95–102.
7. Centers for Disease Control and Prevention. Guidelines for treatment of sexually transmitted diseases. MMWR Morb Mortal Wkly Rep 1998; 47 (RR-1):1–118.
8. Low N, Cohen F. Genital chlamydia infection. In: Barton S, ed. Clinical Evidence, Issue 5. London: BMJ Publishing Group, 2001: 1083–1088.
9. Chlamydial STD treatment [comment]. Bandolier 1996; 28: 4–6.
10. Martin DH, Mroczkowski TF, Dalu ZA, et al. A controlled trial of single dose azithromycin for the treatment of chlamydial urethritis and cervicitis. N Engl J Med 1992; 327: 921–925.
11. Thorpe EM Jr, Stamm WE, Hook EW III, et al. Chlamydial cervicitis and urethritis: single dose treatment compared with doxycycline for seven days in community based practices. Genitourin Med 1996; 72: 93–97.
12. Hedges L, Olkin I. Statistical Methods for Meta-Analysis. San Diego, CA: Academic Press, 1985.
13. Greenland S. Quantitative methods in the review of epidemiologic literature. Epidemiol Rev 1987; 9: 1–29.
14. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1985; 6: 306–317.
15. Egger M. Tests for publication bias in meta-analysis. Stata Tech Bull 1998; 44: 3–4.
16. Carlin EM, Barton SE. Azithromycin as the first-line treatment of nongonococcal urethritis (NGU): a study of follow-up rates, contact attendance and patients’ treatment preference. Int J STD AIDS 1996; 7: 185–189.
17. Steingrimsson O, Olafsson JH, Thorarinsson H, Ryan RW, Johnson RB, Tilton RC. Azithromycin in the treatment of sexually transmitted disease. J Antimicrob Chemother 1990; 25 (suppl A): 109–114.
18. Nilsen A, Halsos A, Johansen A, et al. A double blind study of single dose azithromycin and doxycycline in the treatment of chlamydial urethritis in males. Genitourin Med 1992; 68: 325–327.
19. Ossewaarde JM, Plantema FHF, Rieffe M, Nawrocki RP, de Vries A, van Loon AM. Efficacy of single-dose azithromycin versus doxycycline in the treatment of cervical infections caused by Chlamydia trachomatis. Eur J Clin Microbiol Infect Dis 1992; 11: 693–697.
20. Hammerschlag MR, Golden NH, Oh MK, et al. Single dose of azithromycin for the treatment of genital chlamydial infections in adolescents. J Pediatr 1993; 122: 961–965.
21. Lauharanta J, Saarinen K, Mustonen MT, Happonen HP. Single-dose oral azithromycin versus seven day doxycycline in the treatment of nongonococcal urethritis in males. J Antimicrob Chemother 1993; 31 (Suppl E): 177–183.
22. Lister PJ, Balechandran T, Ridgway GL, Robinson AJ. Comparison of azithromycin and doxycycline in the treatment of nongonococcal urethritis in men. J Antimicrob Chemother 1993; 31 (suppl E): 185–192.
23. Steingrimsson O, Olafsson JH, Thorarinsson H, Ryan RW, Johnson RB, Tilton RC. Single dose azithromycin treatment of gonorrhea and infections caused by C. trachomatis and U. urealyticum in men. Sex Transm Dis 1994; 21: 43–46.
24. Stamm WE, Hicks CB, Martin DH, et al. Azithromycin for empirical treatment of the nongonococcal urethritis syndrome in men. JAMA 1995; 274: 545–549.
25. Hillis SD, Coles FB, Litchfield B, Black CM, Mojica B, Schmitt D. Doxycycline and azithromycin for prevention of chlamydial persistence of recurrence one month after treatment in women: a use-effectiveness study in public health settings. Sex Transm Dis 1998; 25: 5–11.
26. Tan HH, Chan RKW. An open label comparative study of Azithromycin and doxycycline in the treatment of nongonococcal urethritis in males and Chlamydia trachomatis cervicitis in female sex workers in an STD clinic in Singapore. Singapore Med J 1999; 40: 519–523.
27. Cook DJ, Mulrow MD, Haynes RB. Systematic reviews: synthesis of best evidence for clinical decisions. Ann Intern Med 1997; 126: 376–380.
28. Sifton DW, Murray L, Westley GJ, Mazur JM, Woerner RN, eds. Physicians’ Desk Reference, 54th ed. New Jersey: Medical Economics Company, 2000.
29. Csango PA, Gunderson T, Anestad G. Doxycycline in the treatment of chlamydial urethritis: a therapeutic study. Pharmatherapeutica 1980; 2: 341–345.
30. Kovacs GT, Wescott M, Rusden J, et al. A prospective single blind trial of minocycline and doxycycline in the treatment of genital Chlamydia trachomatis infection in women. Med J Aust 1989; 150: 483–485.
31. Sanders LL, Harrison HR, Washington AE. Treatment of sexually transmitted chlamydial infections. JAMA 1986; 255: 1750–1756.
32. Lassus AB, Virrankoski T, Reitamo SJ, et al. Pivampicillin versus doxycycline in the treatment of chlamydial urethritis in men. Sex Transm Dis 1990; 17: 20–22.
33. Jeskanen L, Karppinen L, Ingervo L, Reitamo S, Happonen HP, Lassus A. Ciprofloxacin versus doxycycline in the treatment of uncomplicated urogenital Chlamydia trachomatis infections: a double-blind comparative study. Scand J Infect Dis 1989; (suppl 60):62–65.
34. Johannisson G, Sernryd A, Lycke E. Susceptibility of Chlamydia trachomatis to antibiotics in vitro and in vivo. Sex Transm Dis 1979; 6: 50–57.
35. Csango PA, Gunderson T, Anestad G. A double-blind comparison of erythromycin and doxycycline therapy for nongonococcal urethritis. In: Mardh P-A, Holmes KK, Oriel JD, Piot P, Schachter J, eds. Chlamydia Infections. New York: Elsevier Biomedical Press, 1982: 257–261.
36. Augenbraun M, Bachmann L, Wallace T, Dubouchet L, McCormack W, Hook E. Compliance with doxycycline therapy in sexually transmitted diseases clinics. Sex Transm Dis 1998; 25: 1–4.
37. Bachmann LH, Stephens J, Richey CM, Hook EW. Measured versus self-reported compliance with doxycycline therapy for chlamydia associated syndromes: high therapeutic success despite poor compliance. Sex Transm Dis 1999; 26: 272–278.
38. Schachter J. What is the minimally effective treatment for Chlamydia trachomatis infection? The compliance paradoxycycline. Sex Transm Dis 1999; 26: 279–280.
39. Handsfield HH, Stamm WE. Treating chlamydial infection: compliance versus cost. Sex Transm Dis 1998; 25: 12–13.
40. Magid D, Douglas JM, Schwartz JS. Doxycycline compared with azithromycin for treating women with genital chlamydia trachomatis infections: an incremental cost-effectiveness analysis. Ann Intern Med 1996; 124: 389–399.
41. Genc M, Mardh P-A. A cost-effectiveness analysis of screening, treatment for Chlamydia trachomatis infection in asymptomatic women. Ann Intern Med 1996; 124: 1–7.