Pelvic inflammatory disease (PID) affects more than 800,000 American women during their reproductive years,1,2 and the prevalence in Latin America and Caribbean is 36.8/100,000.3 A single intramuscular injection of 250 mg of ceftriaxone and 200 mg per day of doxycycline for 14 days has been shown to be effective for outpatient treatment of mild pelvic inflammatory disease,4 and has been recommended by the Centers for Disease Control and Prevention.5 This regimen is inconvenient, requiring twice per day dosing for 14 days. Azithromycin, an azalide antibiotic with a long (68 hours) half-life in tissues, has a spectrum of activity against microorganisms similar to that of doxycycline, but it can be used as single dose, improving compliance, and it has few side effects. Such properties would overcome the disadvantages associated with doxycycline. Largely for these reasons, azithromycin is considered the treatment of choice for genital tract infection with Chlamydia trachomatis.4 Recent investigations have evaluated the use of azithromycin in the treatment of PID; however, they were not designed for outpatient treatment or have some methodologic bias.6,7 Thus, currently available data are insufficient to recommend this agent as a component of any of the oral treatment regimens for PID.5 The purpose of this randomized, controlled trial was to test the hypothesis that a single intramuscular injection of 250 mg of ceftriaxone followed by either 200 mg of doxycycline per day for 14 days or 1g of azithromycin once per week for 2 weeks are equivalent for treating outpatient (mild) cases of PID.
PARTICIPANTS AND METHODS
All patients attending the Medical Emergency of Hospital de Clínicas de Porto Alegre (a university teaching hospital) with a complaint of pelvic pain were evaluated for the presence of PID and referred to one of the researchers (R.F.S., L.M.T., or T.G.T.). All patients with a history of pelvic discomfort for less than 30 days, with findings of pelvic organ tenderness (uterine or adnexal) on bimanual examination, and leukorrhea or mucopurulent cervicitis were invited to participate in the doxycycline–azithromycin protocol. Leukorrhea was defined as an excess number of white blood cells (10 white blood cells or more per high power field) as viewed microscopically in the cul-de-sac secretions mixed with normal saline solution.8 Mucopurulent cervicitis was defined as the presence of grossly yellow or green exudates observed on the cervix. Pelvic or abdominal ultrasound and urine analysis were performed on the first visit to rule out other causes, at discretion of the attendant physician. Patients with current urinary infection, pregnancy demonstrated by β-hCG or ultrasonography, with the presence of tuboovarian abscess, endometriosis, appendicitis, diverticulitis, hemorrhagic ovarian cysts or torsion, or abdominal hernia, confirmed by ultrasound or laparoscopy were excluded. Other exclusion criteria included homelessness, fever above 38°C, abdominal rebound tenderness, pelvic pain more than 30 days duration, allergy to ceftriaxone, azithromycin, or doxycycline, history of antimicrobial therapy within 7 days of recruitment, delivery, abortion or gynecologic surgery within 30 days, prior hysterectomy or bilateral salpingectomy, and oral intolerance for the antibiotics. Oral intolerance was checked in patients who experienced one episode of vomiting, after taking the first dose of medication. The patient was observed for 1 hour, after an intravenous injection of 10 mg of metoclopramide. Intolerance for an outpatient oral regimen was defined by the vomiting of 230 mL of water within 1 hour.9 Those exclusion criteria were used because of the nature of PID. The diagnosis for PID is clinical, and we should start the treatment before laboratory examinations are ready. Other diagnosis related to pelvic pain, with the exception of infection of the upper genital tract, exclude the diagnosis of PID.10 Furthermore, fever, homelessness, and abdominal rebound tenderness preclude outpatient treatment.10 Human immunodeficiency virus (HIV)–positive patients were not excluded from the sample, because the available evidence does not show a difference in the clinical course of PID without tuboovarian abscess when compared with seronegative patients.11,12
Eligible patients were invited to participate in the doxycycline–azithromycin protocol. Details of the management and follow-up for each group are shown in Table 1. One of three researchers (R.F.S., L.M.T., or T.G.T.) performed the physical examination. Whenever possible, the same researcher examined the same patient on the subsequent follow-up visits. A visual analog scale (VAS) of pain13 and a modified McCormack pain scale14 were used to assess the degree of pain at all visits. The VAS is a 100-mm straight line with anchors placed at both poles (0 cm indicated no symptom and 10 cm indicated the worst possible pain). Patients were asked to place a mark somewhere along the line that best described the actual status of their symptoms before pelvic examination. The modified McCormack pain scale assesses direct and rebound tenderness range from 0=tenderness absent, 1=tenderness referred by the patient, 2=tenderness causing observable distress, and 3=rebound tenderness. Total score was defined as the sum of individual scores for 12 abdominal and pelvic regions (maximum score=36). Written instructions were read and given to the patient about the treatment. For the duration of the study, patients were asked to avoid sexual intercourse, or have their partner use a condom if they engaged in sexual intercourse. Use of additional antibiotics or analgesics were not permitted during the study. In case of worsening of the pain, the patient was instructed to return for inpatient intravenous antibiotics and reevaluation.
The primary outcome for the trial was clinical cure, defined as 70% or greater reduction in the total tenderness score at day 14 compared with baseline, for both VAS and McCormack pain scales as described in the PEACH protocol.15 For secondary outcomes, clinical cure was defined by the absence of or minimal pelvic tenderness, a temperature of 37.7°C or less (axillary temperature more than 37.7°C was considered a treatment failure) and a white blood cell count of less than 10,000/mm, if a minimum of 4 days of treatment had been completed. Clinical improvement was defined as resolution of two of the three symptoms mentioned above. Therapy was considered a failure when one of the following was noted after more than 48 hours on protocol: signs and symptoms remained unchanged or worsened; some signs and symptoms initially abated, but later worsened or failed to dissipate further; additional antibiotic therapy was required to treat the PID; pelvic surgery was required because of deteriorating status; or protocol therapy was required for more than 14 days as defined in the original article of Arredondo et al.4 Those criteria are in accordance to the original article that established doxycycline as a therapy for pelvic inflammatory disease. Histologic and microbiologic cure were defined as absence of plasma cell per ×120 field in the endometrial stroma and fewer than 5 neutrophils per ×400 field in the endometrial surface (if present at baseline), and absence of Neisseria gonorrhoeae and C trachomatis in endometrial samples on day 30 by nucleic acid amplification test (if present at baseline).
The study protocol was approved by the ethics committee of Hospital de Clínicas de Porto Alegre, and registered at http://isrctn.org under #ISRCTN46117662. To compare equivalence between the two treatments we calculated the sample size according to Blackwelder,16 considering an alpha error of 0.05, a β error of 0.1, and difference between the two groups of no more than 10%. Based on the literature, we postulated a 97% clinical cure for azithromycin,6 and 95% for doxycycline.4 If the clinical cure for azithromycin was 97% we would need at least 56 patients in each group. If the cure rate was 96% or less, we would need a minimum of 75 patients in each group. Interim analysis was calculated according to Fang.17 We considered two stages of analysis, an adjusted critical Z′ value (derived from adjusted α for 0.05, and β of 0.1), a rate of cure for one drug as 98% and for the other as 79% for possible early stopping, if clinical cure with azithromycin was more than 97%, or if clinical cure was less than 80% (chosen arbitrarily). This figures elicited 33 patients per group, in each stage, yielding a total of 132 patients.
A table generated by computer was used for allocation sequence. Subjects were allocated in blocks of four.18 If the patient was eligible for the study, she was blindly allocated to one of the two treatments by an investigator not related to the cases. The allocation was concealed, coded, and kept away from the other investigators. The other investigators were not aware of the sequence of the treatment assignment. Patients and those who assessed the outcomes were blind to group assignment. To avoid bias, both medications were manipulated by the hospital pharmacy and put in identically coded blisters and capsules. Because of the difference in the number of capsules in each treatment, the empty azithromycin blisters were filled with placebo. To confirm compliance, only 7 days of treatment were given to the patient, forcing her to return in the 8th day to receive the rest of the treatment. All patients with an intrauterine device had to have the device removed before initiation of treatment.
At the first visit, an endometrial biopsy was performed at the office using a Karman plastic curette (MedGyn Products Inc., Lombard, IL) and repeated after 30 days. Histopathologic endometritis was defined by the presence of 1 plasma cell or more per ×120 field in the endometrial stroma plus five neutrophils or more per ×400 field in the endometrial surface.19 Specimens were frozen at –60°C until nucleic acid amplification testing could be performed. Commercially available APTIMA Combo 2 Assay (AC-2) (Gen-Probe Incorporated, LaJolla, CA) was used to detect the presence of N gonorrhoeae and C trachomatis in the endometrial samples as described by others.20 The APTIMA test format was also used to detect Mycoplasma genitalium and Trichomonas vaginalis using research reagents developed by the manufacturer.
Student t test, Mann–Whitney test, and Fisher exact test were used for statistical analysis. The secondary outcomes were used for rates of cure in modified intention-to-treat21 and per protocol analysis with 95% confidence intervals. In the modified intention-to-treat analysis, the patients who have an exclusion criterion found after the initial allocation treatment are excluded. To remain unbiased, decisions to exclude participants were made 1) by researchers blinded to treatment allocation and 2) on the basis of information not related to either the allocated treatment or to events or outcomes that occur after random allocation.22 In per protocol analysis, only patients who sufficiently complied with the trials were considered in the analysis. Compliance covers exposure to treatment, availability of measurements, and absence of protocol violations.21
From April 2003 to March 2004, 163 patients with pelvic pain were enrolled in the protocol after signing the informed consent. From these, 133 were eligible for the study and were randomly assigned to one of the treatment regimens. Details of the randomization can be found on Figure 1. Seven patients lost on follow-up in the doxycycline group and four in the azithromycin group were analyzed as failure. Nine and four cases from the doxycycline and azithromycin group were excluded from the final analysis after they had initiated the treatment because they did not have PID. Total compliance was achieved in 106 patients (49 and 57 in the doxycycline and azithromycin groups, respectively). The average age of participants was 29.3±1.1 years in the doxycycline group and 28.3±0.8 years in the azithromycin group. No significant difference was found between the groups, when age, and pain score on day 0 and 14, were compared (Table 2). There were no cases of gonorrhea detected by AC-2 in endometrial biopsies. Only six cases of Chlamydia were found in the first endometrial biopsies, and all had 100% of cure on day 30. There was one case positive for Chlamydia infection on the second endometrial biopsy, despite its absence on the first biopsy. The patient confirmed the use of a condom in all sexual intercourse and also had a clinical cure using the 70% reduction or more of the pain criterion in the VAS and McCormack pain scale. There were seven patients positive for M genitalium and two patients with T vaginalis. Failure of treatment was seen in one case each with M genitalium and T vaginalis, both in the azithromycin group. The histologic findings in both groups are seen in Table 3. When we considered only the cases with confirmed endometritis in the first biopsy, the cure rate for the azithromycin group was 66.6% (95% confidence interval [CI] 0.5–0.8) and for the doxycycline group was 53.1% (95% CI 0.34–0.7). Four cases in each group presented endometritis in the second biopsy, despite having no endometritis in the first biopsy. In the doxycycline group, two cases had clinical cure, ie, “0” on the VAS and McCormack pain scales on days 14 and 30 of treatment. In the azithromycin group, these four patients were considered cured by the VAS on day 30 of treatment, with the 70% or more reduction of initial pain. No serious adverse reaction was observed in either group.
Our initial hypothesis was that azithromycin and doxycycline were equivalent drugs for treating mild pelvic inflammatory disease. To conduct such a trial, we followed the CONSORT recommendations.23 We mirrored the original work of Arredondo et al4 as closely as possible. Patient losses and other deviations from the protocol were less than 10% from the total. In all clinical and laboratory terms, azithromycin is equivalent or superior to doxycycline. The rates of cure are within the 95% CI. When clinical cure was defined as the original work from Arredondo et al,4 ie, by the absence or reduction of pelvic tenderness, considering the initial pain, the rate of cure per protocol was 98.2% (95% CI 0.90–0.99) in the azithromycin group, and 85.7% (95% CI 0.72–0.93) in the doxycycline group. These rates of cure for azithromycin are comparable to those published in the literature, varying around 94%.6 Nevertheless, when the 70% or more reduction was used, these rates of cure dropped below 80%, which was a criterion to stop the study. Those patients received a course of metronidazole after the second endometrial biopsy. Such low levels of cure could be explained by the peculiarities of the pain scale. If the initial pain score was 10, and it ended up with 2, clinical cured was established. Nevertheless, this same final score of pain would be considered as clinical failure, if the initial pain score was 5. However, in all clinical and laboratory comparisons, azithromycin was equivalent or superior to doxycycline. The results of clinical cure, by modified intention to treat, were 90.3% (56 of 62) for azithromycin, and 72.4% (42 of 58) for doxycycline. The former was within the acceptable range of preestablished rate of cure, whereas the latter was not. We used the modified intention-to-treat analysis because we routinely commence the antibiotic treatment before laboratory confirmation of the diagnosis. Of note, in the modified intention-to-treat analysis, we considered the cases lost on follow-up as failures and not as cure. Such cases are depicted in Figure 1. The rates of cure are even lower when histologic findings are considered, but the equivalence remains, 53.1% (17 of 32) for doxycycline, compared with 66.6% (28 of 42) for azithromycin. However, the histologic criteria must be seen with caution. Traditionally, the criteria proposed by Kiviat et al19 have been used as the criterion standard for diagnosing endometritis. But there are a number of different criteria used in the literature to diagnose upper genital tract inflammation, ranging from two plasma cells per low power field15 to even one plasma cell in the entire specimen.11 However, new data show that endometritis is present in asymptomatic patients.24 Furthermore, almost one half of the patients from the PEACH trial had endometritis after the end of treatment, whereas the clinical cure was around 80%.15 These findings are in agreement with ours. In the beginning of the study, 43 patients had no endometritis, and, after the treatment, five developed it (Table 4). The explanations for this finding remain unclear, but raise a concern whether these criteria should be used as diagnosing PID, or reflect ongoing inflammation after treatment. The later seems to be unlikely. Haggerty et al25 found that endometritis was not associated with reproductive morbidity. From the point of view of microbiology, the prevalence of C trachomatis and N gonorrhoeae is lower than that observed in North American studies. Nevertheless, these prevalences are in accord with other findings of South American population.4,25 Such findings support the hypothesis of the polymicrobial nature of PID infection coming from the cervix.26 Research in PID is difficult because of our difficulty diagnosing the condition accurately. Clinical symptoms and signs detect only 60–90% of cases subsequently confirmed at laparoscopy and also have poor specificity.10 Despite using highly sensitive methods for diagnosing common sexually transmitted diseases, we were not able to identify a significant pathogen in the majority. Anaerobes and Herpes simplex virus type-2 could be involved in the pathogenesis of PID as mentioned by others,27,28 but this awaits further confirmation. It is important to remember that the use of a shorter treatment course is related to a better compliance.29 This trial has shown that 1g of azithromycin repeated after 1 week is effective in an antibiotic regimen for mild PID, and we suggest that this treatment should be considered for replacing the doxycycline one.
1. Kelly JE, Mosher WD, Duffer AP, Kinsey SH. Plan and operation of the 1995 National Survey of Family Growth. Vital Health Stat 1 1997;(36):1–89.
2. McNeeley SG Jr. Pelvic inflammatory disease. Curr Opin Obstet Gynecol 1992;4:682–6.
3. Murray CJ, Lopez AD. Global health statistics: a compendium of incidence, prevalence, and mortality estimates for over 200 conditions. Cambridge (MA): Harvard School of Public Health; 1996.
4. Arredondo JL, Diaz V, Gaitan H, Maradiegue E, Oyarzun E, Paz R, et al. Oral clindamycin and ciprofloxacin versus intramuscular ceftriaxone and oral doxycycline in the treatment of mild-to-moderate pelvic inflammatory disease in outpatients. Clin Infect Dis 1997;24:170–8.
5. Sexually transmitted diseases treatment guidelines 2002. Centers for Disease Control and Prevention. MMWR Recomm Rep 2002;51:1–78.
6. Bevan CD, Ridgway GL, Rothermel CD. Efficacy and safety of azithromycin as monotherapy or combined with metronidazole compared with two standard multidrug regimens for the treatment of acute pelvic inflammatory disease. J Int Med Res 2003;31:45–54.
7. Malhotra M, Sharma JB, Batra S, Arora R, Sharma S. Ciprofloxacin-tinidazole combination, fluconazole- azithromicin-secnidazole-kit and doxycycline- metronidazole combination therapy in syndromic management of pelvic inflammatory disease: a prospective randomized controlled trial. Indian J Med Sci 2003;57:549–55.
8. Davies HD, Wang EE. Periodic health examination, 1996 update: 2. Screening for chlamydial infections. Canadian Task Force on the Periodic Health Examination. CMAJ 1996;154:1631–44.
9. Ness RB, Soper DE, Peipert J, Sondheimer SJ, Holley RL, Sweet RL, et al. Design of the PID Evaluation and Clinical Health (PEACH) Study. Control Clin Trials 1998;19:499–514.
10. Workowski KA, Berman SM. CDC sexually transmitted diseases treatment guidelines. Clin Infect Dis 2002;35:S135–7.
11. Korn AP, Landers DV, Green JR, Sweet RL. Pelvic inflammatory disease in human immunodeficiency virus–infected women. Obstet Gynecol 1993;82:765–8.
12. Mugo NR, Kiehlbauch JA, Nguti R, Meier A, Gichuhi JW, Stamm WE, et al. Effect of human immunodeficiency virus-1 infection on treatment outcome of acute salpingitis. Obstet Gynecol 2006;107:807–12.
13. Lai IR, Wu MS, Lin JT. Prospective, randomized, and active controlled study of the efficacy of alginic acid and antacid in the treatment of patients with endoscopy-negative reflux disease. World J Gastroenterol 2006;12:747–54.
14. McCormack WM, Nowroozi K, Alpert S, Sackel SG, Lee YH, Lowe EW, et al. Acute pelvic inflammatory disease: characteristics of patients with gonococcal and nongonococcal infection and evaluation of their response to treatment with aqueous procaine penicillin G and spectinomycin hydrochloride. Sex Transm Dis 1977;4:125–31.
15. Ness RB, Soper DE, Holley RL, Peipert J, Randall H, Sweet RL, et al. Effectiveness of inpatient and outpatient treatment strategies for women with pelvic inflammatory disease: results from the Pelvic Inflammatory Disease Evaluation and Clinical Health (PEACH) Randomized Trial. Am J Obstet Gynecol 2002;186:929–37.
16. Blackwelder WC. “Proving the null hypothesis” in clinical trials. Control Clin Trials 1982;3:345–53.
17. Fang J. Design and analysis of sequential experiments. In: Fang J, editor. Medical statistics and computer experiments. 3rd ed. Singapore: World Scientific Publishing Co. Pte. Ltd.; 2005. p. 563–78.
18. Altman DG, Bland JM. How to randomise. BMJ 1999;319:703–4.
19. Kiviat NB, Wolner-Hanssen P, Eschenbach DA, Wasserheit JN, Paavonen JA, Bell TA, et al. Endometrial histopathology in patients with culture-proved upper genital tract infection and laparoscopically diagnosed acute salpingitis. Am J Surg Pathol 1990;14:167–75.
20. Gaydos CA, Quinn TC, Willis D, Weissfeld A, Hook EW, Martin DH, et al. Performance of the APTIMA Combo 2 assay for detection of Chlamydia trachomatis and Neisseria gonorrhoeae in female urine and endocervical swab specimens. J Clin Microbiol 2003;41:304–9.
21. Heritier SR, Gebski VJ, Keech AC. Inclusion of patients in clinical trial analysis: the intention-to-treat principle. Med J Aust 2003;179:438–40.
22. Montori VM, Guyatt GH. Intention-to-treat principle. CMAJ 2001;165:1339–41.
23. Grimes DA. The “CONSORT” Guidelines for randomized controlled trials in Obstetrics & Gynecology [editorial]. Obstet Gynecol 2002;100:631–2.
24. Achilles SL, Amortegui AJ, Wiesenfeld HC. Endometrial plasma cells: do they indicate subclinical pelvic inflammatory disease? Sex Transm Dis 2005;32:185–8.
25. Haggerty CL, Ness RB, Amortegui A, Hendrix SL, Hillier SL, Holley RL, et al. Endometritis does not predict reproductive morbidity after pelvic inflammatory disease. Am J Obstet Gynecol 2003;188:141–8.
26. Audu BM, Kudi AA. Microbial isolates and antibiogram from endocervical swabs of patients with pelvic inflammatory disease. J Obstet Gynaecol 2004;24:161–4.
27. Ness RB, Kip KE, Hillier SL, Soper DE, Stamm CA, Sweet RL, et al. A cluster analysis of bacterial vaginosis–associated microflora and pelvic inflammatory disease. Am J Epidemiol 2005;162:585–90.
28. Cherpes TL, Wiesenfeld HC, Melan MA, Kant JA, Cosentino LA, Meyn LA, et al The associations between pelvic inflammatory disease, Trichomonas vaginalis infection, and positive herpes simplex virus type 2 serology. Sex Transm Dis 2006;33:747–52.
© 2007 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
29. Dunbar-Jacob J, Sereika SM, Foley SM, Bass DC, Ness RB. Adherence to oral therapies in pelvic inflammatory disease. J Womens Health (Larchmt) 2004;13:285–91.