Post–cesarean delivery endometritis is the most common complication of cesarean delivery and is generally believed to have a polymicrobial etiology resulting from ascension of bacteria from the lower to the upper genital tract. 1 Numerous different microorganisms including gram-positive and gram-negative aerobic and anaerobic bacteria have been recovered from the upper genital tract of women with postpartum endometritis. 1 Earlier reports implicated Ureaplasma urealyticum as having a potential role in the pathophysiology of endometritis. 2,3 This microorganism is indigenous to the vaginal flora and has been reported to be present in the lower genital tract of approximately 70–90% of pregnant women. 4 Perhaps because of its high prevalence in the lower genital tract during pregnancy, several studies have failed to demonstrate a conclusive link between U urealyticum in cervicovaginal secretions and postpartum fever 5,6 or adverse pregnancy outcomes. 4 However, U urealyticum has been isolated from maternal blood within minutes of delivery 7 and from blood cultures in women with postpartum infections. 8,9 Serum antibody responses to U urealyticum have also been reported in women with clinical chorioamnionitis, 10 and this organism has been implicated as a significant potential pathogen in post–cesarean delivery wound infections. 11 Ureaplasma urealyticum has been reported as the most common microorganism isolated from the upper genital tract of women having a spontaneous preterm birth, 12,13 and preterm birth is associated with an increased risk for postpartum endometritis. 14–16
Because of the increasing evidence indicating the potential pathogenicity of U urealyticum in upper genital tract infections, we conducted a study to determine if antenatal chorioamnion colonization with this microorganism was associated with post–cesarean delivery endometritis. 14 In that study, we evaluated 575 women without clinically overt evidence of chorioamnionitis and who had intact membranes at the time of cesarean delivery. 14 The chorioamnion was cultured at cesarean delivery for aerobic and anaerobic bacteria, mycoplasmas, Chlamydia trachomatis, and Trichomonas vaginalis. Ten percent of these women developed post–cesarean delivery clinical endometritis. Chorioamnion colonization with U urealyticum was associated with a three-fold increased risk of post–cesarean delivery endometritis and, in a subgroup analysis, an eight-fold increased risk in women in whom the onset of labor was spontaneous. 14
The association between U urealyticum colonization of the chorioamnion at cesarean delivery and an increased risk of post–cesarean delivery endometritis occurred despite the routine use of intraoperative prophylactic antibiotics. The substantial benefit of antibiotic prophylaxis to reduce post–cesarean delivery endometritis is well established. 17 Numerous antimicrobial agents have been reported to have similar efficacy in reducing post–cesarean delivery febrile morbidity by approximately 50%. However, the antimicrobials commonly used as prophylaxis do not have established efficacy against U urealyticum. Thus, the objective of the current randomized clinical trial was to determine if extended spectrum antimicrobial prophylaxis, including presumed efficacy against U urealyticum, further reduces post–cesarean delivery clinical endometritis compared with prophylaxis with antibiotics without efficacy against U urealyticum.
MATERIALS AND METHODS
Between October 1, 1996, and September 9, 1999, consenting women who were undergoing cesarean delivery were enrolled in this randomized clinical trial. Women were approached for consent upon admission to the labor and delivery suite, and randomization occurred in consenting women at the time of cesarean delivery. All women undergoing cesarean delivery at our institution were potentially eligible. Exclusion criteria included allergy to study medications (including penicillin), fever (100F or higher), clinical signs or symptoms of chorioamnionitis, impaired hepatic (total bilirubin 3.0 mg/dL or greater, or aspartate aminotransferase 93 μL or greater) or renal (serum creatinine greater than 1.5 mg/ dL) function, immunocompromised condition, on medications metabolized by the cytochrome P450 system (such as carbamazepine or phenytoin), language barrier, need for hysterectomy at cesarean, need for emergent cesarean precluding consent or availability of study medication, and prior participation in the trial. After clamping of the umbilical cord at cesarean delivery, all subjects immediately received intravenous prophylaxis with 2 g of cefotetan. Subjects were then randomized in a double-blind trial to receive additional prophylaxis with doxycyline 100 mg intravenously (initiated simultaneously with the cefotetan) plus azithromycin 1 g given orally 6–12 hours postoperatively or corresponding identical-appearing intravenous and oral placebos. Double-blinded randomization was accomplished using a computer-generated random sequence that was maintained in the Investigational Pharmacy at the University of Alabama at Birmingham. A system of opaque envelopes was used to make randomized group assignments in the operating room. Parenteral doxycycline and identical placebo infusions were prepared in the Investigational Pharmacy and stored at 4C in a refrigerator maintained by the pharmacy in the labor and delivery suite until assignment for use in the study. Azithromycin 250-mg capsules and identical-appearing placebo capsules were provided by Pfizer Pharmaceuticals (Groton, CT). The Investigational Pharmacy dispensed four 250-mg azithromycin capsules (or corresponding placebos) to the post–cesarean delivery ward nursing unit, and subjects were administered these four capsules with sips of water within 6–12 hours after the cesarean delivery.
Study personnel were not involved in the clinical management of enrolled women but followed their clinical course, verifying that the oral azithromycin capsules (or placebo) were administered and abstracting the antenatal, intraoperative, and postoperative medical record for detailed data reported in the Results section. Post–cesarean delivery endometritis was defined clinically as fever 100.4F or higher with one or more supporting clinical signs or a physician diagnosis of endometritis plus the absence of a nonpelvic source of fever. Supporting clinical signs of endometritis included maternal tachycardia (more than 100 beats per minute on two occasions at least 30 minutes apart), foul-smelling or purulent lochia, tender uterus (greater than anticipated after a cesarean delivery), and maternal leukocytosis (more than 12,000 per mm3). The study was approved by the Institutional Review Board at our center, and all participants gave written informed consent.
The sample size was calculated assuming a baseline post–cesarean delivery endometritis rate of 22% and an anticipated effect size of a 40% reduction in this outcome (22% to 13%). A two-sided test with a power of 80% and an α of 0.05 yielded a necessary sample size of 301 women in each group.
Comparisons between treatment groups were made using χ2 or Fisher exact tests where appropriate. Continuous variables were compared using a two-tailed Student t test for independent group analysis. Adjustments for variables that differed significantly between groups were performed using logistic regression analysis. Data were analyzed using SAS software (SAS Institute Inc., Cary, NC). Statistical significance was defined as P < .05.
During the study period, 1928 cesarean deliveries occurred at our institution. A summary of the women approached and recruited for the trial is depicted in Figure 1. Among these women, 455 refused participation, and 485 were determined to be ineligible. Because research personnel were not available to recruit women 24 hours a day, especially at night and on weekends, 391 women were not approached for consent. A total of 597 women were enrolled in the study including 301 women in the group randomized to receive doxycycline plus azithromycin and 296 women assigned to placebo. Selected characteristics of the study population are depicted in Table 1. The two groups were similar in regard to ethnicity, maternal age, parity, multiple gestations, prevalence of sexually transmitted diseases, and prior cesarean deliveries. Intrapartum factors that might influence the risk of developing post–cesarean delivery endometritis were also evaluated. No statistically significant difference in any of these factors was observed between the two groups with the exception that maternal leukocytosis (more than 12,000 white blood cells per mm3) was more common in the group assigned to doxycycline plus azithromycin compared with the placebo-treated group (Table 2).
Intraoperative characteristics of the two groups are depicted in Table 3. The groups were similar for the type of anesthesia administered, the primary indication for the cesarean delivery, and a variety of other intraoperative factors that might influence the risk of post–cesarean delivery endometritis (Table 3). Women randomized to receive doxycycline plus azithromycin were slightly more likely to have a low cervical hysterotomy incision than women assigned to the placebo group (P = .048, Table 3).
The overall incidence of clinical post–cesarean delivery endometritis in the study population was 20.4% (122 of 597 women). All women with endometritis had a temperature at diagnosis of at least 101F or higher and no potential extrapelvic source of their fever. Women randomized to receive doxycycline plus azithromycin (in addition to the cefotetan given to all participants) were significantly less likely to develop clinical post–cesarean delivery endometritis (n = 51, 16.9%) compared with women assigned to placebo (n = 73, 24.7%, P = .020, Table 4). Adjusting for maternal leukocytosis and uterine incision type did not alter the risk ratio for post–cesarean delivery endometritis in the doxycycline plus azithromycin versus the placebo-treated group (relative risk 0.65, 95% confidence interval 0.43, 0.98). Post-hospital discharge follow-up was available for all subjects using a computerized hospital readmission database and was achieved for 85% of the enrolled subjects by direct telephone contact. This follow-up identified two additional women with post–cesarean delivery endometritis who were both in the placebo-treated group (Table 4). Nine of the 11 women (1.8%) in the study who developed postoperative wound infections were identified during follow-up. Women in the doxycycline/azithromycin group were over four times less likely (0.8% versus 3.6%) to develop postoperative wound infections (P = .033, Table 4). When combining both endometritis and wound infections as a single outcome variable, the frequency was significantly lower in the doxycycline/ azithromycin group compared with the placebo group (P = .016, Table 4). The mean length of hospitalization was significantly shorter in the doxycycline/azithromycin group compared with the placebo group (P = .019, Table 4). Among the 122 women who developed post–cesarean delivery endometritis, antibiotic prophylaxis with doxycycline plus azithromycin was associated with an 18-hour shortening of the length of hospitalization (P = .047, Table 4). After adjusting for maternal leukocytosis and uterine incision type, the overall reduced length of stay (8 hours) in the doxycycline/azithromycin group remained significant (P = .028), and the reduced length of stay among women with endometritis and treated with doxycycline/azithromycin was 17 hours (P = .068).
Post–cesarean delivery endometritis was lower in the doxycycline/azithromycin-treated versus placebo-treated women among subjects at higher risk for post–cesarean delivery endometritis, such as those whose cesarean was indicated for dystocia (n = 84, 22.0% versus 39.5%, P = .081), a nonreassuring fetal heart rate pattern (n = 108, 19.2% versus 26.2%, P = .387), or breech presentation (n = 133, 16.7% versus 29.5%, P = .079), and among subjects with a lower risk for this outcome, such as elective repeat cesarean (n = 116, 6.8% versus 10.5%, P = .472). However, none of the differences observed on these secondary analyses were statistically significant.
We 14 and others 2,3 have reported an association between U urealyticum colonization of the genital tract and an increased risk of post–cesarean delivery endometritis. We specifically demonstrated that isolation of U urealyticum from the chorioamnion at the time of cesarean delivery is associated with a three- to eight-fold increased risk of subsequent clinical endometritis. 14 Notably, although antibiotic prophylaxis at cesarean is associated with a decreased risk of postoperative endometritis, the agents typically used for prophylaxis demonstrate little or no efficacy against U urealyticum. In this randomized clinical trial, use of adjunctive antimicrobial therapy to expand the activity spectrum of prophylactic antibiotics to include coverage for U urealyticum was associated with a significant one-third reduction in the frequency of post–cesarean delivery endometritis, a greater than fourfold decrease in post–cesarean delivery wound infections, and an approximate 1-day shortening of hospital stay among women with post–cesarean delivery endometritis.
The active drug regimen used in this study (doxycycline plus azithromycin) was chosen to maximize coverage for U urealyticum. Doxycycline was chosen because it could be given immediately as intravenous prophylaxis at the time of umbilical cord clamping along with the standard use of cefotetan. However, because of the indolent nature of infection with U urealyticum, antibiotic prophylaxis for a longer period was desired. Therefore, oral azithromycin was selected as a second agent because of its efficacy against U urealyticum and because of its long half-life resulting in prolonged antibiotic coverage with a single dose. The parenteral form of azithromycin had not yet been approved for clinical use by the United States Food and Drug Administration at the time this trial was initiated and, therefore, was not available for use in this study.
The two study groups were closely balanced for most factors that may have influenced the potential for development of post–cesarean delivery endometritis. Maternal leukocytosis (more than 12,000 white blood cells per mm3) was significantly more common in the group treated with doxycycline plus azithromycin compared with the placebo-treated group. If this were an indicator of increased risk for post–cesarean delivery endometritis, it would have biased the study so that it was less likely to observe a reduced frequency of endometritis in the doxycycline/azithromycin group. Classic uterine hysterotomy incisions are believed to have an increased risk for post–cesarean delivery infection, and the significantly higher frequency of this incision type observed in the placebo-treated group might have biased the results in favor of a reduced incidence of post–cesarean delivery endometritis in the group treated with doxycycline plus azithromycin. However, although statistically significant, the observed difference in classic uterine incisions affected only 2% of the women enrolled in the trial. Additionally, logistic regression analysis adjusting for both maternal leukocytosis and uterine incision type did not alter the reduced risk ratio for post–cesarean delivery endometritis in the doxycycline plus azithromycin group versus the placebo-treated group.
During this trial, chorioamnion cultures were not obtained to determine which women were or were not colonized with U urealyticum at the time of cesarean delivery. Thus, we do not know whether the nature of intrauterine microbial colonization at the time of cesarean delivery affected the likelihood of developing post–cesarean delivery endometritis in the two study groups. However, in the usual clinical setting in which extended spectrum antibiotic prophylaxis might be employed in the future, the microbial colonization status of the chorioamnion will not be known before administration of prophylactic antibiotics. Therefore, we conclude that the absence of culture information does not undermine the clinical observations noted in the trial. Additionally, because chorioamnion cultures were not obtained, we cannot conclude that the positive results in this study were attributed to specific expanded antibiotic efficacy against U urealyticum. Rather, it can only be concluded that the expanded spectrum antibiotic prophylaxis was more effective than cefotetan plus placebo.
All women in this study received standard prophylaxis with cefotetan that was the established antibiotic for routine cesarean prophylaxis at our institution before initiation of this study. However, it should be recognized that first-generation cephalosporins are likely to have similar efficacy compared with second-generation cephalosporins when used for cesarean antibiotic prophylaxis. Notably, however, neither first- nor second-generation cephalosporins have efficacy against U urealyticum. Thus, although cefotetan plus doxycycline and azithromycin was more effective in this study compared with cefotetan plus placebo, we do not find that our data compel a recommendation for second-generation cephalosporin use as cesarean prophylaxis.
In summary, extension of the prophylactic antibiotic efficacy spectrum at cesarean delivery beyond the commonly administered cephalosporin coverage significantly reduced morbidity associated with both post–cesarean delivery endometritis and wound infections. This reduction in morbidity was associated with a nearly 1-day shortening of hospital stay among women who developed post–cesarean delivery endometritis. Thus, the use of extended spectrum antibiotic prophylaxis should be strongly considered, especially in obstetric populations at high risk for post–cesarean delivery infection-related morbidity such as those managed in tertiary care centers. Whether such an approach would be equally effective and justifiable in a low-risk obstetric population remains a subject for future research.
1. Gilstrap LC, Cunningham FG. The bacterial pathogenesis of infection following cesarean section. Obstet Gynecol 1979;53:545–9.
2. Eschenbach DA. Ureaplasma urealyticum as a cause of postpartum fever. Pediatr Infect Dis 1986;5:S258–61.
3. Lavery JP, Marcell CC, Walker R. An association between Ureaplasma urealyticum
and endomyometritis after cesarean. J Ky Med Assoc 1985;83:359–62.
4. Carey JC, Blackwelder WC, Nugent RP, Matteson MA, Rao AV, Eschenbach DA, et al. Antepartum cultures of Ureaplasma urealyticum are not useful in predicting pregnancy outcome. Am J Obstet Gynecol 1991;164:728–33.
5. Harrison HR, Alexander ER, Weinstein L, Lewis M, Nash M, Sim DA. Cervical Chlamydia trachomatis and mycoplasmal infections in pregnancy. JAMA 1983;250:1721–7.
6. McCormack WM, Lee YH, Lin JS, Rankin JS. Genital mycoplasmas in postpartum fever. J Infect Dis 1973;127:193–6.
7. McCormack WM, Rosner B, Lee YH, Lin JS. Isolation of genital mycoplasmas from blood obtained shortly after vaginal delivery. Lancet 1975;I:596–9.
8. Kelly VN, Garland SM, Gilbert GL. Isolation of genital mycoplasmas from the blood of neonates and women with pelvic infection using conventional SPS-free blood culture media. Pathology 1977;19:277–80.
9. Neman-Simha V, Renaudin H, de Barbeyrac B, Leng JJ, Horovitz J, Dallay D, et al. Isolation of genital mycoplasmas from blood of febrile obstetrical-gynecological patients and neonates. Scand J Infect Dis 1992;24:317–21.
10. Gibbs RS, Cassell GH, Davis JK, St. Clair PJ. Further studies on genital mycoplasmas in intra-amniotic infection: Blood cultures and serologic response. Am J Obstet Gynecol 1986;154:717–26.
11. Roberts S, Maccato M, Faro S, Pinell P. The microbiology of post-cesarean wound morbidity. Obstet Gynecol 1993; 81:383–6.
12. Hillier SL, Martius J, Krohn M, Kiviat N, Holms KK, Eschenbach DA. A case-controlled study of chorioamniotic infection and histologic chorioamnionitis in prematurity. N Engl J Med 1988;319:972–8.
13. Goldenberg RL, Hauth JC, Andrews WW. Intrauterine infection and preterm delivery. N Engl J Med 2000;342:1500–7.
14. Andrews WW, Shah SR, Goldenberg RL, Cliver SP, Hauth JC, Cassell GH. Association of post-cesarean delivery endometritis with colonization of the chorioamnion by Ureaplasma urealyticum. Obstet Gynecol 1995;85:509–14.
15. Evans LC, Combs CA. Increased maternal morbidity after cesarean delivery before 28 weeks of gestation. Int J Gynecol Obstet 1993;40:227–33.
16. Seo K, McGregor JA, French JI. Preterm birth is associated with increased risk of maternal and neonatal infection. Obstet Gynecol 1992;79:75–80.
17. Duff P. Antibiotic prophylaxis. In: Phelan JP, Clark SL, eds. Cesarean delivery. New York: Elsevier, 1988:283–97.