- Email: [email protected]
Association of post-cesarean delivery endometritis with colonization of the chorioamnion by Ureaplasma urealyticum
Association of Post-Cesarean Delivery Endometritis With Colonization of the Chorioamnion by Ureaplasma urealyticum WILLIAM
W. ANDRE WS, PhD, MD, SHEILA X. SHAH, BS,
ROBERT L. GOLDENBERG, GAIL H. CASSELL, PhD
MD, SUZANNE
Objective: To determine if asymptomatic antenatal colonization of the chorioamnion with Ureaplasma urealyticum is a risk factor for the development of post-cesarean delivery endometritis. Methods: The chorioamnion was cultured at cesarean delivery for aerobic and anaerobic bacteria, mycoplasmas, Chlamydia trachomatis, and Trichomonas vaginalis in 575 singleton gestations with intact membranes. Culture results were compared with the clinical outcome. Postoperative endometritis was defined as a temperature of 38C with uterine tenderness and without other nonpelvic sources of fever. Results: Fifty-eight (10%) of the 575 women developed endometritis. Women with spontaneous labor developed endometritis twice as often as those delivered for medical or obstetric indications (17 versus 8%, P = .002). Endometritis occurred in 28% of women with U urealyticum present in the chorioamnion at cesarean delivery, compared with only 8.4% if the culture was negative and 8.8% if only bacteria other than U urealyticum were isolated (P < .OOll. Gestational age less than 34 weeks, spontaneous labor, and a vertical uterine incision were all associated with endometritis (P 5 .002). Regression analysis controlling for gestational age and incision type revealed a threefold increased risk of endometritis if the chorioamnion was colonized with U urealyticum at cesarean (odds ratio [OR] 3.0, 95% confidence interval [CI] 1.1-8.31 and an eightfold risk (OR 7.7, 95% CI 1.9-31.5) in women in whom the onset of labor was spontaneous. Conclusion: Colonization of the chorioamnion with U urealyticum in women with intact membranes being delivered by cesarean is a significant, independent predictor of subsequent endometritis. (Obstet GynecoZ1995;85:509-24)
From the Departments of Obstetrics and Gynecology and Microhiology, Unioersify of AIaburna at Birmingham, Birmingham, Alabama. This investigation was supported in part by a grant from the National Institutes of Health (NIH) (HD 209281 to Dr. Cassell and the Agency for Health Care Policy Research Contract (DHHS 282-92-0055). Ms. Shah was supported by an NIH Training Grant to Dr. Cassell (T35HLO7473).
VOL.
85,N0.
4,APRIL
lY9S
P. CLIVER, BS, JOHN C. HAUTH,
MD, AND
Endometritis has a polymicrobial etiology and is believed to occur primarily via an ascending route from the lower to upper genital tract.’ Numerous different microorganisms, including gram-positive and -negative aerobes and anaerobes, have been recovered from the upper genital tract in women diagnosed with endometritis.’ The potential role of Ureaplasma urealyticum in the pathophysiology of endometritis remains poorly defined2r3 but is of interest because these organisms are indigenous to the vaginal flora and have been reported to be present in cervicovaginal specimensfrom 68-89% of women.4 Despite the frequent occurrence of this organism in the lower genital tract, no conclusive evidence has established a relationship between its presence in the lower genital tract and postpartum feve?’ or adverse pregnancy outcome.4 However, U urealyticum has been isolated from maternal blood within minutes of delivery7 and from blood in women with postpartum infection.sr9 Gibbs et all” reported a statistically significant maternal serum U urealyticum antibody response as well as isolation of the organism from the blood of 13.6% of women with clinically evident chorioamnionitis. In addition, U urealyficum has been demonstrated to play a prominent role as a potential pathogen in post-cesarean delivery wound infections.‘i Ureaplasmaurealyticum has been reported to be the most common organism isolated by culture from the chorioamnion of women with spontaneous preterm labor and without clinical signs or symptoms of pelvic infection.12 Becauseof the increasing evidence for the pathogenicity of U urealyticum in upper genital tract infections, we initiated this investigation to determine if antenatal colonization of the chorioamnion with U urealyticum detected by culture obtained at cesarean
0029.7844/95/$9.50 0029-7844(94)00436-H
509
delivery in women with intact membranes is a risk factor for post-cesarean delivery endometritis.
Materials and Mefhods Between October 1987 and January 1991, there were 8215 live births at our institution; of these, 778 were delivered by cesarean with intact membranes. During a study designed to evaluate the association between microbial invasion of the upper genital tract and spontaneous preterm labor, research personnel were able to obtain placental cultures in 643 of these 778 women. Six hundred nine of the 643 women enrolled had singleton gestations, and the remainder had multiple gestations. Thirteen of the 609 women were excluded because they had a hysterectomy performed in conjunction with the cesarean, precluding their ability to develop endometritis. The hospital records of an additional 21 (3.4%) women were not available to establish the diagnosis of post-cesarean delivery endometritis. Therefore, the study population consisted of 575 singleton gravidas with intact membranes at cesarean delivery. The chorioamnion interface was cultured for aerobic and anaerobic bacteria, mycoplasmas, Chlamydia trachomatis, and Trickomonas vaginalis. Limitation of the investigation to women delivered by cesarean with intact membranes provided for a microbiologic evaluation of the upper genital tract to include placental tissue obtained under sterile conditions at surgery. Our intent was to preclude gross contamination by microorganisms from the lower genital tract either by the ascending route or as a result of passage of the placenta through the birth canal. The 575 gravidas were categorized as to the presence or absence of spontaneous labor. One hundred fifty-one were delivered after the spontaneous onset of labor. The primary indications for cesarean in this group were malpresentation (37%), repeat cesarean (17%), abnormal fetal heart rate (FHR) tracing (ll%), supracervical bleeding (lo%), fetal anomaly or macrosomia (lo%), and other (15%). The comparison group consisted of 424 women delivered by cesarean with intact membranes who did not have spontaneous labor and in whom delivery was indicated for specific maternal medical or obstetric complications. The primary indications for cesarean delivery in this group were repeat cesarean (26%), malpresentation (22%), abnormal FHR tracing (15%), fetal anomaly or macrosomia (13%), supracervical bleeding (3%), and other (22%). Five of the 575 women in the study population demonstrated overt clinical evidence of chorioamnionitis, such as a fever, tender uterus, or foul-smelling amniotic fluid (AF); four of these women had spontaneous labor and one had an indicated delivery. Four hundred seventy-eight (83.1%) of the 575 women received prophylactic antimicrobial
510 Andrews
et al
Ureaplasma
and Endometritis
therapy at the time of umbilical cord clamping, usually a single dose of a first-generation cephalosporin. Gestational ages were assigned by one of two maternalfetal medicine specialists after a review of the available obstetric and ultrasound data and without knowledge of culture results. Eighty-five percent of the population had at least one ultrasound examination during pregnancy, and 80% of these examinations were performed before 30 weeks’ gestational age. The predominant indication for ultrasound examination was confirmation of gestational age. Gestational age was dichotomized at 34 weeks in the multivariate analysis in order to present outcomes using odds ratios and because other? have noted an increase in microbial colonization of the chorioamnion before this gestational age. Although all subjects were enrolled and cultured prospectively, the diagnosis of endometritis was established by retrospective review of the medical records by research personnel without knowledge of the culture results. Assignment of the diagnosis of postoperative endometritis required documentation of all of the following criteria: 1) a temperature of at least 38C, 2) the presence of uterine tenderness beyond that felt to be appropriate after a cesarean, 3) the absence of other nonpelvic sources of fever, and 4) the clinical diagnosis of endometritis by the attending physicians with initiation of parenteral antimicrobial therapy. This investigation was approved by the institutional review board of the University of Alabama at Birmingham. A research nurse was on call at all times to avoid enrollment bias and insure immediate collection and processing of specimens. The placentas were collected under sterile conditions at cesarean delivery and processed in the microbiology research laboratory within 1 hour of delivery. Research personnel managed all specimen collection and processing. The placenta was placed on a sterile surface and examined grossly for anatomic malformations or obvious lesions suggestive of infection. Five sets of parallel, equally spaced incisions were made on the fetal surface of the placenta. From the first incision of each set, three swabs were inserted between the chorion and the amnion and rotated 360”. A slide for Gram stain was prepared with the first swab. The second and third swabs were placed in 2 mL of sucrose phosphate buffer containing 10% fetal bovine serum albumin, in which they were rotated several times. The fluid was then expressed, and the swabs were discarded. This suspension represented culture of the chorion and amnion and was used to inoculate all culture media. Only calcium alginate, Dacron, and polyester swabs were used. From the second incision of each set, five tissue blocks of approximately 1 mL were teased out with sterile forceps and scalpel blades. Using a sterile blade,
Obstetrics
t? Gynecology
we minced the tissues in 1 mL of sterile sucrose phosphate buffer containing 10%’ fetal bovine serum albumin and processed them for culture immediately. Media (lOB, A8, and SP4 broth and agar) were prepared and underwent quality control as described previously I3 for cultivation of mycoplasmas. The placental swab suspension and tissue specimen were thoroughly mixed on a vortex mixer. Five tenfold dilutions were made from a lOO-PL aliquot of each specimen type in 10B and SP4 broths to minimize mycoplasmacidal tissue factors known to interfere with culture recovery of mycoplasmas.‘4 An aliquot (20 PL) of the undiluted specimen and each dilution were inoculated onto A8 and SP4 agar. All plated media were incubated at 37C under 5% carbon dioxide and 95% nitrogen in a humidified incubator for a minimum of 30 days before they were designated negative. All broths were incubated at 37C under atmospheric conditions. We kept 10B broths for 2 weeks; SP4 broths were held for 8 weeks and subcultured periodically to increase the chance of isolating Mycoplasmagenitalium. Any broth tube showing a color change suggestive of mycoplasmal growth was subcultured to solid media and incubated further. Colonies of U wealyticum were identified on A8 agar by urease production in the presence of calcium chloride indicator. We identified large colony mycoplasma isolates as Mycoplasma kominis by immunoblot using species-specific monoclonal antibodies and by polymerase chain reaction @‘CR) as described previous15,16 Positive and negative controls were included in *Y. each immunoblot and PCR assay. A Gram stain of an impression smear of the cut surface of each placenta was examined systematically to make certain that any bacteria were recovered by culture. The following media were inoculated for aerobic bacterial isolation: brain-heart infusion broth, chocolate agar, Sabouraud agar (for yeasts), two Campylobacter agar plates (one incubated at 37C, and one at 42C). Human blood Tween Bilayer Medium BBL (BectonDickinson Microbiology Services, Cockeysville, MD) was used to facilitate the isolation of Gardnerellavaginalis. The chocolate Campylobacterand human blood agar Tween plates were held for 48 hours before being designated negative. The brain-heart infusion broth was held for 7 days and the Sabouraud agar for 30 days. Colonies were identified on the basis of Gram stain reaction and standard biochemical procedures. Identification procedures have been described elsewhere.17 Anaerobic media used for primary isolation included prereduced Columbia agar with 5% sheep blood and a chopped meat broth in a prereduced, anaerobically sterilized tube. Cultures were examined after 48-72 hours of incubation in an atmosphere of 90% nitrogen, 5% carbon dioxide, and 5% hydrogen at 37C for growth.
VOL.
85,
NO.
4, APRIL
1995
Table
1. Incidence of Post-Cesarean Versus Gestational Age
Delivery
Gestational Study
group
Spontaneous
Indicated Total
n
530
31-33
(n = 95)
(n = 47)
151 424
11/35 9/60
(31%) (15%)
575
20/95(21%)
3/10 5/37
(30%) (14%)
S/47(17%)
Endometritis
age (wk) 34-36
(n = 82)
237
(n = 351)
6/29 3/53
(21%) (6%)
S/77(6%) 16/274 (6%)
9/82
(11%)
21/351(6%)
Negative primary isolation plates were discarded at 5 days. Broths were observed for up to 14 days, and any broth showing evidence of growth received Gram stain and was subcultured to solid media. Identification, performed as outlined elsewhere,18,19 included biochemical testing using prereduced, anaerobically sterilized sugars, gas liquid chromatography, and antibiotic susceptibility testing. Trichosel BBL (Becton-Dickinson Microbiology Services) broth incubated at 37C was used for cultivation of T vaginalis.Aliquots of broth were examined microscopically for the presence of motile trichomonads after 3 and 5 days of incubation. For culture of C frackomatis, minced placental tissue was sonicated to disrupt cellular membranes, stored at 4C for a maximum of 24 hours, and used to inoculate (0.1 mL per well in three wells) microtiter plate cultures of cycloheximide-treated McCoy cells. Known positive and negative specimens were inoculated simultaneously for controls. After 48 hours incubation in 5% carbon dioxide at 37C, we evaluated the cultures by immunofluorescence using a specific monoclonal antibody (Syva Company, San Jose, CA). The data were analyzed with two-tailed ,$ tests for univariate analyses of dichotomous variables. Multiple logistic regression techniques performed using SAS statistical software (SAS Institute, Inc., Cary, NC) were used for multivariate analyses. Statistical significance was set at 01 = 0.05.
Results Fifty-eight (10%) of the 575 women in the study population developed post-cesarean delivery endometritis. By univariate analyses, women in spontaneous labor at the time of cesarean delivery developed endometritis twice as often as those without spontaneous labor who were delivered for medical or obstetric indications (17 versus 8%, P = .002). A significant inverse relationship was observed overall between the incidence of postcesarean delivery endometritis and gestational age at delivery (P < .OOl; Table 1). Although this relationship was more pronounced in women with spontaneous
Andrews
et al
Ureaplasma
andEndometritis
511
GestatIonal
age
Spontaneous
Vertical
labor
uterine
Duration Use
< 34 weeks
lnclsion
of surgery
of prophylactic
Primary
_;
> 1 hr
I_
antibiotics
Pelvic
adhesive
Uterlne
disease
lacerations
Btlateral
tubal
Blood
loss
I
-i---
cesarean
--A-
ligation
z 1000
I_ II
mL 0:1
0:5 Odds
IlO Ratio
210
and
95%
Id.0 Cl.
Figure 1. Results of a regression analysis of factors with a potential impact on post-cesarean delivery endometritis. Closed circles represent the odds ratio and horizontal lines depict the 95% confidence interval (CI).
ria, and 4) culture positive for U urealyfictfm only. When U urealyficum was recovered from the chorioamnion at cesareandelivery, either as a single isolate or in combination with other bacteria, the incidence of postcesarean endometritis (28 and 22.5%, respectively) was significantly increased compared with women with negative cultures or from whom bacteria only were isolated (8.4 and 8.8%, respectively; P < ,001; Table 2). Regression analysis controlling for gestational age and uterine incision type revealed that in women from whom ll urealyficum was recovered in pure culture, there was a threefold increased risk for endometritis overall (odds ratio [OR] 3.0, 95% confidence interval [CI] 1.1-8.3) and a nearly eightfold increase (OR 7.7, 95% CI 1.9-31.5) in women who also had spontaneous labor.
Discussion labor (P = .005), it was also observed in the absence (P = .05) of spontaneous labor. A single multiple logistic regression analysis of potential risk factors for post-cesarean delivery endometritis revealed that gestational age less than 34 weeks, spontaneous labor, and a vertical uterine incision were each significantly associated with the development of postoperative endometritis (Figure 1). Duration of surgery, use of prophylactic antibiotics, primary cesarean, presence of adhesions or uterine lacerations, performance of a tubal ligation, and intraoperative blood loss were not statistically associated with the development of postoperative endometritis in this population. In the overall study population, endometritis occurred in 16.8% of women whose original chorioamnion culture was positive, compared with 8.4% in those with a negative culture at cesarean delivery (P = .OOB). The most common organism isolated from the chorioamnion was U tlrealyticum 00%) followed by M kominis (3%) and one or more of 25 species of gram-positive and -negative aerobic and anaerobic bacteria (15%). For subsequent analyses, the women were subdivided into four groups according to the chorioamnion culture results: 1) negative culture, 2) culture positive for all bacteria other than U urea!yficum (bacteria only), 3) culture positive for U urenlyficum and any other bacteTable 2. Incidence by Culture Culture Negative Bacteria Bacteria
of PostkCesarean Results
results
Delivery
Endometritis
Endometritis 39/462 5/57 7/31 7/25
only and U urealyticum
U urealyticum
u = Ureaplnsma.
512 Andrews et al
Ureaplasma
and Erzdomrfritis
Incidence 8.4% 8.8% 22.5% 28%
Limiting the study population to women delivered by cesarean with intact membranes offers several advantages for microbiologic study. First, each placenta was obtained under sterile conditions for chorioamnion culture, thus avoiding gross contamination with lower genital tract microorganisms by passage of the specimen through the birth canal. Second, although women with ruptured membranes are at increased risk for developing post-cesarean delivery endometritis, limiting the study population to those with intact membranes prevented gross microbial contamination of the chorioamnion in situ via an ascending route from the lower genital tract. Thus, several of the disadvantages inherent in the study of microbial invasion of the upper genital tract in women with ruptured membranes and/or delivered vaginally were specifically avoided by the study design. In addition, less than 1% of the study population demonstrated overt clinical evidence of intrauterine infection, such as fever, a tender uterus, or foul-smelling AF before or during the cesarean delivery. Therefore, isolation of microorganisms from the chorioamnion at cesarean generally represented asymptomatic antenatal colonization of the upper genital tract. The overall incidence of post-cesarean delivery endometritis in the study population (10%) was somewhat lower than the incidence in the general population delivered at our institution (15-20%). This lower incidence was expected because women with ruptured membranes, one of the highest risk factors for development of post-cesarean endometritis, were excluded. However, despite the low incidence of endometritis in this study population, women with spontaneous labor developed this postoperative infectious complication twice as often as women without spontaneous labor.
Obstetrics 6 Gynecology
The striking inverse relationship between the incidence of post-cesarean endometritis and the gestational age at delivery (Table 1) confirms other reports.20,21 We observed that earlier gestational age at delivery was associated with an increasingly greater risk of developing post-cesarean delivery endometritis. Although the inverse relationship between gestational age and endometritis was observed regardless of the presence or absence of labor, this relationship was strongest in women with spontaneous labor. Indeed, women with spontaneous labor who delivered before 31 weeks’ gestation were almost five times more likely to develop post-cesarean endometritis than women with spontaneous labor who were delivered at term. We have reported previously (Cassell G, Hauth JC, Andrews W, Cutter G, Goldenberg R. Chorioamnion colonization: Correlation with gestational age in women delivered following spontaneous labor versus indicated delivery [abstract]. Am J Obstet Gynecol 1993;168:425) that asymptomatic chorioamnion colonization is more frequent and inversely proportional to gestational age in women with the spontaneous onset of labor, and have speculated that such colonization may play an important role in the etiology of preterm labor. Therefore, we hypothesize that an important contributing factor to the increased incidence of post-ces;r’ean delivery endometritis in women with spontaneous labor and early gestational age at delivery is clinically silent microbial colonization of the chorioamnion already present at the time of cesarean. This hypothesis is supported by our observation that women with a positive chorioamnion culture at cesarean were twice as likely to develop endometritis as those whose culture was negative. Perhaps the most important finding in this investigation was the strong association between chorioamnion colonization with U urealyticum at delivery and subsequent development of post-cesarean delivery endometritis. Indeed, women from whom U uvealyticum was isolated in pure culture were over three times more likely to develop post-cesarean endometritis than were women with negative chorioamnion cultures or with cultures positive for bacteria other than U urealyticum. A regression analysis controlling for the important confounding risk factors of gestational age and uterine incision type revealed a threefold increased risk for endometritis overall if the chorioamnion was colonized with U uvealyticum as a single isolate and an eightfold increased risk in women with spontaneous labor. We should emphasize that despite the strong statistical association, only 14 of the 58 cases of endometritis (24.1%) were associated with antenatal chorioamnion colonization with U urealyticum isolated either in pure culture or in combination with other bacteria. These data, together with the preliminary reports by
VOL.
85, NO.
4, APRIL
1995
Eschenbach2 and Lavery et al,” strongly support the conclusion that asymptomatic colonization of the chorioamnion with U urealyticum in women with intact membranes undergoing cesarean delivery is a significant independent predictor of subsequent endometritis. The mechanism by which such colonization of the chorioamnion may predispose to subsequent endometritis remains speculative. Available data suggest that the organisms reach the chorioamnion via an ascending route.22 Virulence mechanisms which might facilitate invasion of intact membranes have not been identified but may involve production of several enzymes, including immunoglobulin-A, protease, urease,24 and phospholipase A and C.25 Once in the upper genital tract, U urealyticum may play a direct role as an important pathogen in post-cesarean endometritis as has been proposed for this microorganism in the pathophysiology of post-cesarean wound infections.” The previously reported increased frequency of U uvealyticum isolation as a predominant organism from the chorioamnion in women with spontaneous labor (Cassell G, Andrews W, Hauth J, et al. Isolation of microorganisms from the chorioamnion is twice that from amniotic fluid at cesarean delivery in women with intact membranes [abstract]. Am J Obstet Gynecol 1993;168:424)i2 along with the strong association between the presence of this organism in the chorioamnion and subsequent endometritis, may offer an explanation of the potential source of U urealyticum in post-cesarean delivery wound infections. An alternative role for U urealyticum in the pathogenesis of post-cesarean endometritis may be as a promoter for growth of gram-negative and grampositive organisms long known to be associated with this infection.’ However, it is important to note that the highest incidence of endometritis among women with positive chorioamnion cultures was observed when U urealyticum was isolated in pure culture. Regardless of the mechanism, it is important to recognize the association between this organism and post-cesarean endometritis. A greater understanding of this association may lead to improved modalities for prevention and treatment of this common complication of cesarean delivery. The association between U urealyticum colonization of the chorioamnion at cesarean with an increased incidence of post-cesarean endometritis occurred despite the routine use of intraoperative prophylactic antimicrobial therapy. The substantial benefit derived from the use of prophylactic therapy at cesarean delivery is well established and results in a 50-60% reduction in postoperative endometritis.26 Numerous antimicrobial agents have been reported to have similar efficacy in reducing the post-cesarean febrile morbidity rate.26 However, the antimicrobials commonly used do not
Andrews
et al
Ureaplasma
and Endomefvifis
513
have established efficacy against U urealyticum. Because in most the cervix is colonized with U urealyticum women’ and genital mycoplasmas are frequently isolated from the uterine cavity in women with ruptured membranes,12,27,28 the relationship between this organism and post-cesarean delivery endometritis might be even higher in women with ruptured membranes. Onequarter of the casesof endometritis in this population of women with intact membranes were associated with chorioamnion colonization with U urealyticum. Thus, expansion of current prophylactic antimicrobial regimens to include coverage for U urealyticum may further reduce the incidence of post-cesarean endometritis by 25% or more. The efficacy of this expanded prophylactic therapy should be investigated in a randomized controlled trial.
20.
References
21.
1. Gilstrap LC, Cunningham FG. The bacterial pathogenesis of infection following cesarean section. Obstet GynecoI 1979;53:545-9. 2. Eschenbach DA. Ureaplasma urealyticum as a cause of postpartum fever. Pediatr Infect Dis 1986;5:5258-61. 3. Lavery JP, Marcel1 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 RI’, 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. Epidemiology and outcomes. 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, Rankin JS, 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, et al. Isolation of genital mycoplasmas from blood of febrile obstetrical-gynecologic patients and neonates. Stand 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, Pine11 P. The microbiology of post-cesarean wound morbidity. Obstet Gynecol 1993;81:383-6. J, Krohn M, Kiviat N, Holms KK, Eschenbach 12. Hillier SL, Martius DA. A case-controlled study of chorioamniotic infection and histologic chorioamnionitis in prematurity. N Engl J Med 1988;319: 972-8. A, Duffy L, Crabb D, Waites KB. Mycoplas13. Cassell GH, Blanchard mas in clinical and pathogenic microbiology. In: Howard BJ, ed. Mycoplasmas. St. Louis: Mosby-Year Book, Inc., 1993:491-502. D, Chen TA. Methods in mycoplasmology. In: 14. Taylor-Robinson
514
Andrews
et al
Ureaplasma
and Endometrifis
15.
16.
17.
18.
19.
22. 23.
24.
25. 26. 27.
28.
Razin S, Tully JG, eds. Growth inhibitory factors in animal and plant tissues. Vol 1. New York: Academic Press, Inc., 1983:109-14. Cassell GH, Watson HL, Blalock DK, Horowitz SA, Duffy LB. Protein antigens of genital mycoplasmas. Rev Infect Dis 1988;lO: S391-8. Blanchard A, Dybvig YK, Watson HL, Griffiths G, Cassell GH. Evaluation of intraspecies genetic variation within the 16s rRNA gene of Mycoplasma hominis and detection by polymerase chain reaction. J Clin Microbial 1993;31:1358-61, Balows A, Hausler WJ, Herrmann KL, Isenberg HD, Shadomy HJ. Manual of clinical microbiology. 5th ed., Washington, DC: American Society for Microbiology, 1991. Sutter VL, Citron DM, Edelstein MAC, Finegold SM. Wadesworth anaerobic bacteriology manual. 4th ed. Belmont, California: Star Publishing Company, 1985. Holdeman LV, Cato El’, Moore WEC. Anaerobic laboratory manual. 4th ed. Blacksburg, Virginia: Virginia Polytechnic Institute and State University, 1977. Evans LC, Combs CA. Increased maternal morbidity after cesarean delivery before 28 weeks of gestation. Int J Gynecol Obstet 1993; 40:227-33. Seo K, McGregor JA, French JI. Preterm birth is associated with increased risk of maternal and neonatal infection. Obstet Gynecol 1992;79:75-80. Romer R, Mazor M. Infection and preterm labor. Clin Obstet Gynecol 1988;31:553-84. Kilian M, Brown MB, Brown TA, Cassell GH. Immunoglobulin A protease activity of Ureaplasma urealyticum. Acta Path01 Microbiol Stand [BJ 1984;92:61-4. Takebe S, Numata A, Kobashi K. Stone formation by Ureaplasma urealyticum in human urine and its prevention by urease inhibitors. J Clin Microbial 1984;20:869-73. Desilva NS, Quinn PA. Phospholipase A and C activity in Wrenplasma urealyticum. J Clin Microbial 1986;23:354-9. Duff I’. Antibiotic prophylaxis. In: Phelan JP, Clark SL, eds. Cesarean delivery. New York: Elsevier, 1988:283-97. Hillier SL, Krohn MA, Kiviat NB, Watts DH, Eschenbach DA. Microbiologic causes and neonatal outcomes associated with chorioamnion infection. Am J Obstet Gynecol 1991;165:955-61. Romero R, Yoon BH, Mazur M, et al. A comparative study of the diagnostic performance of amniotic fluid glucose, white blood cell count, interleukin-6, and Gram stain in the detection of microbial invasion in patients with preterm premature rupture of membranes. Am J Obstet Gynecol 1993;169:839-51.
Addressreprint requeststo: Gail K. Cassell,PhD University of Alabama at Birmingham Department of Microbiology University Station Birmingham, AL 35233-7333
Received August 29, 1994. Received in revised form November 14, 1994. Accepted November 29, 2994. Copyright 0 1995 by The American Gynecologists.
College
of Obstetricians
and
Obstetrics G Gynecology
W. ANDRE WS, PhD, MD, SHEILA X. SHAH, BS,
ROBERT L. GOLDENBERG, GAIL H. CASSELL, PhD
MD, SUZANNE
Objective: To determine if asymptomatic antenatal colonization of the chorioamnion with Ureaplasma urealyticum is a risk factor for the development of post-cesarean delivery endometritis. Methods: The chorioamnion was cultured at cesarean delivery for aerobic and anaerobic bacteria, mycoplasmas, Chlamydia trachomatis, and Trichomonas vaginalis in 575 singleton gestations with intact membranes. Culture results were compared with the clinical outcome. Postoperative endometritis was defined as a temperature of 38C with uterine tenderness and without other nonpelvic sources of fever. Results: Fifty-eight (10%) of the 575 women developed endometritis. Women with spontaneous labor developed endometritis twice as often as those delivered for medical or obstetric indications (17 versus 8%, P = .002). Endometritis occurred in 28% of women with U urealyticum present in the chorioamnion at cesarean delivery, compared with only 8.4% if the culture was negative and 8.8% if only bacteria other than U urealyticum were isolated (P < .OOll. Gestational age less than 34 weeks, spontaneous labor, and a vertical uterine incision were all associated with endometritis (P 5 .002). Regression analysis controlling for gestational age and incision type revealed a threefold increased risk of endometritis if the chorioamnion was colonized with U urealyticum at cesarean (odds ratio [OR] 3.0, 95% confidence interval [CI] 1.1-8.31 and an eightfold risk (OR 7.7, 95% CI 1.9-31.5) in women in whom the onset of labor was spontaneous. Conclusion: Colonization of the chorioamnion with U urealyticum in women with intact membranes being delivered by cesarean is a significant, independent predictor of subsequent endometritis. (Obstet GynecoZ1995;85:509-24)
From the Departments of Obstetrics and Gynecology and Microhiology, Unioersify of AIaburna at Birmingham, Birmingham, Alabama. This investigation was supported in part by a grant from the National Institutes of Health (NIH) (HD 209281 to Dr. Cassell and the Agency for Health Care Policy Research Contract (DHHS 282-92-0055). Ms. Shah was supported by an NIH Training Grant to Dr. Cassell (T35HLO7473).
VOL.
85,N0.
4,APRIL
lY9S
P. CLIVER, BS, JOHN C. HAUTH,
MD, AND
Endometritis has a polymicrobial etiology and is believed to occur primarily via an ascending route from the lower to upper genital tract.’ Numerous different microorganisms, including gram-positive and -negative aerobes and anaerobes, have been recovered from the upper genital tract in women diagnosed with endometritis.’ The potential role of Ureaplasma urealyticum in the pathophysiology of endometritis remains poorly defined2r3 but is of interest because these organisms are indigenous to the vaginal flora and have been reported to be present in cervicovaginal specimensfrom 68-89% of women.4 Despite the frequent occurrence of this organism in the lower genital tract, no conclusive evidence has established a relationship between its presence in the lower genital tract and postpartum feve?’ or adverse pregnancy outcome.4 However, U urealyticum has been isolated from maternal blood within minutes of delivery7 and from blood in women with postpartum infection.sr9 Gibbs et all” reported a statistically significant maternal serum U urealyticum antibody response as well as isolation of the organism from the blood of 13.6% of women with clinically evident chorioamnionitis. In addition, U urealyficum has been demonstrated to play a prominent role as a potential pathogen in post-cesarean delivery wound infections.‘i Ureaplasmaurealyticum has been reported to be the most common organism isolated by culture from the chorioamnion of women with spontaneous preterm labor and without clinical signs or symptoms of pelvic infection.12 Becauseof the increasing evidence for the pathogenicity of U urealyticum in upper genital tract infections, we initiated this investigation to determine if antenatal colonization of the chorioamnion with U urealyticum detected by culture obtained at cesarean
0029.7844/95/$9.50 0029-7844(94)00436-H
509
delivery in women with intact membranes is a risk factor for post-cesarean delivery endometritis.
Materials and Mefhods Between October 1987 and January 1991, there were 8215 live births at our institution; of these, 778 were delivered by cesarean with intact membranes. During a study designed to evaluate the association between microbial invasion of the upper genital tract and spontaneous preterm labor, research personnel were able to obtain placental cultures in 643 of these 778 women. Six hundred nine of the 643 women enrolled had singleton gestations, and the remainder had multiple gestations. Thirteen of the 609 women were excluded because they had a hysterectomy performed in conjunction with the cesarean, precluding their ability to develop endometritis. The hospital records of an additional 21 (3.4%) women were not available to establish the diagnosis of post-cesarean delivery endometritis. Therefore, the study population consisted of 575 singleton gravidas with intact membranes at cesarean delivery. The chorioamnion interface was cultured for aerobic and anaerobic bacteria, mycoplasmas, Chlamydia trachomatis, and Trickomonas vaginalis. Limitation of the investigation to women delivered by cesarean with intact membranes provided for a microbiologic evaluation of the upper genital tract to include placental tissue obtained under sterile conditions at surgery. Our intent was to preclude gross contamination by microorganisms from the lower genital tract either by the ascending route or as a result of passage of the placenta through the birth canal. The 575 gravidas were categorized as to the presence or absence of spontaneous labor. One hundred fifty-one were delivered after the spontaneous onset of labor. The primary indications for cesarean in this group were malpresentation (37%), repeat cesarean (17%), abnormal fetal heart rate (FHR) tracing (ll%), supracervical bleeding (lo%), fetal anomaly or macrosomia (lo%), and other (15%). The comparison group consisted of 424 women delivered by cesarean with intact membranes who did not have spontaneous labor and in whom delivery was indicated for specific maternal medical or obstetric complications. The primary indications for cesarean delivery in this group were repeat cesarean (26%), malpresentation (22%), abnormal FHR tracing (15%), fetal anomaly or macrosomia (13%), supracervical bleeding (3%), and other (22%). Five of the 575 women in the study population demonstrated overt clinical evidence of chorioamnionitis, such as a fever, tender uterus, or foul-smelling amniotic fluid (AF); four of these women had spontaneous labor and one had an indicated delivery. Four hundred seventy-eight (83.1%) of the 575 women received prophylactic antimicrobial
510 Andrews
et al
Ureaplasma
and Endometritis
therapy at the time of umbilical cord clamping, usually a single dose of a first-generation cephalosporin. Gestational ages were assigned by one of two maternalfetal medicine specialists after a review of the available obstetric and ultrasound data and without knowledge of culture results. Eighty-five percent of the population had at least one ultrasound examination during pregnancy, and 80% of these examinations were performed before 30 weeks’ gestational age. The predominant indication for ultrasound examination was confirmation of gestational age. Gestational age was dichotomized at 34 weeks in the multivariate analysis in order to present outcomes using odds ratios and because other? have noted an increase in microbial colonization of the chorioamnion before this gestational age. Although all subjects were enrolled and cultured prospectively, the diagnosis of endometritis was established by retrospective review of the medical records by research personnel without knowledge of the culture results. Assignment of the diagnosis of postoperative endometritis required documentation of all of the following criteria: 1) a temperature of at least 38C, 2) the presence of uterine tenderness beyond that felt to be appropriate after a cesarean, 3) the absence of other nonpelvic sources of fever, and 4) the clinical diagnosis of endometritis by the attending physicians with initiation of parenteral antimicrobial therapy. This investigation was approved by the institutional review board of the University of Alabama at Birmingham. A research nurse was on call at all times to avoid enrollment bias and insure immediate collection and processing of specimens. The placentas were collected under sterile conditions at cesarean delivery and processed in the microbiology research laboratory within 1 hour of delivery. Research personnel managed all specimen collection and processing. The placenta was placed on a sterile surface and examined grossly for anatomic malformations or obvious lesions suggestive of infection. Five sets of parallel, equally spaced incisions were made on the fetal surface of the placenta. From the first incision of each set, three swabs were inserted between the chorion and the amnion and rotated 360”. A slide for Gram stain was prepared with the first swab. The second and third swabs were placed in 2 mL of sucrose phosphate buffer containing 10% fetal bovine serum albumin, in which they were rotated several times. The fluid was then expressed, and the swabs were discarded. This suspension represented culture of the chorion and amnion and was used to inoculate all culture media. Only calcium alginate, Dacron, and polyester swabs were used. From the second incision of each set, five tissue blocks of approximately 1 mL were teased out with sterile forceps and scalpel blades. Using a sterile blade,
Obstetrics
t? Gynecology
we minced the tissues in 1 mL of sterile sucrose phosphate buffer containing 10%’ fetal bovine serum albumin and processed them for culture immediately. Media (lOB, A8, and SP4 broth and agar) were prepared and underwent quality control as described previously I3 for cultivation of mycoplasmas. The placental swab suspension and tissue specimen were thoroughly mixed on a vortex mixer. Five tenfold dilutions were made from a lOO-PL aliquot of each specimen type in 10B and SP4 broths to minimize mycoplasmacidal tissue factors known to interfere with culture recovery of mycoplasmas.‘4 An aliquot (20 PL) of the undiluted specimen and each dilution were inoculated onto A8 and SP4 agar. All plated media were incubated at 37C under 5% carbon dioxide and 95% nitrogen in a humidified incubator for a minimum of 30 days before they were designated negative. All broths were incubated at 37C under atmospheric conditions. We kept 10B broths for 2 weeks; SP4 broths were held for 8 weeks and subcultured periodically to increase the chance of isolating Mycoplasmagenitalium. Any broth tube showing a color change suggestive of mycoplasmal growth was subcultured to solid media and incubated further. Colonies of U wealyticum were identified on A8 agar by urease production in the presence of calcium chloride indicator. We identified large colony mycoplasma isolates as Mycoplasma kominis by immunoblot using species-specific monoclonal antibodies and by polymerase chain reaction @‘CR) as described previous15,16 Positive and negative controls were included in *Y. each immunoblot and PCR assay. A Gram stain of an impression smear of the cut surface of each placenta was examined systematically to make certain that any bacteria were recovered by culture. The following media were inoculated for aerobic bacterial isolation: brain-heart infusion broth, chocolate agar, Sabouraud agar (for yeasts), two Campylobacter agar plates (one incubated at 37C, and one at 42C). Human blood Tween Bilayer Medium BBL (BectonDickinson Microbiology Services, Cockeysville, MD) was used to facilitate the isolation of Gardnerellavaginalis. The chocolate Campylobacterand human blood agar Tween plates were held for 48 hours before being designated negative. The brain-heart infusion broth was held for 7 days and the Sabouraud agar for 30 days. Colonies were identified on the basis of Gram stain reaction and standard biochemical procedures. Identification procedures have been described elsewhere.17 Anaerobic media used for primary isolation included prereduced Columbia agar with 5% sheep blood and a chopped meat broth in a prereduced, anaerobically sterilized tube. Cultures were examined after 48-72 hours of incubation in an atmosphere of 90% nitrogen, 5% carbon dioxide, and 5% hydrogen at 37C for growth.
VOL.
85,
NO.
4, APRIL
1995
Table
1. Incidence of Post-Cesarean Versus Gestational Age
Delivery
Gestational Study
group
Spontaneous
Indicated Total
n
530
31-33
(n = 95)
(n = 47)
151 424
11/35 9/60
(31%) (15%)
575
20/95(21%)
3/10 5/37
(30%) (14%)
S/47(17%)
Endometritis
age (wk) 34-36
(n = 82)
237
(n = 351)
6/29 3/53
(21%) (6%)
S/77(6%) 16/274 (6%)
9/82
(11%)
21/351(6%)
Negative primary isolation plates were discarded at 5 days. Broths were observed for up to 14 days, and any broth showing evidence of growth received Gram stain and was subcultured to solid media. Identification, performed as outlined elsewhere,18,19 included biochemical testing using prereduced, anaerobically sterilized sugars, gas liquid chromatography, and antibiotic susceptibility testing. Trichosel BBL (Becton-Dickinson Microbiology Services) broth incubated at 37C was used for cultivation of T vaginalis.Aliquots of broth were examined microscopically for the presence of motile trichomonads after 3 and 5 days of incubation. For culture of C frackomatis, minced placental tissue was sonicated to disrupt cellular membranes, stored at 4C for a maximum of 24 hours, and used to inoculate (0.1 mL per well in three wells) microtiter plate cultures of cycloheximide-treated McCoy cells. Known positive and negative specimens were inoculated simultaneously for controls. After 48 hours incubation in 5% carbon dioxide at 37C, we evaluated the cultures by immunofluorescence using a specific monoclonal antibody (Syva Company, San Jose, CA). The data were analyzed with two-tailed ,$ tests for univariate analyses of dichotomous variables. Multiple logistic regression techniques performed using SAS statistical software (SAS Institute, Inc., Cary, NC) were used for multivariate analyses. Statistical significance was set at 01 = 0.05.
Results Fifty-eight (10%) of the 575 women in the study population developed post-cesarean delivery endometritis. By univariate analyses, women in spontaneous labor at the time of cesarean delivery developed endometritis twice as often as those without spontaneous labor who were delivered for medical or obstetric indications (17 versus 8%, P = .002). A significant inverse relationship was observed overall between the incidence of postcesarean delivery endometritis and gestational age at delivery (P < .OOl; Table 1). Although this relationship was more pronounced in women with spontaneous
Andrews
et al
Ureaplasma
andEndometritis
511
GestatIonal
age
Spontaneous
Vertical
labor
uterine
Duration Use
< 34 weeks
lnclsion
of surgery
of prophylactic
Primary
_;
> 1 hr
I_
antibiotics
Pelvic
adhesive
Uterlne
disease
lacerations
Btlateral
tubal
Blood
loss
I
-i---
cesarean
--A-
ligation
z 1000
I_ II
mL 0:1
0:5 Odds
IlO Ratio
210
and
95%
Id.0 Cl.
Figure 1. Results of a regression analysis of factors with a potential impact on post-cesarean delivery endometritis. Closed circles represent the odds ratio and horizontal lines depict the 95% confidence interval (CI).
ria, and 4) culture positive for U urealyfictfm only. When U urealyficum was recovered from the chorioamnion at cesareandelivery, either as a single isolate or in combination with other bacteria, the incidence of postcesarean endometritis (28 and 22.5%, respectively) was significantly increased compared with women with negative cultures or from whom bacteria only were isolated (8.4 and 8.8%, respectively; P < ,001; Table 2). Regression analysis controlling for gestational age and uterine incision type revealed that in women from whom ll urealyficum was recovered in pure culture, there was a threefold increased risk for endometritis overall (odds ratio [OR] 3.0, 95% confidence interval [CI] 1.1-8.3) and a nearly eightfold increase (OR 7.7, 95% CI 1.9-31.5) in women who also had spontaneous labor.
Discussion labor (P = .005), it was also observed in the absence (P = .05) of spontaneous labor. A single multiple logistic regression analysis of potential risk factors for post-cesarean delivery endometritis revealed that gestational age less than 34 weeks, spontaneous labor, and a vertical uterine incision were each significantly associated with the development of postoperative endometritis (Figure 1). Duration of surgery, use of prophylactic antibiotics, primary cesarean, presence of adhesions or uterine lacerations, performance of a tubal ligation, and intraoperative blood loss were not statistically associated with the development of postoperative endometritis in this population. In the overall study population, endometritis occurred in 16.8% of women whose original chorioamnion culture was positive, compared with 8.4% in those with a negative culture at cesarean delivery (P = .OOB). The most common organism isolated from the chorioamnion was U tlrealyticum 00%) followed by M kominis (3%) and one or more of 25 species of gram-positive and -negative aerobic and anaerobic bacteria (15%). For subsequent analyses, the women were subdivided into four groups according to the chorioamnion culture results: 1) negative culture, 2) culture positive for all bacteria other than U urea!yficum (bacteria only), 3) culture positive for U urenlyficum and any other bacteTable 2. Incidence by Culture Culture Negative Bacteria Bacteria
of PostkCesarean Results
results
Delivery
Endometritis
Endometritis 39/462 5/57 7/31 7/25
only and U urealyticum
U urealyticum
u = Ureaplnsma.
512 Andrews et al
Ureaplasma
and Erzdomrfritis
Incidence 8.4% 8.8% 22.5% 28%
Limiting the study population to women delivered by cesarean with intact membranes offers several advantages for microbiologic study. First, each placenta was obtained under sterile conditions for chorioamnion culture, thus avoiding gross contamination with lower genital tract microorganisms by passage of the specimen through the birth canal. Second, although women with ruptured membranes are at increased risk for developing post-cesarean delivery endometritis, limiting the study population to those with intact membranes prevented gross microbial contamination of the chorioamnion in situ via an ascending route from the lower genital tract. Thus, several of the disadvantages inherent in the study of microbial invasion of the upper genital tract in women with ruptured membranes and/or delivered vaginally were specifically avoided by the study design. In addition, less than 1% of the study population demonstrated overt clinical evidence of intrauterine infection, such as fever, a tender uterus, or foul-smelling AF before or during the cesarean delivery. Therefore, isolation of microorganisms from the chorioamnion at cesarean generally represented asymptomatic antenatal colonization of the upper genital tract. The overall incidence of post-cesarean delivery endometritis in the study population (10%) was somewhat lower than the incidence in the general population delivered at our institution (15-20%). This lower incidence was expected because women with ruptured membranes, one of the highest risk factors for development of post-cesarean endometritis, were excluded. However, despite the low incidence of endometritis in this study population, women with spontaneous labor developed this postoperative infectious complication twice as often as women without spontaneous labor.
Obstetrics 6 Gynecology
The striking inverse relationship between the incidence of post-cesarean endometritis and the gestational age at delivery (Table 1) confirms other reports.20,21 We observed that earlier gestational age at delivery was associated with an increasingly greater risk of developing post-cesarean delivery endometritis. Although the inverse relationship between gestational age and endometritis was observed regardless of the presence or absence of labor, this relationship was strongest in women with spontaneous labor. Indeed, women with spontaneous labor who delivered before 31 weeks’ gestation were almost five times more likely to develop post-cesarean endometritis than women with spontaneous labor who were delivered at term. We have reported previously (Cassell G, Hauth JC, Andrews W, Cutter G, Goldenberg R. Chorioamnion colonization: Correlation with gestational age in women delivered following spontaneous labor versus indicated delivery [abstract]. Am J Obstet Gynecol 1993;168:425) that asymptomatic chorioamnion colonization is more frequent and inversely proportional to gestational age in women with the spontaneous onset of labor, and have speculated that such colonization may play an important role in the etiology of preterm labor. Therefore, we hypothesize that an important contributing factor to the increased incidence of post-ces;r’ean delivery endometritis in women with spontaneous labor and early gestational age at delivery is clinically silent microbial colonization of the chorioamnion already present at the time of cesarean. This hypothesis is supported by our observation that women with a positive chorioamnion culture at cesarean were twice as likely to develop endometritis as those whose culture was negative. Perhaps the most important finding in this investigation was the strong association between chorioamnion colonization with U urealyticum at delivery and subsequent development of post-cesarean delivery endometritis. Indeed, women from whom U uvealyticum was isolated in pure culture were over three times more likely to develop post-cesarean endometritis than were women with negative chorioamnion cultures or with cultures positive for bacteria other than U urealyticum. A regression analysis controlling for the important confounding risk factors of gestational age and uterine incision type revealed a threefold increased risk for endometritis overall if the chorioamnion was colonized with U uvealyticum as a single isolate and an eightfold increased risk in women with spontaneous labor. We should emphasize that despite the strong statistical association, only 14 of the 58 cases of endometritis (24.1%) were associated with antenatal chorioamnion colonization with U urealyticum isolated either in pure culture or in combination with other bacteria. These data, together with the preliminary reports by
VOL.
85, NO.
4, APRIL
1995
Eschenbach2 and Lavery et al,” strongly support the conclusion that asymptomatic colonization of the chorioamnion with U urealyticum in women with intact membranes undergoing cesarean delivery is a significant independent predictor of subsequent endometritis. The mechanism by which such colonization of the chorioamnion may predispose to subsequent endometritis remains speculative. Available data suggest that the organisms reach the chorioamnion via an ascending route.22 Virulence mechanisms which might facilitate invasion of intact membranes have not been identified but may involve production of several enzymes, including immunoglobulin-A, protease, urease,24 and phospholipase A and C.25 Once in the upper genital tract, U urealyticum may play a direct role as an important pathogen in post-cesarean endometritis as has been proposed for this microorganism in the pathophysiology of post-cesarean wound infections.” The previously reported increased frequency of U uvealyticum isolation as a predominant organism from the chorioamnion in women with spontaneous labor (Cassell G, Andrews W, Hauth J, et al. Isolation of microorganisms from the chorioamnion is twice that from amniotic fluid at cesarean delivery in women with intact membranes [abstract]. Am J Obstet Gynecol 1993;168:424)i2 along with the strong association between the presence of this organism in the chorioamnion and subsequent endometritis, may offer an explanation of the potential source of U urealyticum in post-cesarean delivery wound infections. An alternative role for U urealyticum in the pathogenesis of post-cesarean endometritis may be as a promoter for growth of gram-negative and grampositive organisms long known to be associated with this infection.’ However, it is important to note that the highest incidence of endometritis among women with positive chorioamnion cultures was observed when U urealyticum was isolated in pure culture. Regardless of the mechanism, it is important to recognize the association between this organism and post-cesarean endometritis. A greater understanding of this association may lead to improved modalities for prevention and treatment of this common complication of cesarean delivery. The association between U urealyticum colonization of the chorioamnion at cesarean with an increased incidence of post-cesarean endometritis occurred despite the routine use of intraoperative prophylactic antimicrobial therapy. The substantial benefit derived from the use of prophylactic therapy at cesarean delivery is well established and results in a 50-60% reduction in postoperative endometritis.26 Numerous antimicrobial agents have been reported to have similar efficacy in reducing the post-cesarean febrile morbidity rate.26 However, the antimicrobials commonly used do not
Andrews
et al
Ureaplasma
and Endomefvifis
513
have established efficacy against U urealyticum. Because in most the cervix is colonized with U urealyticum women’ and genital mycoplasmas are frequently isolated from the uterine cavity in women with ruptured membranes,12,27,28 the relationship between this organism and post-cesarean delivery endometritis might be even higher in women with ruptured membranes. Onequarter of the casesof endometritis in this population of women with intact membranes were associated with chorioamnion colonization with U urealyticum. Thus, expansion of current prophylactic antimicrobial regimens to include coverage for U urealyticum may further reduce the incidence of post-cesarean endometritis by 25% or more. The efficacy of this expanded prophylactic therapy should be investigated in a randomized controlled trial.
20.
References
21.
1. Gilstrap LC, Cunningham FG. The bacterial pathogenesis of infection following cesarean section. Obstet GynecoI 1979;53:545-9. 2. Eschenbach DA. Ureaplasma urealyticum as a cause of postpartum fever. Pediatr Infect Dis 1986;5:5258-61. 3. Lavery JP, Marcel1 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 RI’, 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. Epidemiology and outcomes. 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, Rankin JS, 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, et al. Isolation of genital mycoplasmas from blood of febrile obstetrical-gynecologic patients and neonates. Stand 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, Pine11 P. The microbiology of post-cesarean wound morbidity. Obstet Gynecol 1993;81:383-6. J, Krohn M, Kiviat N, Holms KK, Eschenbach 12. Hillier SL, Martius DA. A case-controlled study of chorioamniotic infection and histologic chorioamnionitis in prematurity. N Engl J Med 1988;319: 972-8. A, Duffy L, Crabb D, Waites KB. Mycoplas13. Cassell GH, Blanchard mas in clinical and pathogenic microbiology. In: Howard BJ, ed. Mycoplasmas. St. Louis: Mosby-Year Book, Inc., 1993:491-502. D, Chen TA. Methods in mycoplasmology. In: 14. Taylor-Robinson
514
Andrews
et al
Ureaplasma
and Endometrifis
15.
16.
17.
18.
19.
22. 23.
24.
25. 26. 27.
28.
Razin S, Tully JG, eds. Growth inhibitory factors in animal and plant tissues. Vol 1. New York: Academic Press, Inc., 1983:109-14. Cassell GH, Watson HL, Blalock DK, Horowitz SA, Duffy LB. Protein antigens of genital mycoplasmas. Rev Infect Dis 1988;lO: S391-8. Blanchard A, Dybvig YK, Watson HL, Griffiths G, Cassell GH. Evaluation of intraspecies genetic variation within the 16s rRNA gene of Mycoplasma hominis and detection by polymerase chain reaction. J Clin Microbial 1993;31:1358-61, Balows A, Hausler WJ, Herrmann KL, Isenberg HD, Shadomy HJ. Manual of clinical microbiology. 5th ed., Washington, DC: American Society for Microbiology, 1991. Sutter VL, Citron DM, Edelstein MAC, Finegold SM. Wadesworth anaerobic bacteriology manual. 4th ed. Belmont, California: Star Publishing Company, 1985. Holdeman LV, Cato El’, Moore WEC. Anaerobic laboratory manual. 4th ed. Blacksburg, Virginia: Virginia Polytechnic Institute and State University, 1977. Evans LC, Combs CA. Increased maternal morbidity after cesarean delivery before 28 weeks of gestation. Int J Gynecol Obstet 1993; 40:227-33. Seo K, McGregor JA, French JI. Preterm birth is associated with increased risk of maternal and neonatal infection. Obstet Gynecol 1992;79:75-80. Romer R, Mazor M. Infection and preterm labor. Clin Obstet Gynecol 1988;31:553-84. Kilian M, Brown MB, Brown TA, Cassell GH. Immunoglobulin A protease activity of Ureaplasma urealyticum. Acta Path01 Microbiol Stand [BJ 1984;92:61-4. Takebe S, Numata A, Kobashi K. Stone formation by Ureaplasma urealyticum in human urine and its prevention by urease inhibitors. J Clin Microbial 1984;20:869-73. Desilva NS, Quinn PA. Phospholipase A and C activity in Wrenplasma urealyticum. J Clin Microbial 1986;23:354-9. Duff I’. Antibiotic prophylaxis. In: Phelan JP, Clark SL, eds. Cesarean delivery. New York: Elsevier, 1988:283-97. Hillier SL, Krohn MA, Kiviat NB, Watts DH, Eschenbach DA. Microbiologic causes and neonatal outcomes associated with chorioamnion infection. Am J Obstet Gynecol 1991;165:955-61. Romero R, Yoon BH, Mazur M, et al. A comparative study of the diagnostic performance of amniotic fluid glucose, white blood cell count, interleukin-6, and Gram stain in the detection of microbial invasion in patients with preterm premature rupture of membranes. Am J Obstet Gynecol 1993;169:839-51.
Addressreprint requeststo: Gail K. Cassell,PhD University of Alabama at Birmingham Department of Microbiology University Station Birmingham, AL 35233-7333
Received August 29, 1994. Received in revised form November 14, 1994. Accepted November 29, 2994. Copyright 0 1995 by The American Gynecologists.
College
of Obstetricians
and
Obstetrics G Gynecology