Pelvic inflammatory disease (PID) is believed to be a polymicrobial infection affecting over 1 million women annually in the United States.1 Both sexually transmitted and endogenous vaginal microorganisms are commonly isolated from the upper genital tract in women with acute PID, suggesting that PID most likely results from ascension of microorganisms from the lower genital tract (vagina and cervix) to the endometrium and fallopian tubes. Women with clinically apparent PID are at risk for fallopian tube damage and subsequent adverse reproductive sequelae. Among over 1300 women attempting to conceive following a laparoscopic-confirmed episode of acute PID, Westrom et al documented that 16.0% of women with acute PID were subsequently infertile, compared with a 2.7% infertility rate in a control group of women without PID.2 Similarly, women diagnosed with PID are at risk for ectopic pregnancy, as 9.1% of first pregnancies following an episode of acute PID were ectopic pregnancies, compared to a rate of 1.4% in women without PID.2
While most women with tubal factor infertility have no prior history of symptomatic PID, several retrospective studies have demonstrated strong associations between tubal infertility and previous chlamydial and gonococcal infections among women without a prior history of acute symptomatic PID.3–7 These data suggest that damage to the fallopian tubes may be due to an inflammatory process that is clinically inapparent, a condition termed subclinical, unrecognized, or silent PID. Although there are substantial historical data supporting the hypothesis that subclinical infections of the endometrium and fallopian tubes can lead to tubal infertility, little is known about the relationship between lower genital tract infections and subclinical upper genital tract infections. Preliminary work suggests that women with lower genital tract infections are at risk for upper genital tract inflammation in the absence of classic signs or symptoms of acute PID.8,9 Our study was performed to comprehensively evaluate the relationship between lower genital tract infections and subclinical PID.
MATERIALS AND METHODS
In this cross-sectional study, women between the ages of 15 and 30 with lower genital tract infections or who were determined to be at risk for these infections were approached for enrollment between 1998 and 2000. This study was approved by Magee-Womens Hospital's Institutional Review Board. Women were recruited from the following ambulatory care sites: Allegheny County Health Department's Sexually Transmitted Diseases Clinic, Magee-Womens Hospital Ambulatory Care Clinics, and clinics of the Family Health Council, a provider of reproductive health care in western Pennsylvania. Women were eligible for participation if they had clinical evidence of mucopurulent cervicitis (purulent cervical discharge), if they were recently diagnosed with gonorrhea or chlamydia and were not yet treated, if they were diagnosed with bacterial vaginosis, or if they reported sexual contact with a male diagnosed with gonorrhea, chlamydia, or nongonococcal urethritis. The clinical diagnosis of bacterial vaginosis required the presence of three of the four following findings: homogeneous vaginal discharge, vaginal pH > 4.5, positive amine test with addition of potassium hydroxide to vaginal fluid, and clue cells on microscopy of vaginal fluid.10 We excluded symptomatic women meeting the minimal diagnostic criteria for PID by the Centers for Disease Control and Prevention (CDC; lower abdominal pain, cervical motion tenderness, and adnexal tenderness).11 Symptomatic women not meeting these minimal criteria were included in the study. Exclusion criteria also included pregnancy, antibiotic use within the preceding 14 days, gynecologic surgery within the previous month, use of an intrauterine contraceptive device, and allergy to any of the antibiotics used in this trial. As this report is part of a longitudinal study of adverse reproductive outcomes, we also excluded women with a possibility of pre-existing tubal damage, such as those women with a history of PID or tuboovarian abscess, ectopic pregnancy, tubal ligation or other adnexal surgery, endometriosis, and ruptured appendicitis.
After providing informed consent, women underwent a standardized interview and physical examination. Particular attention was paid to symptoms and signs of vaginal, cervical, and upper genital tract infections. Vaginal fluid from the lateral vaginal sidewall was collected for pH measurement, whiff test for the presence of amines, microscopy (normal saline and potassium hydroxide), and Gram stain for BV scoring.12 Vaginal fluid was cultured for Trichomonas vaginalis. Endocervical samples were obtained for culture for Neisseria gonorrhoeae and polymerase chain reaction (PCR) testing for Chlamydia trachomatis.
All women then underwent endometrial sampling for histologic analysis. After cleansing the endocervix and exocervix with betadine, a sterile endometrial sampler (Unimar Pipelle de Cornier, CooperSurgical, Shelton, CT) was placed into the endometrial cavity and a sample of endometrial tissue was aspirated into the cannula. Endometrial tissue was placed into 10% formalin fixative for histologic processing. Regardless of the reason for enrollment, all participants then underwent uniform treatment with single-dose agents for gonorrhea (cefixime 400 mg), chlamydia (azithromycin 1 g), and bacterial vaginosis (metronidazole 2 g).11
A Gram stain of vaginal fluid was prepared and interpreted following previously published criteria, with a score of 0–3, 4–6, and ≥ 7 representing normal vaginal flora, intermediate flora, and bacterial vaginosis, respectively.12 For data analysis, we used Gram stain of vaginal fluid for the diagnosis of bacterial vaginosis in lieu of clinical criteria, given its greater reliability for the diagnosis of BV.13 Vaginal specimens for T vaginalis culture were placed into modified Diamond media, incubated at 37C, and examined every second day, to a maximum of 1 week, for the presence of motile trichomonads. Endocervical swabs were streaked onto modified Thayer-Martin and chocolate media for the recovery of N gonorrhoeae, with identification confirmed using Gram stain examination, oxidase test, and Gonochek-II (EY Laboratories, Inc., San Mateo, CA). Chlamydia trachomatis was identified by PCR amplification of endocervical samples following the manufacturers' recommendations (Roche Diagnostics, Branchburg, NJ)
Endometrial tissue was embedded in paraffin, sectioned, and stained with hematoxylin and eosin and with methylgreen pyronine. Methylgreen pyronine was used to aid in the identification of plasma cells. A single pathologist experienced in prior studies of upper genital tract infection interpreted all specimens, and was blinded to the clinical presentation of the participants. Endometrial specimens were graded based on previously established and validated criteria for endometrial inflammation, and the inflammatory cells (neutrophils and plasma cells) were quantified.14 Histologic endometritis with neutrophilic infiltrate, indicating the presence of subclinical PID, was defined as the presence of at least five neutrophils in the superficial endometrial epithelium per 400× field and at least one plasma cell per 120× field of endometrial tissue.14 Plasma cell endometritis was defined as the presence of at least one plasma cell per 120× field of endometrial tissue.
Analysis included Yates continuity corrected χ2 statistics for 2 × 2 categorized variables, Fisher exact test for cross-tabulations for 2 × 2 tables with expected cell frequencies less than five, and Mann-Whitney U tests or exact statistics for ranked nonparametric variables. When variables such as categorized bacterial vaginosis score were evaluated for hypothesis testing, the appropriate test was a χ2 test for trend because bacterial vaginosis has the characteristics of an ordered increasing variable. The risk of subclinical PID associated with lower genital tract infection was estimated using polychotomous logistic regression models with three dependent variable categories, after controlling for confounding factors. The potentially confounding variables were identified by listing the characteristics of women that were associated with endometritis. Factors that fit these criteria were tested in a polychotomous logistic regression and retained in the model if the 95% confidence interval derived from the Wald test did not include one and the P value from the likelihood ratio test was less than .05.15,16
Among the 556 women enrolled in this study, 57 women (10%) were diagnosed with cervical N gonorrhoeae infections, 103 women (18.5%) had cervical C trachomatis infections, and 377 (68%) were diagnosed with bacterial vaginosis by Gram stain. The median age of our cohort was 22 years. As women enrolled in our study were known to have bacterial vaginosis, chlamydia, or gonorrhea, or suspected of the latter two infections, most participants reported lower genital tract symptoms. Abnormal vaginal discharge or vaginal odor was reported by 294 (53%) and 265 (48%) women, respectively. Subtle symptoms of upper genital tract infection were mostly absent among women in our cohort. Only 48 women (9%) had pelvic discomfort at enrollment, and 145 (26%) reported recent abnormal menstrual pattern, either inter-menstrual bleeding or last menses heavier or longer than normal. On physical examination, cervical motion tenderness was present in 22 women (4%), while uterine tenderness was noted in 31 participants (6%) and adnexal tenderness in 32 (6%). Women meeting the CDC's minimal criteria for the diagnosis of PID were excluded from participation.11
Subclinical PID, defined histologically as the presence of at least 5 neutrophils per 400× field and at least one plasma cell per 120× field of endometrial tissue, was present in 74 women (13%). One hundred and twenty-six women (23%) had evidence of plasma cell endometritis, defined as the presence of at least one plasma cell per 120× field of endometrial tissue. Subclinical PID and plasma cell endometritis were more common among black women, women who were in the proliferative phase of the menstrual cycle (within 14 days of the onset of the last menses), those who were previously pregnant, and in women who douched (Table 1). Subclinical PID, but not plasma cell endometritis, was more common among women with a prior history of chlamydial infection, while plasma cell endometritis was more common among women with a prior history of gonorrhea. After controlling for confounding factors, oral contraceptive use remained associated with plasma cell endometritis (odds ratio 2.2, 95% confidence interval 1.2, 3.7, P < .01). There was no relationship between oral contraceptive use and subclinical PID.
Subclinical PID was more commonly found in women with abnormal menstrual pattern (intermenstrual bleeding or heavier or prolonged last menses) and in women with pelvic pain (Table 2). Similarly, although infrequent, cervical motion tenderness, uterine tenderness, and adnexal tenderness were all more commonly present among women with subclinical PID than among women without endometritis. Plasma cell endometritis was not associated with vaginal symptoms, pelvic pain, abnormal menstrual pattern, or pelvic tenderness.
The relationship between lower genital tract infection and endometritis, by univariate analysis, is shown in Table 3. Subclinical PID was present in 15/57 (26%) women with cervical N gonorrhoeae infections compared with 57/496 (11%) women without gonorrhea (P < .01). Similarly, cervical C trachomatis infections were associated with higher rates of subclinical PID (27% in women infected with chlamydia versus 11% in chlamydia-negative women, P < .01). Endometrial plasma cells were associated with cervical C trachomatis but not N gonorrhoeae. Subclinical PID was present in 9/23 (39%) women coinfected with N gonorrhoeae and C trachomatis, while 39/413 (9%) women with neither of these infections had subclinical PID (P < .001). However, subclinical PID was not more common in women coinfected with these two STDs than among those infected solely with one of these organisms; the rate of subclinical PID among women infected with N gonorrhoeae (and not C trachomatis) was 18% (P = .13) while the rate among women solely infected with C trachomatis was 23% (P = .18). Among the 377 women with bacterial vaginosis, 58 (15%) had evidence of subclinical PID while plasma cell endometritis was present in 90 (24%). Subclinical PID, but not plasma cell endometritis, was more common in women with bacterial vaginosis than in women without bacterial vaginosis. Women with intermediate vaginal flora on Gram stain were at intermediate risk for subclinical PID (P < .05, test for trend), while the same was not found when studying vaginal flora and plasma cell endometritis.
Logistic regression analysis was then performed to further evaluate the relationship between lower genital tract infections and endometritis. The following factors were identified and evaluated as potentially confounding variables: proliferative phase of the menstrual cycle, previous pregnancy, black race, and being positive for N gonorrhoeae, C trachomatis, bacterial vaginosis, or T vaginalis. Oral contraceptive use was placed in the model for plasma cell endometritis given the strong association by univariate analysis. After adjustment for these potentially confounding factors, chlamydia, gonorrhea, and bacterial vaginosis remained significantly associated with subclinical PID, with adjusted odds ratios of 3.4, 2.4, and 2.7, respectively (Table 4). Following logistic regression T vaginalis was not significantly associated with subclinical PID. Gonorrhea, chlamydia, and bacterial vaginosis were not associated with plasma cell endometritis, although a trend was noted with chlamydia infection (odds ratio 1.7, 95% confidence interval 0.96, 3.1). Plasma cell endometritis was more commonly found among black women and women in the first half of the menstrual cycle.
Findings from this study indicate that in the absence of signs or symptoms of acute PID, one in four women with gonorrhea or chlamydia, and one in seven women with bacterial vaginosis have subclinical PID. Subclinical PID is believed to cause similar long-term reproductive sequelae as acute PID. Tubal factor infertility, ectopic pregnancy, and chronic pelvic pain are common sequelae of PID, yet most women with these conditions have no history of acute PID.17 Evidence supporting the important role of subclinical PID in the development of fallopian tube damage has evolved from several retrospective studies of women with tubal factor infertility documenting serologic evidence of prior chlamydial infections in 23–91% of women and prior gonococcal infections in over 60%, the majority of whom had no prior known history of acute PID.3–7 Further, subclinical upper genital tract inflammation causes fallopian tube damage akin to that seen in acute PID. Using light and electron microscopy, Patton et al demonstrated similar morphologic changes and alterations in ciliary function in the mucosa of fallopian tubes in women with acute and silent PID.18 Results of our current study may explain retrospective studies linking tubal factor infertility with prior chlamydial or gonococcal infections in women without a history of acute PID. With estimates of over 3 million cases of chlamydia and nearly 1 million cases of gonorrhea occurring annually in the United States, and that bacterial vaginosis is the most common vaginal infection in reproductive-aged women, our study indicates that a substantial number of women are at risk for fallopian tube damage and subsequent adverse reproductive sequelae.19
Pelvic inflammatory disease is often an infection of polymicrobial etiology, involving N gonorrhoeae, C trachomatis, anaerobic organisms, and/or facultative organisms.20,21 While gonorrhea and chlamydia have long been associated with acute PID, bacterial vaginosis has emerged as another risk factor for upper tract infection.22,23 Among women without PID, ascension of lower genital tract microorganisms to the endometrium occurs frequently; however, the organisms are likely cleared by the immune system. In women with bacterial vaginosis, the microbial load and the pathogenicity of the organisms are greater, which may overcome host defense mechanisms. Preliminary data have suggested that, similar to acute PID, bacterial vaginosis is associated with subclinical upper genital tract inflammation. Korn et al documented an increased risk of plasma cell endometritis among women with bacterial vaginosis than among women without bacterial vaginosis (10/22 versus 1/19).9 In our study of 556 women without acute PID, the independent associations between gonorrhea, chlamydia, bacterial vaginosis, and subclinical PID are demonstrated, supporting the link between these lower genital tract infections and subclinical PID.
In addition to an increased risk of subclinical PID among women with bacterial vaginosis, we also identified a greater frequency of subclinical PID among women with intermediate vaginal flora than among women with lactobacillus-predominant vaginal flora. These findings indicate that an environment with reduced lactobacilli may represent a potentially pathogenic condition, and not merely an intermediate step between normal flora and bacterial vaginosis. The hypothesis that abnormal vaginal flora is a risk factor for ascending pelvic infection is supported by the findings of a Scandinavian study of over 1000 women undergoing first trimester surgical abortion, indicating that abnormal flora is associated with postabortal endometritis.24 The data of our study indicate that, similar to acute PID, abnormal vaginal flora is important in the pathogenesis of subclinical PID.
While women meeting the criteria of acute PID were excluded, we found that women with subclinical PID were more likely to report symptoms of pelvic infection, particularly pelvic pain and abnormal menstrual pattern (intermenstrual bleeding or prolonged or heavier menses). Similarly, pelvic tenderness was more commonly detected in women with subclinical PID than in women without endometritis. Women with plasma cells alone on endometrial biopsy, however, were no more likely to have symptoms than were the women without endometrial plasma cells. These findings provide further evidence that subclinical PID, as determined by the presence of both neutrophils and plasma cells in the endometrium, represents an inflammatory process of the upper genital tract with similar characteristics as acute PID. In contrast, plasma cell endometritis was not associated with any of the signs or symptoms of upper genital tract infection.
Upper genital tract infection, either acute PID or subclinical PID, is difficult to diagnose with certainty. The gold standard diagnostic test for acute PID is direct visualization of the pelvic organs by laparoscopy or laparotomy, techniques that are invasive, costly, and not uniformly available. There is also concern that visualization of the pelvic organs may not detect an inflammatory response restricted to the mucosa of the uterus or fallopian tube. Endometrial biopsy to identify histologic endometritis has been evaluated as a less invasive diagnostic tool for acute PID.14,25 Kiviat et al comprehensively studied 67 women with upper genital tract infection and laparoscopic-confirmed salpingitis.14 Using criteria of the presence of five or more neutrophils per 400× field and one or more plasma cells per 120× field of endometrial tissue, the sensitivity and specificity of histologic endometritis for the diagnosis of upper genital tract infection and salpingitis were 92% and 87%, respectively. This validated definition of histologic endometritis was used in the present study as indicative of subclinical PID. Plasma cell endometritis, defined as the presence of at least one plasma cell per low power field of endometrial tissue, has been used as a marker of acute PID by other researchers, but has a lower specificity.14,25 Nevertheless, plasma cells alone have been used as a marker of subclinical PID in two small studies, with plasma cell endometritis identified in 14/35 women with cervicitis and 10/22 women with bacterial vaginosis.8,9 However, the low specificity of plasma cell endometritis as a marker of true upper genital tract infection is a concern when relying on the presence of plasma cells alone to diagnose subclinical PID. Stern et al analyzed the endometria of women with plasma cell endometritis for the presence of C trachomatis infection using PCR assay for the major outer membrane protein (MOMP) gene.26 The authors identified C trachomatis in only 1/38 cases, concluding that there is little evidence linking mild or moderate plasma cell endometritis and chlamydial infection. We did not identify an association between plasma cell endometritis and either gonorrhea, chlamydia, or bacterial vaginosis, the common lower genital tract infections involved in the pathogenesis of acute PID. The association between oral contraceptive use and plasma cell endometritis is noted with interest but not readily explained, requiring the further investigation of the role of exogenous hormones in endometrial inflammation. The aforementioned findings, coupled with the absence of associations between plasma cell endometritis and subtle signs or symptoms of upper genital tract infection, suggest that the combined presence of neutrophils and plasma cells, rather than plasma cells alone, correlates with histologic evidence of true upper genital tract infection.
We observed a trend towards associations between subclinical PID and both black race and proliferative phase of the menstrual cycle (within 14 days of the onset of the last menstrual period), consistent with findings identified among women with acute PID by Hillier et al.23 Menses has been associated with the development gonococcal or chlamydial PID, possibly due to compromise of cervical barriers or other host defenses during the menstrual flow allowing for ascension of organisms to the upper genital tract.27 Douching has been reported to be a risk factor for PID.28,29 We found that douching within the preceding month was related to endometritis, supporting the findings of Cates et al associating douching with atypical PID.17 The role of douching in facilitating PID has not been characterized. Plausible mechanisms include alteration of the microbiologic or immunologic environment of the vagina or cervix, flushing of lower genital tract organisms into the upper genital tract, and/or a disruption of other host defense mechanisms, any of which may facilitate ascension of microorganisms into the upper genital tract. As subclinical PID is believed to be an inflammatory process with similar pathogenesis as acute PID, the association of douching with subclinical PID is not surprising. While risky sexual behavior (multiple sexual partners, frequent intercourse, infrequent condom use) has been previously reported to be associated with PID, we did not find such a relationship in our cohort.30 This discrepancy may be due to lack of enrollment of low-risk women needed to identify these associations.
According to a search of the literature via MEDLINE (1966 to February 2002; all languages; search terms: “pelvic inflammatory disease” and “Chlamydia trachomatis,” “Neisseria gonorrhoeae,” or “bacterial vaginosis”; “silent PID”; “sublinical PID”; “atypical PID”; “asymptomatic PID”; and “endometritis”), this is the largest study to date elucidating the relationship between lower genital tract infection and subclinical upper genital tract inflammation. Over 25% of women with either gonorrhea or chlamydia and 15% of women with bacterial vaginosis have subclinical PID. Our data are fully consistent with the substantial retrospective data linking tubal factor infertility (among women without a history of PID) with prior chlamydial and gonococcal infections. We hypothesize that, if left untreated, N gonorrhoeae, C trachomatis, or the organisms associated with bacterial vaginosis may invade the upper genital tract and cause subclinical PID. These infections may then spontaneously resolve, progress to acute symptomatic PID, or lead to the development of asymptomatic, clinically inapparent fallopian tube damage. Prospective research is needed to document that subclinical PID causes infertility with or without the intermediate stage of acute PID. Further, studies would be needed to determine whether women with lower genital tract infections should be treated with longer duration antibiotic regimens, as administered for acute PID, to reduce the long-term reproductive risks associated with upper genital tract infection.
1. Rein DB, Kassler WJ, Irwin KL, Rabiee L. Direct medical cost of pelvic inflammatory disease and its sequelae: Decreasing but still substantial. Obstet Gynecol 2000;95:397–402.
2. Westrom L, Joesoef R, Reynolds G, Hagdu A, Thompson SE. Pelvic inflammatory disease and fertility. A cohort study of 1844 women with laparoscopically verified disease and 657 control women with normal laparoscopic results. Sex Transm Dis 1992;19:185–92.
3. Punnonen R, Terho P, Nikkanen V, Meurman O. Chlamydial serology in infertile women by immunofluorescence. Fertil Steril 1979;31:656–9.
4. Jones RB, Ardery BR, Hui SL, Cleary RE. Correlation between serum antichlamydial antibodies and tubal factor as a cause of infertility. Fertil Steril 1982;38:553–8.
5. Tjiam KH, Zeilmaker GH, Alberda AT, van Heijst BY, de Roo JC, Polak-Vogelzang AA, et al. Prevalence of antibodies to Chlamydia trachomatis, Neisseria gonorrhoeae
, and Mycoplasma hominis
in infertile women. Genitourin Med 1985;61:175–8.
6. Sellors JW, Mahony JB, Chernesky MA, Rath DJ. Tubal factor infertility: An association with prior chlamydial and asymptomatic salpingitis. Fertil Steril 1988;49:451–7.
7. World Health Organization Task Force on the Prevention and Management of Infertility. Tubal infertility: Serologic relationship to post chlamydial and gonococcal infection. Sex Transm Dis 1995;22:71–7.
8. Paavonen J, Kiviat N, Brunham RC, Stevens CE, Kuo CC, Stamm WE, et al. Prevalence and manifestations of endometritis among women with cervicitis. Am J Obstet Gynecol 1985;152:280–6.
9. Korn AP, Bolan G, Padian N, Ohm-Smith M, Schachter J, Landers DV. Plasma cell endometritis in women with symptomatic bacterial vaginosis. Obstet Gynecol 1995;85:387–90.
10. Amsel R, Totten PA, Spiegel CA, Chen KCS, Eschenbach DA, Holmes KK. Nonspecific vaginitis: Diagnostic criteria and microbial epidemiological associations. Am J Med 1983;74:14–22.
11. Centers for Disease Control and Prevention. 1998 Guidelines for treatment of sexually transmitted diseases. MMWR Recomm Rep 1998;47:1–111.
12. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of Gram stain interpretation. J Clin Microbiol 1991;29:297–301.
13. Schwebke JR, Hillier SL, Sobel JD, McGregor JA, Sweet RL. Validity of the vaginal gram stain for the diagnosis of bacterial vaginosis. Obstet Gynecol 1996;88:573–6.
14. 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.
15. Hosmer DW, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons; 1989.
16. Altman DG. Practical statistics for medical research. London: CRC Press; 1992.
17. Cates W Jr, Joesoef MR, Goldman MB. Atypical pelvic inflammatory disease: Can we identify clinical predictors? Am J Obstet Gynecol 1993;169:341–6.
18. Patton DL, Moore DE, Spadoni LR, Soules MR, Halbert SA, Wang SP. A comparison of the fallopian tube's response to overt and silent salpingitis. Obstet Gynecol 1989;73:622–30.
19. American Social Health Association. Sexually transmitted diseases in America: How many cases and at what cost? Menlo Park, CA: Kaiser Family Foundation, 1998.
20. Paavonen J, Teisala K, Heinonen PK, Aine R, Laine S, Lehtinen M, et al. Microbiological and histopathological findings in acute pelvic inflammatory disease. Br J Obstet Gynaecol 1987;94:454–60.
21. Jossens MO, Schachter J, Sweet RL. Risk factors associated with pelvic inflammatory disease of differing microbial etiologies. Obstet Gynecol 1994;83:989–97.
22. Soper DE, Brockwell NJ, Dalton HP, Johnson D. Observations concerning the microbial etiology of acute salpingitis. Am J Obstet Gynecol 1994;170:1008–14.
23. Hillier SL, Kiviat NB, Hawes SE, Hasselquist MB, Hanssen PW, Eschenbach DA, et al. Role of bacterial vaginosis-associated microorganisms in endometritis. Am J Obstet Gynecol 1996;175:435–41.
24. Larsson P, Platz-Christensen J, Dalaker K, Eriksson K, Fahraeus L, Irminger K, et al. Treatment with 2% clindamycin vaginal cream prior to first trimester surgical abortion to reduce signs of postoperative infection: A prospective, double-blinded, placebo-controlled, multicenter study. Acta Obstet Gynecol Scand 2000;79:390–96.
25. Paavonen J, Aine R, Teisala K, Heinonen PK, Punnonen R. Comparison of endometrial biopsy and peritoneal fluid cytologic testing with laparoscopy in the diagnosis of acute pelvic inflammatory disease. Am J Obstet Gynecol 1985; 151:645–50.
26. Stern RA, Svoboda-Newman SM, Frank TS. Analysis of chronic endometritis for Chlamydia trachomatis
by polymerase chain reaction. Hum Pathol 1996;27:1085–8.
27. Sweet RL, Blankfort-Doyle M, Robbie MO, Schacter J. The occurrence of chlamydial and gonococcal salpingitis during the menstrual cycle. JAMA 1986;255:2062–4.
28. Wolner-Hanssen P, Eschenbach DA, Paavonen J, Stevens CE, Kiviat NB, Critchlow C, et al. Association between vaginal douching and acute pelvic inflammatory disease. JAMA 1990;263:1936–41.
29. Ness RB, Soper DE, Holley RL, Peipert J, Randall H, Sweet, RL, et al. Douching and endometritis: Results from the PID evaluation and clinical health (PEACH) study. Sex Transm Dis 2001;28:240–5.
30. Lee NC, Rubin GL, Grimes DA. Measures of sexual behavior and the risk of pelvic inflammatory disease. Obstet Gynecol 1991;77:425–30.