Secondary Logo

Journal Logo


Presence and Concentrations of Select Bacterial Vaginosis-Associated Bacteria Are Associated With Increased Risk of Pelvic Inflammatory Disease

Haggerty, Catherine L. PhD, MPH; Ness, Roberta B. MD, MPH; Totten, Patricia A. PhD; Farooq, Fouzia MPH; Tang, Gong PhD§; Ko, Daisy BS; Hou, Xuezhou PhD; Fiedler, Tina L. BS; Srinivasan, Sujatha PhD; Astete, Sabina G. PhD; Fredricks, David N. MD

Author Information
Sexually Transmitted Diseases: May 2020 - Volume 47 - Issue 5 - p 344-346
doi: 10.1097/OLQ.0000000000001164
  • Free

Pelvic inflammatory disease (PID), infection and inflammation of the female upper genital tract, is a common condition among young women that often results in infertility, chronic pelvic pain, and recurrent PID.1 Although PID has a polymicrobial etiology, with Chlamydia trachomatis and/or Neisseria gonorrhoeae accounting for approximately one third to one half of the cases,1 up to 70% of PID cases have an unidentified etiology. Bacterial vaginosis (BV), determined by analysis of Gram-stained vaginal smears, and specific cultivable BV-associated species including anaerobic gram-negative rods have also been associated with PID.1

Bacterial vaginosis is a polymicrobial condition characterized by a shift from a lactobacilli predominant vaginal microbiota to one with high concentrations and diversity of facultative and anaerobic bacteria. The BV-associated microorganisms have been cultured from upper tract samples from women with PID.1 However, the associations between individual BV-associated bacteria and risk of PID have not been determined using highly sensitive polymerase chain reaction (PCR) methods. Cultivation-independent studies using 16S rRNA gene sequence PCR amplified from vaginal DNA have revealed previously unrecognized bacterial genera associated with BV.2 We sought to determine the associations among key BV-associated bacteria and PID incidence among a population of women at high risk for sexually transmitted infection.


We conducted an ancillary study of 34 women enrolled in the Gynecologic Infections Follow-Through (GIFT) study, which has been described in detail elsewhere.3 Briefly, 1199 women 13 to 36 years of age were recruited into the parent GIFT Study from family planning clinics, university health clinics, gynecology clinics, and STD units at five clinical sites in the United States between May 1999 and June 2001 and were followed approximately 3 years for the development of PID. Women were eligible for the GIFT study if they were not specifically seeking care for an STI but were considered at elevated risk for having chlamydial cervicitis, according to a modification of the Stergachis et al4 risk paradigm. Approximately two thirds of women were aged 19 to 24 years of age and 75% were black. As part of the parent study, participants collected vaginal specimens at baseline, 6, 12, 24, and 36 months after being educated by study staff on a standardized method for self-collection using the BD CultureSwab collection and transport system (Becton Dickinson, NJ). DNA amplification for C. trachomatis and N. gonorrhoeae was performed by using a strand displacement DNA amplification assay (Becton Dickinson, NJ) from self-obtained vaginal swabs and residual specimens were archived and frozen at −80°C. Women were educated on the signs and symptoms of PID and advised to contact study staff at any point during follow-up to report pelvic pain, abnormal bleeding or urethritis, or a diagnosis of chlamydial or gonococcal cervicitis. To detect PID, women who reported pelvic pain at any point in the study and women who tested positive on C. trachomatis and N. gonorrhoeae screening were scheduled for an additional, symptomatic visit involving a pelvic examination, and an endometrial biopsy.

For the current ancillary nested pilot study utilizing a case to control ratio of 1:1, we randomly selected 17 women who experienced histologically confirmed PID over follow-up and 17 controls selected randomly among all study participants who did not experience PID signs or symptoms, matched by follow-up visit and race. DNA was extracted from 200 μL of the archived vaginal material using the MoBio BiOstic Bacteremia DNA Isolation Kit (MoBio Laboratories, Carlsbad, CA). Mycoplasmal bacteria previously associated with PID in cross-sectional studies1 and key BV-associated bacteria5–8 were selected for analysis. Species-specific 16S ribosomal RNA (rRNA) gene quantitative PCR (qPCR) assays targeting Atopobium vaginae,5Gardnerella vaginalis,5Sneathia spp.,5 BV-associated bacterium 1 (BVAB1),5 BVAB2,5Mageeibacillus indolicus,5Megasphaera spp.,5Eggerthella-like bacterium,8Mobiluncus spp.,6Prevotella timonensis,8Prevotella amnii,8Ureaplasma urealyticum,7Ureaplasma parvum,7 and Mycoplasma genitalium9 were applied to DNA from vaginal samples from visits occurring immediately preceding and within 3 months of PID using methods which have been previously described.5–9 All these assays used a StepOne Plus thermal cycler, with 45 cycles of amplification, input DNA in 3 μL in a 30-μL reaction. In addition, an assay was developed targeting BVAB-TM7 for this study. For that assay, each 30-μL qPCR assay contained 1× Buffer A (Life Technologies), 3-mM magnesium chloride, 0.8-μM forward primer (TM7-992F; 5′TGACATCCCTAGAATTTCTCC-3′), 0.8 μM reverse primer (TM7-1051R; 5′-GGATCTGTCACCTAGTTCT-3′), 150 nM probe (TM7_1015; 5′-6FAM-AAGGAGAGAGTGCTTTTTA-MGBNFQ-3′), 0.05 units uracil-N-glycosylase (UNG), 1.0 unit AmpliTaq Gold DNA polymerase (Life Technologies) and 3-μL sample DNA. Assays were run on the StepOne Plus qPCR instrument (Life Technologies) using the following amplification conditions: 50°C for 2 minutes (UNG activation), 95°C for 10 minutes (premelt), 45 cycles of 95°C for 15 seconds (melt), 59°C for 39 seconds (annealing), 72°C for 30 seconds (extension). Relationships between the presence and concentrations of individual bacteria and PID were determined using conditional logistic regression models. As all cases and no controls tested positive for C. trachomatis, we were unable to include chlamydia as a covariate in this pilot study. Variables measuring 16S rRNA gene copies were log transformed (base 10) and negative samples were assigned a value equal to half the lower limit of detection for each respective bacterium. Analyses were performed using R version 3.6.0.


Cases were more likely to have BV determined by Gram stain as compared with controls, although the differences in proportions were not statistically significant (45.5% vs 29.4%, P = 0.119). Generally, the bacteria we assayed were frequently identified in the lower genital tract before the development of PID (see Table 1). Mycoplasma genitalium was the least prevalent and P. timonensis was the most prevalent, identified in 8% and 94%, respectively, of vaginal samples collected among cases before PID diagnosis. In unadjusted analyses, several BV-associated bacteria were significantly associated with subsequent PID (Table 1). Women who tested positive for A. vaginae (ORadj, 13.7; 95% confidence interval [CI], 2.7–108.5), Sneathia spp. (ORadj, 5.8; 95% CI, 1.4–27.7), Megasphaera spp. (ORadj, 6.0; 95% CI, 1.4–29.7]), Eggerthella-like bacterium (ORadj, 10.6; 95% CI, 2.4–59.0), and Prevotella amnii (ORadj, 4.6; 95% CI, 1.1–22.5) were significantly more likely to develop PID. There was a trend that women who tested positive for G. vaginalis, BVAB1, BVAB2, M. indolicus, BVAB-TM7, P. timonensis, and Mobiluncus spp. were also more likely to develop PID. Women who tested positive for M. genitalium, U. urealyticum, or U. parvum did not have an elevated risk of subsequent PID.

The Risk of PID for Women Testing Positive for Vaginal Bacteria Preceding and Within 3 Months of Diagnosis by Species-Specific Quantitative 16S Ribosomal RNA Gene Polymerase Chain Reaction Assays

Cases had significantly higher mean 16S rRNA gene copies/mL as compared with controls for the following bacteria: A. vaginae (2.6 E6 ± 8.1E6 vs. 1.9E6 ± 5.7E6, P = 0.01), Megasphaera spp. (9.8E5 ± 2.0E6 vs. 1.1E5 ± 3.2E5, P = 0.02), Eggerthella-like bacterium (1.6E6 ± 3.0E6 vs. 1.3E6 ± 1.6E6, P = 0.009), and P. timonensis (2.2E6 ± 4.4E6 vs. 4.0E5 ± 1.4E6, P = 0.02). Mean 16S rRNA gene copies/mL of Sneathia spp. (1.6E6 ± 4.6 E6 vs. 5.3E5 ± 1.7E6, P = 0.05) and BVAB2 (9.7E4 ± 2.2E5 vs. 5.2E4 ± 1.3E5, P = 0.07) were also higher among cases as compared with controls, although differences were of borderline statistical significance. Higher bacterial concentrations of G. vaginalis, BVAB1, M. indolicus, BVAB-TM7, U. urealyticum, U. parvum, M. genitalium, Mobiluncus spp., and P. amnii were not predictive of PID.


In our targeted study of women considered at high risk for sexually transmitted infection, women who tested positive for BV-associated bacteria including A. vaginae, Sneathia spp., Megasphaera spp., Eggerthella-like bacterium, and P. amnii in the vagina were at significant risk for developing subsequent PID within 3 months. Further, greater bacterial load of A. vaginae, Megasphaera spp., Eggerthella-like bacterium, P. timonensis, Sneathia spp., and BVAB2 was predictive of subsequent PID. All women with PID were coinfected with C. trachomatis, thus the listed bacteria could modulate PID risk among C. trachomatis infected women but we do not have evidence from this study that BV-bacteria alone are associated with PID risk.

To our knowledge, this is the first qPCR study to prospectively demonstrate that the presence and quantity of fastidious BV-associated bacteria are associated with subsequent PID in women with Chlamydial infection. Results are consistent with a prior GIFT study finding, where we demonstrated that a cluster of cultured vaginal BV-associated organisms (absence of hydrogen peroxide-producing lactobacilli, presence of G. vaginalis, Mycoplasma hominis, anaerobic Gram-negative rods, and undifferentiated ureaplasmas) was associated with a 2-fold risk of incident PID.10

Our study has a number of strengths, including quantitative measurements of specific vaginal bacterial concentrations and prospectively assessed histologically confirmed incident PID. The selection of vaginal specimens within a critical exposure window of 3 months preceding PID diagnosis supports a temporal association between the bacteria and PID. Our pilot study was limited by its sample size, which resulted in large confidence intervals and may have biased some of our models toward the null. As women without signs and symptoms of PID over follow-up did not receive pelvic examinations, it is possible that some women in the control group may have had subclinical PID, potentially biasing models toward the null. Although M. genitalium has been cross-sectionally associated with endometritis and clinically suspected PID in prior studies,1 it was infrequently detected among samples from women in our study. As no controls tested positive for M. genitalium, we were unable to model the relationship between this bacterium and PID. Additional well-powered studies are needed to confirm these preliminary findings, improve precision, examine a broader range of BV-associated bacteria, and allow for additional confounder adjustment and examination of bacterial interactions. All cases in our study tested positive for C. trachomatis, highlighting the need for future studies with repeated vaginal sampling to explore the potential for BV-associated bacteria to increase the risk of chlamydial infection and ascension to the upper genital tract. Our findings raise the question of whether screening and treatment of women harboring bacteria associated with high risk of PID has the potential to reduce the incidence of PID.


1. Haggerty CL, Ness RB. Epidemiology, pathogenesis and treatment of pelvic inflammatory disease. Expert Rev Anti Infect Ther 2006; 4:235–247.
2. Fredricks DN, Fiedler TL, Marrazzo JM. Molecular identification of bacteria associated with bacterial vaginosis. NEJM 2005; 353:1899–1911.
3. Ness RB, Hillier SL, Richter HE, et al. Douching in relation to bacterial vaginosis, lactobacilli, and facultative bacteria in the vagina. Obstet Gynecol 2002; 100:765.
4. Stergachis A, Scholes D, Heidrich FE, et al. Selective screening for Chlamydia trachomatis infection in a primary care population of women. Am J Epidemiol 1993; 138:143–153.
5. McClelland RS, Lingappa JR, Srinivasan S, et al. Evaluation of the association between the concentrations of key vaginal bacteria and the increased risk of HIV acquisition in African women from five cohorts: A nested case-control study. Lancet Infect Dis 2018; 18:554–564.
6. Fredricks DN, Fiedler TL, Thomas KK, et al. Targeted PCR for detection of vaginal bacteria associated with bacterial vaginosis. J Clin Microbiol 2007; 45:3270–3276.
7. Xiao L, Glass JI, Paralanov V, et al. Detection and characterization of human Ureaplasma species and serovars by real-time PCR. J Clin Microbiol 2010; 48:2715–2723.
8. Srinivasan S, Morgan MT, Fiedler TL, et al. Metabolic signatures of bacterial vaginosis. MBio 2015; 6.
9. Jensen JS, Bjornelius E, Dohn B, et al. Use of TazMan 5′ nuclease real-time PCR for quantitative detection of Mycoplasma genitalium DNA in males with and without urethritis who were attendees at a sexually transmitted disease clinic. J Clin Microbiol 2004; 42:683–692.
10. Ness RB, Kip KE, Hillier SL, et al. A cluster analysis of bacterial vaginosis-associated microflora and pelvic inflammatory disease. Am J Epidemiol 2005; 162:585–590.
Copyright © 2020 American Sexually Transmitted Diseases Association. All rights reserved.