Objectives: To determine the effects of vaginal, cervical, and endometrial infections on shedding of HIV-1 RNA in the female genital tract.
Methods: Antiretroviral-naive women from Nairobi, Kenya with CD4 cell counts ≥ 350 cells/μl had plasma and endocervical wick samples collected for HIV quantification by real-time RNA reverse transcriptase-polymerase chain reaction. Vaginal and cervical Gram stains and endometrial biopsies were obtained. Vaginal Gram stain was used to diagnose bacterial vaginosis and to quantify Lactobacillus levels.
Results: Twenty-six of 50 (52%) women had detectable endocervical HIV-1 RNA with a median endocervical viral load of 1760 copies/ml (range: undetectable to 1 030 000 copies/ml). Women with decreased Lactobacillus had 15.8-fold [95% confidenceinterval (CI), 2.0–123] greater endocervical HIV-1 RNA than women with normal Lactobacillus levels. Women with plasma cell (PC) endometritis [≥ 1 PC/high-power field (hpf)] had a 15.8-fold (95% CI, 2.0–120) higher endocervical HIV RNA level than women without PC endometritis. Both these associations remained after controlling for plasma viral load. Cervicitis (≥ 30 polymorphonuclear leukocytes/hpf), however, was not associated with endocervical HIV-1 RNA shedding (P = 0.81).
Conclusions: In HIV-1-infected, antiretroviral-naive women without symptoms of pelvic inflammatory disease infection, abnormal vaginal flora and inflammatory cells in the endometrium affected HIV-1 shedding from the lower genital tract. These data suggest that both the upper and lower genital tracts contribute to female HIV-1 genital shedding.
From the aDepartment of Obstetrics, Gynecology, and Reproductive Sciences, University of California, San Francisco, USA
bDepartment of Obstetrics and Gynecology, USA
cDepartment of Laboratory Medicine, University of Washington, Seattle, Washington, USA
dCenter for Respiratory Disease Research, Kenya
eMicrobiology Research, Kenya Medical Research Institute, Nairobi, Kenya
fDepartment of Statistics, University of Nairobi, Kenya.
*Current affiliation: Department of Obstetrics and Gynecology, University of Pennsylvania, Philadelphia, Pennsylvania, USA.
Received 28 February, 2006
Revised 9 October, 2006
Accepted 30 October, 2006
Correspondence to Craig R. Cohen MD, MPH, UCSF, 50 Beale Street, Suite 1200, San Francisco, CA 94105, USA. E-mail: firstname.lastname@example.org
Up to 90% of HIV-1-infected women shed infectious virus in genital tract secretions either continuously or intermittently [1,2]. The concentration of HIV-1 in female genital secretions is determined by multiple factors, including plasma viral load, antiretroviral therapy, and hormonal effects . In addition, recent evidence suggests that lower reproductive tract infection and inflammation may be important determinants of genital HIV-1 shedding [4,5].
Upper genital tract infections, such as pelvic inflammatory disease (PID) and endometritis, describe a continuum of infection from the lower to upper genital tract. Eckert et al. demonstrated a 38% prevalence of asymptomatic endometritis in HIV-1-infected women . A similar prevalence (49%) of histologically confirmed endometritis has also been shown in symptomatic HIV-1-infected women . Both of these studies outlined risk factors for endometritis as well as the clinical severity and course, but did not assess the effect of upper tract infections on shedding of HIV-1. The purpose of this investigation was to determine whether upper genital tract infections, specifically endometritis, are related to HIV-1 genital shedding.
One hundred and sixty-five women infected with HIV-1 were screened for participation in this study in Nairobi, Kenya. Eligibility criteria included: reproductive age (18–50 years old); antiretroviral treatment naive; CD4 lymphocyte counts ≥ 350 cells/μl; intact uterus and cervix; not pregnant; and willing to adhere to the study schedule. The women were enrolled between March and October 2004 at the Kenya Medical Research Institute. All gave written, informed consent in English or Kiswahili. This study was reviewed and approved by the institutional review boards of the University of Washington, the University of California San Francisco and the Kenya Medical Research Institute.
The women were screened within 30 days before enrollment. Screening included a targeted interview and blood collection. Whole blood was processed for complete blood count, CD4 T-cell count, rapid plasma reagin syphilis screening, and confirmation of HIV-1 serostatus. Enrollment visits were scheduled during the second week of the menstrual cycle and included collection of demographic and medical information, a gynecologic examination, urine pregnancy test and plasma viral load (HIV-1 RNA) measurements.
Women were instructed to refrain from sexual intercourse, tampon use or douching for 48 h preceding each visit. Samples were collected in the following order: four vaginal swabs were taken from the posterior fornix for Gram stain, saline and KOH wet mount, pH, Trichomonas vaginalis culture using InPouch (Biomed Diagnostics, White City, Oregon, USA); two endocervical swabs for detection of Chlamydia trachomatis and Neisseria gonorrhoeae and cervicitis; endocervical fluid collection using Sno-strip wicks (Chauvin Pharmaceuticals Ltd., Romford, Essex, England); papanicolaou smear; and endometrial biopsy for histology and anaerobic and aerobic bacterial culture using a Pipelle endometrial suction curette (Unimar, Wilton, Connecticut, USA). Three-quarters of the endometrial biopsy specimen was expelled into a sterile container with 10% buffered formalin and the remainder was placed in a sterile culture tube.
Lower genital tract infections were diagnosed according to the following criteria. Candida colonization was diagnosed by Gram stain; bacterial vaginosis and Lactobacillus levels were diagnosed according to Nugent's criteria ; trichomoniasis was evaluated by wet mount and InPouch; chlamydia and gonorrhea were assessed by polymerase chain reaction (PCR) (AMPLICOR; Roche Diagnostics, Inc., Branchburg, New Jersey, USA); and cervicitis by ≥ 30 polymorphonuclear leukocytes (PMNs) per high-power field (hpf) on Gram stain .
Genital tract HIV-1 RNA was measured as previously described  using endocervical Sno-strips, which were vortexed and frozen at −70°C.
HIV-1 RNA was measured using a validated real-time reverse transcriptase (RT)-PCR amplification assay . The lower limit of detection of HIV-1-RNA in endocervical canal fluid was 600 HIV-1 RNA copies/ml. The CD4 T-lymphocyte enumeration was performed using standard flow cytometric methods .
Histology was assessed from formalin-fixed, paraffin-embedded sections stained with methyl green pyronine (MGP). Slides were coded and read by one observer blinded to the clinical and laboratory status of the women. Five non-adjacent areas of endometrial tissue were examined at 400× to count lymphocytes, polymorphonuclear cells and plasma cells. Endometrial biopsies were characterized as normal, plasma cell endometritis (PCE) (≥ 1 plasma cell/hpf) or subclinical PID (≥ 1 plasma cell/hpf plus ≥ 5 PMNs/hpf) .
Fisher exact tests were used for categorical variables and Student's t-test for continuous variables. Correlations were computed using Pearson product-moment correlation. Plasma and Sno-strip endocervical HIV-1 RNA levels were log-transformed to stabilize their variance and are presented here in the original scale as the multiplicative difference. Univariate and multiple linear regression were used to estimate the effect of genital tract infections on endocervical RNA measurements. Multiple logistic regression was used to test for associations between the covariates and HIV-1 genital shedding levels after adjustment for plasma HIV-1 RNA levels. All regression analyses used jackknifed robust standard error estimates . Analyses were carried out using SPSS 13.0 statistical software (SPSS, Inc., Chicago, Illinois, USA) and Stata 7.0 (Stata Corp., College Station, Texas, USA).
Sixty-eight of the 165 (41%) HIV-infected women screened met study eligibility and fifty-one (31%) were enrolled in the cohort. One woman was subsequently excluded secondary to an enlarged uterus leaving 50 evaluable women. In comparison with the 17 women who were eligible but not enrolled, the 50 women enrolled had similar mean age (32.5 years), and mean CD4 T-cell count (601 versus 555 cells/μl; P = 0.42).
Plasma HIV-1 RNA was detected in 42 (84%) women with a median of 5695 copies/ml (range, undetectable to 588 000 copies/ml). Twenty-six (52%) women had detectable endocervical HIV-1 RNA with a median of 1760 copies/ml (range, undetectable to 1030 copies/ml). All women had CD4 cell counts ≥ 350 cells/μl. Detectable plasma viral load did not correlate to endocervical HIV-1 shedding (r = 0.30; P = 0.038).
There were no statistically significant sociodemographic differences between the women who tested positive for endocervical HIV-1 RNA shedding and those who tested negative (Table 1). Candida colonization was detected in nine women, one woman had C. trachomatis and 17 (35%) had cervicitis. PCE was detected in 28 (62%) womens and none of the endometrial histologic results met the criteria for subclinical PID. Furthermore, the endometrial culture results were similar among those with and without PCE: Gardnerella vaginalis (10 versus 11%), Enterococcus spp. (14 versus 33%), Corynebacterium spp. (25 versus 22%), Micrococcus spp. (18 versus 33%), Streptococcus viridans (25 versus 22%) and Bifidobacterium group E (25 versus 28%; all P > 0.1).
In the univariate analysis of lower genital tract infections, endocervical HIV-1 RNA shedding was 5.2-fold [95% confidenceinterval (CI), 0.64–43; P = 0.12] higher for women with abnormal vaginal flora (Nugent's score 4–10) than for women with normal flora. Women with diminished Lactobacillus (< 3+ on Gram stain) had 15.8-fold (95% CI, 2.0–123-fold; P = 0.010) greater endocervical HIV-1 RNA than women with normal Lactobacillus levels. These associations remained after controlling for plasma HIV-1 viral load (Fig. 1). Cervicitis and Candida colonization were not associated with endocervical HIV-1 RNA shedding (P = 0.81 and P = 0.34, respectively).
Next, we considered the influence of upper genital tract markers of inflammation and infection. Women with PCE had a 15.8-fold (95% CI, 2.0–120-fold; P = 0.01) higher endocervical HIV-1 RNA level than women without endometritis in univariate analysis. In multivariate analysis, the association remained significant after controlling for plasma HIV-1 viral load (Fig. 1). None of the bacteria detected in endometrial specimens was associated with endocervical HIV-1 RNA shedding (data not shown).
This study is the first to demonstrate a correlation between upper genital tract inflammation and increased HIV-1 RNA genital tract shedding. The presence of plasma cells in the endometrium is recognized as a major diagnostic criterion of endometrial inflammation . Prior investigations have demonstrated that the endometrium contains HIV-1-infectable cells that have the capacity to support viral replication [14,15]. A high prevalence of PCE (62%) was found in asymptomatic HIV-1-infected women. As a comparison, PCE was detected in 49% of HIV-1-infected and 30% of HIV-1-uninfected women with clinical PID in Nairobi . However, this study did not investigate the effect of upper genital tract inflammation on female genital tract HIV-1 RNA load.
The precise cause of the endometritis remains unclear. It is possible that HIV-1 itself leads to an increased amount of plasma cells within the endometrium. Alternatively, HIV-1-infected women may be more susceptible to infection caused by other organisms, such as bacterial vaginosis-associated bacteria or Mycoplasma genitalium, due to altered immunity, higher risk sexual behavior or both . We attempted to address this question by performing endometrial cultures. Both aerobes and anaerobes were isolated from the endometrium; however, it was difficult to distinguish between an endometrial pathogen versus contamination.
There are several potential implications of increased shedding of HIV-1 RNA among women with asymptomatic plasma cell endometritis. First, higher levels of genital shedding in women with endometritis may lead to increased female-to-male sexual transmission of HIV-1. Second, given the high prevalence of PCE in HIV-1-infected women, there may be a role for clinicians to routinely screen their HIV-1-infected women for upper tract disease. Future studies need to investigate the causes of PCE in HIV-1-infected women, especially those potentially amenable to intervention such as bacterial and some viral infections such as herpes simplex virus type-2. Such strategies could lead to reduced sexual transmission of HIV-1.
Our study confirms, in an African population, prior investigations that described the effects of abnormal vaginal flora and loss of Lactobacillus on increased concentrations of genital HIV RNA shedding [4,5,17]. Although hydrogen peroxide-producing Lactobacillus was not quantified, diminished levels based on Gram stain further confirms the inverse relationship with genital HIV-1 load as described by Sha and colleagues . We were somewhat, but not entirely, surprised not to find an association between non-specific cervicitis and HIV-1 shedding. In one study, effective treatment of gonococcal and chlamydial cervicitis resulted in significant decreases in shedding of HIV-1; however, treatment of non-specific cervicitis did not alter HIV-1 RNA shedding in cervical secretions .
The present study has several limitations. First, the population had few genital tract infections. The absence of subclinical endometritis in the study population and the low prevalence of other sexually transmitted infections, such as chlamydia, gonorrhea and trichomoniasis mean that it was not possible to determine the effect of these conditions on HIV-1 RNA genital shedding per se. Furthermore, the detection of plasma cells plus PMNs has been reported to be more specific for the diagnosis of endometritis . We could not apply this diagnosis since no participant in the current study fulfilled the criteria. Second, causal relationships could not be determined because of the cross-sectional design. As such, we do not know if treatment for the PCE would affect genital HIV-1 load. Nevertheless, the finding that plasma cells found in the endometrium were independently associated with increased genital tract HIV-1 RNA shedding suggests the need to further explore the function of the upper genital tract in HIV-1 replication and shedding.
The authors would like to thank the Director KEMRI Dr. D. Koech, the research team including Jesca Sande, Naomi Mwachari, Lucy Sanguli, Lawrence Thiongo, James Mwangi, Michael Kahara, Martin Matu, and the study participants.
Sponsorship: This study was funded by the National Institutes of Health Women's HIV Interdisciplinary Network (WHIN) at the University of Washington (HD40540). J.C. was a fellow supported by the National Institute of Child Health and Human Development grant (5T32HD40671).
1. Kovacs A, Wasserman SS, Burns D, Wright DJ, Cohn J, Landay A, et al
. Determinants of HIV-1 shedding in the genital tract of women. Lancet 2001; 358:1593–1601.
2. Coombs RW, Wright DJ, Reichelderfer PS, Burns DN, Cohn J, Cu-Uvin S, et al
. Variation of human immunodeficiency virus type 1 viral RNA levels in the female genital tract: implications for applying measurements to individual women. J Infect Dis 2001; 184:1187–1191.
3. Reichelderfer PS, Coombs RW, Wright DJ, Cohn J, Burns DN, Cu-Uvin S, et al
. Effect of menstrual cycle on human immunodeficiency virus type 1 levels in the peripheral blood and genital tract. AIDS 2000; 14:2101–2107.
4. Hashemi FB, Ghassemi M, Faro S, Aroutcheva A, Spear GT. Induction of human immunodeficiency virus type 1 expression by anaerobes associated with bacterial vaginosis. J Infect Dis 2000; 181:574–580.
5. Hashemi FB, Ghassemi M, Roebuck KA, Spear GT. Activation of human immunodeficiency virus type 1 expression by Gardnerella vaginalis
. J Infect Dis 1999; 179:924–930.
6. Eckert LO, Watts DH, Thwin SS, Kiviat N, Agnew KJ, Eschenbach DA. Histologic endometritis in asymptomatic human Immunodeficiency virus-infected women: characterization and effect of antimicrobial therapy. Obstet Gynecol 2003; 102:962–969.
7. Bukusi EA, Cohen CR, Stevens CE, Sinei S, Reilly M, Grieco V, et al
. Effects of human immunodeficiency virus 1 infection on microbial origins of pelvic inflammatory disease and on efficacy of ambulatory oral therapy. Am J Obstet Gynecol 1999; 181:1374–1381.
8. 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.
9. Holmes KK, Stamm WE. Lower genital tract infection syndromes in women
. In: Holmes KK, editor. Sexually Transmitted Disease
, 3d edn. New York: McGraw-Hill; 1999. pp. 769–778.
10. Li CC, Seidel KD, Coombs RW, Frenkel LM. Detection and quantification of human immunodeficiency virus type 1 p24 antigen in dried whole blood and plasma on filter paper stored under various conditions. J Clin Microbiol 2005; 43:3901–3905.
11. Calvelli T, Denny TN, Paxton H, Gelman R, Kagan J. Guidelines for flow cytometric immunophenotyping: a report from the National Institute of Allergy and Infectious Diseases, Division of AIDS. Cytometry 1993; 14:702–715.
12. Long JS, Ervin LH. Using heteroskedastic consistent standard errors in the linear regression model. Am Statistician 2000; 54:217–224.
13. Kiviat NB, Wolner-Hanssen P, Eschenbach DA, Wasserheit JN, Raavonen JA, Bell TA, et al
. Endometrial histopathology in patients with culture-proven upper genital tract infection and laparoscopically diagnosed acute salpingitis. Am J Surg Pathol 1990; 14:167–175.
14. Asin SN, Fanger MW, Wildt-Perinic D, Ware PL, Wira CR, Howell AL. Transmission of HIV-1 by primary human uterine epithelial cells and stromal fibroblasts
. J Infect Dis
:236–245 [Epub 2004 Jun 11].
15. Yeaman GR, Howell AL, Weldon S, Demian DJ, Collins JE, O'Connel DM, et al
. Human immunodeficiency virus receptor and coreceptor expression on human uterine epithelial cells: regulation of expression during the menstrual cycle and implications for human immunodeficiency virus infection. Immunology 2003; 109:137–146.
16. Cohen CR, Mugo NR, Astete SG, Odondo R, Manhart LE, Kiehlbauch JA, et al
. Detection of Mycoplasma genitalium
in women with laparoscopically diagnosed acute salpingitis. Sex Trans Infect 2005; 81:463–466.
17. Cu-Uvin S, Hogan JW, Caliendo AM, Harwell J, Mayer KH, Carpenter CC, HIV Epidemiology Research Study. Association between bacterial vaginosis and expression of human immunodeficiency virus type 1 RNA in the female genital tract. Clin Infect Dis 2001; 33:894–896.
18. Sha BE, Zariffard MR, Wang QJ, Chen HY, Bremer J, Cohen MH, Spear GT. Female genital-tract HIV load correlates inversely with Lactobacillus
species but positively with bacterial vaginosis and Mycoplasma hominis
. J Infect Dis 2005; 191:25–32.
19. McClelland RS, Wang CC, Mandaliya K, Overbaugh J, Reiner MT, Panteleeff DD, et al
. Treatment of cervicitis is associated with decreased cervical shedding of HIV-1. AIDS 2001; 15:105–110.