The vaginal environment, including the microbiome and immune system, plays an important role in the defense against external pathogens. Vaginal health is associated with a lactobacilli-dominated bacterial microbiota with low levels of proinflammatory cytokines.1,2 On the other hand, a dysbiotic microbiota is characterized by a polybacterial overgrowth and increased proinflammatory cytokines; this condition when symptomatic is known clinically as bacterial vaginosis.1,3 Unbalanced microbiota conditions with increased bacterial vaginosis–related pathogens are implicated in neonatal infections, miscarriage and preterm birth,4 pelvic inflammatory disease,5 and increased susceptibility to sexually transmitted infections including human immunodeficiency virus.6
The protective lactobacilli-dominated vaginal microbiota, postulated to originate from the rectum, have been shown to appear during pubertal transition; they constitute a minority of the vaginal microbiota in premenarcheal girls.7–9 In young women sexual debut was associated with the vaginal colonization by bacterial vaginosis-associated bacteria.10,11 Rectal Lactobacillus crispatus and colonization of the vagina and anorectum by lactobacilli have been shown to be protective against bacterial vaginosis acquisition in adult women.12,13 Furthermore, the presence of a semen marker has been associated with a negative influence on the Lactobacillus species and with bacterial vaginosis in adult and young women.14–17 Despite the wealth of information on adults and young women, knowledge is missing for adolescent girls at the time of sexual debut.
We conducted a school-based study in adolescent girls in Antwerp, Belgium, with the purpose to study the association of sexual debut with the composition of the vaginal, anorectal, and oral microbiota and vaginal inflammatory markers.
MATERIALS AND METHODS
Schools in Antwerp that would be supportive of a study among adolescent girls were identified. Before the main study, formative focus group discussions with adolescent school girls were held to inform the studies' recruitment strategies and the study questionnaire (eg, use of appropriate terms for this age group) and assess the acceptability of self-collection of specimens. At each collaborating school an information session was held for grades 4 (usually 14- to 16-year-old pupils) and 5 (17–19 years). Those showing an interest received an information leaflet and consent form for parents to sign. Girls from the same class were allowed to participate independent of their age and sexual orientation. The study included three visits, at baseline and 2 and 4 months. The girls were seen individually by the study staff at a quiet location at the schools' premises. During the first visit, the parents' written informed consent forms were collected. Informed consent procedure included information on the study's objectives, an overview of the study visits, and types of data to be collected. The right to withdraw at any time was stressed. At each visit the following procedures were performed: general information on the participant's medical history related to reproductive and sexual health was collected by the study staff and more private sexual behavior questions were answered on a tablet computer. Contraceptive counseling and safe sex counseling were offered on an individual basis after questioning by the participant. Next, one oral swab, three vaginal swabs, and one anorectal swab were self-collected by the participant after detailed explanation. The protocol was approved by the institutional review boards of the Institute of Tropical Medicine and of the University Teaching Hospital in Antwerp, Belgium.
The collected swabs were stored and transported on dry ice to the HIV/STI Reference Laboratory, where they were frozen at −80°C until testing. At the school site, a vaginal smear slide was made and the vaginal pH (color-fixed indicator strips pH-Fix 3.6–6.1) was assessed. Vaginal Gram stains were scored using the Nugent method.18 The eluted vaginal swab suspension was centrifuged and 120 microliters were tested for the presence of the semen marker prostate-specific antigen. Test bands of any intensity were defined as positive. Quantitative polymerase chain reaction, with concentrations of organisms expressed as genome equivalents per milliliter, was performed in duplicate for L crispatus, Lactobacillus iners, Lactobacillus jensenii, Lactobacillus gasseri, Lactobacillus vaginalis, Lactobacillus genus, Gardnerella vaginalis, and Atopobium vaginae for the vaginal swab and for Lactobacillus genus, G vaginalis, and A vaginae for the oral and anorectal swab. The swab handling, polymerase chain reaction mixture, and primers for the bacterial species and the amplification reactions have been described before.16,19 Standard curves in duplicate were constructed using DNA extracts as described before except for G vaginalis (LMG 7832T for DNA source). Concentrations of interleukin (IL)-1α, IL-1β, IL-6, IL-8, macrophage inflammatory protein-1β, interferon γ-induced protein-10, and regulated on activation, normal T cell expressed and secreted were measured in duplicate using Millipore fluorescent bead-based multiplex kits on a Luminex platform. Below out-of-range values were assigned half the value of the lowest standard.
The sample size estimate of 110 girls was based on detecting a difference of 0.65 log10Lactobacillus genus concentration between 55 adolescents who had not passed sexual debut (standard deviation [SD] 0.8) and 55 adolescents who had passed sexual debut (SD 1.5) with 80% power. Data analysis was performed using Stata 11 and R 3.1.2 and the significance level was set at P<.05. The study population characteristics at the first visit were described in terms of numbers and proportions and means with SDs and characteristics were compared by sexual experience: 1) no sexual experience and semen marker negative; 2) vaginal touching but no penile–vaginal penetration and semen marker negative; and 3) penile–vaginal penetration or semen marker positive using Fisher exact tests and linear regression analysis. The sexual experience definition was identical for adolescents identifying as homosexual or bisexual. The microbiota results were expressed categorically as present or absent and concentrations of species and inflammatory markers were log10-transformed and medians were calculated. A comparison of the presence and concentrations (when present) of species and the inflammatory markers by sexual experience as described previously was performed with the no sexual experience group as the reference. Multivariable regression stepwise deletion analysis including the population characteristics showing an association with sexual experience at the first visit (age, education, smoking, age at first menses, contraceptive use) was performed for the species and markers that showed an association. We modeled the longitudinal species presence and inflammatory concentrations for associations with sexual debut (girls who remained without sexual intercourse during the follow-up, girls who initiated penile–vaginal penetration, and girls who had passed sexual debut at the first visit during the study) in two ways. First we studied whether an association existed over the three visits for the three groups by logistic or linear mixed-effects models with random intercept. This first model aimed to reflect the changes over time in all three groups. Second, we focused on a before–after comparison between the group that stayed with no sexual intercourse and the group that started sexual intercourse during the study and used two visits for each participant. The models are different insofar as the second model describes a recent start of sexual penile–vaginal penetration whereas the first model describes differences including girls who have passed sexual debut for a few months or years.
We next performed a mediation analysis to explore whether the vaginal species that showed a significant higher odds with the start of sexual debut (ie, G vaginalis and A vaginae) mediated the association of sexual debut with each of those vaginal inflammatory markers showing a positive association at the first visit (ie, IL-1α, IL-1β, IL-8).20 Mediation analysis seeks to explain the underlying mechanism of an association between an independent variable and a dependent variable (ie, the inflammatory marker). This analysis allowed for splitting the total association with sexual debut into two pathways: first the sexual debut (no sexual intercourse compared with sexual intercourse) association with the outcome by changes in the sum of the logs of G vaginalis and A vaginae (ie, indirect association) and second the direct association.
Between February 2012 and March 2013, 134 of the 187 eligible girls out of the four participating schools (of the 11 schools approached) showed an interest in participating in the study. Of those 134, 93 were enrolled in the study after providing written consent from their parents and giving consent themselves. Seven participants were lost to follow-up by visit 3 mostly because they had changed schools after the holiday period (Fig. 1). The participants were on average 16.2 year old (SD 1.3 years) (Table 1). Of the total group, 36 (38.7%) were following vocational education, 13 (14%) were enrolled in technical college, and just under 50% was attending general secondary education. Although data were not collected, no major differences in socioeconomic status are to be expected as a result of the democratic Belgian school system. The majority of girls (84%) originated in Belgium or Europe; 11 were born in an African country (ie, Sierra Leone, Mali, Guinea, Congo, Ghana, and Morocco) and the remaining four in other countries. Smoking was common; one in five currently smoked and one of four had ever smoked. Approximately half of the participants were using contraceptives; 38% were using a combined pill. The practice of vaginal washing (douching) was common: water was used by 47% and 11% applied lactic acid intravaginally. A bacterial vaginosis Nugent score of 7–10 was present in six girls (7%), and an intermediate score of 4–6 was detected in 12 (14%). The seminal marker was present in 13 (14.4%) of the adolescents. Of the 93 participants at the first visit, 53 (57%) said they never had a sexual relationship but 12 of the 53 (22.6%) currently had a nonsexual relationship with a boyfriend. A small number of the adolescents (n=5 [5.4%]) did not report penetrative sexual intercourse but had experience with kissing their partner, touching of the vaginal area by the partner (ie, finger play, cunnilingus, and use of a vaginal stimulator), and performing oral sex on their partner. The remaining 35 (37.6%) girls said they had passed their sexual debut by the time of the first visit and reported vaginal penetration with a mean age at sexual debut of 14.8 years. Other sexual behaviors reported by this group were: kissing their partner, touching of the vaginal area by the partner, anal finger play and anal sex, and performing oral sex on their partner. Finally, six of the participants declaring no sexual experience at the first visit had a positive semen marker result at one or more visits and were therefore considered past sexual debut for the remainder of the analysis. Over the follow-up period, 10 adolescents reported to have passed sexual debut, four adolescents by visit 2 and six by visit 3. At visit 3, 40 adolescents had not yet passed sexual debut (Fig. 1).
Girls who had started coital sex were older than girls without sexual experience (16.6 compared with 15.8 years; P=.005) (Table 1). Furthermore, adolescents enrolled in vocational education had more sexual experience compared with those following technical and general education (P=.001). Smoking and sexual experience was positively related (P=.003). Seventy-six percent of girls who had passed sexual debut used hormonal contraceptives with half of the remainder using either condoms or no contraception. Five girls not engaged in a sexual relationship used hormonal contraceptives for other reasons, of which treatment of acne was common. A bacterial vaginosis Nugent score of 7–10 was not present among girls without sexual experience; a bacterial vaginosis score of 7–10 was present in five girls (12.5%) engaged in coital sex and one girl (20%) in the noncoital sexual experience group (P=.014). An intermediate Nugent score of 4–6, however, was present for four of the girls without sexual experience.
At the first visit, the presence of vaginal species did not significantly differ by sexual experience (Table 2) and sexual experience was not significantly associated with species concentrations (data not shown). Over the period of the study, no association of sexual debut with the presence of vaginal Lactobacillus species was seen in either model (model 1 including all girls and model 2 including the girls who had started sexual intercourse compared with girls remaining without sexual intercourse) (Tables 3 and 4 ). Sexual debut was associated with increased odds for the presence of vaginal G vaginalis (P=.021) and A vaginae (P=.041) in model 1 (Table 3). For girls who initiated sexual intercourse over the 8-month period, the odds of vaginal G vaginalis presence was 6.3 times higher (confidence interval [CI] 0.9–45.9) and, for girls who had passed sexual debut, the odds were six times higher (confidence interval 1.6–22.5) compared with girls who remained without sexual intercourse. The odds for vaginal A vaginae presence were nine times higher (CI 0.3–258.5) for girls starting sexual intercourse and 21 times higher (CI 2.0–221.3) for girls who had passed sexual debut compared with girls remaining without sexual intercourse. When modeling only those girls initiating sexual intercourse compared with girls remaining without sexual intercourse, this association was not found (Table 4).
At the first visit, in the anorectal compartment, the presence and concentrations (data not shown) of Lactobacillus genus, G vaginalis, and A vaginae did not significantly differ by sexual experience (Table 2). Over the 8-month period, anorectal presence of G vaginalis (P=.030) and A vaginae (P=.012) was associated with sexual debut. The odds for G vaginalis were 11 times higher (CI 0.4–296.8) and for A vaginae seven times higher (CI 0.0–2,210.6) for girls starting sexual intercourse compared with girls who remained without sexual intercourse (Table 3). Girls who had passed sexual debut had an odds ratio (OR) of 25.8 (CI 2.3–287.7) for G vaginalis and an OR above 100 (CI 4.9–5,424.6) for A vaginae compared with girls who remained without sexual intercourse. When modeling only those girls initiating sexual intercourse compared with girls who remained without sexual intercourse, no associations were found (Table 4).
At the first visit, sexual experience did not significantly affect the presence and concentrations (data not shown) of oral Lactobacillus genus present in all adolescents (Table 2). Oral G vaginalis was not detected at visit 1 and was detected on one occasion only at visit 3 in an adolescent who had passed sexual debut. A vaginae was only present in two girls (2.2%). For these two girls, A vaginae was present in the oral cavity at all three visits. The first adolescent, 18 years old, performed coital sex, had previous female and male partners, had vaginal complaints and discharge, had a Nugent score of 7, had vaginal G vaginalis and A vaginae present (no anorectal sample available), and reported sexual behavior with a male partner the day before (ie, vaginal sex, finger play, and cunnilingus). The second adolescent, 19 years old, had genital touching experience with a Nugent score 4, was negative for any G vaginalis, and had no A vaginae anorectally or vaginally. She had last sexual contact more than 1 week previously when vaginal finger play was performed.
For the inflammatory markers at the first visit, an inflamed vaginal pattern with higher concentrations of the markers IL-1α (P<.001), IL-1β (P=.030), and IL-8 (P=.009) was noted for girls with sexual experience compared with girls without sexual experience (Table 2). Multivariable regression analysis reduced to the simple unadjusted model for IL-1α and IL-1β, whereas for IL-8, next to sexual experience (P=.002), only age remained significant (P=.041). The association of sexual debut at the first visit was mediated for more than 70% by the sum of log concentrations of G vaginalis and A vaginae: IL-1α (86.2%; total effect 4.4, direct effect 0.8, mediated indirect effect 3.5), IL-1 β (74.2%; total effect 2.3, direct effect −0.2, mediated indirect effect 2.1), and IL-8 (84.1%; total effect 4.9, direct effect 1.0, mediated indirect effect 3.9) (Fig. 2). Over the 8-month period, an increment in interleukin markers IL-1α (P<.001), IL-1β (P=.046), and IL-8 (P=.033), and the chemokines regulated on activation, normal T cell expressed and secreted (P<.001), and macrophage inflammatory protein-1β (P=.040), was estimated for girls who initiated sexual intercourse and for girls who had passed sexual debut compared with the girls remaining without sexual intercourse (Table 3). When modeling only those girls who commenced sexual intercourse compared with girls who remained without sexual intercourse, the increments were not significant (Table 4).
This school-based study in adolescent girls in Antwerp showed that over a period of 8 months sexual debut was associated with the composition of the vaginal and anorectal microbiota and vaginal inflammatory markers when modeling girls who remained without sexual intercourse, girls initiating sexual debut, and girls who had passed sexual debut at the first visit. Sexual debut was associated with increased odds for the bacterial vaginosis–related species G vaginalis and A vaginae to be present vaginally and anorectally. Furthermore, sexual debut was associated with raised inflammatory markers in the vagina with higher concentrations over the 8-month period. We could not demonstrate differences when comparing the evolution over time (first and last visit) for girls initiating sexual intercourse and girls staying without intercourse.
Our data show that around the time of initiating penetrative intercourse, colonization of the vagina and anorectum with G vaginalis and A vaginae takes place. Only a limited number of studies have looked at dysbiotic microbiota changes related to sexual debut and only a very small number of adolescents have been included in any study. One recent study in young (17–21 years) Australian women identified bacterial vaginosis by Nugent scoring in girls performing vaginal-penetrative and nonpenetrative sexual activities, whereas no dysbiosis was detected before sexual debut.10 The initiation of penile–vaginal sex, in another study in a low sexually transmitted infection setting enrolling 18- to 22-year-old virgins in the United States, was associated with G vaginalis colonization.11 In this study no association was seen with the quantitative polymerase chain reaction presence of Lactobacillus species L crispatus, L jensenii, and L iners. This observation is confirmed by our data: no immediate change in Lactobacillus species presence or concentrations is found around the time of sexual debut. We cannot conclude that with sexual debut a loss of Lactobacillus species presence or concentrations would occur. Nevertheless, several studies have shown that continued sexual activity was associated with bacterial vaginosis and negatively affected the Lactobacillus species.14–17 An explanation for this observation may be as follows: whereas the bacterial vaginosis–related species increase early on when initiating sexual activity, the Lactobacillus are resilient and only fail to recover or sustain themselves after longstanding sexual experience with high frequency of coitus or frequent partner change possibly including exposure to unprotected sex and to semen. This hypothesis would fit with our observation in African women in whom the semen marker was negatively associated with the presence of L crispatus, L vaginalis, and with Lactobacillus genus concentrations.16
Sexual debut was associated with an inflammatory vaginal status in this study with significant increments for the markers IL-1α, IL-1β, and IL-8 in the cross-sectional and longitudinal data including all girls. We elucidated that the presence of higher concentrations of G vaginalis or A vaginae was causing most of the increments of the aforementioned markers. Part of the association of sexual debut and the levels of interleukin markers was explained by the higher concentrations of the two bacterial vaginosis–related species. In other words, sexual debut possibly provokes an inflammatory reaction, first directly, and second through an increment in the presence and concentrations of G vaginalis or A vaginae. We conclude with the hypothesis that part of the inflammatory reaction with sexual debut is the result of a reaction against bacterial vaginosis-associated bacteria, but moreover interleukin marker increments may reflect mechanical insults on the vaginal mucosal barrier.
This study has several limitations. We did not reach the total number of planned participants, and therefore the study was underpowered to detect a difference in Lactobacillus genus concentration and sexual experience at the first visit. A larger sample size may have improved the power of the second model. Second, multiple testing was performed for associations with species and markers and this could have led to false-positive results. Interpretation should therefore be done with caution and ideally the study should be repeated. We acknowledge that we only tested for a limited number of species known to be related to bacterial vaginosis. Nevertheless, the selection of G vaginalis and A vaginae was based on the literature showing their importance in a dysbiotic microbiota.21
In this study of adolescent girls in Antwerp schools, sexual debut including nonpenetrative sexual activities was linked to the presence of vaginal and anorectal bacterial vaginosis-associated bacteria and this was accompanied by an inflammatory vaginal reaction. As a result, adolescent women around this time of life are possibly more prone to acquiring sexually transmitted infections. Strategies to prevent the colonization of bacterial vaginosis–related organisms during early sexual debut are urgently needed and will help to sustain vaginal health of adolescents into adulthood.
1. Kyongo JK, Crucitti T, Menten J, Hardy L, Cools P, Michiels J, et al. Cross-sectional analysis of selected genital tract immunological markers and molecular vaginal microbiota in Sub-Saharan African women, with relevance to HIV risk and prevention. Clin Vaccine Immunol 2015;22:526–38.
2. van de Wijgert JH, Borgdorff H, Verhelst R, Crucitti T, Francis S, Verstraelen H, et al. The vaginal microbiota: what have we learned after a decade of molecular characterization? PLoS One 2014;9:e105998.
3. Mitchell C, Marrazzo J. Bacterial vaginosis and the cervicovaginal immune response. Am J Reprod Immunol 2014;71:555–63.
4. Li J, McCormick J, Bocking A, Reid G. Importance of vaginal microbes in reproductive health. Reprod Sci 2012;19:235–42.
5. Taylor BD, Darville T, Haggerty CL. Does bacterial vaginosis cause pelvic inflammatory disease? Sex Transm Dis 2013;40:117–22.
6. Buve A, Jespers V, Crucitti T, Fichorova RN. The vaginal microbiota and susceptibility to HIV. AIDS 2014;28:2333–44.
7. Gerstner GJ, Grünberger W, Boschitsch E, Rotter M. Vaginal organisms in prepubertal children with and without vulvovaginitis. A vaginoscopic study. Arch Gynecol 1982;231:247–52.
8. Hickey RJ, Zhou X, Settles ML, Erb J, Malone K, Hansmann MA, et al. Vaginal microbiota of adolescent girls prior to the onset of menarche resemble those of reproductive-age women. MBio 2015;6:e00097–15. doi: 10.1128/mBio.00097-15.
9. Myhre AK, Bevanger LS, Berntzen K, Bratlid D. Anogenital bacteriology in non-abused preschool children: a descriptive study of the aerobic genital flora and the isolation of anogenital Gardnerella vaginalis. Acta Paediatr 2002;91:885–91.
10. Fethers K, Twin J, Fairley CK, Fowkes FJ, Garland SM, Fehler G, et al. Bacterial vaginosis (BV) candidate bacteria: associations with BV and behavioural practices in sexually-experienced and inexperienced women. PLoS One 2012;7:e30633.
11. Mitchell CM, Fredricks DN, Winer RL, Koutsky L. Effect of sexual debut on vaginal microbiota in a cohort of young women. Obstet Gynecol 2012;120:1306–13.
12. Antonio MA, Rabe LK, Hillier SL. Colonization of the rectum by Lactobacillus species and decreased risk of bacterial vaginosis. J Infect Dis 2005;192:394–8.
13. Marrazzo JM, Fiedler TL, Srinivasan S, Thomas KK, Liu C, Ko D, et al. Extravaginal reservoirs of vaginal bacteria as risk factors for incident bacterial vaginosis. J Infect Dis 2012;205:1580–8.
14. Fethers KA, Fairley CK, Hocking JS, Gurrin LC, Bradshaw CS. Sexual risk factors and bacterial vaginosis: a systematic review and meta-analysis. Clin Infect Dis 2008;47:1426–35.
15. Fethers KA, Fairley CK, Morton A, Hocking JS, Fehler G, Kennedy LJ, et al. Low incidence of bacterial vaginosis in cohort of young Australian women. Sex Transm Dis 2011;38:124–6.
16. Jespers V, van de Wijgert J, Cools P, Verhelst R, Verstraelen H, Delany-Moretlwe S, et al. The significance of Lactobacillus crispatus and L. vaginalis for vaginal health and the negative effect of recent sex: a cross-sectional descriptive study across groups of African women. BMC Infect Dis 2015;15:115.
17. Verstraelen H, Verhelst R, Vaneechoutte M, Temmerman M. The epidemiology of bacterial vaginosis in relation to sexual behaviour. BMC Infect Dis 2010;10:81.
18. 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.
19. Jespers V, Menten J, Smet H, Poradosu S, Abdellati S, Verhelst R, et al. Quantification of bacterial species of the vaginal microbiome in different groups of women, using nucleic acid amplification tests. BMC Microbiol 2012;12: 83.
20. Hardnett FP, Pals SL, Borkowf CB, Parsons J, Gomez C, O'Leary A. Assessing mediation in HIV intervention studies. Public Health Rep 2009;124:288–94.
© 2016 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
21. Ravel J, Gajer P, Abdo Z, Schneider GM, Koenig SS, McCulle SL, et al. Vaginal microbiome of reproductive-age women. Proc Natl Acad Sci USA 2011;108(suppl 1):4680–7.