Group B Streptococcus (GBS) causes morbidity and mortality among newborns and young infants, as well as among pregnant women and nonpregnant adults. Infected mothers have about a 50% chance of infecting their newborns during labor and delivery. 1 Such transmission can be minimized by intrapartum chemoprophylaxis. 2 GBS disease among newborns has a case-fatality rate of 4% to 6%. 3
Among adults, transmission is hypothesized to be by the fecal-oral route or by person-to-person direct contact. 4 The elderly or those with underlying medical problems are susceptible to a range of conditions, from urinary tract infections (UTI) to meningitis, caused by GBS. 5,6 Young, healthy adults are usually asymptomatic. The prevalence of GBS colonization has been estimated at 10% to 40%2,7,8; little is known about duration of GBS colonization.
Although ever engaging in sexual activity increases the risk of colonization by GBS, 9 individuals who have never engaged in sexual activity are often colonized. In an early report, four out of 20 women entering college who reported no previous genital contact were GBS colonized. 9 Nonetheless, evidence from several studies suggests that sexual activity may be an important mode of transmission for GBS. 4 For example, 18 husbands of 40 GBS-positive women were colonized with GBS of the same serotype. 10 Similar serotype matches were found in 80 father-mother pairs of colonized and uncolonized babies. 11 Although these results are suggestive, serotype does not assess the identity of GBS strains. Thus, colonization with the same serotype might have occurred independently rather than via sexual transmission in the pairs studied.
A better understanding of GBS transmission is needed to prevent GBS carriage in women and thus to reduce GBS disease in newborns. Our focus is on GBS carriage rather than GBS disease because carriers are essential for maintaining transmission; colonization is the first step toward GBS disease. We test the hypothesis that intimate contact is associated with GBS transmission by determining co-colonization rates with the identical GBS strain (as assessed by pulsed-field gel electrophoresis, or PFGE) among heterosexual couples in which one or both partners carried GBS.
These data are from a study of the heterosexual transmission of UTI conducted between September 1996 and April 1999 at the University of Michigan Student Health Service. We recruited women with clinically diagnosed, culture-confirmed UTIs and their most recent sex partners, and women without UTI and their most recent sex partners. Women without UTIs either had a sore throat (25% of those enrolled) or responded to a study advertisement received while filling a prescription. Enrollment was restricted to nonpregnant women between 18 and 39 years of age who were not diabetic, who had engaged in sexual activity within the past 2 weeks, and who were willing to recruit their most recent male sex partner. The women could not have had a gynecologic exam the day of enrollment. Each male partner had to enroll within 7 days after the female visit in order for the couple to be eligible. Only participants enrolled during the second 2 years of the study were screened for GBS and included in the present analysis.
After giving informed consent, all participants independently completed a self-administered questionnaire. Men collected an initial void urine specimen, and women collected a clean-catch, mid-stream urine specimen, as instructed by the study personnel. All participants self-collected rectal specimens by gently swabbing the anal orifice, and women collected a vaginal specimen using a tampon. Each couple received payment for participating in the study.
The study protocol was approved by the Institutional Review Board at the University of Michigan.
The tampons, anal swabs, and 1 μl of urine were inoculated onto trypticase soy agar with 5% sheep blood (BBL; Becton Dickinson, Sparks, MD) and incubated overnight at 37°C with CO2. Plates were inspected for colony morphology and hemolysis typical of GBS and then confirmed serologically using the Slidex Strepto B kit (BioMerieux Vitek; Hazlewood, MO).
Pulsed-Field Gel Electrophoresis
PFGE was performed on all GBS isolates as described previously. 12 ImageQuant (Molecular Dynamics, Amersham Biosciences, Uppsala, Sweden) was used for the gel analyses. PFGE banding patterns within a couple or person were considered identical if the strains differed by no more than one band.
Serotyping was performed on unique GBS isolates within a couple or person as determined by PFGE; isolates with identical PFGE patterns have the same serotypes. 13 Serotyping using hyperimmune rabbit antisera to GBS polysaccharide types Ia, Ib, and II-VIII and the c protein antigen was performed as described previously. 12
Some strains initially identified as nontypable were identical by PFGE to a different serotype within a person or couple. We considered these to have the same serotype (with less capsule production) as the other strain. 14 For example, one female had two genetically identical strains in her anal and vaginal samples; the anal isolate was serotype Ia/c, whereas the vaginal isolate was nontypable/c. Three of 28 nontypable GBS isolates were changed to a typable category using these criteria.
This analysis was limited to participating couples in which at least one partner carried GBS in at least one site. We tested associations with GBS co-colonization using t-tests for continuous variables and chi-square tests for categorical variables. To assess the impact of confounding, we stratified the associations by the variables that yielded the strongest crude associations and calculated adjusted odds ratios (ORs). Finally, we fit a logistic regression model to adjust for factors associated with co-colonization, and estimated ORs and 95% confidence intervals (CIs). 15 Variables were chosen for inclusion based on the stratified analysis and biological plausibility. Because partners did not always give identical reports of partnership characteristics (variables within the partnership that should be identical, eg, frequency of vaginal intercourse), results are given for the male and female responses separately. Using McNemar's test, we determined whether the frequency of co-colonization given that the male had GBS was the same as the frequency given that the female had GBS.
Two hundred and eighty-five women with UTI participated in the study. Sixty-two percent (177/285) of these women recruited their most recent sex partner. One hundred and eighty-two women without UTI participated, with 67% (122/182) recruiting their sex partner.
Data analysis was restricted to the 120 couples in which at least one partner was colonized with GBS at one or more sites, resulting in 77 UTI couples and 43 non-UTI couples. Among the women without a UTI, 13 were enrolled after a throat culture and 30 responded to the study advertisement.
Most female participants were 20 to 22 years of age, whereas males were somewhat older (Table 1). The majority of participants self-identified as white, and 59% of women and 4% of men self-reported a previous history of UTI. Several reported previous history of sexually transmitted diseases (gonorrhea [1% of women, 0% of men], chlamydia [4%, 3%], herpes [7%, 3%], genital warts or human papilloma virus [7%, 1%], trichomoniasis [1%, 1%], and HIV [1%, 0%]). History of bacterial vaginosis or Gardnerella was reported by 4% of women, and 45% and 9% of women and men, respectively, reported history of a yeast infection or “jock itch.” The mean lifetime number of sex partners was 4.5 for women and 6 for men. The mean age at first sex was 17 for both men and women. Seventeen per cent of men and 12% of women reported having a concurrent sex partner during their most recent partnership.
Most couples had engaged in vaginal sex more than once within the past 2 weeks, and 79% of women had performed oral sex, as had 63% of men (Table 1). The median length of partnership was 9 months for women (mean, 1.4 years; range, 0 to 9.6 years). Within the same partnership, men were more likely than their sex partners to report condom use and less likely to report receptive oral sex.
Among the 120 couples in which at least one partner was colonized, the most common colonization site was the anal orifice (54% of men, 71% of women). Overall, 14% of men and 29% of women had GBS in their urine, whereas 52% of females had GBS in their vaginal sample. Only 2 women with UTI and 5 male partners of women with UTI had GBS solely in their urine. Colonization at more than one site occurred in 43% of subjects.
Pulsed-Field Gel Electrophoresis
Among the 57 couples in which both partners carried GBS, 49 (86%) were colonized with the identical strain, whereas 8 couples carried genetically different strains (Figure 1). Twenty-seven out of 34 couples (79%) in which the woman had a UTI shared the identical GBS strain, as did 22 out of 23 couples (96%) in which the woman did not have a UTI.
Among the 76 men colonized with GBS, 49 of their partners (64%) were colonized with the identical strain. Among the 101 colonized women, only 49 of their partners (49%) were colonized with the same strain. In both UTI and non-UTI couples, the men were more likely to share identical GBS with their sex partners (27/45 UTI, 22/31 non-UTI) than the women (27/66 UTI, 22/35 non-UTI); this difference was stronger among couples in which the women had a UTI (paired analysis OR = 2.2; CI = 1.2–3.8) than among couples in which the women did not have a UTI (paired analysis OR = 1.4; CI = 0.6–3.4).
Individual, Sociodemographic, and Behavioral Characteristics
We did not observe any strong associations between individual characteristics and the likelihood of both sex partners carrying the identical GBS strain (Table 1).
Those couples engaging in male-to-female oral sex were more than twice as likely to be co-colonized with GBS (for women, OR = 2.9 [CI = 1.2–6.9]; for men, OR = 2.5 [CI = 1.1–5.6]), but female-to-male oral sex had a weaker association. Crude associations are reported separately for men and women in Table 1 because there were differences in responses by gender.
Logistic Regression Analysis
Because of our small sample size, we fit a series of logistic regression models predicting co-colonization (Table 2). Each model included male-to-female oral sex in addition to the predictor of interest. Models were fit for each variable that was biologically plausible or for which the crude and stratified analyses had odds ratios of interest. Separate models were fit for men and women. Results were consistent by gender; where there were major differences in point estimates, the confidence intervals are wide. There is a suggestion of the role of immunity in both genders; increasing numbers of sex partners were associated with decreased co-colonization, as was earlier age at first sexual activity.
Among the co-colonized couples, no particular serotype predominated. The numbers are too small to make inferences regarding differential transmission rates by serotype (Table 3).
GBS has been characterized as a sexually transmitted disease for almost 30 years, but direct evidence of sexual transmission has been lacking. Understanding GBS transmission in asymptomatic, healthy individuals is needed to prevent GBS colonization in women and the subsequent transmission to newborns. Among heterosexual college couples in which at least one partner carried GBS, we found high co-colonization rates with the identical strain of GBS as determined by PFGE. The primary risk factor for co-colonization was male-to-female oral sex.
We know of no other studies that have assessed the effects of oral sex on co-colonization with GBS.
We found no association between vaginal intercourse frequency and co-colonization, despite evidence that sexual activity is important for GBS colonization. 9 We asked participants whether they had engaged in anal-oral or anal-penile sex, but this was reported too infrequently to assess an effect. Because all our participants engaged in vaginal-penile intercourse, we cannot determine whether this activity confers an increased risk of co-colonization over intimate, nonsexual contact that might occur during cohabitation.
Oral sex most likely contributes to the transmission of GBS by the fecal-oral route. This may be less likely with female-to-male than with male-to-female oral sex. The male urethral opening probably contains a low density of GBS even in men colonized at another site (eg, lower gastrointestinal tract). Consequently, women are less likely to acquire GBS from men during female-to-male oral sex. Because GBS readily colonizes the female mucosal tissues (ie, urethra, vagina, and rectum), transmission to men via male-to-female oral sex is more likely. The higher transmission with male-to-female oral sex might also occur if men were more likely to be colonized with GBS in their throat. GBS can be isolated from throat swabs, but limited data are available to confirm this organism as a cause of pharyngitis. 16
Men were more likely to report condom use and less likely to report receptive oral sex, consistent with previous reports. 17 If the differential reporting by gender was associated with co-colonization, then a reporting bias might explain some of the observed gender differences.
Individuals engaging in high-risk sexual behaviors such as having more than two lifetime sex partners and those with an earlier age at first sex were less likely to be co-colonized. This is consistent with the hypothesis of sexual transmission. It has been demonstrated that continuous exposure to different GBS strains over time leads to the development of mucosal or systemic immunity. 18 Our findings also suggest that men who had only one sex partner were more likely to be part of a co-colonized couple implying that they had yet to acquire immunity to GBS.
Our study has several limitations in addition to the small sample size. First, our outcome was colonization, not disease. Although colonization is the first step in disease progression, many other factors are required before symptoms appear. Because we observed similar effects on co-colonization with male-to-female oral sex regardless of the health of the female, it seems likely that the effect is real. The insensitive sampling technique (anal rather than rectal swabs, and direct inoculation of swabs onto solid culture media rather than the more sensitive selective broth media) may have selected for individuals with the greatest density of GBS colonization. If so, the rate of GBS transmission associated with sexual activity may be overestimated because more densely colonized individuals may be more likely to transmit GBS to their partners. Finally, the study is cross-sectional; thus, we cannot ascertain the direction of transmission. For example, it is possible that male-to-female oral sex results in the colonization of the male bowel flora with GBS, rather than infection of the female from the male pharynx.
The response rate was somewhat lower than is desirable, probably reflecting the requirement that both partners in a couple participate. Because one of the most common reasons for refusal was that the woman was hesitant to recruit her sex partner, we suspect that participation was biased toward either more committed partnerships or partnerships in which the financial incentive was more attractive. This might affect the prevalence of the behaviors under study (eg, the variety or frequency of sexual behaviors), but should not bias the association between these behaviors and GBS status.
We chose to measure co-colonization using PFGE rather than serotyping because PFGE is more definitive. Further, there was a relatively high prevalence (10%–11%) of nontypeable strains among women and their male partners that could not be serotyped. This was probably related to poor capsule production, 19 as suggested by our PFGE analysis. Although more sensitive and specific than serotyping, PFGE has some limitations. First, if the background circulation of a particular pattern is high, we might infer that the couple passed the strain between them when in fact it was acquired independently. Given the high level of diversity we observed in banding patterns, and given that there were several different serotypes circulating during the course of the study, this is unlikely. Second, as banding patterns are dependent on the enzyme used, we might observe different patterns if a different enzyme was used. However, given that the level of diversity is high, it again seems unlikely that this would lead us to a different result.
GBS rarely causes disease among otherwise healthy young adults. It is therefore unlikely that any individual's participation in the study was based on GBS status, or that participants answered questions differently because of their knowledge of GBS risk factors. There is the possibility, albeit slight, that some women recruited a man who was not their most recent sex partner, thereby decreasing our estimates of co-colonization.
In conclusion, co-colonization is common among heterosexual college couples in which one partner carries GBS. Male-to-female oral sex resulted in an increase in the sharing between sex partners of identical GBS strains assessed by PFGE. Thus, the role of the pharynx in the overall transmission dynamics of GBS needs to be evaluated. Larger studies should be conducted to identify the effects of both bacterial characteristics and host response on transmission. A better understanding of the transmission dynamics and the natural history of GBS is critical for the design and implementation of an immunization program to prevent GBS disease in pregnant and nonpregnant adults.
We thank Charlotte Williams and the laboratory staff at the University of Michigan Student Health Service for collecting bacterial specimens, and Caesar Briefer for his support of the project. We also thank Katie Neighbors, Bonnie Andree, Lexie Bopp, and the many study recruiters for their efforts that led to excellent data quality and subject recruitment. In addition, we thank Melissa Hickman from Baylor College of Medicine for serotyping the GBS isolates, and Mark Pearlman and Janet R. Gilsdorf for offering comments on this manuscript.
1. Baker CJ, Edwards MS. Group B Streptococcal Infections. Infectious Disease of the Fetus and Newborn Infant. Philadelphia: WB Saunders, 1983.
2. Centers for Disease Control and Prevention. Prevention of perinatal group B streptococcal disease: a public health perspective. MMWR Recomm Rep 1996; 45: 1–24.
3. Schrag SJ, Zywicki S, Farley MM, et al
. Group B streptococcal disease in the era of intrapartum antibiotic prophylaxis. N Engl J Med 2000; 342: 15–20.
4. Hill HR. Group B Streptococcal Infections. Sexually Transmitted Diseases. 2nd ed. New York: McGraw-Hill, 1990.
5. Schuchat A. Group B streptococcus. Lancet 1999; 353: 51–56.
6. Schuchat A, Wenger JD. Epidemiology of group B streptococcal disease. Risk factors, prevention strategies, and vaccine development. Epidemiol Rev 1994; 16: 374–402.
7. Schuchat A, Oxtoby M, Cochi S, et al
. Population-based risk factors for neonatal group B streptococcal disease: results of a cohort study in metropolitan Atlanta. J Infect Dis 1990; 162: 672–677.
8. Dillon HC Jr, Gray E, Pass MA, Gray BM. Anorectal and vaginal carriage of group B streptococci during pregnancy. J Infect Dis 1982; 145: 794–799.
9. Baker CJ, Goroff DK, Alpert S, et al
. Vaginal colonization with group B streptococcus: a study in college women. J Infect Dis 1977; 135: 392–397.
10. Franciosi RA, Knostman JD, Zimmerman RA. Group B streptococcal neonatal and infant infections. J Pediatr 1973; 82: 707–718.
11. Weindling AM, Hawkins JM, Coombes MA, Stringer J. Colonisation of babies and their families by group B streptococci. Br Med J (Clin Res Ed) 1981; 283: 1503–1505.
12. Bliss SJ, Manning SD, Tallman P, et al
. Group B Streptococcus colonization in male and non-pregnant female university students: a cross-sectional prevalence study. Clin Infect Dis 2002; 34: 184–190.
13. Gordillo ME, Singh KV, Baker CJ, Murray BE. Typing of group B streptococci: comparison of pulsed-field gel electrophoresis and conventional electrophoresis. J Clin Microbiol 1993; 31: 1430–1434.
14. Benson JA, Flores AE, Baker CJ, Hillier SL, Ferrieri P. Upregulation of Polysaccharide Expression and Pulsed-Field Gel Electrophoresis Characterization of Nontypeable Group B Streptococcal Isolates. Streptococci and Streptococcal Diseases Entering the New Millenium. Proceedings of the XIV Lancefield International Symposium on Streptococci and Streptococcal Diseases. Auckland: Securacopy, 2000.
15. Schlesselman JJ, with contributions by Stolley PD. Case-Control Studies: Design, Conduct, Analysis. New York: Oxford University Press, 1982.
16. Chretien JH, McGinniss CG, Thompson J, Delaha E, Garagusi VF. Group B beta-hemolytic streptococci causing pharyngitis. J Clin Microbiol 1979; 10: 263–266.
17. Seal DW. Interpartner concordance of self-reported behavior among college dating couples. J Sex Res 1997; 34: 39–55.
18. Campbell JR, Hillier SL, Krohn MA, Ferrieri P, Zaleznik DF, Baker CJ. Group B streptococcal colonization and serotype-specific immunity in pregnant women at delivery. Obstet Gynecol 2000; 96: 498–503.
19. Palacios GC, Eskew EK, Solorzano F, Mattingly SJ. Decreased capacity for type-specific-antigen synthesis accounts for high prevalence of nontypeable strains of group B streptococci in Mexico. J Clin Microbiol 1997; 35: 2923–2926.