Chlamydia trachomatis (CT) was added to the United States (US) Nationally Notifiable Conditions List in 1995 and case reports have increased over time, with the highest burden of infection among 15–24 year olds.¹ Most prior research on CT has been conducted among cisgender adolescent and young women who have sex with men (WSM) and cisgender men who have sex with men (MSM), leaving cisgender women and other individuals assigned female sex at birth who have sex with women (WSW) understudied.^2,3

In people with a uterus, untreated CT infections can lead to serious complications, including pelvic inflammatory disease (PID), which may lead to ectopic pregnancy, tubal factor infertility, and chronic pelvic pain.⁴ Despite the potential for severe sequelae, CT infection is often asymptomatic; therefore, the US Preventive Services Task Force recommends annual screening for all “sexually active women younger than 25 and women ≥25 years with increased risk of infection”.⁴ WSW are often missed in CT screening efforts because of perceptions that their risk of exposure is lower than cisgender women who have sex with men (WSM).^2,5 However, in internet and venue-based surveys among WSW, self-reported lifetime STI prevalence in WSW was approximately 20%.^3,6 In a separate assessment of nearly 10,000 family planning clinic visits, CT infection detected by nucleic acid amplification testing (NAAT) was the same in women who have sex with women only (WSW-only) as in women who had sex with women and men (WSWM) (7.1% for both), and slightly higher than in women who had sex with men only (WSM-only, 5.3%).⁷

Although most data on sex of sex partners are over a decade old, existing studies indicate that 75–97% of WSW have ever had a male sexual partner.^2,8,9 Additionally, like WSM, WSW engage in behaviors that may increase the likelihood of STI exposure, including transactional sex, substance use, inconsistent condom use, and sharing sex toys.^3,5,10,11 STI transmission has also been documented between female sex partners regardless of history with male sex partners,² although the transmission probability is likely lower.

Serologic assays detect the presence of anti-CT antibody, capturing prior CT infection,¹² and can provide a better estimate of lifetime CT infection than self-report or single point-in-time NAAT testing. The National Health and Nutrition Examination Survey (NHANES) includes data on the presence of antibodies to CT in serum specimens from women aged 14–39 years participating in the 2013–2014 and 2015–2016 cycles (neither gender identity nor birth sex were elicited so we use the terms male or men and female or women to represent the response option available to the respondent). Although two prior analyses of these data reported population-level CT seroprevalence in women,^13,14 no detailed information about WSW and factors associated with the presence of anti-CT antibody were included. We analyzed NHANES data from 2013–2016 to estimate and compare CT seroprevalence in women who reported ever having sex with a woman to that in women who had male sex partners only (WSM-only). We also identified characteristics associated with prior CT infection in each group.

Overall, 19,357 people participated in the MEC examinations and completed questionnaires in the 2013–2014 and 2015–2016 NHANES cycles, of whom 9,850 reported female sex. CT serology was performed on serum samples from 1,724 women aged 18–39 years who had CT NAAT results. We excluded 101 women who reported never having any type of sex, 198 women who either refused to answer or were missing data on whether they had ever had sex with a woman, and the seven WSW-only for whom estimates would have to be suppressed, leaving 1,418 sexually-experienced women aged 18–39 years in this analysis (none were missing data on whether they had sex with men).

In this nationally-representative, population-based survey, CT seroprevalence in WSWM (38.9%) was higher than in WSM-only (28.6%) Both groups had common correlates of seropositivity including Black non-Hispanic race/ethnicity, Hispanic ethnicity, more lifetime male sex partners, and CT infection in the past year. The strongest predictor of CT seropositivity for WSWM was the lifetime number of male sex partners (4.5 times higher in WSWM with ≥5 lifetime male partners than WSWM with 1–2 partners). Additionally, bisexual identity was associated with lower seroprevalence in WSWM. Among WSM-only, low income, and young age at sexual debut were associated with higher CT seroprevalence and the strongest predictor of CT seropositivity was young age at sexual debut (3.8 times higher among WSM-only who were ≤13 years than WSM-only who were ≥18 years of age).

WSWM and WSM-only in NHANES 2013–2016 differed in several ways. WSWM were younger and a lower proportion had a college education than WSM-only. WSWM had more lifetime male partners and were significantly younger at sexual debut than WSM-only. This is consistent with reports among WSWM in other populations^8,10,11 and women who report only female partners may differ. The high CT seroprevalence that we observed in WSWM with more lifetime male sex partners emphasizes the role that sex with men has in acquiring CT infection. Several reports have demonstrated higher STI prevalence among WSWM than WSW-only, including among women attending urban sexual health clinics in Baltimore, MD¹¹, and Black women attending a Mississippi sexual health clinic.²⁰ More male partners have been associated with higher STI risk in other settings as well,²⁴ likely attributable to more efficient transmission during penile-vaginal sex, higher background CT prevalence in WSWM sexual-partner pools, or a mixture of both.

Black non-Hispanic race/ethnicity was reported equally by WSWM and WSM-only and associated with CT seropositivity in both groups; seroprevalence in Black non-Hispanic women was 2.2 to 2.9 times higher than seroprevalence in White non-Hispanic women. Dramatically higher prevalence of both acute and lifetime CT among Black women in general¹ and among Black women who report recent sex with women has been previously reported,^3,22,25 and could partly be due to high CT prevalence in sexual networks. Importantly, it highlights disparities in sexual healthcare delivery, often due to institutionalized racism, sexism, and homophobia faced by Black WSW which manifest in unequal access to screening and preventive health services, decreased access to affordable health insurance, and stigma from healthcare providers.^3,6

Although few studies of CT infection or other STIs among WSW have measured lifetime prevalence using serology, several have measured lifetime STI risk using self-reported diagnoses. In these studies, the lifetime STI prevalence in WSW, many of whom also reported male partners, was around 20%.¹⁴ This is lower than our CT seroprevalence estimates based on detection of Pgp3 antibodies and is likely explained by the additional undiagnosed infections that serology captures. Earlier analyses of NHANES 2013–2016 that considered sexual identity/orientation used an alternative seroassay, the Pgp3 multiplex bead array (MBA) assay, and stratified results by high and low median fluorescence intensity as a marker of Pgp3 antibody levels. Among all women (WSW-only and WSWM together) in this earlier analysis, both weighted low- and weighted high-seropositivity (30.6% and 10.5%, respectively)¹⁴ were lower than the seroprevalence that we observed among WSWM using the Pgp3 ELISA assay (weighted seropositivity=38.9%). This difference may be due to the variability between the two assays,^13,26 or to the greater number of WSM-only in the population, attenuating the seroprevalence estimate for the combined group.

Although CT seroprevalence increased with age in British women participating in the British National Surveys of Sexual Attitudes and Lifestyles (Natsal),¹⁹ we did not observe a significant relationship between seropositivity and age in either group, similar to prior analyses of these data among all women.¹³ This may reflect the more limited age range in NHANES 2013–2016 (18–39 years) than in Natsal (16–44 years). Alternatively, CT antibody levels may wane with time and/or disappear (e.g., seroreversion). Öhman and colleagues found that the proportion of people with IgG antibodies from urogenital CT declined from 65.5% shortly after an incident infection to 34.5% 3–10 years later.²⁷ CT incidence is highest in women under 25,¹ and the CT Pgp3 ELISA may have preferentially identified more recent CT infections in younger women, failing to identify prior infections in some older women for whom over a decade had passed since their most recent infection.

Paradoxically, while a higher proportion of WSWM were under age 13 when they initiated sexual activity, age at sexual debut was only associated with CT seropositivity in WSM-only. Despite our stratified results, age at debut was associated with CT seroprevalence when all women were considered together¹³, consistent with findings among British women overall¹⁹ and reflects the larger number of WSM-only than WSWM in our NHANES population. The absence of association between condomless sex in the past year and CT seropositivity is likely because condom use was reported for the past 12 months, while seroprevalence is a long-term indicator of infection.

Almost half of WSWM in our population reported their sexual identity as straight, demonstrating the discordance between sexual behavior and sexual identity that has been characterized in other literature.^8,28 Sexual identity has a complex and indirect relationship with sexual experience and many women who have had sexual experiences with other women identify as heterosexual or straight. In Australia, only 31% of women who reported ever having sexual experiences with women and feeling attraction towards women identified as bisexual or lesbian.²⁸ Consistent with this discordance, data on the association of sexual identity and likelihood of STI are conflicting. We observed lower CT seroprevalence in WSWM who identified as bisexual than those who identified as straight or lesbian, yet Canadian WSW who identified as bisexual were more likely to report ever having had an STI than WSW who identified as lesbian (whether women also had male partners was not specified).³ These differences could reflect the inclusion of asymptomatic infections captured by serology in NHANES, and the less frequent screening undergone by lesbian women in routine clinical care, perhaps due to perceived low risk. Clinicians assessing potential for CT exposure in their patients should use an inclusive approach, focusing on recent sexual history with partners of any gender/sex, irrespective of sexual identity.²⁹ STI prevention strategies and messaging for WSW should address the variety of sexual practices and identities in this group and clinicians should be trained to provide sexual healthcare free of stigma that might otherwise discourage WSW from seeking care.

The primary strengths of this study were the large, nationally-representative nature of NHANES and the ability to estimate lifetime infection using serology. Most prior studies of lifetime prevalence of CT in WSW have been conducted in specialized populations (clinics or convenience surveys) and have used self-reported data to define prior CT infection. The serology assay that we used is more sensitive and less biased than self-reported data.^3,20 Pgp3 antigens are highly conserved across CT isolates and the corresponding IgG antibodies can persist for over 10 years in many cisgender women,^12,27 enhancing their reliability as a measure of lifetime CT infection.

Our study also had limitations. First, seroreversion can occur and some women classified as CT seronegative may have previously had CT infection. Second, statistical power was limited for analyses of WSWM and the small number of WSW-only required us to exclude them from our analysis. The latter prevented us from examining differences between WSW-only and WSWM and from inferring whether WSWM were infected by male or female sex partners. However, the strong association with more male but not more female partners in WSWM suggests that substantial CT exposure occurred from male partners. Third, sex of sex partners and CT seropositivity were measured over a lifetime, yet condom use, income, and recent STI history were measured over more proximal timeframes. Sociobehavioral characteristics measured at one timepoint are often, but not always, good proxies for lifetime measures and associations with seropositivity should be interpreted in light of the timeframe defining them. Fourth, the lack of distinction between gender and sex assigned at birth in the survey design may have resulted in some misclassification. Transgender, non-binary and gender non-conforming individuals may not have been asked relevant sexual behavior questions due to sex-specific skip patterns. This misclassification has been present in other national surveys and may have resulted in some bias.³⁰ Fifth, our analyses were limited to non-institutionalized, civilian women ages 18–39 years. The exclusion of adolescents and women above 40 limited generalizability. Sixth, if people who provided serum samples fundamentally differed from the general population, our estimates may be biased. Seventh, the cross-sectional design of NHANES prevents us from making conclusions related to causality.

In summary, WSWM had a substantial burden of lifetime CT infection that was higher than the lifetime burden of CT in WSM-only. The considerable differences in CT seroprevalence in WSWM by race/Hispanic ethnicity highlight the need to address the disparities that occur at the intersection of race/Hispanic ethnicity and sex/gender of sex partner. Clinicians should consider the heterogeneity in women described as WSW and assess common correlates of lifetime CT infection when evaluating the likelihood of exposure. Future research will need to oversample women who exclusively have sex with women to fully understand risk for CT infection in this group.