Urethral sexual exposures: relationship with prevalent nongonococcal urethritis, the urethral microbiota, and resolution of symptoms
Non-gonococcal urethritis (NGU), the most common male genital tract syndrome, is often caused by Chlamydia trachomatis (CT) and Mycoplasma genitalium (MG). However, up to 50% of cases are of unknown etiology, and yet-unidentified pathogens and polymicrobial communities likely play a role in some of these cases. Studies suggest that specific urethral sexual exposures influence the composition of the male urethral microbiota, in addition to facilitating transmission of known pathogens, although this hypothesis has not been tested. Clinical management of NGU is challenged by its diverse etiologies and syndromic management, and the efficacy of syndromic management may vary depending on the etiology. At NGU diagnosis, men are advised to abstain from sex for ≥7 days and until their urethral symptoms resolve, although the efficacy of and adherence to these recommendations is unknown. An improved understanding of the risk factors and etiologies of NGU, as well as the effectiveness of current NGU management approaches, would advance prevention, diagnosis, and treatment of the syndrome. We enrolled patients assigned male sex at birth and age ≥16 years who were attending the Public Health – Seattle and King County STD Clinic into cross-sectional and cohort studies. The cross-sectional study included patients with and without NGU. The cohort study included only patients with NGU who reported sex with exclusively with men in the past year. We defined NGU as ≥5 polymorphonuclear leukocytes (PMNs) per high-power filed (HPF) with either urethral symptoms or visible urethral discharge on examination, in the absence of Neisseria gonorrhoeae (GC). Cross-sectional study participants had a single study visit, while cohort study participants returned for follow-up visits every three weeks over three months, collected urine at home weekly, and completed a web-based sex diary weekly. Study visits included an examination, collection of urethral swab and urine specimens, and a computer-assisted self-interview. We tested clinic-collected urine specimens for CT, MG, and GC (Aptima, Hologic, Inc., Marlborough, Massachusetts). We applied PCR and sequencing to MG-positive specimens to detect macrolide resistance mediating mutations (MRMM). We applied broad-range 16S rRNA gene PCR with deep sequencing to all clinic- and home-collected urine specimens. Among cross-sectional study participants, we used logistic regression to estimate the association between urethral sexual exposures at last sexual episode and NGU and non-CT/non-MG NGU, separately among cisgender men and transgender women who have sex with men (MSM/TGWSM) and cisgender men who have sex with women (MSW). Among MSM in the cohort study, we estimated the association between the diversity (Shannon Index) and log10-richness (i.e., number of bacterial species) of the urethral microbiota and participants’ urethral sexual exposures in seven 3-day time windows before specimen collection using generalized estimating equations, adjusting for recent antibiotics, age, race/ethnicity, HIV status, and HIV pre-exposure prophylaxis use. For each exposure category, we tested whether all seven window coefficients equaled zero (i.e., no overall association) using a Wald test. Finally, among MSM in the cohort study who received azithromycin, we estimated median time until resolution of symptoms and resumption of sex after presumptive azithromycin therapy for NGU using the Kaplan-Meier method, overall and by etiology. We also fit a Cox proportion hazards model to estimate the association between NGU etiology and time of symptom resolution, adjusting for confounders. From 08/2014-11/2017, we enrolled 432 patients in the cross-sectional study (118 NGU+ MSM/TGWSM, 65 NGU- MSM/TGWSM, 126 NGU+ MSW, 123 NGU- MSW). Seventy-two (30%) and 49 (20%) participants with NGU had CT and MG, respectively. Compared to MSM/TGWSM reporting only non-urethral exposures at last sex, those reporting insertive anal intercourse (IAI) alone (adjusted odds ratio [AOR]=4.46, 95% confidence interval [CI]=1.09-18.19) and IAI with insertive oral sex (IOS) (AOR=7.88, 95%CI=2.67-23.26) had higher odds of NGU, while those reporting IOS alone had similar odds of NGU. Compared to MSW whose only urethral exposure at last sex was vaginal sex (VS), MSW reporting IOS and VS had similar odds of NGU. Results were similar for non-CT/non-MG NGU. From 12/2014-5/2018, we enrolled 92 MSM with NGU in the cohort study. They contributed 1,095 person-weeks of behavioral data (median=12 diaries/man, interquartile range [IQR]=12-13). Among 894 clinic- and home-collected urine specimens (median=10 specimens/man, IQR=8-12), median diversity was 1.33 (IQR=0.76-1.99), and median richness was 14 species (IQR=9-23). Overall, diversity and log10-richness were associated with condomless IAI alone (both P<0.01) but not IOS alone or IOS with condomless IAI in the 21 days prior. Diversity and log10-richness were lower 1-3 days after condomless IAI alone and higher 16-18 days after condomless IAI alone. Finally, between 12/2014-7/2018, 103 MSM in the cohort study received presumptive azithromycin therapy for CT-NGU (35%), MG-NGU (21%), and non-CT/non-MG NGU (41%); 3% had CT/MG co-infection. Among MSM with MG-NGU, 94% had MRMM. Overall, median time to symptom resolution after azithromycin was seven days (95%CI=5-9), and 37% had symptoms lasting >7 days. For CT-NGU, MG-NGU, and non-CT/non-MG NGU, median time to symptom resolution was four (95%CI=2-6, 16% >7 days), undefined (95%CI=7-undefined, 60% >7 days), and seven (95%CI=5-11, 46% >7 days) days, respectively. Men with MG-NGU (hazard ratio [HR]=0.29, 95%CI=0.13-0.68) and non-CT/non-MG NGU (HR=0.56, 95%CI=0.36-0.89) had a decreased “hazard” for symptom resolution (i.e., longer time to symptom resolution) after azithromycin therapy compared to those with CT-NGU. Median time to urethral sexual exposure after treatment was 16 days (95%CI=12-18); 27% did not avoid exposure for ≥7 days and until symptoms resolved. Results from these studies suggest that, among MSM/TGWSM, IAI may lead to transmission of yet-unidentified rectal micro-organisms that cause non-CT/non-MG NGU, in addition to transmission of known pathogens. Among MSM with NGU at baseline, condomless IAI in the prior 21 days was independently associated with diversity and richness of the urethral microbiota during follow-up, suggesting that some urethral sexual exposures influence at least certain dimensions of the composition of the urethral microbiota. In contrast, sites of urethral sexual exposure appear less important for understanding NGU risk among MSW due to minimal variation in sexual behavior. Finally, among MSM, NGU symptoms often persist for >7 days following presumptive azithromycin, particularly for those with MG-NGU and non-CT/non-MG NGU. The very high prevalence of MRMM among men with MG-NGU likely led to their long duration of symptoms after azithromycin, while inadequate treatment of unidentified pathogens or polymicrobial communities may have contributed to the long duration of symptoms for non-CT/non-MG NGU. Counseling at NGU diagnosis should educate patients that symptoms may persist for >7 days and emphasize the rationale for the 7-day abstinence period.