Relationship between bacterial diversity, specific urethral bacteria and incident NGU in men who have sex with women

Nongonococcal urethritis (NGU) is a common syndrome in men that is not well understood. Prior studies have investigated factors associated with prevalent NGU, but factors associated with incident NGU have rarely been examined. We evaluated potential precursors of incident NGU in a group of men who have sex with women (MSW) attending a sexual health clinic in Seattle from August 2014 to July 2018. Methods: We conducted a nested case-control study, evaluating a subset of MSW enrolled in a cohort study designed to investigate the relationship between the male urethral microbiota and NGU. At enrollment and monthly follow-up visits, participants had a clinical examination performed at which they provided a urethral swab specimen for Gram staining and a first-void urine sample. Participants also completed a computer-assisted self-interview (CASI) to assess history of sexually transmitted infections (STIs), sexual behaviors, and other sociobehavioral data at each visit. Case visits were defined as the visit at which an individual had an incident case of NGU. NGU was defined as elevated levels of polymorphonuclear leukocytes (PMNs) when urethral discharge was examined on a Gram-stained slide in the absence of Neisseria gonorrhoeae (GC), with or without symptoms such as painful urination, itching, or urethral discharge, and was considered incident if the participant did not have NGU at the prior monthly visit. We selected the visit when the case was diagnosed with incident NGU, and the preceding visit when they were NGU-negative. Controls were matched to cases on follow-up time for both time points using a 1:1 matching ratio, matching by visit number and interval between visits. Controls were NGU-negative at both matched visits. Nucleic acid amplification testing (NAAT) for GC, Chlamydia trachomatis (CT) and Mycoplasma genitalium (MG) was performed on the urine samples using Aptima assays (Hologic, Inc., San Diego, CA). To characterize the urethral microbiota, we applied broad-range 16S ribosomal RNA (rRNA) gene polymerase chain reaction (PCR) and sequencing to urine samples. We used Fisher’s exact test for comparisons of categorical characteristics and a paired t-test for continuous characteristics. Conditional logistic regression models were used to estimate odds ratios (ORs) for the association of bacterial diversity, as measured by the Shannon Diversity Index (SDI) and species richness, as well as detection of each of four bacterial species (Haemophilus influenzae, Fannyhessea vaginae, Lactobacillus iners, Streptococcus mitis) at the visit prior to diagnosis with incident NGU. Results: The 262 enrolled persons with follow-up data contributed 63 incident NGU events. Although cases and controls did not differ with respect to age, race/ethnicity, or NGU at enrollment, cases had a significantly lower level of education, used condoms less frequently with their most recent sexual partner, and were more likely to have a prior history of NGU and/or CT than controls (p≤0.03 for all). After adjusting for condom use, participants with one-unit higher SDI at the visit prior to NGU diagnosis had 10.29 times the odds of incident NGU (aOR=10.29, 95% CI: 1.49-73.16, p=0.02) compared to participants with a one-unit lower SDI. Similarly, species richness at the visit prior to NGU diagnosis was associated with increased odds of incident NGU (aOR per species = 1.08, 95% CI: 1.001-1.169, p=0.047). In a multivariate model that included all four bacterial species detected at the visit prior to NGU diagnosis, F. vaginae was associated with 29.7 times the odds of incident NGU (aOR=29.68, 95% CI: 1.55-568.26, p=0.02) compared to those without F. vaginae. Neither H. influenzae, L. iners, nor S. mitis were significantly associated with incident NGU, although MSW with S. mitis were somewhat less likely to have incident NGU. Conclusions: Higher SDI, higher species richness, and presence of F. vaginae were significantly associated with higher odds of incident NGU. Future work should involve larger longitudinal studies re-examining these factors’ associations with incident NGU to better understand these relationships and other potential risk factors for the syndrome.