Incidence of Nongonococcal Urethritis in Men Who Have Sex With Women and Associated Risk Factors

Background Incidence and risk factors for nongonococcal urethritis (NGU) remain poorly defined. We conducted a cohort study to estimate the incidence of NGU and identify risk factors in men who have sex with women. Methods We enrolled cisgender male sexually transmitted disease clinic attendees 16 years or older who reported exclusively female partners. At enrollment and 6 monthly follow-up visits, men underwent a clinical examination, provided urethral swab and urine specimens, completed a sexual behavior survey and biweekly diaries, and were tested for Chlamydia trachomatis (CT) and Mycoplasma genitalium (MG) using Aptima assays (Hologic, Inc). Nongonococcal urethritis was defined as >= 5 polymorphonuclear leukocytes per high-power field plus either urethral symptoms or visible discharge. We estimated the incidence of NGU overall, asymptomatic and symptomatic NGU, non-CT/non-MG NGU, and CT/MG-associated NGU using Poisson regression for clustered outcomes. We performed relative risk binomial regression for clustered data to identify characteristics associated with incident NGU. Results From August 2014 to July 2018, 307 participants at risk for NGU contributed 109.4 person-years. Median age was 32 years, and 52% were White. At enrollment, 107 men had NGU; of these, 88% were symptomatic, 27% had CT, and 22% had MG. Fifty men had 60 cases of incident NGU (incidence rate, 56 per 100 person-years; 95% confidence interval, 43-74). Unlike prevalent NGU at enrollment, CT/MG-associated incident NGU was rare (incidence rate, 7; 95% confidence interval [CI], 4-15), and most (78%) incident NGU was asymptomatic. Risk factors for incident NGU were <= high school education (adjusted rate ratio [ARR], 2.45; 95% CI, 1.19-5.00), history of CT (ARR, 2.15; 95% CI, 1.08-4.27), history of NGU (ARR, 2.67; 95% CI, 1.27-5.62), and NGU at enrollment (ARR, 2.03; 95% CI, 1.04-3.98). Neither condom use nor having a new partner was associated with incident NGU; Black race was only associated with incident symptomatic and non-CT/non-MG NGU. Conclusions Incidence of NGU was high, predominantly non-CT/non-MG and asymptomatic. Future studies should investigate the etiology and clinical significance of asymptomatic NGU.