Patients and methods

A total of 292 male subjects enrolled at STD clinics to evaluate microbial aetiologies of urethritis with written informed consent. The study was approved by the Western Institutional Review Board (IRB) for Johns Hopkins University and the Baltimore City Health Department IRB. Subjects were approached using random assignment of every fifth patient to the research clinician. The acceptance rate was estimated to be >85%. Sociodemographic information, symptoms, behavioural risk factors and STD history were obtained.

All subjects received clinical standard of care. Urine and two urethral swabs were collected. A Gram stain smear was prepared and urine was tested by Aptima Combo2, (GenProbe, San Diego, CA, USA) for CT and NG.4 Research TV and MG polymerase chain reactions (PCR) were performed on urine,5 6 7 as well as GenProbe TMA targeting TV rRNA5 and Gen-Probe TMA research assay targeting MG rRNA.8 Both research and TMA assays were required to be positive for the patient to be considered infected with TV or MG.

Urethritis was diagnosed if the Gram stain demonstrated ⩾5 polymorphonuclear leucocytes per oil-immersion field. Data analyses were performed using SAS (9.1, Cary, NC, USA). χ² Tests (p value <0.05 for significance) and multiple logistic regression models were used to examine the association of CT, NG, TV and MG with urethritis. Possible a priori confounders included age and infected partner contact. Two-way interactions were evaluated. (See STI website for supplemental details.)

Results

Of 292 men enrolled, specimens were available for analysis for 290: 153 (52.8%) were classified as having urethritis and 137 (47.2%) without urethritis; 56.6% (164/290) reported symptoms. Reasons for visit included symptoms (68%), “check-up” (23%) and “infected partner contact” (14%).

The mean age of the study participants was 26.9 (SD 7.6) years; 96.9% were black. Men attending one clinic were more likely have urethritis (59.2% vs 47.5%, p = 0.05). Men referred as “contacts” were more likely to have urethritis (18.3% vs 10.2%, p = 0.05). Men with urethritis were more likely to be symptomatic than those without (66.7% vs 45.3%, p<0.001). All men with gonorrhoea were symptomatic. Reported symptoms (discharge, dysuria and/or itching) were significantly associated with urethritis (OR, 2.41; 95% CI 1.50 to 3.89, p<0.001). Education, employment status, tobacco use, condom use, new partner and previous STDs were not significant.

The overall prevalences for CT, NG, TV and MG were 20.3%, 12.8%, 3.4% and 15.2%, respectively. The prevalence for each organism was always greater in those participants with urethritis (32.7%, 24.2%, 5.2% and 22.4% for CT, NG, TV and MG, respectively) than those without urethritis (6.6%, 0%, 1.5% and 7.3% for CT, NG, TV and MG, respectively.

Co-infections with additional organisms ranged from 30–40%. The percentage with co-infections was greater in men with urethritis. At least one organism was detected in 67.3% of men with urethritis versus 14.6% without urethritis. In men with urethritis, 5.9% were co-infected with CT and MG. The same percentage was co-infected with CT and NG. Co-infection rates of other combinations were <1%. Without urethritis, co-infection was 0.7%.

Univariate analysis demonstrated the association of organisms with urethritis (table 1). Men infected with CT were almost seven times as likely to have urethritis as those not infected with CT. Infection with MG was three times as likely in men with urethritis. There was no significant association of TV with urethritis (table 1) or age. Participants who were “contacts” were twice as likely to have urethritis (table 1).

Two models were fitted using multiple logistic regression (table 1). One model contained the organisms CT, NG, TV and MG as the independent variables; the other included age and contact referral as potential confounders. No significant interactions were found. NG could not be included in modelling owing to a zero in a response cell; all NG infected men had urethritis. CT and MG were strongly associated with urethritis (OR, 6.92 and OR, 3.67, respectively), but TV was not (p = 0.08). Because of possible impact of co-infections, similar analyses were performed using data from those infected with one organism, but provided the same conclusions. A model containing two-way interaction terms for all of the organism pairs demonstrated no terms were significant. (See STI website for supplemental results.)

Discussion

This study is unusual in that the prevalence of CT, NG, TV, MG and co-infections, and their association with urethritis were examined in the same study population using prospectively collected specimens. The prevalence of CT, TV and MG in men with urethritis was similar to that reported by others in this and other STD clinics.9 10

We also found high co-infection rates of MG with NG and CT. Our co-infection rate was lower than that of Mena et al, but similar or higher than that reported by others.11 12 The rate of co-infection in subjects with urethritis with at least one pathogen was >30%, compared to 10% for those without urethritis. Our gold standard of having two positive tests for MG did not appear to affect the overall positivity rate, since there were only five positive tests that were not positive in both assays. (kappa, 0.937).

In men with urethritis, 18.3% were “contacts”, whereas in men without urethritis 10.2% were “contacts” (p = 0.05). Participants who were contacts may have been asymptomatically infected with the organism. Men with CT, NG or MG were more likely to have urethritis in univariate analysis. Bradshaw et al reported an increased likelihood of CT and MG in men with non-gonococcal urethritis (NGU) (20% versus 3% for CT and 9% versus 1% for MG).13 Our study corroborated those findings of significant associations of CT and MG with urethritis or NGU. We did not find an association of TV with urethritis, perhaps due to our low prevalence (3.4%). Requiring two positive tests for TV did not appear to affect the overall positivity rate, as only four tests were not positive in both assays (kappa, 0.826). Our multiple logistic regression model showed CT and MG were significantly associated with urethritis (p<0.001). All GC infected men had urethritis. Our data confirm other studies reporting the association of CT and MG with urethritis and NGU.10 11 12 13 Because men with MG were almost four times as likely to have urethritis, the important implication of our study may be that men need to be screened for MG, as well as NG and CT. The high prevalence of MG as a single pathogen in 14.4% of men with urethritis also conveyed significant public health implications for making diagnostic changes in the way clinics deal with urethritis and treatment.14 15 Our limitations included not being able to enrol larger numbers of men with urethritis. Strengths included using the NAATs for four organisms concurrently, thereby identifying co-infections. In conclusion, all of the pathogens except TV were associated with urethritis in men. The association of MG with urethritis in this study confirms the importance of routinely screening men for MG as a cause of urethritis.