Ureaplasma urealyticum is significantly associated with non-gonococcal urethritis in heterosexual Sydney men

Abstract

We investigated the prevalence of various genital organisms in 268 men with (cases) and 237 men without (controls) urethral symptoms/signs (urethral discharge, dysuria and/or urethral irritation) from two sexual health clinics in Sydney between April 2006 and November 2007. The presence of urethral symptoms/signs was defined as non-gonococcal urethritis (NGU) for this study. Specific aims were to investigate the role of Ureaplasma urealyticum in NGU and the prevalence of Mycoplasma genitalium in our population. Multiplex polymerase chain reaction-based reverse line blot (mPCR/RLB) assay was performed to detect 14 recognized or putative genital pathogens, including Chlamydia trachomatis, M. genitalium, U. urealyticum and U. parvum. U. urealyticum was associated with NGU in men without another urethral pathogen (odds ratio [OR] 2.0, 95% confidence interval [CI] 1.1–3.8; P = 0.04); this association remained after controlling for potential confounding by age and history of unprotected vaginal sex in the last four weeks (OR 2.0, 95% CI: 1.1–3.9; P = 0.03). C. trachomatis (OR 7.5, P < 0.001) and M. genitalium (OR 5.5, P = 0.027) were significantly associated with NGU. The prevalence of M. genitalium was low (4.5% cases, 0.8% controls). U. urealyticum is independently associated with NGU in men without other recognized urethral pathogens. Further research should investigate the role of U. urealyticum subtypes among heterosexual men with NGU.

Keywords

non-gonococcal urethritis urethral symptoms genital pathogens multiplex PCR-based reverse line blot assay

INTRODUCTION

Questions about the aetiology of non-gonococcal urethritis (NGU) remain unanswered. Neisseria gonorrhoeae, Chlamydia trachomatis and Mycoplasma genitalium are recognized sexually transmitted pathogens and independent causes of male urethritis. However, there is continuing controversy over the role of ureaplasmas in NGU.

Prior to the recognition of two separate species of ureaplasma¹ – Ureaplasma parvum and U. urealyticum – all human ureaplasmas were classified as U. urealyticum, and studies had reported variable levels of association with NGU.^2–4 Recent studies in which the two Ureaplasma species were distinguished, using nucleic acid amplification techniques (NAAT), have shown that U. parvum is a non-pathogenic genital commensal organism in the male urethra,^5–7 probably accounting for the conflicting results of early studies. Nonetheless, the role of U. urealyticum in NGU remains unclear. Recent studies have provided conflicting results,^5–7 even though experimental intraurethral inoculation, into human volunteers, of two different strains of U. urealyticum serovar 5 resulted in symptomatic urethritis.⁸ It is possible that only certain subtypes of U. urealyticum are associated with urethritis, and subtype 1 (serovars 2, 5, 8 and 9) has been implicated in two reports.^8,9

NGU is a clinical syndrome consisting of urethral symptoms – discharge, dysuria, irritation – and/or signs – urethral discharge and inflammation, in the absence of gonococcal infection. Varying diagnostic criteria for NGU influence study results, with one recent study defining NGU solely on the basis of a laboratory finding of microscopic urethritis,⁵ another on the presence of urethral discharge and microscopic criteria⁶ and yet another on the basis of symptoms only.⁷ Many men with urethral pathogens detected do not have microscopic urethritis,⁷ and clinical diagnostic criteria^7,10,11 are often used in the clinical setting.^10,11

NAAT such as the polymerase chain reaction (PCR) provide an opportunity to identify multiple genital organisms in a single assay. We have previously developed a multiplex PCR-based reverse line blot (mPCR/RLB) assay.¹² We used this assay to determine the prevalence of a range of potential genital pathogens in men, with and without urethral symptoms/signs, with reference to their sexual practices and partners, attending two sexual health clinics in Sydney, Australia. Specifically, we wanted to investigate the role of U. urealyticum in NGU and the prevalence of M. genitalium in our population.

METHODS

Study population

Men with and without urethral symptoms who attended Sydney Sexual Health Centre (SSHC) or Parramatta Sexual Health Clinic (PSHC) in metropolitan Sydney, Australia, between April 2006 and November 2007 were invited to participate in the study, which was approved by the Sydney West Area Health Service Human Research Ethics Committee. Cases and controls were opportunistically recruited by all clinical staff. Cases were men with urethral symptoms or signs: urethral discharge (reported or on examination), dysuria and/or urethral irritation, defined as NGU, and controls were men with none of these symptoms or signs. The exclusion criteria included a history of, or sexual partner with, genital chlamydia, gonorrhoea or NGU within the last three months, antibiotic use within the past eight weeks or conditions likely to affect the diagnosis of NGU, including gonorrhoea, severe balanitis, meatal warts and overt genital ulcers.

Clinical and laboratory methods

Medical and sexual history included the presence of urethral symptoms (urethral discharge, dysuria and/or urethral irritation), sexual practices and condom use in the last four weeks, and gender and number of sexual partners in the last four weeks and three months. Urethral swab for Gram stain and bacterial culture, and first-voided urine specimen were collected from all men. Genital examination noted circumcision status and the presence or absence of abnormal urethral discharge after milking the urethra. Men were treated according to standard clinic protocols, including contact tracing and treatment.

Urethral specimens were collected using a rayon-tipped swab passed 0.5–1.0 cm into the distal urethra. Samples were smeared onto a glass slide, heat-fixed and retained for Gram stain and microscopy by one independent microscopist. Microscopic urethritis was defined as ≥5 polymorphonuclear leukocytes (PMNL) per high-power field (HPF) (×1000 magnification) in at least five random fields. Symptomatic men had an additional smear sample prepared for immediate examination.

First-voided urine samples were stored and transported at 2–8°C, and processed within seven days of collection. One half of the sample was sent for routine C. trachomatis testing using Cobas Amplicor PCR (Roche Diagnostics, Australia, Castle Hill, NSW), and the other half retained and later tested by mPCR/RLB for C. trachomatis, M. genitalium, M. hominis, Trichomonas vaginalis, Gardnerella vaginalis, U. parvum, U. urealyticum, Neisseria gonorrhoeae, N. meningitidis, Streptococcus pneumoniae, Haemophilus influenzae, adenovirus and herpes simplex virus (HSV) types 1 and 2.

Details of mPCR/RLB development and validation are reported separately.¹² The assay was performed according to a well-described protocol.¹³ Briefly, two pairs of biotin-labelled, species-specific primers for each species were used in a single mPCR, designed and optimized to allow simultaneous amplification of targets, from DNA extracts, without cross-reactions or interference. Amplification products were hybridized to species-specific probes fixed to a nylon membrane and detected by chemiluminescence of the peroxidase-labelled streptavidin conjugate.

Statistical methods and analysis

A sample size of 259 men in each group was chosen in order to detect a significant difference in U. urealyticum prevalence between those men with and without NGU (assumed 10% and 3% U. urealyticum positive, respectively), with a power of 90% at a significance value of 0.05. However, this calculation was limited by the lack of data on the prevalence of U. urealyticum at the time of study design. Data were analysed using SPSS version 15.1 for windows. Univariate associations were assessed using odds ratios (OR) and the chi-square test for categorical variables, and the Mann-Whitney U-test for continuous variables. A multivariate logistic regression analysis was undertaken to identify whether U. urealyticum was an independent risk factor for NGU.

Subjects who tested positive for N. gonorrhoeae by any method were excluded from the data analysis. A positive mPCR/RLB result for a study organism was considered as detection of that organism, and subjects with mixed infections were included in the analysis.

RESULTS

Study population

A total of 540 men were enrolled, 284 (55%) at SSHC and 232 (45%) at PSHC. Twenty-four men were ineligible according to the exclusion criteria including one in whom N. gonorrhoeae was detected by mPCR/RLB and confirmatory PCR despite negative microscopy and culture, and one whose microscopy was negative, but culture was positive. Urine samples for mPCR/RLB were missing for 11 subjects, leaving 505 available for analysis. Urethral smear microscopy results were available for 491 of these. Compared with PSHC, men enrolled at SSHC were significantly younger (median age 32.5 versus 35.9 years; P = 0.004) and had more male partners in the last four weeks (P < 0.001). However, the proportions of men with symptoms/signs, microscopic urethritis and with positive test results for any study microorganism/s did not differ between sites (data not shown) and therefore the results for both clinics were analysed together.

Patient characteristics associated with cases and controls

Demographic and behavioural characteristics of all eligible cases (268) and controls (237) are presented in Table 1. Cases were significantly older than controls and more than twice as likely to report vaginal sex with a new, or any, partner in the last four weeks, with the association stronger for unprotected (<100% condom use) vaginal sex than protected (100% condom use) vaginal sex.

Table 1

Demographic and behavioural characteristics of the study population

Characteristic	Controls (%) n = 237	Cases (%) n = 268	Crude OR (95% CI)	P value
Age (median)	31.5 years	35.9 years		0.002
Study site
PSHC	103 (43.5)	123 (45.9)
SSHC	134 (56.5)	145 (54.1)	0.9 (0.6–1.3)	0.592
Circumcised*
No	127 (53.8)	130 (48.9)
Yes	109 (46.2)	136 (51.1)	1.2 (0.9–1.7)	0.284
NSP last four weeks^†
No	98 (41.4)	75 (28.0)
Yes	139 (58.6)	193 (72.0)	1.8 (1.3–2.6)	0.002
NSP last three months^†
No	52 (21.9)	46 (17.2)
Yes	185 (78.1)	222 (82.8)	1.4 (0.9–2.1)	0.178
≥1 FSP last four weeks^†
No	146 (61.6)	102 (38.1)
Yes	91 (38.4)	166 (61.9)	2.6 (1.8–3.7)	<0.001
≥1 MSP last four weeks^†
No	125 (52.7)	181 (67.5)
Yes	112 (47.3)	87 (32.5)	0.5 (0.4–0.8)	0.001
VS with NSP last four weeks^†
No	191 (80.6)	157 (58.6)	1
Protected	19 (8.0)	32 (11.9)	2.0 (1.1–3.8)	0.020
Unprotected	27 (11.4)	79 (29.5)	3.6 (2.2–5.8)	<0.001
VS with ASP last four weeks^†
No	147 (62.0)	105 (39.2)	1
Protected	28 (11.8)	30 (11.2)	1.5 (0.8–2.7)	0.187
Unprotected	62 (26.2)	133 (49.6)	3.0 (2.0–4.4)	<0.001
IAS with NSP last four weeks^†
No	168 (70.9)	229 (82.8)
Yes	69 (29.1)	46 (17.2)	0.5 (0.3–0.8)	0.002
IAS with ASP last four weeks^†
No	145 (61.4)	204 (76.4)
Protected	45 (19.1)	31 (11.6)	0.5 (0.3–0.8)	0.007
Unprotected	46 (19.5)	32 (12.0)	0.5 (0.3–0.8)	0.006
IOS with ASP last four weeks^†
No	66 (27.8)	64 (23.9)
Yes	171 (72.2)	204 (76.1)	1.2 (0.8–1.8)	0.309

OR: odds ratio; CI: confidence interval; PSHC: Parramatta Sexual Health Clinic; SSHC: Sydney Sexual Health Centre; ASP: any sexual partner; FSP: female sexual partner; IOS: insertive oral sex; IAS: insertive anal sex; MSP: male sexual partner; NSP: new sexual partner; VS: vaginal sex; Protected: 100% condom use; Unprotected: <100% condom use

*Circumcision status unknown for one control and two cases

^†Last four weeks and last three months refer to the period prior to presentation

Insertive oral sex with any partner in the last four weeks was common among both cases (76.1%) and controls (72.2%); there was no significant association with NGU (Table 1), and few men always used a condom (13/375; 3.5%). In contrast, insertive anal sex with a new or any partner was significantly less likely among cases than controls, with similar results for unprotected versus protected anal sex (Table 1). Circumcised men were significantly older than uncircumcised men (median age 36.0 versus 31.2 years, P < 0.001). More cases than controls were circumcised, but the difference was not significant.

Urethral microorganisms and NGU

Associations between selected study microorganisms and NGU are shown in Table 2.

Table 2

Association of selected urethral organisms* with NGU

	No. (%) of
Organism detected	Controls n = 237	Cases n = 268	OR (95% CI)	P value
Chlamydia trachomatis	7 (3.0)	50 (18.7)	7.5 (3.3–17.0)	<0.001
Mycoplasma genitalium	2 (0.8)	12 (4.5)	5.5 (1.2–24.9)	0.027
Ureaplasma urealyticum	20 (8.4)	33 (12.3)	1.5 (0.9–2.7)	0.158
Ureaplasma parvum	15 (6.3)	15 (5.6)	0.9 (0.4–1.8)	0.728
Haemophilus influenzae	12 (5.1)	19 (7.1)	1.4 (0.7–3.0)	0.361
Mycoplasma hominis	8 (3.4)	5 (1.9)	0.5 (0.2–1.7)	0.292
HSV-1	2 (0.8)	5 (1.9)	2.2 (0.4–11.6)	0.340

NGU: non-gonococcal urethritis; OR: odds ratio; CI: confidence interval; HSV: herpes simplex virus

*Organisms identified in >5 subjects

C. trachomatis was the most common organism detected, and had the strongest association with NGU. Associations between chlamydia and all the clinical indicators of urethritis were consistently highly significant (Table 3). Only seven (3.0%) C. trachomatis infections were asymptomatic.

Table 3

Associations between major pathogens and each clinical feature

Symptom or sign	Absent n (%)	Present n (%)	OR (95% CI)	P value
Dysuria
C. trachomatis positive	25 (7.3)	32 (19.6)	3.1 (1.8–5.4)	<0.001
M. genitalium positive	4 (1.2)	10 (6.1)	5.5 (1.7–17.9)	0.004
U. urealyticum positive	28 (8.2)	25 (15.3)	2.0 (1.1–3.6)	0.016
Urethral discharge (sign)
C. trachomatis positive	17 (4.5)	40 (31.0)	9.5 (5.1–17.5)	<0.001
M. genitalium positive	4 (1.1)	10 (7.8)	7.8 (2.4–25.4)	0.001
U. urealyticum positive	31 (8.2)	22 (17.1)	2.3 (1.3–4.1)	0.006
Urethral discharge (symptom)
C. trachomatis positive	20 (5.5)	37 (26.2)	6.1 (3.4–11.0)	<0.001
M. genitalium positive	8 (2.2)	6 (4.3)	2.0 (0.7–5.8)	0.215
U. urealyticum positive	33 (9.1)	20 (14.2)	1.7 (0.9–3.0)	0.095
Urethral irritation*
C. trachomatis positive	21 (6.9)	36 (18.0)	3.0 (1.6–5.2)	<0.001
M. genitalium positive	5 (1.6)	9 (4.5)	2.8 (0.9–8.5)	0.067
U. urealyticum positive	29 (9.5)	24 (12.0)	1.3 (0.7–2.3)	0.379

OR: odds ratio; CI: confidence interval

*Urethral irritation unknown for one case

The prevalence of M. genitalium was low (Table 2), but detection was also significantly associated with NGU, and with dysuria and urethral discharge on examination (Table 3). Twelve of 14 M. genitalium infections were symptomatic and 12 occurred in men with female sexual partners.

U. urealyticum was associated with NGU, but the association was not statistically significant in the univariate analysis (Table 2). However, detection of U. urealyticum was statistically significantly associated with urethral discharge on clinical examination and dysuria (Table 3).

U. parvum, M. hominis, H. influenzae and HSV-1 were found in similar proportions of cases and controls (Table 2). All other target organisms were identified in five or fewer subjects: HSV-2 in no cases or controls; adenovirus in four cases and no controls; S. pneumoniae and T. vaginalis each in one case; N. meningitidis in one case and one control and G. vaginalis in one case and two controls.

Independent association between U. urealyticum and NGU

Age and unprotected vaginal sex were associated with NGU in the univariate analysis (Table 1) and could potentially confound the association between U. urealyticum and NGU. Cases were older than controls, and compared with men who had no vaginal sex, U. urealyticum was statistically more likely to be detected in men with a history of unprotected but not protected vaginal sex in the last four weeks (OR 2.1, P = 0.022; OR 1.4, P = 0.481, respectively). To control for these potential confounders and determine whether U. urealyticum was independently associated with NGU, we performed a multivariate analysis. However, this analysis was restricted to the 424 men who did not have another urethral pathogen, including C. trachomatis, M. genitalium, HSV or adenovirus. This was because U. urealyticum was positively associated with NGU in men without another pathogen (OR 2.0, 95% CI: 1.1–3.8; P = 0.04), but less likely among the 81 men with another pathogen (OR 0.2; P = 0.07). Because of this interaction, we excluded men with another pathogen from the multivariate analysis.

After adjusting for age and unprotected vaginal sex in the last four weeks, U. urealyticum remained significantly associated with NGU in men without another pathogen detected (OR 2.0; 95% CI: 1.1–3.9; P = 0.03).

Association between microscopic urethritis and detection of microorganisms

Although 159/266 (59.8%) cases did not have ≥5 PMNL per HPF on urethral Gram stain, there was a significant association between NGU and microscopic urethritis (Table 4). Each individual symptom and sign was significantly associated with microscopic urethritis. While detection of C. trachomatis, U. urealyticum and M. genitalium was also associated with microscopic urethritis, the association was only statistically significant for C. trachomatis and U. urealyticum. When men with another urethral pathogen were excluded from the analysis, the association between U. urealyticum and microscopic urethritis remained and was in fact stronger than when mixed infections were also included (OR 2.6, 95% CI: 1.3–4.9; P = 0.008).

Table 4

Association between Gram stain findings, detection of major pathogens, NGU and symptoms/signs*

	<5 PMNL^† n (%)	≥5 PMNL^† n (%)	OR (95% CI)	P value
Chlamydia trachomatis detected	20 (5.6)	37 (27.0)	6.2 (3.4–11.1)	<0.001
Ureaplasma urealyticum detected	31 (8.8)	21 (15.3)	1.9 (1.0–3.4)	0.036
Mycoplasma genitalium detected	7 (2.0)	7 (5.1)	2.7 (0.9–7.8)	0.071
NGU	159 (44.9)	107 (78.1)	4.4 (2.8–6.9)	<0.001
Urethral discharge (symptom or sign)^‡	72 (20.3)	93 (67.9)	8.3 (5.3–12.9)	<0.001
Dysuria	83 (23.4)	79 (57.7)	4.4 (2.9–6.8)	<0.001
Urethral irritation^§	121 (34.3)	77 (56.2)	2.5 (1.6–3.7)	<0.001

NGU: non-gonococcal urethritis; OR: odds ratio; CI: confidence interval

*Gram stain results available for 491 subjects

^†PMNL = polymorphonuclear leukocytes in ≥5 oil-immersion microscopic fields of a Gram-stained urethral smear

^‡ P value for urethral discharge was also <0.001 for symptoms and signs separately

^§Urethral irritation unknown for one case

DISCUSSION

This study provides evidence supporting a role for U. urealyticum as a causative organism of NGU. An independent association between U. urealyticum and NGU was found among men without another pathogen detected. There were also statistically significant relationships between detection of U. urealyticum and urethral discharge on examination, dysuria and microscopic urethritis features of urethritis caused by recognized urethral pathogens. As expected, C. trachomatis and M. genitalium were detected significantly more often among cases than controls. However, the prevalence of M. genitalium was low, perhaps because this organism was more common among men with female rather than male sexual partners, and the study population included both groups. There was no evidence that the other oropharyngeal bacteria we studied (S. pneumoniae, H. influenzae and N. meningitidis) are associated with NGU. However, our study design may have led to underestimation of the role of N. meningitidis in NGU, as men with clinical features suggesting gonococcal infection were not enrolled.

It has been thought that male urethral infections with U. urealyticum are related to vaginal intercourse, as ureaplasmas are common vaginal organisms. Although the organism was detected among men whose sexual partners were reported to be exclusively male, we found that U. urealyticum was more likely to be associated with NGU in men with female sexual partners, especially those having unprotected vaginal sex. We did not study U. urealyticum subtypes in this study and the question of whether some strains are more virulent and more likely to cause symptoms than others remains.

Case definitions for NGU vary among reported studies. Few studies have separately examined the value of individual symptoms and signs in relation to detection of urethral pathogens. We found a significant association between microscopic urethritis and our case definition of NGU. However, 59% of cases did not have microscopic urethritis, and an association was not statistically significant for M. genitalium, although our urethral swab method may have led to underestimation of microscopic urethritis. Our results support the view that the absence of microscopic urethritis in men with clinical features of urethritis should not preclude the diagnosis of NGU.^7,11 The symptom of urethral discharge reported by the patient was not as robust a predictor of infection with either M. genitalium or U. urealyticum as dysuria, although it was the symptom most predictive of C. trachomatis infection. For all three organisms, the sign of urethral discharge on examination was the strongest indicator of infection, providing evidence that genital examination of men is a fundamental investigation.

Ureaplasma research now spans more than three decades, and although some questions have been answered, definitive evidence of a pathological role for U. urealyticum in NGU remains elusive. Our results indicate that further study of U. urealyticum in NGU is warranted, especially to examine the differential role of subtypes of U. urealyticum in heterosexual men with urethritis.

Footnotes

ACKNOWLEDGEMENTS

We would like to thank Basil Donovan, Katerina Lagios and Anna McNulty for their contributions and support of this research. We acknowledge the contributions of staff of the Centre for Infectious Diseases and Microbiology Laboratory Services, Sydney Sexual Health Centre, Parramatta Sexual Health Clinic and the men who participated in this study. Laboratory work was supported by a PhD scholarship through a capacity building infrastructure grant from NSW Health.

References

Robertson

, Stemke

, Davis

Jr , Proposal of Ureaplasma parvum sp. nov. and emended description of Ureaplasma urealyticum (Shepard et al. 1974) Robertson et al. 2001. Int J Syst Evol Microbiol 2002;52:587–97

Bowie

, Wang

S-P

, Alexander

, Etiology of nongonococcal urethritis. Evidence for Chlamydia trachomatis and Ureaplasma urealyticum . J Clin Invest 1977;59:735–42

Taylor-Robinson

. The history of nongonococcal urethritis: Thomas Parran Award Lecture. Sex Transm Dis 1996;23:86–91

Horner

, Thomas

, Gilroy

, Role of Mycoplasma genitalium and Ureaplasma urealyticum in acute and chronic nongonococcal urethritis. Clin Infect Dis 2001;32:995–1003

Povlsen

, Bjornelius

, Lidbrink

, Relationship of Ureaplasma urealyticum biovar 2 to nongonococcal urethritis. Eur J Clin Microbiol Infect Dis 2002;21:97–101

Deguchi

, Yoshida

, Miyazawa

, Association of Ureaplasma urealyticum (biovar 2) with nongonococcal urethritis. Sex Transm Dis 2004;31:192–5

Bradshaw

, Tabrizi

, Read

TRH

, Etiologies of nongonococcal urethritis: bacteria, viruses, and the association with orogenital exposure. J Infect Dis 2006;193:336–45

Taylor-Robinson

, Csonka

, Prentice

. Human intra-urethral inoculation of ureaplasmas. Q J Med 1977;46:309–26

Yoshida

, Ishiko

, Yasuda

, Polymerase chain reaction-based subtyping of Ureaplasma parvum and Ureaplasma urealyticum in first-pass urine samples from men with and without urethritis. Sex Transm Dis 2005;32:454–7

10.

Haddow

, Bunn

, Copas

, Polymorph count for predicting non-gonococcal urethral infection: a model using Chlamydia trachomatis diagnosed by ligase chain reaction. Sex Transm Infect 2004;80:198–200

11.

Handsfield

. Nongonococcal urethritis: a few answers but mostly questions. J Infect Dis 2006;193:333–5

12.

McKechnie

, Hillman

, Couldwell

, Simultaneous identification of 14 genital micro-organisms in urine using a multiplex PCR-based reverse line blot (mPCR-RLB) assay. J Clin Microbiol 2009;47:1871–7

13.

Kong

, Gilbert

. Multiplex PCR-based reverse line blot hybridisation assay (mPCR-RLB) – a practical epidemiological and diagnostic tool. Nat Protoc 2006;1:2668–80