336 • JID 2006:193 (1 February) • Bradshaw et al.
M A J O R A R T I C L E
Etiologies of Nongonococcal Urethritis:
Bacteria, Viruses, and the Association
with Orogenital Exposure
Catriona S. Bradshaw,1,2Sepehr N. Tabrizi,3,4Timothy R. H. Read,1Suzanne M. Garland,3,4Carol A. Hopkins,1
Lorna M. Moss,1and Christopher K. Fairley1,2
1Melbourne Sexual Health Centre, The Alfred Hospital,
of Melbourne, and
2School of Population Health and
3Department of Obstetrics and Gynaecology, University
4Department of Microbiology and Infectious Diseases, The Royal Women’s Hospital, Victoria, Australia
(See the editorial commentary by Handsfield, on pages 333–5.)
The purpose of the present study was to determine pathogens and behaviors associated with
nongonococcal urethritis (NGU) and the usefulness of the urethral smear in predicting the presence of pathogens.
We conducted a case-control study of men with and without symptoms of NGU. Sexual practices
were measured by questionnaire. First-stream urine was tested for Chlamydia trachomatis, Mycoplasma genitalium,
Ureaplasma parvum, U. urealyticum, herpes simplex virus (HSV)–1, HSV-2, adenoviruses, and Gardnerellavaginalis
by polymerase chain reaction.
C. trachomatis (20%), M. genitalium (9%), adenoviruses (4%), and HSV-1 (2%) were more common
in cases with NGU () after age and sexual risk were adjusted for (n p 329
and G. vaginalis were not. Infection with adenoviruses or HSV-1 was associated with distinct clinical features, oral
sex, and male partners, whereas infection with M. genitalium or C. trachomatis was associated with unprotected
vaginal sex. Oral sex was associated with NGU in which no pathogen was detected (
polymorphonuclear leukocytes (PMNLs) per high-power field (HPF) on urethral smear were present in 32%, 37%,
38%, and 44% of cases with C. trachomatis, M. genitalium, adenoviruses, and HSV, respectively.
We identified adenoviruses and HSV-1 as significant causes of NGU with distinct clinical and
behavioral characteristics and highlighted the association between insertive oral sex and NGU. A urethral PMNL
count of ?5 PMNLs/HPF is not sufficiently sensitive to exclude pathogens in men with urethral symptoms.
); U. urealyticum, U. parvum,P ? .01
). Fewer than 5P ? .001
Acute nongonococcal urethritis (NGU) is one of the
commonest sexually transmitted infections affecting
men, yet a pathogen is not identified in a significant
proportion of cases (20%–50%). Chlamydia tracho-
matis typically accounts for 30%–50% of cases of NGU
[1, 2], and Mycoplasma genitalium accounts for 10%–
30% [2, 3]; Ureaplasma urealyticum, Haemophilus spe-
cies, Streptococcus species, and Gardnerella vaginalis
have been associated with NGU, but their role is un-
Received 2 June 2005; accepted 3 August 2005; electronically published 28
Potential conflicts of interest: none reported.
Financial support: C.S.B. holds a National Medical and Research Council
Research Scholarship. No external sources of funding were used to support this
Reprints or correspondence: Dr. Catriona S. Bradshaw, Melbourne Sexual Health
Centre, 580 Swanston St., Carlton, 3053, Victoria, Australia (cbradshaw@mshc
The Journal of Infectious Diseases
? 2005 by the Infectious Diseases Society of America. All rights reserved.
proven [4, 5]. Few studies have investigated potential
viral causes of acute NGU, such as herpes simplex virus
(HSV) [6–8] and adenoviruses [9–12].
The diagnosis of NGU has traditionally required mi-
croscopic evidence of urethritis, defined as ?5 poly-
morphonuclear leukocytes (PMNLs) per high-power
field (HPF) (?1000 magnification) in ?5 fields of a
urethral Gram stain. Urethral pathogens, however, are
often detected in symptomatic men with !5 PMNLs/
HPF [13–16], particularly when sensitive nucleic acid
amplification tests are used. Studies of NGU frequently
restrict enrollment to men with urethral discharge or
dysuria and ?5 PMNLs/HPF, and the prevalence of
urethral pathogens in such populations has been well
described. However, a significant proportion of men
with urethral symptoms do not have microscopic evi-
dence of urethritis or experience other urethral symp-
toms, including burning or irritation. We conducted a
case-control study of men with symptoms of NGU—
regardless of urethralPMNLcount—andasymptomatic
Viral and Bacterial Causes of NGU • JID 2006:193 (1 February) • 337
control subjects, to establish the prevalence of known and sus-
pected bacterial and viral pathogens, to examine their associ-
ations with sexual practices, and to investigate the usefulness
of the urethral smear cutoff of ?5 PMNLs/HPF in predicting
the presence of pathogens.
SUBJECTS AND METHODS
Centre (MSHC), Australia, between March 2004 and March
2005 were eligible for enrollment. MSHC, the principle public
sexual health service in Melbourne, provided 10,122 consul-
tations to men during this period. Cases included men with
urethral symptoms (discharge, dysuria, and urethral burning
or irritation) for ?1 month, and controls included men with
no current urethral symptoms. Men with urethral gonorrhoea
(gram-negative intracellular diplococci on urethral Gram stain
and/or N. gonorrhoeae on culture) or genital herpeswithlesions
were excluded at enrollment or on chart review. Clinicianswere
asked to recruit all eligible cases, starting in March 2004; triage
nurses marked files of eligible patients; and a computerized
study reminder activated with the diagnosis of acute NGU.
There was no electronic reminder for controls, but enrollment
was coordinated to match recruitment of cases with monthly
e-mail reminders to the staff.
Clinical and laboratory methods.
a questionnaire regarding symptoms (rating scale, 1–5) and
sexual and behavioral practices and underwent genital exam-
ination. Clinicians recorded clinical and laboratory findings on
standardized data-collection sheets. Cases underwent aurethral
of poor acceptability of the procedure in asymptomatic men.
All participants provided a first-stream urine specimen and
recorded time since last voiding. Two experienced members of
the on-site laboratory staff evaluated urethral Gram-stained
smears from cases for PMNLs in 5 oil-immersion fields(?1000
magnification). Cases were treated presumptively with 1 g of
First-stream urine from all participants was tested for C.
trachomatis by strand-displacement amplification (ProbeTec-
ETCT-Amplified DNA-Assay; Becton Dickinson) and for M.
genitalium, HSV-1 and -2, Trichomonas vaginalis, U. urealyti-
cum, U. parvum, G. vaginalis, and adenoviruses by polymerase
chain reaction (PCR). One milliliter of urine was centrifuged,
and the pellet was extracted using the automated MagNA Pure
LC (Roche) with the DNA Isolation Kit I protocol. Extracted
DNA was amplified by PCR for 7 targets (5-mL aliquots of
DNA for each reaction). Amplification and detection was per-
formed for the b-globin gene (as a positive control) , M.
genitalium , T. vaginalis , U. urealyticum and U.parvum
, G. vaginalis , adenoviruses, and HSV, using an ad-
aptation of a previously described assay targeting a 239-bp
sequence of the glycoprotein D gene and a 206-bp sequence
of the hexon gene [22, 23]. All PCR assays have been shown
to be highly specific for the target amplified, with analytical
sensitivity of 10 copies/reaction (200 copies/mL of urine). Ad-
enovirus serovars were determined by sequence comparison
between amplicons from positive samples and known nucle-
otide sequences in the GenBank database, using the BLAST
Cases were tested for Neisseria gonorrhoeae by culture in
modified Thayer-Martin medium. Controls were not screened
for urethral N. gonorrhoeae, because of the very low prevalence
of N. gonorrhoeae in asymptomatic men . The Human Re-
search and Ethics Committee of the Alfred Hospital, Victoria,
approved this study.
Data were entered and stored in Mi-
crosoft Access and analyzed using SPSS (version 12; SPSS).
Clinical, behavioral, and laboratory findings in cases werecom-
pared with those in controls. Proportions were comparedusing
x2and Fisher’s exact tests; unpaired t tests, analysis of variance,
and nonparametric tests (Mann-Whitney U test and Kruskal-
Wallis test) were used for continuous variables, where appro-
priate. Univariate analysis was used to calculate crude odds
ratios (ORs) and 95% confidence intervals (CIs). Logistic re-
gression was used to control for confounding factors. Variables
included in the model were those found to be significant in
the univariate analysis (P ! .05
important on the basis of published literature; variables were
examined for multicolinearity. Patients were excluded from the
analysis when clinical information or specimens were not avail-
able. With 300 cases and controls, this study has 80% power
to detect an OR of ?1.7 for factors present in 20% of controls.
) and those considered to be
Characteristics of the study population.
thirty-six heterosexual men and men who have sex with men
(MSMs) with aage of mean?SD
rolled; 329 had urethral symptoms (cases) and 307 did not
(controls). During the study period, 479 men with acute NGU
attended MSHC, 69% of whom were enrolled; those not en-
rolled included men who were not eligible, declined, or were
not asked to participate. Controls were enrolled during the
same period as cases; however, we could not determine their
participation rate, because the study definition of a control
(male without urethral symptoms) is not captured in the clinic
Clinical features in cases included a history of acute urethral
discharge (62%), dysuria (73%), urethral itch (56%), and ure-
thral burning (62%), and, on examination, urethral discharge
macroscopically evident mucous threads in first-stream urine
(34%). Thetime since last urine passage for themean?SD
Six hundred and
years were en-32.3?9.1
338 • JID 2006:193 (1 February) • Bradshaw et al.
Demographic and behavioral characteristics of the
No. (%) of
n p 307()
n p 329()
Age 131 years
NSP in past month
RSP in past month
11 FSP in past month
?1 MSP in past month
UPVS with CSP in past month
IOS with CSP in past month
UPAS with CSP in past month
partner; IOS, !100% condom use for insertive oral sex; MSP , male sex partner;
NSP , new sex partner; OR, odds ratio; RSP , regular sex partner; UPAS, !100%
condom use for insertive anal sex; UPVS, !100% condom use for vaginal sex.
“Past month” denotes the month before the onset of symptoms.
CI, confidence interval; CSP , casual sex partner; FSP , female sex
study population was
tween cases and controls (
C. trachomatis, M. genitalium, HSV-1, and adenoviruses were
detected more frequently in cases than controls (
2). C. trachomatis occurred in 20% (95% CI, 15%–24%) of
cases, and M. genitalium occurred in 9% (95% CI, 7%–13%),
whereas adenoviruses (4% [95% CI, 2%–6%]) and HSV-1 and
-2 (3% [95% CI, 1%–5%]) were significant but less common
causes of NGU. T. vaginalis occurred in only 1 case. Herpes
serologic testing was not performed; 1 case with HSV-2 had a
history of genital herpes. One case was coinfected with C. tra-
chomatis and adenovirus, but no other participant had 11 ure-
thral pathogen. G. vaginalis, U. parvum, and U. urealyticum
(biovar 2) were detected more frequently in controls, although
this difference was significant only for G. vaginalis and U. par-
vum ( ). Associations between organisms and NGU re-
P ! .01
mained unchanged after age and unprotected vaginal, oral, and
anal sex with a casual partner in the past month were adjusted
for, with the exception of U. parvum, which nolongerremained
associated with being a control. Serotypes of 9 adenovirussam-
ples were available: 6 were subgenus D (3 serotype 9 and 3
serotype 37), 2 were subgenus B2 (serotype 35), and 1 was
subgenus E (serotype 4).
Clinical characteristics of cases with and cases without ure-
Table 3 shows the clinical characteristics of
cases infected with specific urethral pathogens (hereafter, the
wording “with a pathogen” indicates infection with C. tra-
chomatis, M. genitalium, T. vaginalis, HSV-1 or -2, or adeno-
viruses but not with G. vaginalis or ureaplasmas). Self-reported
urethral discharge, itch, burning, balanitis, epididymitis, or ?5
PMNLs/HPF on urethral Gram stain were not associated with
specific pathogens. Cases with C. trachomatis infection were
more likely to have macroscopically evident mucous threads in
first-stream urine than were other cases (
Cases with HSV or adenoviruses were significantly more
likely than other cases to present with meatitis and moderate
to severe dysuria: 20 (91%) had meatitis, 16 (73%) had mod-
erate to severe dysuria, and 15 (68%) had both findings, in
contrast to 33%, 30%, and 11% of cases without viral patho-
gens, respectively (). Meatitis was a sensitive indicator
P ! .001
of the presence of viral pathogens in NGU (91%), with a high
negative predictive value (99%), but was not specific (67%).
Moderate to severe dysuria had a high negative predictivevalue
(97%) but was not specific (71%) or sensitive (73%). The com-
bination of both clinical characteristics had a high specificity
(89%) and negative predictive value (98%) for the presence of
viral pathogens but was not sensitive (68%), and it had a pos-
itive predictive value of 31%. Conjunctivitis was reported in 4
cases (39%) with adenovirus but not in cases with other path-
h; there was no difference be-
). Table 1 describes the de-
P p .85
P ! .01
P ! .001
ogens. Adenovirus infections were seasonally clustered; 8 oc-
curred in autumn (comprising 8% of all cases in autumn), 4
occurred in winter/early spring, and 1 occurred in summer
( ).P ! .01
Relationship between urethral PMNL count and detection of
Microscopic evidence of urethritis was re-
lated to the presence of urethral discharge on history (OR, 3.6
[95% CI, 2.2–6.0]) and examination (OR, 11.3 [95% CI, 6.3–
20.4]) but was not associated with other clinical characteristics.
A urethral PMNL count of ?5 PMNLs/HPF was present in 133
cases (42% [95% CI, 36%–47%]), including 41 (68%) caseswith
C. trachomatis, 17 (63%) with M. genitalium, 8 (62%) with ad-
enoviruses, and 5 (56%) with HSV. C. trachomatis(OR,3.9[95%
CI, 2.1–7.1]) and M. genitalium (OR, 2.6 [95% CI, 1.1–5.8])
were significantly associated with ?5 PMNLs/HPF on urethral
Gram stain. A urethral pathogen (C. trachomatis, M. genitalium,
adenoviruses, or HSV)wasdetectedin20%(95%CI,15%–27%)
of men with urethral symptoms but !5 PMNLs/HPF, which was
significantly more common than in asymptomatic controls (4%
[95% CI, 2%–7%]) ().P ! .001
Viral and Bacterial Causes of NGU • JID 2006:193 (1 February) • 339
Table 2.Urethral organisms associated with nongonococcal urethritis.
No. (%) of
OR (95% CI)
n p 307()
n p 329() Crude Adjusted
sex with a casual sex partner in the past month; and infection with C. trachomatis, M. genitalium,
adenoviruses, U. parvum, or G. vaginalis. Data on 617 patients included 19 missing from the multi-
variable model. One case and 3 controls were coinfected with U. parvum and U. urealyticum. Coin-
fection with adenoviruses, C. trachomatis, M. genitalium, and HSV-1 occurred in 15 cases with U.
urealyticum and 3 cases with U. parvum. CI, confidence interval; HSV, herpes simplex virus; NS, not
significant; OR, odds ratio; S, significant.
aOne case was coinfected with C. trachomatis and adenoviruses.
bMen with visible lesions consistent with genital herpes were excluded from the study.
Variables in the adjusted analysis included age; unprotected vaginal sex, oral sex, and anal
Behavioral associations with urethral organisms.
shows the prevalence of urethral organisms in participants,
according to sexual preference. Cases with HSV-1 or adeno-
viruses were more likely to have had male sex partners during
the past month ( ), whereas cases with C. trachomatisP p .03
or M. genitalium were more likely to be heterosexual (P p
); U. parvum and G. vaginalis were also significantly more .07
common in heterosexual cases and controls (
To determine whether urethralpathogenswereassociatedwith
CI, 1.5–4.6]) and M. genitalium (OR, 2.5 [95% CI, 1.1–5.3]),
).P ? .02
insertive oral sex with a casual partner was associated with ad-
enoviruses (OR, 3.9 [95% CI, 1.1–14.6]) and C. trachomatis(OR,
2.1 [95% CI, 1.2–3.6]) and was more common in men with
HSV-1(71%vs.46%; ),andanalsexwithacasualpartnerP p .18
(OR, 3.1 [95% CI, 1.0–9.8]) was associated with adenoviruses.
partner were adjusted for, C. trachomatis (adjusted OR, 2.3[95%
CI, 1.2–4.3]) and M. genitalium (adjusted OR, 2.4 [95% CI, 1.1–
5.2]) remained associated with unprotected vaginal sex with a
casual partner. An adjusted analysis was performed for eachviral
pathogen, but numbers were small, and behavioral associations
were not significant.
Table 5 shows the behavioral associations in pathogen-pos-
340 • JID 2006:193 (1 February) • Bradshaw et al.
Table 3. Clinical features associated with specific organisms in men with nongonococcal urethritis ().n p 328
n p 63
(n p 31)
n p 13
HSV-1 and -2
(n p 9)
(n p 212)P
Age 131 years
Moderate to severe
Self-reported urethral discharge
Urethral discharge on examination
Nature of urethral discharge
?5 PMNLs/HPF on urethral Gram stain
26 (41) 20 (65)7 (54) 3 (33)131 (62).03
aOne case with adenovirus was coinfected with C. trachomatis and is included only in the adenoviruses column.
b“Pathogen-negative” means that cases with C. trachomatis, M. genitalium, adenoviruses, or HSVs were excluded from the analysis.
Data are no. (%) of cases, unless otherwise indicated. HPF , high-power field; HSV, herpes simplex virus; PMNLs, polymorphonuclear leukocytes.
itive and pathogen-negative cases compared with pathogen-
negative controls. Pathogen-positive and pathogen-negative
cases were significantly more likely to have engaged in unpro-
tected vaginal and insertive oral sex with a casual partner than
were pathogen-negative controls. After age and unprotected
vaginal, oral, and anal sex with a casual partner were adjusted
for, casual unprotected vaginal sex remained associated with
pathogen-positive cases, whereas insertive oral sex and, to a
lesser extent, unprotected vaginal sex with a casual partnerwere
associated with pathogen-negative cases. Exclusion of menwho
practiced unprotected vaginal or anal sex from the analysis
strengthened the association between pathogen-negative NGU
and casual insertive oral sex (OR, 8.5 [95% CI, 3.6–20.5]).
To examine behavioral associations with organisms not as-
sociated with NGU, controls with G. vaginalis were compared
with controls without G. vaginalis. Controls with G. vaginalis
were more likely to have had unprotected vaginal sex within
14 days (median, 14 days), to have had a greater number of
female partners, and to have U. parvum and U. urealyticum
(table 6). U. parvum had similar behavioral associations but
was negatively associated with U. urealyticum. In the adjusted
analyses, both G. vaginalis and U. parvum remained morelikely
to be detected in the urine of asymptomatic men within 14
days of unprotected vaginal sex.
In this study of men with urethral symptoms, adenovirusesand
HSV-1 were found to be significantly associated with NGU and
to have different clinical and behavioral associations with C.
trachomatis and M. genitalium. Adenoviruses and HSV-1 were
associated with sex with men and with insertive oralsex,whereas
C. trachomatis and M. genitalium were associated with sex with
women and unprotected vaginal sex. Insertive oral sex was
significantly associated with NGU in which no pathogen was
detected. T. vaginalis and HSV-2 were uncommon causes of
NGU in this population, and U. urealyticum, U. parvum, and
G. vaginalis were not associated with NGU. To our knowledge,
this is the first case-control study of NGU to examine a broad
range of viral and bacterial pathogens in heterosexual men and
MSMs by use of sensitive molecular techniques and to dem-
onstrate a significant association of HSV-1 and adenoviruses
Few studies have examined the prevalence and role of ade-
noviruses as causative agents of urethritis. Subgenus D ade-
noviruses have been reported to manifest an affinity for the
eye and genital tract  and are an established cause of ker-
atoconjunctivitis [12, 26–29]. We previously reported a series
of NGU cases associated with marked dysuria, meatitis, con-
junctivitis, and constitutionalsymptoms,inwhichadenoviruses
were isolated . Insertive oral sex precededonsetofsymptoms
in all cases, in 4 cases the adenoviruses were serotyped and
found to be subgenus D, and there was seasonal clustering
(autumn to spring). Azariah and Reid reported similar asso-
ciations in 6 MSMs who had adenovirus-associated urethritis
, and Harnett et al.  and Swenson et al.  reported
adenoviruses to be uncommon causes of urethritis and to be
associated with conjunctivitis. These studies used viral cultures,
and the majority of the adenoviruses were subgenus D sero-
types. Our data indicate an association between adenoviruses
and insertive oral sex and, possibly, between adenoviruses and
insertive anal sex, although the latter rarely occurred in the
absence of unprotected oral sex. Adenoviruses have uncom-
Viral and Bacterial Causes of NGU • JID 2006:193 (1 February) • 341
Table 4. Prevalence of organisms in cases and controls ().n p 635
(n p 220)
n p 109()P
(n p 231)
n p 75()P
who have sex with men.
Data are no. (%) of subjects, unless otherwise indicated. HSV, herpes simplex virus; MSM, men
monly been isolated from the genital tract in women [12, 29–
31], indicating that vaginal transmission could occur. Despite
limited recognition in the past, adenoviruses—particularlythose
of subgenus D—appear able to be sexually transmitted, asso-
ciated with insertive oral sex, and capable of causing a distinct
clinical syndrome in men with NGU [9, 10].
Our data suggest that HSVs are a cause of NGU in the
absence of visible herpetic lesions and that HSV-1 may be a
more common cause of NGU than HSV-2. Both viruses can
ever, HSV-2 has infrequently been reported as a cause of NGU
without lesions [6–8], and there have been few studies of HSV-
1 in NGU . It is possible that we detected shedding of HSV
in the presence of another urethral pathogen; however, the
strong association with symptoms of urethritis, the similarity
of clinical characteristics with those in cases with adenoviruses,
and the absence of other known coinfections and of HSV in
of urethral symptoms in these cases. The association between
HSV-1, insertive oral sex, and sex with men may exist because
oral sex is a particularly common exposure in MSMs . As
HSV-1 becomes responsible for an increasing proportion of
genital herpes [33, 34], it may become a more significant cause
of NGU in populations in which oral sex is commonly prac-
ticed. The characteristic clinical features of NGU found to be
caused by viruses in the present study may mean that these
findings could be used clinically to distinguish viral from bac-
terial causes of urethritis, although the predictive value of these
signs is dependent on the prevalence of viral pathogens in the
Unprotected insertive oral and vaginal sex with a casualpart-
ner was associated with pathogen-negative NGU, and the as-
sociation between pathogen-negative NGU and insertive oral
sex was greater in men not practicing unprotected vaginal or
anal sex. Unprotected anal sex was not a risk factor for path-
ogen-negative NGU in this study. Few studies have addressed
the role of insertive oral sex in NGU. Insertive oral sex has
been found to be a risk factor in several [35–37], but not all,
studies that have examined the association between oral sex
and NGU . Insertive oral sex is recognized as a mode of
transmission for N. gonorrhoeae, Streptococcus pneumoniae
and N. meningitis [40–43] have been isolated in patients with
NGU after oral sex, and Haemophilus influenzae and H. par-
ainfluenzae have been cultured in patients with NGU .
These findings support transmission of oropharyngeal patho-
gens or commensal oral flora via oral sex to the urethra. Such
organisms may be responsible for a greater proportion of ure-
thral symptoms than has been previously recognized. Ourfind-
ings regarding both adenoviruses and HSV-1 support this
premise and highlight the need for further studies of oro-
pharyngeal viruses and bacteria in NGU.
U. urealyticum (biovar 2), U. parvum, and G. vaginalis was
not found to be associated with NGU in the present study.
Human and animal inoculation studies and some case-control
studies [45, 46] have reported an association between U. ureal-
yticum and NGU, with biovar 2 most recently being implicated
[47, 48]. However, U. urealyticum is common in the genital
tract of sexually active men and women and is associated with
increased numbers of partners , which poses difficulties for
determining its role in NGU. It has been proposed that symp-
toms are related to higher bacterial loads [50, 51], specific
serotypes of biovar 2 [52, 53], or initial exposure  but
subside with colonization. Although we found no association
between biovar 2 and NGU, we did not perform serotyping or
Neither U. parvum nor G. vaginalis were found to be as-
sociated with urethral symptoms in the present study but were
more likely to be detected within 14 daysof unprotectedvaginal
sex. Because both organisms form part of the normal vaginal
flora , detection may be a consequence of transient colo-
nization of the male urethra after vaginal sex. G. vaginalis has
been reported as being more common in asymptomatic men
342 • JID 2006:193 (1 February) • Bradshaw et al.
Table 5. Behavioral associations for cases with and without urethral pathogens, compared with controls without urethral pathogens.
controls, no. (%)
(n p 294)
Pathogen-positive cases ()n p 117
Pathogen-negativeacases ()n p 212
(95% CI) No. (%)
Age 131 years
11 FSP in past month
?1 MSP in past month
RSP in past month
CSP in past month
UPVS with CSP in past month
IOS with CSP in past month
UPAS with CSP in past month
oral sex, and anal sex with a casual sex partner (CSP). Data on as many as 12 subjects were missing in the adjusted analysis. CI, confidence interval; FSP , female
sex partner; IOS, !100% condom use for insertive oral sex; MSP , male sex partner; OR, odds ratio; RSP , regular sex partner; SP , sex partner; UPAS, !100%
condom use for insertive anal sex; UPVS, !100% condom use for vaginal sex. “Past month” denotes the month before the onset of symptoms.
a“Pathogen negative” means that subjects with Chlamydia trachomatis, Mycoplasma genitalium, Trichomonas vaginalis, herpes simplex virus, or adenoviruses
were excluded from the analysis.
Data on as many as 10 subjects were missing in the univariate analysis. Variables in the adjusted analysis included age and unprotected vaginal sex,
than in men with NGU , but it has also been isolated from
men with urethritis [5, 56]. Although G. vaginalis is a com-
mensal of the female genital tract, in increased numbers it is
associated with bacterial vaginosis (BV), and an association
between NGU and BV has been reported . We may have
failed to detect an association between G. vaginalis and NGU;
however, if an association with BV and NGU exists, it is also
possible that it is with another BV-associated organism that is
not commonly part of the normal vaginal flora.
Laboratory evidence of urethritis was associated with the
presence of urethral discharge and was predictive of the pres-
enceof C. trachomatis andM.genitalium,butusing?5PMNLs/
HPF as the sole criterion for urethritis in our study would have
resulted in a failure to identify a significant proportion of path-
ogens in cases (C. trachomatis [32%], M. genitalium [37%],
adenoviruses [38%], and HSV [44%]). Previous studies have
also reported that a significant proportion of men with NGU
and urethral pathogens have !5 PMNLs/HPF [13–16]. Al-
though ?5 PMNLs/HPF on urethral Gram stain has been
shown to be associated with isolation of C. trachomatis and N.
gonorrhoeae, the sensitivity for C. trachomatis in some studies
has been as low as 29%  to 63% . A urethral PMNL
count of ?5 PMNLs/HPF in our study was not useful for
predicting the presence of a pathogen in the absence of urethral
discharge, a finding also reported by Janier . Therefore, in
studies of NGU in which enrollment is not limited to those
with urethral discharge, a weaker association between NGU
and the urethral Gram stain may be reported. Although the
urethral Gram stain may provide objective evidence of inflam-
mation, it appears that it cannot reliably be used to exclude a
urethral pathogen in NGU.
Our study did not have sufficient power to examine all be-
havioral risk factors for urethral pathogens, and studies with
greater numbers would be helpful in exploring these associa-
tions. Recruitment on the basis of symptoms may have resulted
in the enrollment of men with symptoms that had noninfec-
tious causes, which may have reduced theprevalenceofurethral
pathogens. However, pathogen-negative cases reported greater
sexual risks than did controls, had symptoms similar to those
in other cases, and had a significant association with casual
Viral and Bacterial Causes of NGU • JID 2006:193 (1 February) • 343
(n p 294
Associations with detection of Gardnerella vaginalis and Ureaplasma parvum in the urine of pathogen-negative controls
G. vaginalisU. parvum
(n p 204)
(n p 90)
(n p 251)
(n p 43)
Any UPVS in past month
?14 days ago
114 days ago
Any oral sex in past month
Any UPAS in past month
U. urealyticum (biovar 2)
………… 185 (74)
denotes the month before the onset of symptoms. “Pathogen negative” means that controls with C. trachomatis, M. genitalium, or adenoviruses (
excluded from the analysis. Median time since last UPVS in controls, 14 days. Data on as many as 5 subjects were missing in the univariate analysis. Variables
in the adjusted analysis included age, UPVS with any partner within 14 days, any oral sex, any UPAS in past month, and U. urealyticum. U. parvum was included
in the analysis for G. vaginalis. Data on 290 subjects were included in the adjusted analysis, and data on 4 were incomplete and, therefore, excluded.
CI, confidence interval; OR, odds ratio; UPAS, !100% condom use for insertive anal sex; UPVS, !100% condom use for vaginal sex. “Past month”
) werenp 12
unprotected oral and vaginal sex, suggesting that infectionwith
potential pathogens not identified in this study is likely. Al-
though cases represented 69% of men with acute NGU who
attended MSHC, we were not able to determine the enrollment
rate for controls. We consider it unlikely that there was a sig-
nificant selection bias, but it is possible that controls were not
representative of all eligible men attending MSHC. A urethral
Gram stain was not performed in controls, because of the poor
acceptance of the urethral smear in asymptomatic men. Astudy
of NGU at MSHC in 2003 (n p 160
of asymptomatic controls had microscopic evidence of ure-
thritis, which was not associated with the detection of urethral
pathogens. It is possible that the inclusion of controls with
asymptomatic urethritis may have reduced our ability to detect
clinical and behavioral associations with NGU.
In conclusion, we have identified adenoviruses and HSV-1
as significant causes of NGU, in addition to C. trachomatis and
M. genitalium. We found that distinctive clinical features were
associated with viral pathogens and that insertive oral sex and
male partners were associated with adenoviruses and HSV-1 in
)  showed that 18%
the crude analysis. Insertive oral sex was also a significant risk
factor in pathogen-negative NGU. These findings raise the po-
tential importance of the oropharynx as a significant source of
bacterial and viral pathogens and indicate that we should
broaden our search for pathogens in NGU. A urethral PMNL
count of ?5 PMNLs/HPF does not appear to be sufficiently
sensitive to excludeurethralinfectioninNGU.Ourdatasuggest
that treatment decisions are best based on clinical features of
urethritis and not solely on microscopic assessment.
We would like to thank all practitioners at Melbourne Sexual Health
Centre for recruitment of participants; Leonie Horvath, Irene Kuzevska,
and Mary Santoro, for their significant contributiontothestudy;andJames
Beeson for his review of the manuscript.
1. Horner PJ, Thomas B, Gilroy CB, Egger M, Taylor-Robinson D. Do
all men attending departments of genitourinary medicine need to be
344 • JID 2006:193 (1 February) • Bradshaw et al.
screened for non-gonococcal urethritis? Int J STD AIDS 2002;13:
2. Horner P, Thomas B, Gilroy CB, Egger M, Taylor-Robinson D. Role
of Mycoplasma genitalium and Ureaplasma urealyticum in acute and
chronic nongonococcal urethritis. Clin Infect Dis 2001;32:995–1003.
3. Taylor-Robinson D. Mycoplasma genitalium—an up-date. Int J STD
4. Schwartz MA, Hooton TM. Etiology of nongonococcal nonchlamydial
urethritis. Dermatol Clin 1998;16:727–33, xi.
5. Lefevre JC, Lepargneur JP, Bauriaud R, Bertrand MA, Blanc C. Clinical
and microbiologic features of urethritis in men in Toulouse, France.
Sex Transm Dis 1991;18:76–9.
6. Uuskula A, Raukas E. Atypical genital herpes: reportoffivecases.Scand
J Infect Dis 2004;36:37–9.
7. Srugo I, Steinberg J, Madeb R, et al. Agents of non-gonococcal ure-
thritis in males attending an Israeli clinic for sexually transmitted dis-
eases. Isr Med Assoc J 2003;5:24–7.
8. Holmes KK, Handsfield HH, Wang SP, et al. Etiologyofnongonococcal
urethritis. N Engl J Med 1975;292:1199–205.
9. Bradshaw CS, Denham IM, Fairley CK. Characteristics of adenovirus
associated urethritis. Sex Transm Infect 2002;78:445–7.
10. Azariah S, Reid M. Adenovirus and non-gonococcal urethritis. Int J
STD AIDS 2000;11:548–50.
11. Harnett GB, Phillips PA, Gollow MM. Associationofgenitaladenovirus
infection with urethritis in men. Med J Aust 1984;141:337–8.
12. Swenson PD, Lowens MS, Celum CL, Hierholzer JC. Adenovirus types
2, 8, and 37 associated with genital infections in patients attending a
sexually transmitted disease clinic. J Clin Microbiol 1995;33:2728–31.
13. Landis SJ, Stewart IO, Chernesky MA, et al. Value of the gram-stained
urethral smear in the management of men with urethritis. Sex Transm
14. Janier M, Lassau F, Casin I, et al. Male urethritis with and without
discharge: a clinical and microbiologicalstudy.SexTransmDis1995;22:
15. Haddow LJ, Bunn A, Copas AJ, et al. Polymorph count for predicting
non-gonococcal urethral infection: a model using Chlamydia tracho-
matis diagnosed by ligase chain reaction. Sex Transm Infect 2004;80:
16. Desai K, Robson HG. Comparison of the Gram-stained urethralsmear
and first-voided urine sediment in the diagnosis of nongonococcal
urethritis. Sex Transm Dis 1982;9:21–5.
17. Tabrizi SN, Chen S, Cohenford MA, et al. Evaluation of real-time
polymerase chain reaction assays for confirmation of Neisseria gon-
orrhoeae in clinical samples tested positive in the Roche Cobas Am-
plicor assay. Sex Transm Infect 2004;80:68–71.
18. Yoshida T, Deguchi T, Ito M, Maeda S, Tamaki M, Ishiko H. Quan-
titative detection of Mycoplasma genitalium from first-pass urine of
men with urethritis and asymptomatic men by real-time PCR. J Clin
19. Sullivan EA, Abel M, Tabrizi S, et al. Prevalence of sexuallytransmitted
infections among antenatal women in Vanuatu, 1999-2000.SexTransm
20. Yoshida T, Maeda S, Deguchi T, Miyazawa T, Ishiko H. Rapiddetection
of Mycoplasma genitalium, Mycoplasma hominis, Ureaplasma parvum,
and Ureaplasma urealyticum organisms in genitourinary samples by
PCR-microtiter plate hybridization assay. J Clin Microbiol 2003;41:
21. Zariffard MR, Saifuddin M, Sha BE, Spear GT. Detection of bacterial
vaginosis-related organisms by real-time PCR for Lactobacilli, Gard-
nerella vaginalis and Mycoplasma hominis. FEMS Immunol Med Mi-
22. Powell KF, Anderson NE, Frith RW, Croxson MC. Non-invasive di-
agnosis of herpes simplex encephalitis. Lancet 1990;335:357–8.
23. Allard A, Albinsson B, Wadell G. Rapid typing of human adenoviruses
by a general PCR combined with restriction endonuclease analysis. J
Clin Microbiol 2001;39:498–505.
24. Handsfield HH, Lipman TO, Harnisch JP, Tronca E, Holmes KK.
Asymptomatic gonorrhea in men: diagnosis, naturalcourse,prevalence
and significance. N Engl J Med 1974;290:117–23.
25. Arnberg N, Mei Y, Wadell G. Fiber genes of adenoviruses with tropism
for the eye and the genital tract. Virology 1997;227:239–44.
26. Elnifro EM, Cooper RJ, Klapper PE, Yeo AC, Tullo AB. Multiplex
polymerase chain reaction for diagnosis of viral and chlamydial ker-
atoconjunctivitis. Invest Ophthalmol Vis Sci 2000;41:1818–22.
27. Schaap GJ, de Jong JC, van Bijsterveld OP, Beekhuis WH. A new
intermediate adenovirus type causing conjunctivitis. Arch Ophthalmol
28. Laverty CR, Russell P, Black J, Kappagoda N, Booth N. Adenovirus
infection of the cervix. Acta Cytol 1977;21:114–7.
29. De Jong JC, Wigand R, Wadell G, et al. Adenovirus 37: identification
and characterization of a medically important new adenovirus type of
subgroup D. J Med Virol 1981;7:105–18.
30. Stanescu D, Copelovici Y, Teleguta L, et al. Investigations on the role
of some viral, chlamydian, rickettsian and mycoplasmic agents in sev-
eral gynecological diseases. Virologie 1989;40:71–7.
31. Pisani S, Gallinelli C, Seganti L, et al. Detection of viral and bacterial
infections in women with normal and abnormal colposcopy. Eur J
Gynaecol Oncol 1999;20:69–73.
32. Grulich AE, de Visser RO, Smith AM, Rissel CE, Richters J. Sex in
Australia: homosexual experience and recent homosexual encounters.
Aust N Z J Public Health 2003;27:155–63.
33. Tran T, Druce JD, Catton MC, Kelly H, Birch CJ. Changing epide-
miology of genital herpes simplex virus infection in Melbourne, Aus-
tralia, between 1980 and 2003. Sex Transm Infect 2004;80:277–9.
34. Roberts CM, Pfister JR, Spear SJ. Increasing proportion of herpes sim-
plex virus type 1 as a cause of genital herpes infection in college stu-
dents. Sex Transm Dis 2003;30:797–800.
35. Lafferty WE, Hughes JP, Handsfield HH. Sexually transmitted diseases
in men who have sex with men: acquisition of gonorrhea and non-
gonococcal urethritis by fellatio and implications for STD/HIV pre-
vention. Sex Transm Dis 1997;24:272–8.
36. Hernandez-Aguado I, Alvarez-Dardet C, Gili M, Perea EJ, Camacho
F. Oral sex as a risk factor for Chlamydia-negativeUreaplasma-negative
nongonococcal urethritis. Sex Transm Dis 1988;15:100–2.
37. McGowan I, Radcliffe KW, Bingham JS, Dencer C, Ridgway GL. Non-
gonococcal urethritis in men practising “safe” sex. Genitourin Med
38. Schwartz MA, Lafferty WE, Hughes JP, Handsfield HH. Risk factors
for urethritis in heterosexual men: the role of fellatio and other sexual
practices. Sex Transm Dis 1997;24:449–55.
39. Noble RC. Colonisation of the urethra with Streptococcus pneumoniae:
a case report. Genitourin Med 1985;61:345–6.
40. Conde-Glez CJ, Calderon E. Urogenital infection due to meningococ-
cus in men and women. Sex Transm Dis 1991;18:72–5.
41. Hagman M, Forslin L, Moi H, Danielsson D. Neisseria meningitidis in
specimens from urogenital sites: is increased awareness necessary? Sex
Transm Dis 1991;18:228–32.
42. Maini M, French P, Prince M, Bingham JS. Urethritis due to Neisseria
meningitidis in a London genitourinary medicine clinic population.Int
J STD AIDS 1992;3:423–5.
43. Wilson AP, Wolff J, Atia W. Acute urethritis due to Neisseria menin-
gitidis group A acquired by orogenital contact: case report. Genitourin
44. Sturm AW. Haemophilus influenzae and Haemophilus parainfluenzaein
nongonococcal urethritis. J Infect Dis 1986;153:165–7.
45. Taylor-Robinson D. The role of mycoplasmas in non-gonococcal ure-
thritis: a review. Yale J Biol Med 1983;56:537–43.
of chimpanzees by Ureaplasma urealyticum. J Med Microbiol 1978;11:
47. Povlsen K, Bjornelius E, Lidbrink P, Lind I. RelationshipofUreaplasma
urealyticum biovar 2 to nongonococcal urethritis. Eur J Clin Microbiol
Infect Dis 2002;21:97–101.
48. Deguchi T, Yoshida T, Miyazawa T, et al. Association of Ureaplasma
Viral and Bacterial Causes of NGU • JID 2006:193 (1 February) • 345
urealyticum (biovar 2) with nongonococcal urethritis. Sex Transm Dis
49. McCormack WM, Lee YH, Zinner SH. Sexual experience and urethral
colonization with genital mycoplasmas: a study in normal men. Ann
Intern Med 1973;78:696–8.
50. Brunner H, Weidner W, Schiefer HG. Quantitative studies on the role
of Ureaplasma urealyticum in non-gonococcal urethritis and chronic
prostatitis. Yale J Biol Med 1983;56:545–50.
51. Bowie WR. Etiology and treatment of nongonococcal urethritis. Sex
Transm Dis 1978;5:27–33.
52. Cracea E, Constantinescu S, Lazar M. Serotypes of Ureaplasma ureal-
yticum isolated from patients with nongonococcal urethritis and gon-
orrhea and from asymptomatic urethral carriers. Sex Transm Dis 1985;
53. Piot P. Distribution of eight serotypes of Ureaplasma urealyticum in
cases of non-gonococcal urethritis and of gonorrhoea, and in healthy
persons. Br J Vener Dis 1976;52:266–8.
54. Bowie WR, Wang SP, Alexander ER, et al. Etiology of nongonococcal
urethritis: evidence for Chlamydia trachomatis and Ureaplasma ureal-
yticum. J Clin Invest 1977;59:735–42.
55. Hillier SL, Krohn MA, Rabe LK, Klebanoff SJ, Eschenbach DA. The
normal vaginal flora, H2O2-producing lactobacilli, and bacterial vagi-
nosis in pregnant women. Clin Infect Dis 1993;16(Suppl 4):S273–81.
56. Iser P, Read TR, Tabrizi SN, et al. Symptoms of non-gonococcal ure-
thritis in heterosexual men—a case control study. Sex Transm Infect
57. Keane FE, Thomas BJ, Whitaker L, Renton A, Taylor-Robinson D. An
association between non-gonococcal urethritis and bacterial vaginosis
and the implications for patients and their sexual partners. Genitourin