Rietmeijer, Cornelis A. MD, PhD*†; Mettenbrink, Christie J. BS, MSPH, MT(ASCP)*
Male urethritis is a common clinical presentation of sexually transmitted infections in men.1 Diagnosis of male urethritis at the day of presentation allows for the expedited treatment of the most common sexually transmitted causes of male urethritis, including Chlamydia trachomatis and Neisseria gonorrhoeae, thus reducing transmission to female partners in whom these pathogens are responsible for a large proportion of genital infections that may be complicated by pelvic inflammatory disease and its sequelae. To illustrate, among men presenting for a new problem at the Denver Metro Health Clinic (DMHC), 17% are diagnosed with urethritis, of whom 33% have gonococcal urethritis and 67% have nongonococcal urethritis (NGU). Of men ultimately diagnosed with urethral chlamydia or gonorrhea infections by nucleic acid amplification testing (NAAT), respectively 39% and 87% are treated at initial presentation on the basis of a urethritis diagnosis (DMHC, data not shown).
Traditionally, the diagnosis of male urethritis is established through the microscopic examination of a Gram-stained smear (GSS) of a urethral discharge sample. The diagnosis of gonococcal urethritis is made when intracellular, gram-negative diplococci are visualized. The diagnosis NGU is typically made when ≥5 polymorphonuclear (white blood) cells (PMN) are seen, averaged over 5 high-power fields (hpf), in the absence of gram-negative diplococci.2 The cutoff value of 5 PMN/hpf for the diagnosis of NGU was established as the best combination of sensitivity and specificity in distinguishing symptomatic from asymptomatic urethritis in an era when chlamydia and gonorrhea cultures were considered the gold standard for diagnosing these infections.3,4 However, NAAT-based assays have superior sensitivity over chlamydia culture with comparable specificity, and have replaced culture as the gold standard over 15 years ago.5,6 It is reasonable to hypothesize that chlamydia infections detected by NAAT, but not by culture, harbor lower levels of organisms at the infection site and that a lower organism burden is associated with a muted inflammatory response, thus explaining why many NAAT-detected chlamydia infections are asymptomatic. Therefore, it is equally conceivable that the current diagnostic cutoff point for GSS-based NGU diagnosis is too stringent as many infections may be accompanied by minimal inflammation.7 To further investigate the latter hypothesis, we conducted a retrospective analysis of male urethral GSS submitted to the DMHC laboratory and evaluated chlamydia and gonorrhea NAAT positivity rates at incremental PMN/hpf levels.
DMHC is the largest clinic for sexually transmitted infections in the Rocky Mountain Region, logging >15,000 patient visits annually. Men with urethral symptoms (dysuria and/or discharge) are evaluated by GSS. Men with evident urethral discharge, either spontaneous or after manual expression by the clinician, will have a small sample of their discharge directly collected on a glass slide. Men without obvious discharge will have a sterile calcium alginate swab gently inserted 1 to 2 cm into their urethra after which the swab is rotated in the urethra and then rolled onto the slide. Slides are sent to clinic stat laboratory where they are Gram-stained using standard laboratory procedure and read by a laboratory technician under high magnification with oil immersion (×1000).
In March 2005, the clinic introduced an electronic medical record (EMR) that established real-time integration of clinical and laboratory activities and also allowed for more detailed data entry by laboratory staff. Thus, where in the past results of the GSS were categorized in combined strata, the EMR allows for more detail: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, >10 PMN/hpf. The EMR also requires the name of the laboratory technician performing the test to be recorded with the test result. While in certain circumstances, male urethral gonorrhea cultures are currently still performed, all men visiting DMHC for a new problem are tested routinely for chlamydia and gonorrhea by urine-based NAAT. Before August 2007, we used the strand displacement assay (ProbeTec, Becton-Dickinson Diagnostic Systems, Sparks, MD); subsequently, we used the transcription-mediated amplification assay (APTIMA COMBO 2, Gen-Probe, San Diego, CA).
For this study, we conducted 2 mutually exclusive analyses to assess chlamydia and gonorrhea positivity rates along the continuum of GSS PMN/hpf strata. For the stratified GSS chlamydia positivity analysis, we excluded all men who were shown to have gonorrhea by NAAT. Likewise, for the stratified GSS gonorrhea positivity analysis, we excluded those men who were NAAT-positive for chlamydia. Men who had both chlamydia and gonorrhea infections were excluded from either analysis.
The main analysis was conducted on all GSS smears submitted between March 1, 2005 and December 31, 2010. To control for variations between laboratory technicians and potential differences associated with switching NAAT technologies, we compared two 24-month time periods, March 1, 2005 to February 28, 2007 and March 1, 2008 to February 28, 2010, respectively, within the overall time frame, when technicians and NAAT technologies were different.
For each of the periods, we evaluated chlamydia and gonorrhea positivity for all GSS PMN/hpf strata (0, 1, 2, 3, etc.).
Finally, to place our findings in context, we determined chlamydia positivity rates for asymptomatic men who were presumptively treated during the overall time frame because they were contacts to patients diagnosed with either chlamydia, gonorrhea, pelvic inflammatory disease (PID), or mucopurulent cervicitis (MPC).
All analyses were conducted using the SAS software package (SAS Institute Inc., Cary, NC).
The study was exempted from overview by the Colorado Multiple Institutional Review Board at the University of Colorado Denver.
During the 70-month period, a total of 36,546 men who visited the clinic were evaluated by urine NAAT for chlamydia or gonorrhea infection. Of these men, 5714 (16.6%) were found to be infected with chlamydia and 2297 (6.3%) with gonorrhea. During the same period, 13,520 GSS were performed for evaluation of male urethritis, representing 36.9% of men who had a NAAT test. Among men with GSS, 3499 (35.9%) were diagnosed with chlamydia and 2098 (15.5%) with gonorrhea by NAAT. Thus, of all men with chlamydia, 3499/5714 (61.2%) had a GSS and of men with gonorrhea, 2098/2297 (91.3%) had a GSS.
After exclusion of men who were diagnosed with gonorrhea based on urine NAAT, 11,422 GSS were available for the chlamydia analysis, with an overall positivity rate of 25.3%. Likewise, after exclusion of men diagnosed with chlamydia, 10,023 GSS were available for the gonorrhea analysis, with an overall positivity rate of 14.8%. Results are presented in Table 1 and Figure 1. In the chlamydia analysis, chlamydia positivity showed a progressive trend along the PMN/hpf strata, increasing from 4.8% at the 0 stratum to 43.8% at the >10 stratum (P < 0.001). Examining the 95% confidence intervals of the positivity rates, a statistically significant increase occurred between the 1 and 2 strata from 6.6% to 16.2%. Of all chlamydia infections diagnosed among men who received a GSS (and did not have gonorrhea), 33.5% had GSS results below the 10 stratum, and 9.5% had GSS results in the 2, 3, and 4 strata. Findings in the gonorrhea analysis were strikingly different; we did not determine a trend of gonorrhea positivity along the PMN/hpf strata. Indeed, of men who had a GSS and were diagnosed with gonorrhea by NAAT, only 2.3% had <10 PMN/hpf.
Controlling for variance between laboratory technicians and NAAT technologies yielded similar results. To illustrate, we conducted a subanalysis of 2 periods when NAAT technologies and laboratory technicians were different. During the first period, we evaluated 8 technicians; of whom, 3 accounted for 81% of GSS tests. During the second period, we evaluated 13 technicians; of whom, 3 accounted for 79% of tests. There were 3 technicians who performed tests in both periods; to avoid overlap, only the tests they performed in the second period were included in the evaluation. Given that the overall analysis did not reveal any trending of gonorrhea positivity by GSS strata, we limited the subperiod analyses to the chlamydia analysis only. Although there were some differences between the 2 periods, the trends were very similar and the increase from the 1 to 2 PMN/hpf strata was significant in both (data not shown, but available from the authors upon request).
Finally, to put our findings in perspective, we assessed chlamydia positivity for asymptomatic men who were treated for chlamydia presumptively. Results were as follows: contacts to patients with chlamydia, 43.9%; contacts to gonorrhea, 15.9%; contacts to PID, 13.4%; contacts to MPC, 8.7%.
Our analysis of 13,520 GSS of male urethral specimens during a 70-month period provided support for our hypothesis that the current diagnostic criteria for male urethritis based on GSS are too stringent. Indeed, there are 3 reasons to lower the cutoff to ≥2 PMN/hpf. First, a significant 2.5-fold increase in chlamydia positivity occurred between the 1 and 2 PMN/hpf strata. Second, by expanding treatment for urethritis to include the 2, 3, and 4 strata, we would have initiated presumptive treatment for an additional 9.5% of men who had a GSS evaluation and who were ultimately diagnosed with chlamydia based on NAAT. Finally, the positivity rate at the 2 stratum (16.2%) compares favorably with the positivity rate among men recommended to receive presumptive chlamydia treatment because they are a contact to a female patient diagnosed with gonorrhea (15.9%), PID (13.4%), or MPC (8.7%).2
An important limitation to our findings is that this is a single-site study, and results might be different at other clinics and laboratories. However, the most likely reasons for such differences are intertechnician variance and the use of other testing technologies as the comparator gold standard. Therefore, the similarity of our findings between the 2 periods when we employed different technicians and NAAT technologies should alleviate these concerns. Nonetheless, our findings would gain in strength and applicability if they could be replicated in different settings, especially those where clinicians prepare and read their own GSS instead of laboratory technicians and where the performance of the GSS may thus be different.
Performance of the GSS may also differ by techniques obtaining the specimen; i.e., direct collection of urethral discharge on the slide or by the rolled swab technique (also mentioned in Methods section). However, this distinction was not recorded in the clinic EMR in a systematic way, and we were thus unable to conduct this analysis.
One might argue that a proposal to change the GSS cutoff is somewhat academic and that, given that chlamydia and gonorrhea positivity among men who undergo GSS testing is high (respectively, 25.3% and 14.8% in our study), there is a rationale to treat all these men and forego a GSS-based diagnosis altogether. In fact, in settings where immediate microscopic diagnosis is not available, this already is standard of practice. However, the GSS is primarily used to diagnose gonococcal urethritis, and thus differentiate gonococcal urethritis from NGU. To cover for a gonococcal etiology, presumptive treatment for male urethritis without GSS assessment should therefore include treatment for NGU as well as gonorrhea2 and will therefore result in considerable overtreatment, given that (in our data) only 11% of men who had a GSS evaluation had gonorrhea.
Finally, it was not our aim in this study to evaluate the GSS as a screening tool for chlamydia or gonorrhea, and we therefore did not include time-honored analytical tools such as receiver-operating characteristic (ROC) curves that could illustrate the relationship of test sensitivity and specificity at different cutoff points of GSS compared to a gold standard, e.g., chlamydia NAAT. Since many men, especially those with chlamydia infection, are asymptomatic, a GSS will not be performed, lowering overall sensitivity when compared to NAAT. Moreover, interpreting specificity of the test when comparing with chlamydia or gonorrhea, NAAT is inappropriate because there may be other causes of urethritis in men that we currently do not screen for (e.g., Mycoplasma genitalium and Ureaplasma urealyticum) and, even if the cause is unknown, urethritis is still a bona fide diagnosis in its own right for which treatment guidelines exist and for which the diagnostic gold standard is the GSS.2
In conclusion, our study suggests that the current diagnostic guidelines for male urethritis may be too stringent. Our data should inform a discussion on the recalibration of this old diagnostic standby.