The optimum approach for detecting Chlamydia trachomatis (CT) in a genitourinary medicine (GUM) setting has been combined cervical and urethral testing. The endocervix has been the preferred anatomic site for specimen collection, although it has been claimed that 10–23% of females will only be infected in the urethra.^1,2 Various studies have quoted an increase of 5–33% in detection of CT with inclusion of a urethral swab.^2–5 However these studies used insensitive detection methods such as cell culture or enzyme immunoassay.

We currently use the highly sensitive Aptima Combo 2 assay (AC2; Gen-Probe, San Diego, CA, USA) to detect CT by both urethral and cervical swabs in female patients. Given the fact that urethral swabbing is painful we wanted to ascertain its contribution in detecting chlamydia before we changed our practice. There has been no study to date looking at the contribution of a urethral swab in females tested by AC2.

METHOD

Urethral and endocervical sampling for chlamydia were performed routinely on all sexually active female patients aged 16 and over within the GUM Department at Macclesfield District General Hospital from October 2005 to November 2006.

Swabs were collected and transported to the laboratory in separate AC2 sample collection tubes and were tested by AC2 assay.

RESULTS

Of the 757 patients tested by both endocervical and urethral swab, 90 had CT identified by either method giving a positivity rate of 11.9% (table 1).

Of the 90 positives, results for urethral and endocervical swabs were concordant in 77 patients (85.6%). Eighty two infections (91.1%) would have been diagnosed by swabbing cervix alone but an additional 8 (8.9%) were picked up by the urethral swab. Urethral symptoms had been mentioned by 1 of these 8 women. The improved detection rate was not statistically significant (OR 1.11, 95% CI 0.80 to 1.54).

DISCUSSION

Our 8.9% (8 of 90) of patients with positive AC2 results only from the urethra is somewhat lower than other studies^1,2 but still indicates that testing from one site alone—the cervix or urethra—may not be optimal. In our study, one woman could have been picked up on presenting urethral symptoms, hence reducing the extra yield to 7.8%.

Thus from 757 urethral swabs taken, the additional yield of 7 represents a detection rate of only 1%. A similar study⁶ found that taking an additional urethral swab, and using the ligase chain reaction assay, increased positives by 6% (reduced to 4.4% if urethral swabs were taken only on patients with urethral symptoms).

Routinely performing a urethral swab along with a cervical swab, together with the AC2 assay leads to only a small (1%) increase in detection rate. This has to be weighed up against increased economic cost of extra resources and discomfort to every patient. What alternatives do we have? A tactic of testing an endocervical swab transported in a specimen of (non-invasive) urine has previously been tried.^7,8 We are not aware of any studies using AC2 to test combined first catch urine (FCU) plus an endocervical swab. Another alternative, with the added advantage of not requiring a speculum investigation, could be to use vaginal swabs. Chernesky et al⁹ found when using AC2 that of 180 positive CT cases (both symptomatic and asymptomatic), 172 (95.5%) were identified on a patient collected vaginal swab (PVS), 174 (96.6%) on a clinician collected vaginal swab (CVS), 165 on FCU (91.6%) and 172 on a clinician collected cervical swab. They also noted that the majority of women found PVS easy and preferred it to FCU or having a pelvic examination to collect an endocervical swab. Schachter et al¹⁰ suggest that vaginal swab specimens probably combine material from two potential sites of infection: urine or discharge coming from the urethra and discharge from the endocervix, thus leading to a higher detection rate than FCU or endocervical swabs alone. Using the ligase chain reaction assay, Shafer et al¹¹ found that testing multiple types of specimen increased the positive yield markedly. However their increased yield from cervical swab plus urine over cervical swab alone (176/205 vs 134/207) was almost matched by vaginal swab alone (167/207). Indeed the proposal to swab the vaginal introitus may perhaps maximise the combined pick up from these two areas.¹²

We have also noticed that in the clinical trial data for AC2 (package insert IN0037-03 Rev A, Table 6b (available at http://www.gen-probe.com/pdfs/pi/501011RevA.pdf)) 16 patients were described where a cervical swab was negative and FCU was positive. Although missed by cervical testing a vaginal swab found most of these cases—14 (95% CI 9.9 to 15.1) of 16 by PVS and 12 (95% CI 7.7 to 14.9) of 16 by CVS.

Vaginal swabs tested by AC2 do not show the problems associated with control of inhibitors of amplification that have been encountered with other test platforms.^13,14

CONCLUSION

Considering only the low yield (only 1%) and the discomfort of urethral swabbing, an additional urethral swab appears unwarranted on grounds of both cost and patient care. However as a small number of cases were detected at the urethra but not the cervix, it may be worthwhile investigating the performance of AC2 when placing an endocervical swab in FCU. An effective and simpler approach may be a switch to testing vaginal swabs by AC2.