Introduction

Nucleic acid amplification methods, such as the polymerase chain reaction (PCR), are significantly more sensitive, and should therefore be used in preference to other tests for the detection of genital Chlamydia trachomatis infection.^1–3 In women, the sensitivity of C trachomatis PCR testing is increased by approximately 12% if both cervical swab and urine specimens are examined as opposed to urine alone.¹ However, this approach is expensive, particularly considering the relatively high cost of PCR tests versus enzyme immunoassays. We investigated the utility of PCR testing of combined urine and cervical swab specimens, for female genitourinary medicine clinic attendees, in order to maximise the sensitivity of C trachomatis PCR testing while minimising specimen processing costs. However, we were concerned to determine whether combining specimens reduced test sensitivity because of naturally occurring inhibitors of the enzyme used to amplify DNA, which may be more prevalent in cervical specimens than in urine.⁴

Patients and methods

Samples were obtained from 100 women attending the genitourinary medicine (GUM) clinic at Leeds General Infirmary who had a high likelihood of genital C trachomatis infection. These comprised women who had tested positive elsewhere for genital C trachomatis infection and who had been referred for treatment and partner notification; women who had been recalled to clinic for follow up with unexpected, untreated genital C trachomatis infection; and female contacts of men with C trachomatis infection. A routine vaginal speculum examination was performed and two endocervical swabs (Medical Wire, UK) were taken simultaneously before other cervical sampling. One swab was placed in transport medium (IDEIA, Dako Diagnostics, UK). The other was placed in one of two aliquots of first void urine (patients voided urine immediately after examination). Specimens were transported to the laboratory within 4 hours of collection.

Specimens (urine + swab, urine alone, and swab alone) were processed by automated C trachomatis PCR (Cobas Amplicor using standard urine or cervical swab protocols (Roche, version 3). Aliquots of 500 μl of urine were used for both the urine and combined urine + swab specimens. The combined urine + swab specimen was tested both before and after freezing at −20°C and then thawing at room temperature. Hence, four PCR assays per patient were performed. An inhibition control supplied by the manufacturer was included with each assay in order to identify specimens containing PCR inhibitors.

Results

Of 100 patients tested, 71 were Amplicor C trachomatis PCR positive (according to the results of at least one specimen). The Amplicor C trachomatis PCR test results were concordant for 95/100 patients. The five discordant result sets (table 1) included only one major discrepancy (1/100, 1%, 95% CI −0.01–3%). In this case (patient 1, table 1) the cervical swab was Amplicor C trachomatis PCR positive, whereas the other three PCR assays were PCR negative. In each of these three, inhibitors of PCR were not detected, as internal control amplification occurred. Hence, the pre- and post-freeze-thaw combined samples yielded false negative results. On further investigation, this patient was found to have received 5 days of treatment with doxycycline immediately before testing, but on examination a purulent cervicitis was still present. In the remaining discordant result sets, one specimen yielded a negative result while the other three were C trachomatis PCR positive.

Inhibitors of PCR were present in 22/400 specimens from 20/100 patients; 20/22 of these specimens were from C trachomatis infected patients (p<0.1). Of these 22 specimens, 16 were cervical swabs (p<0.001). One pre-freeze-thaw combined sample was C trachomatis PCR negative, presumably due to PCR inhibitors as the specimen was positive after freeze-thawing.

Discussion

Although nucleic acid amplification methods permit non-invasive sampling for C trachomatis, additional sampling of the cervix can increase diagnostic sensitivity.¹ Cervical specimens are easily obtained in GUM clinics where female patients are routinely offered a vaginal speculum examination. We have demonstrated that acceptable sensitivity is retained by performing a C trachomatis PCR test on a combined urine and cervical swab specimen (sensitivity 98.6%, specificity 100%), as opposed to testing separate samples. We examined specimens from women with a high probability of C trachomatis infection as these may be more likely to contain PCR inhibitors secondary to tissue inflammation.⁴ We did not aim to prove that testing both cervical and urine samples for C trachomatis would increase the chance of a positive result as this has been shown previously,¹ and requires a considerably larger patient cohort, preferably including individuals with a low C trachomatis burden.

We found no evidence of PCR inhibition as a result of combining cervical and urine specimens. By contrast, in C trachomatis positive women PCR inhibitors were significantly more common in cervical swabs than in other specimens (11/72 versus 4/216, χ² p<0.001). Inhibition was detected almost three times more commonly in specimens from C trachomatis positive women than from uninfected females (p<0.1). This is probably due to the higher concentration of serum/tissue products in cervical swab specimens, particularly those taken from women with cervicitis. PCR inhibitors present on cervical swabs (n=16) appeared to be diluted to non-detectable levels when swabs were placed in urine, as inhibitors were still evident in only 2/16 cases when combined specimens were tested. In these two cases inhibitors could not be detected post-freeze-thawing. Freeze-thawing of combined swab and urine specimens is advantageous in terms of laboratory processing, allowing batching and aiding specimen transport, as specimens can be frozen if delay is expected. and also reduces the prevalence of PCR inhibitors.^{4, 5} Freeze-thawing also reduces the prevalence of PCR inhibitors. For example, C trachomatis PCR inhibitors can be found in 20% of urine specimens from pregnant women, but approximately half of these are rendered non-inhibitory following freeze-thawing.⁵ Alternative methods of removing C trachomatis PCR inhibitors include specimen dilution,^{4, 6, 7} but this is relatively cumbersome, particularly when large numbers of specimens are processed, and may cause DNA contamination of samples and/or false negative results.

Combining a cervical swab with a urine specimen is acceptable for PCR testing for genital C trachomatis infection. Such an approach maximises test sensitivity while minimising costs, and indeed has the potential to increase further the cost effectiveness of DNA based screening for C trachomatis genital infection.²