METHODS

Study subjects were women who participated in the Heterosexual AIDS Transmission Study (HATS). This cohort has been described in detail elsewhere.⁶ Briefly, between March 1990 and December 1993, 665 HIV negative women (18–50 years of age) with no history of injecting drug use were recruited into a prospective study to assess the risk of heterosexual acquisition of HIV and other STIs. The women had to have had at least one male sexual partner within the previous year and were recruited from medical clinics in central Brooklyn. Follow up visits occurred approximately every 6 months, with an interview, laboratory tests, including cultures for C trachomatis, N gonorrhoeae, Trichomonas vaginalis, and physical examination. Urine samples obtained during baseline and at each subsequent visit were processed and then stored at a NIH repository in Bethesda, Maryland at −70°C.

LCR tests for C trachomatis and N gonorrhoeae were performed on urine samples obtained at the time of the positive cultures. In order to determine the prevalence of LCR positive/culture negative C trachomatis and N gonorrhoeae genital infections we tested urine specimens from a subset of 181 women who had consistently negative cultures for these infections and who we defined as “high risk” (less than 40 years old at baseline, and having either T vaginalis at baseline or a history of more than one sexual partner during the 12 months before baseline). We utilised the C trachomatis and N gonorrhoeae LCR assay (Abbott Laboratories, Abbott Park, IL, USA) performed by experienced personnel, who were unaware of the previous culture results. Urine specimens were completely thawed on the day of processing. We pooled urine specimens from all visits for each person (if there were two or more visits) and then performed LCR on the pool. For each pool, a 1 ml aliquot of urine from each visit was transferred to a 15 ml pooling tube. This was then vortexed and 1 ml of the pool was used for processing. The processing and DNA amplification included standard procedures according to the manufacturer's instructions. Pooled samples were evaluated as described in published studies^7–10 using lowered sample cut-off (S/CO) ratios. Specimens from each positive pool were tested individually.

Analysis of health outcomes included development of pelvic inflammatory disease (PID) and ectopic pregnancy, which were reported or diagnosed during follow up visits. The statistical analysis was conducted with spss 10.0 software (SPSS Inc, Chicago, IL, USA).

RESULTS

We tested urine samples obtained simultaneously with endocervical cultures from 58 women who had had positive cultures for C trachomatis and nine women who had had positive N gonorrhoeae cultures. LCR tests were positive in 49 (84.5%) chlamydia infected women and in six (67%) of nine women who had had gonorrhoea.

Of the 181 “high risk” women whose cultures for C trachomatis and N gonorrhoeae were negative, 10 (5.5%) had positive urine LCR tests. Nine women had positive LCR tests for chlamydial infection; one woman had a positive LCR for gonorrhoea.

None of the 10 women had evidence of PID during the follow up period, which ranged from 11 to 46 months. By comparison, PID was diagnosed in two of 66 women who had been treated for gonococcal or chlamydial infection and in eight of 171 women who had neither culture, nor LCR test evidence of these infections. None of these differences are statistically significant using χ² test for contingency table method. No ectopic pregnancies were noted in any of the above mentioned groups.

We calculated the cost savings of pooling. If pooling was not used, 484 LCR tests would have been done. We actually performed 254 tests, which represents a 47% savings.

DISCUSSION

Archived urine samples proved to be useful for retrospective detection of chlamydial and gonococcal infection. The samples were mostly casual (rather than first void) specimens, and had been frozen for up to 10 years. However, we found chlamydial and gonococcal nucleic acids in 49 (84.5%) of 58 and in six (67%) of nine urine specimens collected from women who had positive cultures, for C trachomatis and N gonorrhoeae, respectively.

We assumed that LCR positive/culture negative results represent “low copy” C trachomatis or N gonorrhoeae infections. However, the small number of patients with these infections did not allow us to assess the association of low copy C trachomatis or N gonorrhoeae infections with PID or ectopic pregnancy. Additional studies involving larger numbers of patients with low copy infections are needed.

Pooling of urine specimens for testing C trachomatis and N gonorrhoeae has been shown to be accurate and cost effective compared to testing individual samples.^7–10 Owing to theoretical concern that pooling would dilute the low level positive sample below the limit of detection of the assay we used a lowered S/CO ratio. The lower S/CO ratio allowed us to detect two true chlamydia positive pools, which otherwise would have been considered negative.

Cost savings associated with pooling depend on the size of the pools and prevalence of infections in the population. With the proportion of individuals with detected C trachomatis and N gonorrhoeae infections during the study being 5.5% we performed 47% fewer tests.

CONCLUSIONS

Additional STIs can be identified when using the LCR test on urine specimens. Pooling of urine specimens is a cost saving technique for C trachomatis and N gonorrhoeae testing that can be utilised for research purposes.