EVALUATION OF COMMERCIAL POLYMERASE CHAIN-REACTION AND LIGASE CHAIN-REACTION ASSAYS FOR THE DETECTION OF CHLAMYDIA-TRACHOMATIS IN URINE SPECIMENS

A study comparing two commercially available nucleic acid amplificatio n assays (Roche Amplicor Polymerase Chain Reaction, Abbott LCx Ligase Chain Reaction), and one traditional antigen detection system (Syva Mi crotrak Direct Fluorescent Antibody), was conducted for patients atten ding a metropolitan sexual health clinic between the months of July an d November 1995. 514 consecutive clients attending the Brisbane Sexual Health Clinic for STD screens were enrolled. A routine urethral or ce rvical swab was taken for C. trachomatis and each patient provided a f irst-catch urine specimen of approximately 30 to 40 mls. 346 urine and urethral swab specimens were collected from males, and 168 urine and cervical or urethral swab specimens from females. Swabs were processed by Direct Fluorescent Antibody assay, urines by both PCR and LCR. The client sample comprised both asymptomatic patients and those presenti ng with symptoms of urethritis or cervicitis. An expanded 'gold standa rd' was used to compare the performance of each assay. 'True' positive s were defined as a positive result in any two of the three assays. If either one of an LCR or PCR result was negative for any set of matche d samples, the negative test was repeated. If a retested sample remain ed negative by PCR and/or LCR, a second alternative PCR was carried ou t for confirmation. The confirmatory PCR was performed with an in-hous e assay using primers to the 16S rRNA gene. Calculations of sensitivit y and specificity were based on the results of the first series of LCR and PCR tests. Using the 'gold standard', of the 514 specimen 'sets' analysed, a total of 32 samples were positive by at least one test, of which 29 were 'confirmed' positive by any of the two methods. The sen sitivity and specificity for DFA, PCR and LCR were 72.4% and 99.6%, 86 .2% and 99.8%, and 100% and 99.8% respectively. A comparison of PCR an d LCR for those specimens tested as true positives during the first ru n of tests indicated 86.2% concordance with respect to sensitivity. An alysis of patient clinical histories for discrepant results indicates that both PCR and LCR have a higher sensitivity than DFA for asymptoma tic infections, particularly for those patients suspected to be contac ts of known carriers. False negative PCR tests were found amongst symp tomatic patients providing urine specimens containing heavy precipitat e, whilst false negative DFA results correlated with patients who were either asymptomatic or whose risk factor was identified as being a co ntact of a partner with C. trachomatis infection. This study would ind icate that nucleic acid amplification systems (PCR, LCR) as applied to urine specimens are of greater sensitivity compared with urethral/cer vical swab based DFA.