Objectives To assess the performance of the Roche fully automated cobas® 4800 CT/NG test for the detection of C trachomatis (CT) infection in clinical specimens compared to the current routine practice.

Methods Consecutive clinical specimens sent to the Bacteriology department of the Bordeaux University Hospital, Bordeaux, between July and September 2010 were included. Results of the cobas® 4800 CT/NG test were compared with those obtained with the cobas® TaqMan CT 48 assay (Roche). For the latter, DNA from 200 μl of urine or swab resuspended in transport medium, (2SP or universal transport medium) was extracted on the MagNA Pure using the DNA I isolation kit (Roche) and amplified on the TaqMan 48 automates. The cobas® 4800 CT/NG performed DNA extraction from urine specimens prepared by adding 4.5 ml to 4.5 ml of cobas® PCR media, and from swabs discharged in 1.0 ml of the same media. The cobas® 4800 system loaded extracted DNA, controls and amplification reagents into 96-well amplification plates. Plates were then covered and placed into the cobas® z480 real-time PCR instrument. Retesting in both cobas® 4800 and TaqMan 48 assays was performed to further investigate specimens providing discrepant results.

Results A total of 708 clinical specimens (293 male urines and 415 swab specimens, of which 356 self-collected vaginal swabs, 45 swabs from cervix and 14 swabs from male urethra) were analysed. The results were concordant in 98.5% of cases (697/708). Out of 708 samples, 50 provided positive results (17 men, 33 women). Three urine specimens and eight vaginal swabs provided discrepant results. Out of five specimens providing positive results in the reference CT assay, four were false-negative in the cobas® 4800 CT test. Out of six positive results by the cobas® 4800 assay, five were false-positive. After discrepancy analysis, the prevalence of the CT infection was 7.7% (55/708). The sensitivity and specificity of the cobas® 4800 CT/NG test were 92.7% (urine specimens 94.1%, swab specimens 92.1%) and 99.2%, respectively. The three false-negative results in swabs could be explained by the procedure not consistent with the manufacturer's instructions. Indeed, swabs were not inserted directly into the cobas® media vials.

Conclusion The cobas® 4800 CT/NG test is suitable for high through-put identification of the C trachomatis infection.