Comparison of Detection Methods for Chlamydia Trachomatis in Specimens Obtained From Pediatric Victims of Suspected Sexual Abuse

Chlamydia trachomatis is the most common sexually transmitted bacterium in the US. It infects an estimated 4 million people per year, with ∼70% of the cases being asymptomatic.1 The detection of C. trachomatis in pediatric populations is used for legal purposes as evidence of suspected sexual abuse; however, it is recognized that perinatally transmitted C. trachomatis infection can persist for up to 3 years.2 Currently in the US the only legally acceptable method for the detection of C. trachomatis in this population is isolation of the agent in cell culture.1 This is in part because of the possibility of false positive test results from cross-reacting test reagents and/or human flora, i.e. decreased specificity of other methods compared with culture.2 The need for culture techniques poses particular problems for small hospitals and clinics lacking tissue culture facilities.3 However, as comparison data are collected on the use of PCR for this special circumstance, we feel the courts will accept C. trachomatis detection by PCR. Moreover the ease of both specimen transport and assay procedure, in addition to the problems associated with loss of integrity of specimens for culture, makes the use of the commercially available PCR kit (Amplicor) extremely attractive. However, determination of the performance characteristics of PCR in the juvenile population is necessary in light of data showing problems with reproducibility in the analysis of cervical specimens from adults.4 It is with this view that we have an ongoing comparison between culture, PCR, and an enzyme immunoassay (EIA) currently used in our laboratory for adult patients. Methods. All specimens collected from pediatric patients in cases of suspected sexual abuse and submitted to the diagnostic virology laboratories of the LSU University Hospital of New Orleans and the LSU Children's Hospital in Shreveport for Chlamydia culture, were initially entered into this study. All were cultured immediately on receipt by either laboratory and then frozen at −70°C. Specimens collected in New Orleans were transported on dry ice to the Shreveport laboratory where all PCR and EIA were performed. Of the total number of 245 specimens tested, 36 were excluded from the study for reason of either culture site (eye swabs) or patient's age (18 years or older). Thus the total number of specimens included in this study is 209. Specimens were processed for shell vial culture assay with the Merifluor Chlamydia immunofluorescent antibody kit (Meridian, Cincinnati, OH). PCR for C. trachomatis was performed with the Amplicor system (Roche Diagnostic Systems, Branchburg, NJ). EIA for Chlamydia was performed with the MicroTrak Chlamydia kit (Behring Diagnostics, Inc., San Jose, CA). All three assays were performed according to the manufacturer's instructions. Strict attention to the physical separation of PCR steps and a unidirectional PCR work-flow was observed. As an additional aid in the prevention of carryover contamination from PCR products, uracil N-glycosylase (provided in the Amplicor kit) was used. All positive PCR results were confirmed by repeat PCR testing. For the purposes of this study, shell vial culture results or the combination of concordant PCR and EIA results were considered to be the standard to which the culture, PCR and EIA results were compared. Results. Of the 209 specimens included in the study 33 (15.8%) were from males and 176 (84.2%) were from females. The mean patient age was 8.6 years, ranging from 1 month to 17 years. Specimens were collected as swabs from the following sites: 95 vaginal, 39 oral cavity, 30 rectal, 22 UDS (undesignated swabs), 12 cervical and 10 penile. Only 1 specimen was urine. Assay results for individual specimen sites, along with sensitivity, specificity and predictive values are given at the top of Table 1.TABLE 1: Cumulative data for all specimens by specimen site Cumulative data are shown at the bottom of Table 1. As expected, most of the specimens (187 or 89.5%) tested negative by all 3 techniques. Three rectal specimens were classified as indeterminate by culture because of an overgrowth of bacteria. All three were negative by both PCR and EIA. Five specimens (2.4%) tested positive by all three techniques. Thus the calculated percent agreement between the three assays is 192 of 209 or 91.9%. Nineteen specimens (9.0%) tested positive by either culture or PCR or by both techniques. These 19 specimens were collected from children ranging in age from 4 to 16 years (mean, 10.7) and included 12 vaginal, 4 rectal, 2 cervical and 1 undesignated swab specimens. Only one of these patients was male. Five of the 19 were positive by all 3 methods (26.0% agreement among the specimens testing positive). All PCR results were repeatable. Thirteen of the 19 specimens (68.4%) testing positive yielded discordant results. Six specimens were positive by both culture and PCR, but not by EIA. Two specimens (both rectal sources) were positive by culture alone, even after the DNA was extracted and PCR was repeated. No specimens were positive by EIA alone. Four specimens (one rectal, two vaginal and one cervical) were negative by culture and EIA but positive by PCR. One of these was a rectal specimen from a patient (Patient EL) from whom two other vaginal swabs and one oral swab also were collected. Of the two vaginal swabs from EL, one tested positive by all three methods and one tested positive only by EIA and PCR. EL's oral swab tested negative by all three techniques. Of the other specimens positive by PCR only, one vaginal swab was from a patient with negative oral and rectal specimens. No other information was available on the last two specimens testing positive only by PCR. Discussion. This study was performed to gather data on the use of the commercially available Amplicor PCR kit assay (Roche Diagnostic Systems) in pediatric patients who are victims of suspected sexual abuse. Data on the use of nucleic acid amplification in this setting are limited.2, 3, 5 In our study culture detected 13 of the 15 specimens that were positive by study criteria; two of these were rectal specimens testing negative by the other tests. One culture of a vaginal specimen was contaminated; another, from an undesignated swab, was negative. Both specimens were positive by EIA and PCR. Therefore the sensitivity of culture was determined to be 13 (true positives) of 15 (true positives plus the number of false negatives) or 86.7%. PCR detected 13 of the 15 specimens that were positive by the study criteria. There were two false negative rectal specimens. One PCR-positive specimen (UDS discussed above) was culture-negative but tested positive by EIA. The sensitivity of PCR is calculated to be 13 (true positives) of 15 (true positives plus the number of false negatives) or 86.7%. Four specimens tested positive by PCR only. If we assume that all 4 of these specimens are false positives, then the specificity of the PCR is calculated to be 190 (true negatives) of 194 (true negatives plus the number of false positives) which is 97.9%. However, one of these PCR "false positive" specimens was a rectal specimen obtained from a patient for whom a vaginal specimen tested positive by all three techniques. If we assume that specimen to be a true positive, then the specificity for PCR becomes 190 of 193 or 98.4%. EIA proved to be less sensitive than either culture or PCR. It missed 8 specimens that were culture positive, 3 of which were rectal specimens. However, it detected C. trachomatis antigen in one vaginal specimen which was overgrown in culture, but PCR-positive, and another specimen (UDS) which was culture-negative, but PCR-positive. The sensitivity of EIA with the use of study criteria is calculated to be 7 (true positives) of 15 (true positives plus the number of false negatives) or 46.7%. Low specificity of EIA, especially without a blocking procedure has been reported.6 However, by using blocking antibody in our study, specificity of 100% was achieved. Interpretation of these results is limited by several factors. First, the study may have been biased to favor culture, because the EIA and PCR assays were performed several months after culture on frozen specimens, one-half of which had been transported from New Orleans to Shreveport (a distance of ∼300 miles). However, previous investigation by our laboratory (unpublished data) of the effects of freezing on EIA reactivity suggests that the relative insensitivity of the EIA in this study cannot be attributed to freezing. Secondly, although specimens taken from oral and rectal sites were included in this study, the EIA and PCR protocols are neither approved by the Food and Drug Administration nor recommended by the manufacturer for these specimen sources. None of our oral specimens were positive by any technique. This supports the recommendation that oral specimens not be collected for Chlamydia cultures in this population.2 However, when comparing the performance of culture and PCR on the 30 rectal specimens submitted, we find that: 23 were negative and 1 was positive by both techniques; 2 were culture-positive, yet PCR-negative; 1 was PCR-positive, but culture-negative; and 3 were PCR-negative and indeterminate by culture (because of an overgrowth of bacteria). This suggests that rectal specimens are problematic for either technique and that PCR detection may be as good as culture. On the other hand the only two positive specimens missed by PCR were rectal specimens. PCR inhibitors are found in many laboratory specimens,7 but rectal specimens commonly are inhibitory.5 Further studies are needed to determine the utility of rectal specimens for assay by PCR, especially because the testing of rectal swabs is included in the routine protocol for examination of children presenting to many sexual abuse clinics. These preliminary data suggest that the performance of the Amplicor kit for C. trachomatis detection in pediatric victims of sexual abuse is comparable with culture and should be further investigated. To this end we intend to continue the collection of comparative data. We believe that as further data are evaluated, PCR testing may become acceptable in the courts. And finally we should emphasize that the Amplicor kit carries a disclaimer by Roche Diagnostic Systems for the use of PCR in this population and that its use for assay of rectal swabs is neither recommended nor approved. Janice Matthews-Greer, Ph.D. Gregory Sloop, M.D. Ann Springer, M.D. Kenneth McRae, M.S. Ethelyn LaHaye, M.S. Richard Jamison, Ph.D. Department of Pediatrics; Louisiana State University Medical Center; Shreveport, LA (JMG, AS.