点击显示 收起
Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, Georgia
Received 26 March 2003/ Returned for modification 11 April 2003/ Accepted 23 May 2003
ABSTRACT |
---|
Top Abstract Text References |
---|
TEXT |
---|
Top Abstract Text References |
---|
The BDProbeTec ET (BDPT) system (Becton Dickinson and Company, Franklin Lakes, N.J.) uses strand displacement amplification and fluorescent resonance energy transfer probes to simultaneously amplify and detect the DNAs of C. trachomatis and N. gonorrhoeae. The amplification targets are DNA sequences in the cryptic plasmid of C. trachomatis and in the pilin gene-inverting protein homologue of N. gonorrhoeae (4). A recent multicenter evaluation of the BDPT system demonstrated that it has sensitivity superior to that of chlamydia culture and performance characteristics similar to those of other commercially available NAATs for these organisms (7). According to a recent College of American Pathologists survey (2003 HC6-A), the BDPT system was the most common NAAT used by participants for detection of C. trachomatis and N. gonorrhoeae.
NAATs are more prone than other tests to false-positive results due to specimen cross-contamination. False-positive test results are particularly problematic in low-prevalence patient populations, in which the impact on the positive predictive value of the tests is greatest. However, regardless of the population or health care setting, false-positive results for C. trachomatis and N. gonorrhoeae can have adverse medical, social, and psychological impacts on patients.
The Centers for Disease Control and Prevention has recently issued guidelines for the selection, use, and interpretation of screening tests to detect C. trachomatis and N. gonorrhoeae infections (2). These guidelines suggest several approaches by which to detect false-positive test results. The approaches include (i) testing of a second specimen with a different test that uses a different target, antigen, or phenotype and a different format; (ii) testing of the original specimen with a different test that uses a different target, antigen, or phenotype and a different format; (iii) repetition of the original test of the original specimen with a blocking antibody or a competitive probe; and (iv) repetition of the original test of the original specimen. However, only the last approach is practical in most clinical laboratories using NAATs because of different sample collection devices and requirements for the various tests, lack of confirmatory tests, and logistical problems in obtaining second samples from patients.
Problems with reproducibility of positive test results have been documented with the LCx (Abbott Laboratories, Abbott Park, Ill.) and AMPLICOR (Roche Diagnostics Corp., Indianapolis, Ind.) assays for detection of C. trachomatis and N. gonorrhoeae (1, 3, 6). The reproducibility of the BDPT system in a clinical laboratory setting has not been reported. We developed a repeat testing algorithm in order to document the reproducibility of positive BDPT system test results.
All samples were tested once for both C. trachomatis and N. gonorrhoeae according to the manufacturer's instructions, and samples with method other than acceleration (MOTA) scores of greater than or equal to 2,000 (cutoff value) for either C. trachomatis or N. gonorrhoeae were retested for both organisms in the next run by use of the same sample. The MOTA score is a metric used to assess the magnitude of the signal generated as a result of the reaction. The magnitude of the MOTA score is not indicative of the level of the organism in the specimen, since the amount of target is only one of many factors that influence the MOTA score.
If the MOTA score of the second test was above the cutoff, then the sample was considered positive and no further testing was performed. If the MOTA score of the second test was below the cutoff, then a third test was performed with the original sample in the next run. Samples with MOTA scores below the cutoff in the third test were considered negative, and those with MOTA scores above the cutoff were considered positive. Samples were stored at 2 to 8°C for up to 4 days before testing was completed. Samples stored for more than 6 h between repeat tests were vortexed and reheated before testing, as recommended by the manufacturer.
The laboratory provides service to internal medicine, obstetrics and gynecology, student heath clinics, and emergency departments at two hospitals. The prevalence of patients with test results positive for C. trachomatis or N. gonorrhoeae during the study period was 8.1 or 3.9%, respectively. Specimens were collected and transported to the laboratory in accordance with the manufacturer's instructions.
Testing was performed in a molecular diagnostic laboratory by six experienced medical technologists using multiple lots of reagents from 29 July 2002 to 5 March 2003. The amplification control provided by the manufacturer was included in tests from swab specimens until 1 February 2003 and in all urine tests. All positive and negative controls gave the expected values for all of the runs included in the data analysis. One run in which an obvious break in technique occurred, as indicated by failed negative controls, was excluded.
On the basis of the initial MOTA scores of greater than or equal to 2,000, C. trachomatis DNA was detected in 156 swab and 51 urine samples. The range of MOTA scores for the 207 positive samples was 2,118 to 40,642, with a median of 17,723. The reproducibility of the positive test results by the initial MOTA score category is shown in Table 1. Only 80.8% of the samples with initial MOTA scores between 2,000 and 9,999 were positive when retested. The reproducibility of results for samples with initial scores of greater than or equal to 10,000 was 96.7%. All of the 38 samples with scores greater than or equal to 27,000 were reproducibly positive.
|
|
The initial positive results for N. gonorrhoeae were confirmed for 90 samples (87.4%) with the first retest. The MOTA scores of the 13 samples that required a third test to resolve a discordance between the results of the first two tests are shown in Table 3. The initial MOTA scores of these samples ranged from 2,235 to 23,946, and only two were resolved as positives. One sample, no. 3070, gave discordant results in both the N. gonorrhoeae and C. trachomatis tests. As noted for the C. trachomatis test, discordant results were not associated with specimen type, run date, or operator (Table 3).
|
The clinical performance data for BDPT system tests given in the package insert showed that 28.9% of the low-positive test results for C. trachomatis and 42.3% of the low-positive test results for N. gonorrhoeae were false positives when arbitrated by culture, direct fluorescent-antibody testing, and another NAAT. Although we did not use alternative test methods in our algorithm, the nonrepeat rates we found for low positives were similar to the false-positive rates reported for these tests by the manufacturer. The false-positive rates reported by the manufacturer, together with the poor reproducibility of low-positive test results described here, support the establishment of a gray zone for these tests and supplemental testing of samples with test results in the gray zone. Our data suggest that the gray zone should be 2,000 to 9,999 for both tests.
Reproducibility problems are not unique to the BDPT system. Similar problems have been reported for the other NAATs (3, 5, 6). The sources of the poor reproducibility of low positives in the BDPT system are not clear. Lack of reproducibility was not associated with any one operator, as five of the six technologists performing the tests experienced similar problems (Tables 2 and 3). Degradation of target DNA and generation of amplification inhibitors were considered since some positive lysates were stored for as long as 4 days before testing was completed. We found no relationship between how long a lysate was stored and the reproducibility of the positive results; in fact, some of the MOTA scores were higher in the retests. In addition, the MOTA scores of the amplification control in those samples with poorly reproducible results remained consistent and robust in the retests, eliminating inhibition as an explanation (data not shown). More likely explanations include random events associated with primer, target, and enzyme interactions leading to signal generation in the absence of target DNA, sporadic target DNA carryover during sample transfer from tubes to microwells or between microwells during the procedure, and samples with organism concentrations near the limits of detection of the tests.
Sample cross-contamination is always a concern with NAATs and probably accounts for some of the nonreproducible results in our study. However, it should be stressed that none of the negative controls showed evidence of this problem, and there was no temporal clustering of the nonreproducible test results. These parameters are used in many laboratories to recognize sample cross-contamination problems. Replicate testing has been shown to increase the sensitivity of other nucleic acid amplification tests, particularly for samples with small amounts of target present. This has been attributed to sampling variability. Although our study was not specifically designed to evaluate the effect of replicate testing on the sensitivity of the BDPT system tests, we found that 2 (1%) of the 196 samples repeatedly positive for C. trachomatis and 2 (2.2%) of the 92 samples repeatedly positive for N. gonorrhoeae did not test positive on the first repeat. In addition, we performed replicate testing of the 92 samples that were initially negative for C. trachomatis and the 196 samples that were initially negative for N. gonorrhoeae because of positive results for the other organism. In each test, only one additional positive sample was detected. These results suggest that testing of all samples in duplicate would yield only modest increases in the sensitivities of these tests.
In summary, our data support the establishment of a gray zone for the BDPT system tests for C. trachomatis and N. gonorrhoeae and confirmation of low positives by retesting. The algorithm proposed here is a practical way for laboratories to limit the number of potentially false-positive test results without adding substantial costs to screening programs for these important pathogens.
ACKNOWLEDGMENTS |
---|
REFERENCES |
---|
Top Abstract Text References |
---|