| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
a author for correspondence: Department of Pathology and Laboratory Medicine, UMDNJ-New Jersey Medical School, 185 South Orange Avenue, Newark, NJ 07103, fax 973-972-3199, e-mail mulcahy{at}umdnj.edu
Several investigators have reported that the Roche AmplicorTM PCR assay for detection of infection by Chlamydia trachomatis in endocervical specimens (Roche Diagnostic Systems) has a sensitivity in the range of 89–97% (1)(2)(3)(4)(5)(6)(7)(8), although a few studies have found lower sensitivity, in the range of 64–86% (8)(9)(10). All of the referenced studies have reported a specificity of 99–100%. Recently, however, the reproducibility of the assay has been questioned by investigators who found that 13 of 35 cervical specimens that were tested and then retested with the assay showed great variability in absorbance values (11). Because there is a paucity of information available concerning the reproducibility of the Amplicor assay, we report herein data informative on the issue that were obtained at UMDNJ-University Hospital. The protocol for this study was approved by the Institutional Review Board of UMDNJ-New Jersey Medical School.
The reproducibility studies reported here were derived from two data
sets. The first set, 162 specimens, was drawn from a consecutive series
of 5011 endocervical specimens examined by the Amplicor assay for
detection of C. trachomatis (Table 1
). This series was begun on January 18, 1994, and extended to
May 27, 1994. The second data set, consisting of 717 specimens
collected during various periods extending from May 31, 1994, to
October 31, 1995, provided additional samples with differing
quantitative assay absorbance values (A450) on initial
examination. In all, 879 specimens (162 717) were used to assess
reproducibility of the assay. Three types of reproducibility studies
were performed: 571 specimens that were either positive or negative on
initial testing were subjected to duplicate-repeat testing on a
subsequent day; 79 specimens with initially equivocal results
(A450, 0.2–0.5) were subjected to duplicate-repeat testing
in triplicate (three sets of duplicate-repeat tests) to test the
manufacturer's algorithm for resolving the status of equivocal
results; and a separate group of 229 specimens was subjected to
within-run triplicate testing. Within each of the several absorbance
categories that were established and that are defined in Table 1
, the
specimens tested for reproducibility were randomly chosen.
|
In accordance with the manufacturer's directions, both initial and duplicate-repeat tests were performed within 10 days of specimen collection. One major exception to this protocol was that 47 of 132 specimens yielding initial absorbance values above 2.0 were retested between 11 and 17 days after specimen collection (all duplicate-repeat tests on these 47 specimens gave positive results). Approximately 70% of specimens were submitted by the hospital's obstetrics and gynecology services, but specimens were also received from the emergency room (~20%) and from other clinical units. Specimens were stored before testing at 2–8 °C.
The PCR assay was performed in accordance with instructions in the package insert of the Amplicor C. trachomatis test kit (12), and the results of testing were interpreted in accordance with instructions in the insert: an A450 <0.2 was considered negative; values from 0.2 to 0.5 were considered equivocal; and values >0.5 were regarded as positive. Duplicate-repeat testing was performed on samples with equivocal values, which were then determined to be either positive or negative in accordance with the manufacturer's criteria. Samples with two of three test results (the initial value and the pair of repeat values) showing an absorbance >0.25 were considered to be positive, and samples with two of the three results <0.25 were considered to be negative. The assay was performed in the clinical chemistry laboratory of UMDNJ-University Hospital, taking care to maintain separate working areas for reagent preparation, specimen preparation, and amplification and detection.
Systat 5.03 (Systat, Inc.) was used to calculate the confidence
intervals shown in Table 1
.
Table 1
shows the distribution of results obtained in the series of
5011 specimens and includes a breakdown of positive results based on a
stratification of absorbance values. The overall rate of positive
results for the series was 9.7%.
Table 1
also provides a summary of the results of repeat testing
performed on 650 specimens: 332 initially negative, 239 initially
positive, and 79 initially equivocal. The 332 negative specimens gave
consistent (negative) results in 99.1% of duplicate-repeat tests. As
shown in Table 1
, the rate of reproducibility of positive results on
duplicate-repeat testing varied greatly according to initial absorbance
values, from 47.4% for specimens with initial absorbance values >0.5
to 1.0 to 98.8% for specimens with initial absorbance values >3.5. In
calculating reproducibility rates for initially positive test results,
a repeat test result was considered to be in agreement with the initial
result if the manufacturer's criteria for a positive test result were
met; quantitative agreement of absorbance values was not required.
Initially equivocal results represented a special case in this evaluation. For only 63 (80%) of the 79 specimens that gave initially equivocal results and on which duplicate-repeat testing in triplicate was performed did the manufacturer's algorithm produce consistent results as to whether specimens were positive or negative (49 negative and 14 positive). Discrepant results were obtained for the other 16 specimens (20%). In each of these 16 cases, two sets of duplicate-repeat test results led to one conclusion as to whether the specimen was positive or negative, but the third set of results led to the opposite conclusion.
The overall rate of reproducibility of the Amplicor assay was
calculated to be 98.4%. This rate was determined by summing the
products obtained when the rate of reproducibility for each absorbance
subset of initial results (Table 1
) was multiplied by the frequency of
the respective subset in the consecutive series of 5011 specimens (also
shown in Table 1
). Specimens yielding absorbance values <0.2 and >3.5
contributed most to the high rate of overall reproducibility for the
assay because (a) absorbance values <0.2 and >3.5 together
accounted for 96.6% of all assay results in the consecutive series of
5011 specimens, and (b) both of these absorbance categories
produced consistent results on replicate testing in ~99% of cases.
Specimens with absorbance values ranging from 0.2 to 3.5, which had
lower rates of reproducibility, accounted for only 3.4% of total
accessions, and thus had relatively little effect on the rate for
overall reproducibility of the assay. The overall rate of
reproducibility for initially positive results was calculated to be
94.5%.
The high reproducibility on duplicate-repeat testing was confirmed in the series in which triplicate analyses of specimens were performed on the day of initial testing. Of 229 specimens in this series, 227 (99.1%) gave results that were either consistently positive (18) or consistently negative (209).
Only limited data on the reproducibility of the Roche Amplicor chlamydial assay in testing endocervical specimens have been published. In a series of 501 cervical and urethral specimens for which duplicate assays were performed (13), 10 (2%) gave discrepant results.
The causes of such discrepant results as we observed have not been established. Contamination of specimens by chlamydial DNA is always difficult to rule out completely as a possible cause of false-positive results, and specimen mix-up and technical error are also possible explanations for the discrepancies encountered. In this case, however, other explanations seem more probable. Negative and positive controls have consistently produced expected results in our runs of the assay, and we have found no evidence for a "nearest neighbor" cross-contamination effect causing false-positive results.
For those specimens yielding negative test results initially but positive results on repeat testing, two likely causes of false-negative reactions especially must be considered. One is that the specimens could have contained polymerase inhibitors that were active at the time of initial testing but then became inactivated in the interval between initial and repeat testing. A second possibility is that the specimens could have had low copy numbers of the plasmid target for amplification, and therefore, particular aliquots of the specimens could have had numbers of target plasmids either above or below the threshold of the assay detection limit.
In respect to polymerase inhibition of the Amplicor assay, investigators have reported instances in which specimens yielded negative results, but after storage for a variable number of days (4)(5)(8)(9)(14), after phenol-chloroform or phenol extraction (1)(2)(8), or after specimen dilution (4)(9), gave positive results on repeat testing. Verkooyen et al. (15) found that, in specimens to which inocula of C. trachomatis were added, a combination of 10-fold dilution of specimens with heating to 95 °C for 10 min before testing produced an inhibition rate of only 4.1%, compared with a rate of 19% for untreated samples.
Support for the importance of copy number in detection of chlamydial DNA with the Roche assay was provided by the work of Toye et al. (7). Roche's package insert for the Amplicor chlamydial assay states that the detection limit of the assay is one inclusion-forming unit, or ~10 plasmid copies (12), but Miyashita et al. (16) estimated that the detection limit of the assay is ~20–40 plasmid copies. Another possible cause of false-negative results, loss of chlamydial target DNA from specimens, is improbable, given DNA's well-known resistance to degradation.
For specimens that gave positive results on initial testing and negative results on repeat testing, low copy number is again a plausible explanation. Furthermore, as suggested by Mahony et al. (4), polymerase inhibitors could also cause such a sequence of testing results if an initially inactive inhibitor became activated with aging of a specimen.
The results produced by the Amplicor assay for C. trachomatis are, overall, highly reproducible, but still show room for improvement. Clearly, if the information presented above on the resolution of equivocal results is confirmed by others, the algorithm constructed by the manufacturer for such resolution will require modification. In fact, there may be a small number of specimens for which results should be reported as indeterminate rather than as either positive or negative. In respect to the problem of false-negative results due to polymerase inhibitors, a number of investigators have recommended that an internal control (to flag possibly false-negative results) be incorporated into the assay's protocol (4)(15)(17); pretreatment of specimens to reduce the activity of inhibitors, e.g., by dilution and heating, has also been advocated (4)(15). These measures appear quite reasonable, although in the case of dilution the potential benefit of a reduction in polymerase inhibitors must be weighed against the undesired effect of a reduction in plasmid copy number. It is notable that the Roche Cobas Amplicor automated PCR methodology includes an optional internal control (15)(18) and that the Cobas procedure for detection of C. trachomatis defines a category of "repeatedly equivocal" results that require additional testing on a new specimen for resolution (19).
Footnotes
UMDNJ New Jersey Medical School, Newark, NJ 07103
References
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  E. E. Culler, A. M. Caliendo, and F. S. Nolte Reproducibility of Positive Test Results in the BDProbeTec ET System for Detection of Chlamydia trachomatis and Neisseria gonorrhoeae J. Clin. Microbiol., August 1, 2003; 41(8): 3911 - 3914. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Van Der Pol, D. V. Ferrero, L. Buck-Barrington, E. Hook III, C. Lenderman, T. Quinn, C. A. Gaydos, J. Lovchik, J. Schachter, J. Moncada, et al. Multicenter Evaluation of the BDProbeTec ET System for Detection of Chlamydia trachomatis and Neisseria gonorrhoeae in Urine Specimens, Female Endocervical Swabs, and Male Urethral Swabs J. Clin. Microbiol., March 1, 2001; 39(3): 1008 - 1016. [Abstract] [Full Text]
|
|
|
|
|
|
A. M. Gronowski, S. Copper, D. Baorto, and P. R. Murray Reproducibility Problems with the Abbott Laboratories LCx Assay for Chlamydia trachomatis and Neisseria gonorrhoeae J. Clin. Microbiol., June 1, 2000; 38(6): 2416 - 2418. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
