Evaluation of sequential testing strategies using non-amplified and amplified methods for detection of Chlamydia trachomatis in endocervical and urine specimens from women

Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51

www.elsevier.com/locate/diagmicrobio

Evaluation of sequential testing strategies using non-amplified and amplified methods for detection of Chlamydia trachomatis in endocervical and urine specimens from women Heather Semeniuk, Ali Zentner, Ron Read, Deirdre Church* From Calgary Laboratory Services, (HS, DLC) and the Departments of Pathology & Laboratory Medicine, Microbiology and Infectious Disease and Medicine, University of Calgary (AZ, RR, DLC), Calgary Alberta, Canada Received 2 October 2001; accepted 2 October 2001 These data were presented in poster form at the 99th American Society for Microbiology Meeting, in Chicago, Ill. However, the manuscipt has not been submitted for publication elsewhere.

Abstract Nucleic acid amplification tests (NAAT) are more sensitive than other methods for the diagnosis of Chlamydia trachomatis (CT) genital infections. Two unique sequential testing strategies that employed two different commercial NAAT methods to detect CT in a population of women with widely varying infection risk were evaluated. Specimens from 504 women aged 15 to 75 years were studied. Two endocervical swabs and a urine sample were collected from each woman. One swab was initially tested using the Access enzyme immunoassay (EIA) (Beckman). An aliquot from the EIA extraction was subsequently amplified using the COBAS AMPLICOR CT assay (PCR) (Roche). The second swab was initially tested using the PACE 2 CT hybridization assay (Gen-Probe). An aliquot was pipetted off prior to performing the PACE 2 assay and also amplified using the AMP-CT assay (TMA) (Gen-Probe). Urine samples were tested for CT using both NAAT methods. True CT infections were defined as any woman that was confirmed to be positive on both NAAT results from endocervical swabs. The results of all other CT assays were compared against this expanded gold standard. 28 women were confirmed to have CT infection giving an overall prevalence of 5.6%; low-risk women had a rate of 1.3% while high-risk women had a rate of 9.8%. NAAT methods have a higher sensitivity for detecting CT cervicitis when swabs are tested compared to urine. The positive predictive value of NAAT is decreased when testing low risk women. Limited automation makes it difficult to test a high volume of samples (i.e., ⬎ 100 swabs and/or urines) using either of these NAAT methods and continue to provide same day results. Laboratories performing CT testing must define the female population served so that appropriate diagnostic strategies can be employed. © 2002 Elsevier Science Inc. All rights reserved.

1. Introduction C. trachomatis (CT) is one of the most prevalent bacterial pathogens that are sexually transmitted (Black, 1997; Pearlman et al., 1992). In Canada, the national rate of genital chlamydial infections currently stands at 112.7/ 100,000 and this is substantially higher than that of either gonorrhea or syphilis (Peeling et al., 1999). Females at high risk for acquiring CT genital infection are generally young (i.e., between 15 and 30 years of age) and sexually active (Black, 1997; Centers for Disease Control and Prevention, 1993). It is important to confirm a diagnosis of CT cervicitis * Corresponding author. Tel.: ⫹1-403-209-5281; fax: ⫹1-403-2095347. E-mail address: [email protected] (D. Church).

because undetected genital infection may cause ectopic pregnancy, pelvic inflammatory disease and salpingitis with secondary tubal scarring and infertility (Black, 1997; Cates et al., 1991; Pearlman et al., 1992). Women with CT cervicitis may also transmit the infection to their sexual partners, and have an increased risk of acquiring human immunodeficiency virus infection (Centers for Disease Control and Prevention, 1993; Wasserheit, 1991). Molecular methods, particularly nucleic acid amplification technologies, have been shown to be much more sensitive compared to non-molecular methods for the diagnosis of genital CT infection (Black, 1997; Black, 1998; Crotchfelt et al., 1998; Dean et al., 1998; Ferraro et al., 1998; Lauderdale et al., 1999; Mahony et al., 1994; Morre´ et al., 1999; Newhall et al., 1994; Skulnick et al., 1994; Wu et al., 1992; Wylie et al., 1998). CT amplification methods have

0732-8893/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 2 - 8 8 9 3 ( 0 1 ) 0 0 3 1 5 - 7

44

H. Semeniuk et al., / Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51

therefore been widely implemented in clinical laboratories over the last few years. However, we were concerned about the positive predictive value of highly sensitive CT amplification tests in our region, since we primarily test women with a low risk of infection (i.e., prevalence rate is ⬍5%) (Fletcher et al., 1982). In addition, we questioned the feasibility of introducing this technology as the primary testing method in a high volume laboratory setting (i.e., ⬃300 assays/day). A recent paper by Chiu et al., 1999 described a sequential testing strategy for detection of CT in endocervical swabs where the extracted sample was initially tested using an enzyme immunoassay method, and an aliquot of the extract was subsequently amplified. No patients who initially tested CT negative by the EIA methods were subsequently found to be CT positive by the more sensitive amplification method (Chui et al., 1999). A sequential test strategy that employed an initial non-amplified CT detection method with secondary testing of equivocal samples (i.e., using an expanded gray zone) by an amplified method may be a feasible alternative strategy in a high volume laboratory provided there were no technical or performance issues. This study therefore evaluated the performance and feasibility of two sequential testing strategies using non-amplified and amplified methods for detection of CT in endocervical swabs compared to amplified testing on urine samples from women with varying infection risk.

2. Materials and methods 2.1. Laboratory setting Calgary Laboratory Services (CLS) is a large centralized regional laboratory that performs most microbiology services for an urban population of almost a million people. Microbiology services are provided to all acute care hospitals, long-term care facilities and out-patient locations in the community. Each day over 2,000 microbiology specimens are received for analysis, and ⬃300 of these are genital specimens for detection of Chlamydia trachomatis infection. Most of the CT genital specimens are collected from women (97%) with a much smaller number of swabs being submitted from men (3%). 2.2. Patient population This study was reviewed and given approval by the Conjoint Health Research Ethics Board at the University of Calgary. This study was restricted to women, aged 15 to 75 years being tested for cervical C. trachomatis (CT) infection. 504 women in the Calgary region were enrolled from the following sites: 1) physician’s offices, 2) obstetric, gynecology and family practice clinics at the hospitals, and 3) the Emergency Departments. The Regional Sexually Transmitted Diseases (STD) clinic also enrolled women during the study and sent samples to CLS for testing although the

Provincial Laboratory is their usual service provider. Women attending the STD clinic represent an important control group because their risk for acquisition of genital CT infection is higher that the general female populace. Also, the staff at the STD clinic is proficient at collecting specimens from the endocervix. This was considered to be another important factor to control since the performance of C. trachomatis tests decreases if inadequate cellular material is collected from the cervix (Welsh et al., 1997). 2.3. Study design A brief history was completed that documented each patient’s infection risk for acquiring genital CT infection according to previously published criteria (Black, 1997; Centers for Disease Control and Prevention, 1993). Fig. 1 outlines the overall study design. Separate sequential testing strategies were used to evaluate CT amplification detection methods performed on two separately collected endocervical swab specimens. Physicians’s collection kits were constructed so that the order of collection of endocervical swabs for the two sequential testing strategies was alternated between sequentially enrolled women. The endocervical swab specimen for the PACE 2 and AMP CT assay was placed in Gen-Probe transport medium (Gen-Probe, San Diego, CA), transported at room temperature and then stored at 2 to 8°C until it was processed. The other endocervical swab was placed into the enzyme immunoassay transport tube, transported and stored in a similar manner. One swab was used to initially perform an enzyme immunoassay (EIA) (Access, Beckman, Missisauga, Ontario, Canada), and an aliquot of the extract was then tested using a polymerase chain reaction (PCR) method (COBAS AMPLICOR CT, Roche Diagnostic Systems, Branchburg, NJ, USA). The other swab was first tested using the PACE 2 assay (Gen-Probe, San Diego, CA, USA) and an aliquot taken off prior to performing the PACE 2 assay was additionally tested by the AMPLIFIED CT (AMP-CT) assay (Gen-Probe). A first void urine (FVU) was also collected from each patient and a dipstick chemical urinalysis was done on a urine aliquot using the Chemstrip-10 (Boehringer Mannheim, Ontario, Canada), in addition to performing both CT amplification tests (PCR and AMP-CT). A Pap smear was collected and submitted to the Cytopathology laboratory, and read for inflammatory changes using standard microscopic criteria (Paavonen et al., 1990). All initial molecular CT assays were completed within 48 h and repeat tests within 72–96 h after receipt of the specimens by the laboratory. 2.3.1 Access EIA assay. Specimens were tested the same day they were received. Prior to performing an EIA assay, a 100 ␮L aliquot of the swab extract was pipetted off and refrigerated at 4°C for molecular testing by PCR (Roche Diagnostic Systems, Branchburg, NJ, USA). The Access

H. Semeniuk et al., / Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51

45

Fig. 1.

EIA assay (Beckman) was performed according to the manufacturer’s instructions. All specimens falling within the equivocal zone for the Access EIA assay had a confirmatory blocking assay performed. The confirmatory assay uses a blocking antibody to inhibit anti-chlamydia antibody. The blocking assay was run in parallel with a retest of the specimen by the manufacturer’s protocol. A reduction of at least 50% in the optical density in the blocked assay compared to the unblocked test indicated that the specimen was a true positive. A direct fluorescent antibody assay was also performed as a secondary confirmatory test using a standard procedure (Chan et al., 1994; Kellogg et al., 1993). 2.3.2 COBAS AMPLICORTM CT assay. The COBAS AMPLICOR TM CT assay uses the PCR to coamplify target DNA along with an internal control. The amplified products are then hybridized to oligonucleotide probes specific to the target(s), and the probe bound to the amplified product is detected by color formation. The COBAS AMPLICORTM CT assay was performed according to the manufacturer’s instructions on first void urine samples. The extracted endocervical swab sample that was pipetted off before performing an EIA test, was also assayed using the following modification of the recently published method (Chui et al., 1998); 200 ␮L of lysis buffer (COBAS AMPLICORTM, Roche Diagnostic Systems Inc., Branchburg, NJ) was added to 100 ␮L of the previously pipetted EIA extract and incubated at room temperature for 10 min. 50 ␮L of each processed specimen was used for the assay. For PCR, the specimen absorbency result was interpreted as follows: a)

⬍0.2 (CT assay negative), b) ⱖ0.8 (CT assay positive), and c) ⱖ0.2 to ⬍0.8 (CT assay equivocal). PCR assays were repeated for all equivocal results, and for CT negative assays with a negative internal control. Specimens with initial negative internal control reactions were either diluted or underwent a freeze-thaw procedure before re-assay. 2.3.3 PACE 2 assay. The PACE 2 (Gen-Probe, San Diego, CA) assay was performed on endocervical specimens according to the manufacturer’s instructions. After the 60°C, 10-min. incubation step, 100 ␮L of sample was removed and refrigerated at 4°C for molecular testing by the AMP-CT assay. The remainder of the sample was used for the PACE 2 assay. The PACE 2 assay utilizes a chemiluminescent labeled, single-stranded DNA probe that is complementary to the ribosomal RNA of CT. Hybridized probe was detected by a chemiluminescence reaction with a GenProbe LEADER 450I luminometer. Specimen results are measured in relative light units (RLUs) as follows: a) ⬍350 (CT negative), b) ⱖ600 (CT positive), and c)350 to ⬍600 (equivocal). All specimens with initial equivocal results were repeated. 2.3.4 AMP-CT assay. The AMP-CT (Gen-Probe, San Diego, CA) assay utilizes transcription-mediated amplification (TMA) technology to amplify target CT ribosomal RNA. The AMP-CT assay was performed according to the manufacturer’s instructions on urine samples. The AMP-CT assay was also used to test the endocervical swab sample aliquot that had been pipetted off prior to testing by the

46

H. Semeniuk et al., / Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51

PACE 2 assay. A portion of the diluted specimen was mixed with amplification reagent followed by the addition of the necessary enzymes for TMA. The reaction was terminated following a 1-h incubation at 42°C. A DNA detection probe was added that recognized a sequence within the amplified product and allowed to hybridize for 15 min. at 60°C. A selection reagent was then added to inactivate nonhybridized probe. Hybridized probe bound to amplified product was detected in the same way as the PACE 2 assay. Specimen results in RLU’s were interpreted as follows: a) ⬍40,000 (CT negative), b) ⬎ 500,000 (CT positive), and c) 40,000 –500,000 (CT equivocal). All specimens giving initial equivocal results were repeated. 2.4. Discrepant analysis A true CT infection was defined as any patient that was positive on both endocervical swab nucleic acid amplification tests (PCR and AMP-CT). Since women with a low prevalence of CT infection were being tested, it was important to use this expanded gold standard to confirm true cases (Black, 1997). Repeat testing of endocervical swab and/or urine samples was performed to resolve any discrepant results between the two amplified CT methods. 2.5. Data analysis Data were entered into a statistical software program (SPSS Version 6.0) and analyzed using standard methods to calculate the sensitivity, specificity, positive and negative predictive values of each test method. The performance of all other CT assays was compared to the gold standard outlined above. Patients were designated as being ‘highrisk’ for CT infection if they met previously published criteria (Black, 1997). Women with minimal or no risk of acquiring CT infection were designated as the low-risk group. Clinical symptoms were not useful in evaluating the performance of the various test methods because regardless of CT infection risk most women were asymptomatic.

3. Results A total of 504 women were enrolled in this study; 323 (65.1%) women were enrolled from the regional STD clinic and the other 181 (34.9%) women had no identified risk factors for acquiring CT infection. 28 women were confirmed to have cervical CT infections by the laboratory yielding an overall prevalence rate of 5.6% (95% CI ⫽ 0.0399 – 0.851). However, the prevalence rate of CT infection was approximately eightfold greater in high-risk women (9.8%) compared to low-risk women (1.3%). Most of the CT positive women (26, 93%) had one or more risk factors for acquiring CT infection. Only a few CT positive women (2, 7%) had no identified risk factors for CT infection. Almost all of the laboratory confirmed CT positive

cases (27, 96.4%) occurred in women between the ages of 15 to 30 years and only one women ⬎ 40 years of age had confirmed CT infection. Women with CT infection did not have pyuria as a marker for urethritis. No correlation was also found between the presence of cervical CT infection and any cytological changes on Pap smear. 3.1. CT assay results A total of 504 endocervical swabs in each of the two sequential CT testing arms of the study (Fig. 1) were tested by each of the CT assay methods. Urine samples were also submitted from 491 (97.4%) women and assayed by both amplification methods. A total of 2998 initial CT assays were therefore performed during the study. Table 1 shows the overall performance of the various CT assays including all positive test results prior to the resolution of discrepant results. The Access EIA assay had the lowest sensitivity for detecting CT in endocervical swabs, and the PACE 2 hybridization assay was less sensitive than either of the amplified CT assays. Both the AMP-CT and COBAS AMPLICORTMCT assays had equivalent sensitivity and detected all of the confirmed CT infections from endocervical swabs. Although all of the molecular assays had high specificity and negative predictive value, the positive predictive value was not as good. The COBAS AMPLICORTMCT assay had the lowest Positive Predictive Value because of the assay’s higher false-positivity rate. Urine amplification results were less sensitive than the assay results from endocervical swabs. The COBAS AMPLICORTMCT assay was more sensitive that the AMP-CT assay when testing urine specimens. However, the COBAS AMPLICORTMCT assay also had a higher false positive rate on urine tests as reflected by the lower Positive Predictive Value for this assay. 3.1.1 Resolution of discrepant results. During sequential testing of endocervical swabs, CT positive patients were found to have test results within or lower than the manufacturer’s gray zones for both the Access EIA and PACE 2 assays. Table 2 outlines the clinical information and resolution of all results that were initially negative or equivocal by either of the non-amplified CT detection methods. Access EIA had the highest number of false negative results. A total of eight endocervical samples from seven women initially tested either CT negative or equivocal (D/E) by the EIA assay, and four of these samples were also negative by the PACE 2 assay. 4 (67%) of these women had been previously confirmed to have genital CT infections, and 3 of them had recently completed a course of appropriate antibiotic therapy. One of these women had had prior genital CT infection 2-years before and despite having no recent risk factors for acquiring infection, tested positive for CT by both amplified methods. 2 (33%) women had urethritis but no cervicitis as the primary site of CT infection. Access EIA had no false negative results. The false

H. Semeniuk et al., / Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51

47

Table 1 Performance of various assays for detection of Chlamydia compared in endocervical swabs and urine samplesa Test

Endocervical Swabs: Access EIA PACE 2 AMP-CTb COBAS AMPLICOR™ CT Urine:c AMP-CT COBAS-AMPLICOR CT

True Positivea

False Positive

True Negative

False Negative

Sensitivity (%)

Specificity (%)

Positive Predictive Value (PPV) (%)

Negative Predictive value (NPV) (%)

Number of Repeats

20 24 28 28

0 5 2 7

476 471 474 469

8 4 0 0

71.4 85.7 100 100

100 98.9 99.2 98.5

100 82.8 87.5 80.0

98.3 99.2 100 100

3 (0.6%)d 5 (0.9%)e 4 (0.8%)f 24 (4.8%)g

21 23

0 3

465 462

5 3

80.8 88.5

100 99.4

100 88.5

99.0 99.4

3 (0.6%)h 37 (7.5%)i

a

Performance compared against expanded gold standard of both amplified tests positive from the endocervical swab. Based on 28 women confirmed to have Chlamydia infection out of 504 enrolled. c 491 women submitted urines; 2 women with Chlamydia infection did not submit a urine sample. d Confirmation of equivocal results. e 5 false positive tests occurred at the start of manual testing. f 2 initial tests were false positive; 2 equivocal test were repeated and confirmed to be negative. g 7 initial tests were false positive; 4 of which occurred in one day, 17 tests were repeated because the internal control was negative. 9 of the PCR swab. internal control negative results were obtained using a master mix that was later recalled by the manufacturer. h Confirmation of false-negative results. i 3 initial tests were false positive; 1 false-negative test was confirmed; 33 tests were repeated because the internal control was negative. b

factors for CT infection and was asymptomatic of cervicitis, this result was therefore considered to be false positive. There were 4 endocervical swabs that initially gave positive AMP-CT results, but 2 of these fell within the test gray zone. Repeat testing of these 4 samples confirmed that 3 of them were negative including the 2 that initially gave grayzone results. The AMP-CT assay that remained positive after repeat testing was noted to contain a lot of mucous, and was therefore considered to be a false positive since the

positive PACE 2 results (5 patients) all occurred during the start-up phase of testing and reflect the need to be technically fastidious when performing manual hybridization tests. Repeat testing of 3 endocervical swabs that were initially only positive when tested by the COBAS AMPLICORTMCT assay confirmed that all except one was negative. One endocervical sample was repeatedly positive by PCR but negative by all other methods including PCR on the patient’s urine sample. Since the patient had no risk

Table 2 Clinical characteristics of women initially testing CT negative according to either non-amplified method Patient

1

Age (yrs.)

22

Risk of infection

High

Chlamydia Test Results

Clinical information

EIA

PACE 2

PCR

AMP-CT

⫺

⫺

⫹

⫹

● ● ●

2

19

High

⫺

⫺

⫹

⫹

3

24

High

⫺

⫺

⫹

⫹

● ● ● ● ● ●

4

27

High

⫺

⫺

⫹

⫹

● ● ●

5

29

Low

⫺

⫹

⫹

⫹

6a

23

High

⫺

⫹

⫹

⫹

High

⫺ ⫺

⫹ ⫹

⫹ ⫹

⫹ ⫹

● ● ● ●

7

31

● ● ● ●

a

Patient was repeat tested

Previous Chlamydia positive Cervicitis Treated with Azithromicin Urethritis On antibiotics Previous Chlamydia positive Cervicitis Vaginal discharge On antibiotics Urethritis Vaginal discharge On antibiotics Previous Chlamydia positive (2 years ago) Routine Pap smear visit Routine Pap smear visit Cervicitis On antibiotics Previous Chlamydia positive (4 and 5 years ago) Vaginal discharge On antibiotics

48

H. Semeniuk et al., / Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51

Table 3 Comparison of analysis time for amplified Chlamydia tests Time (hours) Method

Number of samplesa

Work listb

Specimen Preparationc Urine

Swab

Amplificationd,e

Detectionf,g,h

Total assay time

Hands on technologist time

Urine

Swab

Urine

Swab

PCR (Roche) 25 50 100 150

0.3 0.4 0.6 0.8

1.3 1.7 2.7 3.0

0.6 0.8 1.1 1.4

2.3 4.5 9.0 13.5

2.3 2.6 3.0 3.4

6.2 9.2 15.3 20.7

5.5 8.3 13.7 19.1

2.1 2.8 4.6 5.8

1.3 1.8 3.0 4.2

25 50 100 150

0.3 0.4 0.6 0.8

0.8 1.0 1.8 2.7

0.6 0.8 1.3 1.9

1.7 1.7 3.3 5.0

1.3 1.4 2.7 3.9

4.1 4.5 8.4 12.4

3.9 4.3 7.9 11.6

3.1 3.5 6.4 9.4

2.9 3.3 5.9 8.6

TMA (Gen-Probe)

a

Approximate number of specimens. A PCR batch consists of 22 specimens and 2 controls; a TMA batch consists of 47 specimens and 2 controls. Includes instrument maintenance time. c Includes reagent preparation time. d Includes time required to load the instrument for the PCR assay. e PCR Detection times remain relatively constant while Amplification times increase because when the COBAS instrument is programmed in Parellel versus Basic Mode, amplification and detection can occur simultaneously in sequential batches, therefore, decreasing the total time to completion. f Includes time required to reload the instrument when running PCR assay in Parallel g Includes time required to load the luminometer for the Amp CT assay h Includes review of results and clean-up time b

patient had no risk factors for CT infection and was asymptomatic. Most of the false positive amplification results by either PCR or AMP-CT (7/11, 63.6%) from endocervical swab tests occurred in low-risk women. The AMP-CT assay on urine samples gave no falsepositive results, but was less sensitive than the PCR assay when testing urine samples. PCR had a higher rate of falsepositive results on urine samples compared to the AMP-CT assay. There were 3 urine samples that initially tested CT positive using PCR but gave negative results in the AMP-CT assay. Repeat testing confirmed these 3 urine specimens to be CT negative in agreement with the negative cervical CT results. All of the false-positive PCR urine tests occurred in low-risk women. 3.1.2 Repeat tests and the technical feasibility of high volume amplification. The number of tests that had to be repeated for each CT assay is also shown in Table 1. The COBAS-AMPLICORTMCT assay had the highest repeat factor for both endocervical swabs as well as first void urine samples. Repeat PCR tests had to done on 33 urine specimens because the internal control detected the presence of substances that may interfere with the amplification. Repeat PCR testing on either a diluted sample or a freeze-thaw aliquot confirmed that all of these samples were CT negative. One other urine specimen initially tested false-negative by the COBAS AMPLICORTMCT assay (internal control and specimen were negative), but repeat PCR testing confirmed this sample to be positive (internal control and specimen positive) in agreement with all of the other test results.

This was the only first-void urine in which inhibitory elements in the sample caused an erroneous PCR result. All of the other assay methods had a ⬍1% repeat rate. The time required to perform each of the steps in the COBAS AMPLICORTMCT and AMP-CT assays on both endocervical swabs and urine was assessed (Table 3). A maximum of 110 COBAS AMPLICORTMCT assays could be performed using a single instrument and a dedicated technologist who continuously reloaded and reprogrammed the instrument throughout a twelve-hour shift. After sample processing for PCR, the technologist would have to complete five sequential instrument runs on 22-samples. Throughput for PCR was improved if all of the batched samples were prepared at one time. No hands-on-technologist time was required for the amplification and detection steps of the COBAS AMPLICORTM CT once the instrument was loaded. However, a high volume laboratory (i.e., ⬎ 100 tests/day) would need more than one instrument running in parallel in order to complete all of the work in the same day the sample was received. Using the equipment we had, a maximum of forty-seven specimens and 2 controls can be set-up for AMP-CT assay in a batch. During the 60 min incubation period a second batch could be set up. Therefore, up to 141 tests could be completed using the AMP-CT assay by a dedicated technologist in a twelve-hour shift because batch testing of three runs of 49 tests each could be performed in sequence. Sample preparation for both amplification assays (PCR and AMP-CT) was more laborious when processing urine samples than endocervical swab assays. A lower number of

H. Semeniuk et al., / Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51

amplified tests could therefore be performed on urine samples than endocervical swabs in the same shift.

4. Discussion This study evaluated the performance and feasibility of two sequential strategies using non-amplified and amplified methods for detection of CT in women with varying risk of infection. Amplified CT methods detected more infections in high-risk patients due to the substantially improved performance of these assays compared to non-amplified methods, particularly enzyme immunoassay. However, despite the excellent sensitivity and specificity of amplified CT assays in detecting ‘true’ infections, this study also highlights the decreased positive predictive value of this technology in low prevalence populations (i.e., ⱕ 5%) (Black, 1997; Dean et al., 1998; Mahony et al., 1994; Morre´ et al., 1999; Skulnick et al., 1994). Approximately two-thirds of all false-positive CT amplification results on endocervical swabs and all false-positive results on urine samples occurred in women who did not have any risk factors for infection. Although the current guidelines do not recommend confirmation of amplified CT tests, an expanded ‘gold standard’ that required two different amplification methods to be positive was used in our study to confirm true CT infections (Black, 1997; Black, 1998). That is, all falsepositive amplification tests were resolved through repeated testing of the sample not only using the same initial method, but also repeat testing using a second amplified method. However, because of the current high cost of commercial amplified assays, our laboratory would not have the resources to confirm all initially positive amplified CT results by a second different amplified method in routine practice. Because there is frequently little or no patient information provided on most microbiology requisitions, laboratorians cannot determine which women have false-positive CT results. Physicians also cannot easily rectify a false versus true CT result on clinical grounds, since up to 70% of all women with confirmed CT infection are completely asymptomatic (Black, 1997; Peeling et al., 1999). Based on the prevalence rate of CT infection in the population our laboratory currently tests and our current workload, it is estimated that routine use of a single amplified CT method in our high volume laboratory, could potentially lead to 2–3 false-positive CT test results a day being reported on low-risk women. At best this would lead to over-treatment of these women and their partners with antibiotics. However, in the worst-case scenario use of this highly sensitive technology without confirmation may result in litigation due to the psychological, social and medical consequences of falsepositive test results. The feasibility of using a sequential testing strategy for endocervical swabs as an alternate approach to amplifying all samples in a high volume laboratory (i.e., ⬃300 CT assays/day) was also evaluated. A non-amplified CT assay

49

was first performed followed by an amplified CT assay (i.e., EIA followed by PCR or Gen-Probe PACE2 followed by Gen-Probe AMP-CT). Only one other study has evaluated this approach using a sequential EIA to PCR algorithm (Chui et al., 1999). Prior to running the initial non-amplified CT assay in either sequential test algorithm, all of the swab extract aliquots have to pipetted off into separate tubes, labeled and stored. Aliquots cannot be taken off after either the EIA or hybridization tests (PACE 2) are completed because performing secondary amplification assays in this manner gives a much higher number of discrepant results (data not presented). Pipetting an aliquot from every specimen is not only labor intensive, but repetitive manipulation of the sample introduces a potential source of contamination when subsequently performing highly sensitive amplification assays. During this study, dedicated research technologists with molecular biology training and experience pipetted off all of the swab aliquots and performed all of the secondary amplification assays. However, a rotating pool of laboratory assistants and technologists would routinely perform these tasks in the clinical laboratory on our evening shift. Both sequential testing strategies were tried in our routine laboratory operation during the study, and a much higher rate of false-positive test results was obtained (unpublished data). Use of either sequential testing strategy as performed in this study also does not allow for either amplification assay to be performed strictly according to the manufacturer’s instructions. For these technical reasons alone, utilization of a sequential testing strategy as evaluated in this study is not recommended for the clinical laboratory. An alternate approach to using a sequential test strategy in a high volume laboratory would be the use of pooling of multiple samples (i.e., 4 or more) to initially detect CT using an amplified method within the pool (Morre´ et al., 2000; Toye et al., 1998). Only if the pooled sample is positive would additional individual tests be done of individual samples within the pool to determine which patient(s) are actually CT positive. Since all individual samples need to be tested if the pool is positive, this approach only makes sense if one is testing a low-risk population of patients. However, the same technical concerns arise with this approach as using a sequential test strategy. Individual samples need to be fastidiously aliquotted to create a pool in order to minimize the potential for carryover and contamination of a positive sample within the pool. Our study confirms that the current limited automation of commercially available CT amplification assays make it difficult to universally implement this technology as the primary testing approach in a high volume laboratory (i.e., ⬎ 100 samples/day) (Jungkind et al., 1996). In order to routinely perform our current volume of CT tests each day using either amplification method (PCR or AMP-CT) it is estimated that 2–3 highly trained technologists would need to be hired in order to maintain our current level of service using a non-amplified method (i.e., same day turnaround

50

H. Semeniuk et al., / Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51

time). In addition, new space would be needed to perform PCR since our laboratory would need three instruments running full-time during a 12 h shift. A significant number of urine assays would also need to be repeated the next day due to inhibition by the PCR assay as shown by this study and previously published (Toye et al., 1998). In order to run the AMP-CT assay we would need two luminometer instruments and an additional waterbath and heating block. Therefore, it is not currently possible for our high volume laboratory to universally employ amplified methods for detection of genital CT infections due to the substantially higher costs/test that this technology incurs. Prior to implementing amplified CT assays, laboratories need to understand the population being served so that appropriate testing strategies are employed. Highly sensitive and specific CT amplification tests should be used to test high-risk women in order to not miss genital CT infection and the potential clinical sequelae (Black, 1997; Cates et al., 1991; Pearlman et al., 1992). In addition, an amplified method should be used to test urine samples not only from men, but also sub-populations of women where an endocervical swab cannot be easily collected. However, until amplified technology becomes more automated, non-amplified CT assays with confirmation may be the prefered strategy when routinely testing low-risk women in a high volume laboratory. Acknowledgments The authors thank Beckman, Gen-Probe and Roche Diagnostics for their support of this work. The staff at the Sexually Transmitted Diseases Clinic, Family Planning Clinics and physicians offices in the community in Calgary provided tireless support to this study. References Black, C. M. (1997). Current methods of laboratory diagnosis of Chlamydia trachomatis infection. Clinical Microbiology Review, 10, 160 – 184. Black, C. M. (1998). “But Doctor, I’m Celibate!” The potential for, sources and implications of false-positive (and negative) results of tests for Chlamydia trachomatis. Clinical Microbiology News, 120, 12–124. Centers for Disease Control (CDC). (1993). Recommendations for the prevention and management of Chlamydia trachomatis infections. Morbidity and Mortality Weekly Report, 42(RR-12), 1–39. Cates, W. Jr., & Wesserheit, J. N. (1991). Genital chlamydial infections: epidemiology and reproductive sequelae. American Journal of Obstetrics and Gynecology, 164, 1771–1781. Chan, E. L., Brandt, K., & Horsman, G. B. (1994). A 1-year evaluation of Syva Microtrak Chlamydia enzyme immunoassay with selective confirmation by direct fluorescent-antibody assay in a high-volume laboratory. Journal of Clinical Microbiology, 32, 2208 –2211. Chui, L., Kakulphimp, J., Detwiler, B., & Prasad, E. (1999). An algorithm to detect Chlamydia trachomatis by polymerase chain reaction on specimens extracted for enzyme immunoassay. Diagnostic Microbiology and Infectious Disease, 32, 185–190. Crotchfelt, K. A., Pare, B., Gaydos, C., & Quinn, T. C. (1998). Detection of Chlamydia trachomatis by the Gen-Probe AMPLIFIED Chlamydia

trachomatis assay (AMP CT) in urine specimens from men and women and endocervical specimens from women. Journal of Clinical Microbiology, 36, 391–394. Dean, D., Ferraro, D., & McCarthy, M. (1998). Comparison of performance and cost-effectiveness of direct fluorescent-antibody, ligase chain reaction, and PCR assays for verification of Chlamydial enzyme immunoassay results for populations with a low to moderate prevalence of Chlamydia trachomatis infection. Journal of Clinical Microbiology, 36, 94 –99. Ferraro, D. V., Meyers, H. N., Schultz, D. E., & Willis, S. A. (1998). Performance of the Gen-Probe AMPLIFIED Chlamydia trachomatis assay in detecting Chlamydia trachomatis in endocervical and urine specimens from women and urethral and urine specimens from men attending sexually transmitted disease and family planning clinics. Journal of Clinical Microbiology, 36, 3230 –3233. Fletcher, R. H., Fletcher, S. W., & Wagner, E. H. (1982). Diagnostic tests. In Clinical Epidemiology-The Essentials (pp. 41–58). Baltimore: Williams & Wilkins. Jungkind, D., Di Renzo, S., Beavis, K. G., & Silverman, N. S. (1996). Evaluation of automated COBAS AMPLICOR PCR system for detection of several infectious agents and its impact on laboratory management. Journal of Clinical Microbiology, 34, 2778 –2783. Kellogg, J. A., Seiple, J. W., & Stroll, E. S. (1993). Direct fluorescentantibody confirmation of chlamydial antigen below the detection threshold of the Chlamydiazyme enzyme-linked immunosorbent assay. Journal of Clinical Microbiology, 31, 1646 –1647. Lauderdale, T-L., Landers, L., Thorneycroft, I., & Chapin, K. (1999). Comparison of the PACE 2 assay, two amplification assays and clearview EIA for detection of Chlamydia trachomatis in female endocervical and urine specimens. Journal of Clinical Microbiology 37: 2223–2229. Mahony, J. B., Luinstra, K. E., Sellors, J. W., Pickard, L., Chong, S., Jang, D., & Chernesky, M. A. (1994). Role of confirmatory PCRs in determining performance of Chlamydia Amplicor PCR with endocervical specimens from women with a low prevalence of infection. Journal of Clinical Microbiology, 32, 2490 –2493. Morre´ , S. A., Meijer, C. J. L. M., Munk, C., Kruger-Kjaer, S., Winther, J. F., Jorgensens, H. O., Vanden Brule, A. J. C. (1999). Determination of Chlamydia trachomatis prevalence in an asymptomatic screening population: Performance of the LCx and COBAS AMPLICOR tests with urine specimens. Journal of Clinical Microbiology, 37, 3092–3096. Morre´ , S. A., Meijer, C. J. L. M., Munk, C., Kruger-Kjaer, S., Winther, J. F., Jorgensens, H. O., Vanden Brule, A. J. C. (2000). Pooling of urine specimens for detection of asymptomatic Chlamydia trachomatis infections by PCR in a low-prevalence population: cost-saving strategy for epidemiological studies and screening programs. Journal of Clinical Microbiology, 38, 1679 –1680. Newhall, W. J., Delisle, S., Fine, D., Johnson, R. E., Hadgu, A., Matsuda, B., Osmond, D., Campbell, J., & Stamm, W. E. (1994). Head to head evaluation of five different nonculture Chlamydia tests relative to a quality-assured culture standard. Sexually Transmitted Diseases, 21(Suppl. 2), S165–S166. Paavonen, J., Koutsky, L. A., & Kiviat, N. (1990). Cervical neoplasia and other STD-related genital and anal neoplasias. In K. K. Holmes, P-A. Mardh, P. F. Sparling, P. J. Wiesner, W. Cates, Jr., S. M. Lemon & W. E. Stamm, (Eds.), Sexually Transmitted Diseases (2nd ed.) (pp. 561–592). New York: McGraw-Hill, Inc. Pearlman, M. D., McNeeley, S. G. (1992). A review of the microbiology, immunology, and clinical manifestations of Chlamydia trachomatis infections. Obstetrics and Gynecology Surveys, 47, 448 – 461. Peeling, R. W., & King, L. (1999). Sexually transmitted diseases: Impact of molecular laboratory diagnostic methods on disease control and prevention. Canadian Journal of Infectious Disease, 10, 387–390. Skulnick, M., Chua, R., Simor, A. E., Low, D. E., Khosid, H. E., Fraser, S., Lyons, E., Legere, E. A., & Kitching, D. A. (1994). Use of the polymerase chain reaction for the detection of Chlamydia trachomatis from endocervical and urine specimens in an asymptomatic low-prev-

H. Semeniuk et al., / Diagnostic Microbiology and Infectious Disease 42 (2002) 43–51 alence population of women. Diagnostic Microbiology and Infectious Disease, 20, 195–201. Toye, B., Woods, W., Bobrowska, M., & Ramotar, K. (1998). Inhibition of PCR in genital and urine specimens submitted for Chlamydia trachomatis testing. Journal of Clinical Microbiology, 36, 2356 –2358. Wasserheit, J. W. (1991). Epidemiological synergy: interrelationships between HIV infection and other STDs. In L. C. Chen, J. S. Amor & S. J. Segal (Eds.), AIDS and Women’s Reproductive Health. (pp. 47–72). New York: Plenum Press. Welsh, L. E., Quinn, T. C., Gaydos, C. A. (1997). Influence of endocervical specimen adequacy on PCR and direct fluorescent-antibody staining for

51

detection of Chlamydia trachomatis infections. Journal of Clinical Microbiology, 35, 3078 –3081. Wu, C-H., Lee, M-F., Yin, S-C., Yang, D-M., Cheng, S. F. (1992). Comparison of polymerase chain reaction, monoclonal antibody based enzyme immunoassay, and cell culture for detection of Chlamydia trachomatis in genital specimens. Sexually Transmitted Diseases, 19, 193–197. Wylie, J. L., Moses, S., Babcock, R., Jolly, A., Giercke, S., Hammond, G. (1998). Comparative evaluation of Chlamydiazyme, PACE 2, and AMP-CT assays for detection of Chlamydia trachomatis in endocervical specimens. Journal of Clinical Microbiology, 36, 3488 –3491.