From the *Indiana University School of Public Health, Bloomington †Indiana University School of Medicine, Indianapolis, IN; ‡Louisiana State University Health Sciences Center, New Orleans, LA; §Medical and Scientific Affairs, Roche Molecular Systems, Pleasanton, CA ¶University of Alabama at Birmingham, and ∥Jefferson County Health Department, Birmingham, AL
We are grateful to John Duncan for expert statistical support and to the members of the collection and testing sites.
B.V.D.P. has consulted for honorarium for Roche. Roche has also supported B.D.V.P. for travel to meetings for the study or other purposes. B.D.V.P. has also consulted for BD Diagnostics. O.L. is currently employed by Roche Molecular Systems as the Director of Medical and Scientific Affairs. As an employee of Roche Molecular Systems, O.L. also owns stocks of Roche. E.W.H. has consulted for Cempra Pharmaceuticals. He has also been employed by MedHelp.org in the role of providing answers to online questions from persons with and at risk for sexually transmitted disease. E.W.H. has also received royalties from McGraw Hill for his work as a book editor and has received payment for the development of educational presentations on Herpes Testing for BD. S.N.T. and J.A.W. have no conflict of interests.
Correspondence: Barbara Van Der Pol, PhD, MPH, Department of Epidemiology and Biostatistics, Indiana University School of Public Health, 1025 E Seventh St, Suite 112, Bloomington, IN 47405. E-mail: firstname.lastname@example.org.
Received for publication June 5, 2012, and accepted August 28, 2012.
Chlamydia trachomatis and Neisseria gonorrhoeae infections remain highly prevalent in the United States and globally despite long-standing control efforts, which universally emphasize screening recommendations that all women younger than 26 years be screened at least annually.1 Although recommendations for screening women are documented to be appropriately targeted, implementation has been poor, with fewer than 50% of eligible women being screened as of 2008.2 Both clinician and patient reluctance to perform or undergo, respectively, pelvic examination for collection of endocervical samples in the absence of symptoms or other medical indications may contribute to poor screening rates. A strategy that includes more easily obtained specimens such as vaginal swabs, whether self-obtained or clinician collected, may increase uptake of screening recommendations.
Evidence-based recommendations for the detection of chlamydial and gonococcal infections highlight the use of vaginal swabs for use with nucleic acid amplification diagnostic methods.3 This recommendation was based on multiple studies that assessed both the performance of vaginal swab specimens4–12 and the acceptability of self-obtained sampling.5,9,11,12 Here we present the performance characteristics of the Cobas CT/NG (c4800) Test (Roche Diagnostics, Indianapolis, IN) performed on the Cobas 4800 system using vaginal swab specimens both compared with patient infection status (PIS) derived from testing with other assays and comparing the vaginal swab results head-to-head with other sample types used on the c4800 test.
The Vaginal, Endocervical, and Urine Screening trial for CT/NG was a multicenter evaluation of the Cobas specimen collection kit and of the c4800 test (Roche Diagnostics). The test procedures, study design, patient population, and other characteristics have been described elsewhere.13,14 In brief, 2 Food and Drug Administration–cleared nucleic acid amplification tests (NAATs), the Gen-Probe APTIMA Combo2 Assay (AC2; Gen-Probe, San Diego, CA), and BD ProbeTec CT/GC Qx Amplified DNA Assay (CTQ/GCQ; BD Diagnostics, Sparks, MD) were used as comparator assays. The specimen collection sites were geographically diverse and included obstetrics-gynecology practices, family planning clinics, and sexually transmitted disease clinics. Four sites performed all testing on the c4800 platform.
Specimens were collected from each female participant in the following order: first-catch urine; a single vaginal swab (women were randomized to self-collection or clinician collection) using the Cobas collection kit, 3 speculum-guided endocervical swabs using each manufacturer’s sample collection device (in randomized order), and a sample suitable for liquid-based cytology (LBC) placed into PreservCyt medium (Cytyc Corp, Marlborough, MS). In cases where the LBC specimen was needed for Papanicolaou testing as part of routine patient care, the LBC specimen collection followed immediately after the urine and vaginal specimen collection and before the collection of endocervical swabs. Urine specimens were divided into 3 aliquots and placed into each assay’s transport tube. Liquid-based cytology samples were aliquoted into c4800 sample tubes and AC2 transport tubes before processing for Papanicolaou testing (prequot LBC). After Papanicolaou processing, residual specimens were aliquoted again into c4800 sample tubes (postquot LBC). All specimens for comparator assays were stored and tested according to the package insert instructions.
Two analyses of the performance of vaginal swabs were performed. The first compared the results obtained with vaginal swabs to a PIS. Thus, this analysis compared vaginal swab results with results obtained with the other commercially available assays using endocervical swabs or urine. Liquid-based cytology results were not used in the calculation of the PIS because those results were not available for all comparator assays (i.e., they were not run using the BD CTQ/GCQ assay). Infection was defined as follows: participants were designated as being infected with C. trachomatis or N. gonorrhoeae if at least 2 NAATs with different target regions gave positive results in the endocervical swab and/or the urine specimen. This analytic method attempts to identify infections regardless of the site of infection (e.g., endocervical, urethral, or both). Patients with urine positive results from both comparator assays were considered infected even if neither endocervical sample was positive.
Samples from 4279 women were eligible for analysis. Evaluable results for vaginal swab testing were available for 4248 participants: the PIS identified 267 (6.3%) chlamydial infections and 66 (1.5%) gonococcal infections. In this study, women were randomly assigned to either clinician collection or self-collection of vaginal swab specimens. A total of 2083 women collected their own samples, whereas 2165 women had clinician-obtained swabs. The estimated sensitivity of c4800 for C. trachomatis compared with the PIS was 93.9% (123/131) and 91.9% (125/136) for self- and clinician-obtained samples, respectively, with an overall sensitivity of 92.9%. The estimates of sensitivity for the 2 collection methods were not statistically different (P = 0.636). Similarly, of 66 PIS-defined gonococcal infections, 32 (97%) of 33 infections were identified in women using self-obtained vaginal swab specimens and 33 (100%) of 33 using samples collected by clinicians, for an overall, combined sensitivity estimate of 98.5%. As with Chlamydia, these sensitivity estimates were not statistically different (P = 1.000). Specificities ranged from 99.7% to 99.8% for Chlamydia and 100.0% for Gonorrhea, and no statistical differences were observed (all P values > 0.70). As a result of these analyses, all further evaluations of vaginal swab performance were based on pooled vaginal swab results, regardless of sample collection method. Performance estimates did not vary by the presence or absence of symptoms (all P values > 0.1; data not shown).
The second analysis was a head-to-head comparison of vaginal swab performance to each of the other sample types tested on the c4800 (i.e., urine, endocervical, LBC). Positive (PPA), negative (NPA), and overall percent agreements (OPA) were calculated for vaginal samples compared with each of the other clinical specimen types (i.e., endocervical swabs, urine, or PreservCyt specimens). Analyses of PPA and NPA require an a priori definition of “true” positives and negatives. Samples from women identified in the PIS analysis as infected were considered to be in the true-positive category, and all others were considered to be true negatives. Thus, for women defined as infected, the concordant positive results achieved by 2 sample types are divided by the total number of true positives to calculate the PPA. Simple κ coefficients were also calculated. Exact P values of the estimates were determined using the McNemar χ2 test; α was set to 0.05. All analyses were performed using SAS/STAT software 9.2.
For head-to-head comparisons, which required that all sample types were collected and evaluable, c4800 results were available from 4049 participants. In this analysis, there were 248 and 65 infections with Chlamydia and Gonorrhea, respectively. Overall percent agreement between C. trachomatis results obtained using vaginal swab specimens compared head-to-head with other specimen types ranged from 98.8% to 99.2% for Chlamydia and 99.8% to 99.9% for Gonorrhea. Both PPA and NPA are shown in Table 1. κ Coefficients were higher than 0.88 for all Chlamydia comparisons and 0.92 for all Gonorrhea comparisons.
When evaluating test performance, some level of variation is always introduced as a result of collection and testing multiple specimen types. We assessed the distribution of results by sample type visually, as shown in Figures 1 and 2. The Venn diagrams in Figure 1 show the overlapping and unique positives identified by vaginal swab, endocervical swab, or urine specimens. The data are restricted to those participants who were identified as infected using the PIS definition. For example, vaginal swabs detected 8 chlamydial infections that were present only in urine samples and would have been missed by endocervical sampling, while also detecting 16 samples that were isolated to the endocervix and would have been missed if urine were the sample type used for diagnosis. Chlamydia shows a higher level of variability than Gonorrhea: for gonococcal infections, there were no patients who were positive with only a single sample type. This may also be a reflection of the fact that fewer patients were infected with Gonorrhea.
Figure 2 summarizes the total number of infections from the head-to-head comparisons that would have been identified if using only 1 specimen (e.g., vaginal swabs). For the 248 chlamydial infections, 93.5% would have been detected using a single vaginal swab, compared with 92.7% and 89.5% when using endocervical swabs and urine, respectively. Similarly, vaginal swabs detected the highest proportion of the 65 gonococcal infections (98.5%).
Data from the Vaginal, Endocervical, and Urine Screening trial, which enrolled more than 4000 women, demonstrate excellent performance of vaginal swabs, whether collected by clinicians or patients themselves, for detection of C. trachomatis and N. gonorrhoeae in the c4800 test. These results were robust in that they were independent of the comparator method. Whether compared with a PIS defined by AC2 and CTQ/GCQ, or head-to-head with other sample types on the same system, vaginal swabs had excellent sensitivity and specificity estimates for both C. trachomatis and N. gonorrhoeae.
For the last decade or more, experts in sexually transmitted infection control have recognized the high degree of patient acceptability and potential use of vaginal swab testing.15 In a multisite study of women’s attitudes toward self-obtained vaginal swabs, Chernesky and colleagues16 described an overwhelming number (>90%) of participants endorsing willingness to be screened more frequently if this were an available option. In this study, we have demonstrated the equivalence of self-obtained versus clinician-collected samples when using a new platform for screening (the c4800).
Vaginal swabs detect infections that may be localized to either the endocervix or the urethra, as shown in Figure 1. These data continue to support the conclusion made by the CDC that vaginal swabs are the preferred specimen type for screening for Chlamydia and Gonorrhea in women.3 More widespread use of vaginal swabs has the potential to simplify testing and could reduce costs and shorten examination time by reducing the number of pelvic examinations to only those who are medically warranted and, in this way, could lead to increased numbers of patients seen in clinical settings.17,18 Use of self-obtained vaginal swabs, collected when patients provide routine samples for urinalysis, has the potential to improve clinic flow and the time burdens on providers even further while still providing a highly sensitive test result. With pressure on health care services to reduce time/patient, this time reduction may be critical to achieving the benchmarks set by Healthcare Effectiveness Data and Information Set and other clinical practice guidelines. Furthermore, vaginal sampling, particularly when specimens are self-obtained, expands our ability to provide screening in less formal settings such as home collection, school-based clinics, detention centers, and community outreach venues. The combination of excellent test performance, patient preference, and clinical practice advantages make the vaginal swab the optimal sample type for any NAAT including the newly approved c4800 assay.
1. Centers for Disease Control and Prevention. Sexually transmitted diseases treatment guidelines, 2010. Morb Mortal Wkly Rep 2010; 59 (RR-12): 1–112.
4. Hook EI, Smith K, Mullen C, et al.. Diagnosis of genitourinary Chlamydia trachomatis
infections by using ligase chain reaction on patient-obtained vaginal swabs. J Clin Microbiol 1997; 38: 2133–2135.
5. Hook EW III, Ching SF, Stephens J, et al.. Diagnosis of Neisseria gonorrhoeae
infections in women by using the ligase chain reaction on patient-obtained vaginal swabs. J Clin Microbiol 1997; 35: 2129–2132.
6. Blake DR, Maldeis N, Barnes MR, et al.. Cost-effectiveness of screening strategies for Chlamydia trachomatis
using cervical swabs, urine, and self-obtained vaginal swabs in a sexually transmitted disease clinic setting. Sex Transm Dis 2008; 35: 649–655.
7. Hobbs MM, van der Pol B, Totten P, et al.. From the NIH: Proceedings of a workshop on the importance of self-obtained vaginal specimens for detection of sexually transmitted infections. Sex Transm Dis 2008; 35: 8–13.
8. Jones HE, Altini L, de KA, et al.. Home-based versus clinic-based self-sampling and testing for sexually transmitted infections in Gugulethu, South Africa: Randomised controlled trial. Sex Transm Infect 2007; 83: 552–557.
9. Keane FE, Bendall R, Saulsbury N, et al.. A comparison of self-takenvulvovaginal and cervical samples for the diagnosis of Chlamydia trachomatis
infection by polymerase chain reaction. Int J STD AIDS 2007; 18: 98–100.
10. Masek BJ, Arora N, Quinn N, et al.. Performance of three nucleic acid amplification tests for detection of Chlamydia trachomatis
and Neisseria gonorrhoeae
by use of self-collected vaginal swabs obtained via an Internet-based screening program. J Clin Microbiol 2009; 47: 1663–1667.
11. Chernesky MA, Martin DH, Hook EW, et al.. Ability of new APTIMA CT and APTIMA GC assays to detect Chlamydia trachomatis
and Neisseria gonorrhoeae
in male urine and urethral swabs. J Clin Microbiol 2005; 43: 127–131.
12. Skidmore S, Horner P, Herring A, et al.. Vulvovaginal-swab or first-catch urine specimen to detect Chlamydia trachomatis
in women in a community setting? J Clin Microbiol 2006; 44: 4389–4394.
13. Van Der Pol B, Liesenfeld O, Williams J, et al.. Performance of the Cobas CT/NG test compared to the Aptima AC2 and Viper CTQ/GCQ assays for detection of Chlamydia trachomatis
and Neisseria gonorrhoeae
. J Clin Microbiol 2012; 50: 2244–2249.
14. Taylor SN, Liesenfeld O, Lillis RA, et al.. Evaluation of the Roche Cobas CT/NG test for detection of Chlamydia trachomatis
and Neisseria gonorrhoeae
in male urine. Sex Transm Dis 2012; 39: 543–549.
15. Schachter J, Chernesky MA, Willis DE, et al.. Vaginal swabs are the specimens of choice when screening for Chlamydia trachomatis
and Neisseria gonorrhoeae
: Results from a multicenter evaluation of the APTIMA assays for both infections. Sex Transm Dis 2005; 32: 725–728.
16. Chernesky MA, Hook EW III, Martin DH, et al.. Women find it easy and prefer to collect their own vaginal swabs to diagnose Chlamydia trachomatis
or Neisseria gonorrhoeae
infections. Sex Transm Dis 2005; 32: 729–733.
17. Hoebe CJ, Rademaker CW, Brouwers EE, et al.. Acceptability of self-taken vaginal swabs and first-catch urine samples for the diagnosis of urogenital Chlamydia trachomatis
and Neisseria gonorrhoeae
with an amplified DNA assay in young women attending a public health sexually transmitted disease clinic. Sex Transm Dis 2006; 33: 491–495.
18. Rose SB, Lawton BA, Bromhead C, et al.. Self-obtained vaginal swabs for PCR Chlamydia
testing: A practical alternative. Aust N Z J Obstet Gynaecol 2007; 47: 415–418.