|Home | About | Journals | Submit | Contact Us | Français|
Screening women for sexually transmitted diseases (STD) in nonclinic settings is highly desirable because many infections are asymptomatic. This is especially true for military women, for whom logistical, social, and other job-related obstacles present barriers to accessing medical care. We assessed the accuracy of intravaginal swabs transported by mail in a wet versus a dry state for PCR (Amplicor CT/NG test) detection of chlamydia and gonorrhea infections in a cross-sectional study of 793 active-duty military women attending an STD clinic. PCR tests of vaginal swabs (wet and dry) were compared to local clinical methods used on cervical swabs. Standard wet vaginal swab PCR testing detected more chlamydia (11.6%) than cervical enzyme immunoassay (9.3%). For detection of chlamydia using wet swabs, the sensitivity and specificity compared with adjudicated true positives were 94.6% (87 of 92) and 99.3% (696 of 701), respectively. Comparing dry swabs to true-positives for chlamydia, the sensitivity was 91.3% (84 of 92) and the specificity was 99.3% (696 of 701). Standard wet vaginal swab PCR detected more gonorrhea (3.3%) than routine cervical culture (2.1%). The sensitivity and specificity of PCR testing of wet swabs compared to true-positives (infected patients) were 96.3% (26 of 27) and 98.2% (752 of 766) for gonorrhea, respectively. For gonorrhea, the sensitivity and specificity of dry swabs compared to true-positives (infected patients) were 88.9% (24 of 27) and 98.3% (753 of 766), respectively. PCR testing of wet and dry transported intravaginal swabs to detect chlamydia and gonorrhea infections was an accurate diagnostic method for military women.
The recent development of pathogen-specific molecular methods has revolutionized the diagnosis of sexually transmitted diseases (STD). New tests based on nucleic acid amplification technology can utilize urine and other specimens such as vaginal swabs, thereby decreasing the need for physical examinations, and are highly sensitive and specific (4, 8-10, 15, 16, 18, 20, 22). Use of these specimens has expanded the venue in which individuals, especially women, can seek screening for sexually transmitted infections.
Traditionally, approaches to STD diagnosis in women required gynecologic pelvic examination, which is often impractical or unavailable in remote areas where military women may be deployed. Studies have found the prevalence of STD among active-duty military women to be higher than among corresponding civilian female populations (7, 21). Military women have reported a “lack of confidence” in military doctors and embarrassment as reasons for delaying diagnosis and treatment (19). Therefore, prompt, confidential, and readily available methods for the diagnosis and treatment of STD among these high-risk, active-duty soldiers are needed to prevent sequelae, reduce transmissibility, and save costs (2).
This study evaluated the performance of vaginal swabs that could be obtained by the women themselves and sent to a laboratory for simultaneous detection of chlamydia and gonorrhea. The study also assessed the sensitivity and specificity of PCR testing of the vaginal swabs compared to standard clinical diagnostic methods.
From March 1997 to October 1998, consecutive active-duty military women, 18 to 59 years of age, attending the Epidemiology and Disease Control (EDC) clinic at Womack Army Medical Center, Fort Bragg, N.C., were invited to participate in the study. This study was approved by the institutional review boards for research on human subjects at the Johns Hopkins Medical Institution and Womack Army Medical Center, Fort Bragg, N.C., as well as the Human Subjects Research Review Board of the U.S. Army Surgeon General.
Any woman presenting to the EDC clinic for evaluation of genitourinary symptoms, for therapy as a known contact of an individual with a diagnosed STD, or for routine STD screening was approached to volunteer for study enrollment. If the woman agreed to participate and signed the informed consent, she was assigned a unique, confidential study number.
Demographic data and other information, routinely collected for all women by trained clinicians using standardized data collection forms, were abstracted from clinic records. After questionnaire completion, Dacron swabs from the Amplicor PCR collection kit (Roche Molecular Systems, Branchburg, N.J.) were used to collect intravaginal samples, either by the clinician (first 12 months of the study) or by the woman (next 7 months). All participants received a standard pelvic exam.
Providers and study participants were given instructions to insert and rotate the swabs 2 inches (ca. 5 cm) into the vagina before collecting routine urine specimens and before pelvic examination. Wet swabs were placed in specimen transport media (STM; Roche, Branchburg, N.J.) and sent within 4 days by overnight mail to the Johns Hopkins STD Research Laboratory for PCR diagnosis (Amplicor CT/NG; Roche) for Chlamydia trachomatis and Neisseria gonorrhoeae. Dry vaginal swabs were transported in a dry state in a sterile capped tube with no fluid transport medium and similarly sent to the laboratory. Specimens were maintained and shipped at 4°C. The vaginal swab routinely used for the determination of pH was stored in transport buffer for ligase chain reaction (LCR; Abbott Laboratories, Abbott Park, Ill.) as an additional specimen to be used in adjudication procedures.
Following wet and dry swab collection, routine diagnostic tests for infectious agents and other conditions were obtained according to standard operating procedures (SOP) for the EDC clinic. Vaginal swabs were obtained for pH testing, wet preparations were made for diagnosing bacterial vaginosis and trichomonas, and 10% potassium hydroxide preparations were obtained for yeast detection. Cervical swabs were obtained for chlamydia testing by enzyme immunoassay (EIA) (Syva Company, San Jose, Calif.) (23) and gonorrhea culture using modified Thayer-Martin medium (13). Additional specimens collected included a Papanicalaou smear using a cytobrush and an additional vaginal swab for trichomonas culture. Two additional Dacron endocervical swabs, both stored and shipped in STM at 4°C, were obtained for a separate human papillomavirus study and for discrepant analysis testing by PCR for gonorrhea or chlamydia, as necessary.
Treatment decisions for chlamydia and gonorrhea were based on the results of culture for gonorrhea and EIA for chlamydia and were in accordance with the Centers for Disease Control and Prevention STD treatment guidelines (3) and EDC clinic SOP algorithms.
Upon receipt by the laboratory, wet and dry transported swabs were processed according to the manufacturer's instructions for swabs (Roche) by adding processing buffers. For the dry transported swabs, the routine transport buffer (STM) was added to the tube containing the swab before further standard processing. Aliquots of specimens (liquid processing buffers after processing each swab) were made and frozen at −70°C until PCR testing. The two additional endocervical swabs in STM were processed according to the Roche Amplicor instructions, and the vaginal pH swab in LCR transport medium was processed according to the manufacturer's instructions (Abbott).
The PCR procedure was a combination test (Amplicor CT/NG; Roche Molecular Systems, Branchburg, N.J.) that utilized PCR combined with DNA probe hybridization in a colorimetric detection assay. The target sequence for DNA amplification for C. trachomatis was the cryptic plasmid. For N. gonorrhoeae, the target sequence was the DNA methyltransferase gene M:NgoP11. The assay was performed in a 96-well PCR format using a Perkin Elmer Cetus thermocycler according to the manufacturer's directions. This procedure allowed the simultaneous detection of both organisms as well as of an internal control to indicate the presence of inhibitors. Less than 5% of specimens demonstrated inhibition. The tests demonstrating inhibition were repeated in duplicate, both undiluted and diluted 1:10, which served to eliminate the inhibition. Sensitivity and specificity for PCR of endocervical specimens as compared to culture in multicenter trials have been reported to be 89.7 and 99.4%, respectively, for C. trachomatis and 92.4 and 99.5%, respectively, for N. gonorrhoeae (17, 24).
The vaginal swab chlamydia PCR results were compared to the Syva chlamydia EIA of the cervical swab. Specimens discrepant between chlamydia PCR and EIA (PCR positive and EIA negative or PCR negative and EIA positive) were resolved as true-positive or true-negative specimens by LCR (4, 11, 15, 20) testing and by OMP-1 gene PCR testing (Roche). True-positives among samples with discrepant results were defined as samples for which either the original positive PCR or positive EIA was confirmed by another DNA amplification test (either the LCR or PCR for the OMP-1 gene).
The vaginal swab gonorrhea PCR results were compared to the results from gonorrhea cervical culture. For N. gonorrhoeae specimens that were PCR positive and culture negative, LCR (12, 22) was performed on the processed stored vaginal swab specimens that were used for pH determination. Additionally, the stored processed endocervical specimens were also used for adjudication of discrepant specimens using the Roche PCR for both the DNA methyltransferase gene and the 16S rRNA gene. Samples positive by culture or positive by two DNA amplification procedures for two different genes were considered true positives.
The sensitivities, specificities, and 95% confidence intervals for vaginally obtained swabs were calculated using the adjudicated true-positive and true-negative results. Analyses were done using standard methods with the SAS software (SAS Institute, Cary, N.C.).
During the enrollment period, 892 women were approached for study enrollment and 793 (89%) consented to participate. Women who refused enrollment were not statistically different from study participants with regard to race, age, or reason for the clinic visit (data not shown). Overall, 59.8% of participants were African-American, 32.5% were Caucasian, and the remainder were of other racial or ethnic background. The mean age was 25.1 years, and the majority of women presented for evaluation of symptoms (82.7%). Of the 793 vaginal swab specimens collected from study participants, 453 were collected by study clinicians and 340 were self-collected.
The wet vaginal swab PCR detected a chlamydia prevalence of 11.6% (92 of 793) compared to a prevalence of 9.3% (74 of 793) detected by chlamydia EIA. Chlamydia was detected in 11.5% of specimens collected by the clinician and in 11.8% of specimens collected by the patient (P = 0.30). Of the 92 positive samples detected by vaginal swab, 5 could not be confirmed by discrepant testing. Of the 74 positive EIA specimens, 7 could not be confirmed as true-positives. There were 30 EIA-negative, PCR-positive discrepant samples, all but 5 of which were confirmed as true-positives. There were 12 EIA-positive, PCR-negative discrepant specimens, of which only 5 could be confirmed. By adjudication of discrepant specimens, 92 samples were defined as true-positives for chlamydia, giving a final prevalence of 11.6%. The sensitivity and specificity for wet vaginal swab PCR for chlamydia detection was 94.6% (87 of 92) and 99.3% (696 of 701), respectively, while the sensitivity and specificity for EIA was 72.9% (67 of 92) and 99.0% (694 of 701), respectively (Table (Table1).1).
Before adjudication, the dry transported vaginal swab detected a chlamydia prevalence of 10.7% (85 of 793). Two comparisons of laboratory results were made. First, the dry swab was compared to the true-positive infected samples based on the adjudication of the dry transported swabs for PCR. Of the 23 swabs that were PCR positive and EIA negative, 22 could be confirmed as true-positives, and of the 12 specimens that were PCR negative and EIA positive, only 4 could be confirmed as true-positives. After resolution, 88 samples were defined as true-positives, providing a final prevalence of 11.1%, with 84 of the dry swabs being confirmed as true-positives. Thus, after adjudication, the sensitivity and specificity of the dry swab PCR was 95.5% (84 of 88) and 99.8% (704 of 705), respectively. Second, when the dry swab PCR results were compared to true positives (N = 92) identified by the wet swab adjudication, the sensitivity and specificity were 91.3% (84 of 92) and 99.3% (696 of 701), respectively (Table (Table1).1). The overlap of the 95% confidence levels for the wet and dry swabs listed in Table Table11 for both sensitivity and specificity indicates that the dry swabs were as accurate as the wet swabs.
Of 793 specimens examined for gonorrhea, 17 (2.1%) were reported as positive by cervical culture. By PCR of wet transported vaginal swabs, gonorrhea was detected in 5.1% of clinician-collected vaginal swabs and in 5.0% of specimens that were self-collected (P = 0.95). The wet vaginal swabs detected 16 of the 17 culture-positive gonorrhea-infected patients. An additional 24 positives were detected by wet vaginal swab PCR, of which only 10 were confirmed as true gonorrhea infections through discrepant testing (7 by adjudication of the wet vaginal swab and 10 [three additional patients] by adjudication of the cervical swabs, for a total of 10 more true-positives). The true prevalence of gonorrhea-infected patients based on adjudicated results was 3.4% (27 of 793). The wet vaginal swab detected more true-positive gonococcal infections (26 of 793, 3.3%) than did standard cervical culture (17 of 793, 2.1%). However, the vaginal swab PCR detected 14 false-positive results (14 of 793, 1.8%). After adjudication, the sensitivity and specificity for gonorrhea of wet vaginal swabs were 96.3% (26 of 27) and 98.2% (752 of 766), respectively (Table (Table2).2).
Thirty-nine positives (4.9%) were detected in the vaginal swabs that were transported to the laboratory in a dry state and tested by PCR. The dry vaginal swabs detected 16 of the 17 culture-positive gonorrhea-infected patients. An additional 23 positives were detected by the dry vaginal swab PCR, of which only 9 were confirmed (5 by adjudication of the dry swab and 9 by cervical swab PCR testing [a total of nine patients]) as true gonorrhea infections through discrepant testing. The true prevalence of gonorrhea-infected patients, based on adjudication of the dry transported swabs, was 26 of 793 (3.3%).
Two comparisons of laboratory results were made. First, the dry swab was compared to the true-positive infected patients based on the adjudication of the dry transported swabs for PCR. The dry vaginal swab detected more true-positive gonococcal infections (25 of 793, 3.2%) than did standard cervical culture (17 of 793, 2.1%). However, the dry vaginal swab PCR detected 14 false-positive results (14 of 793, 1.8%). Second, the dry swab PCR results were compared to true gonorrhea positives (N = 27) based on the wet swab and cervical swab adjudication (i.e., 27 infected patients). The sensitivity was 88.9% (24 of 27), and the specificity was 98.3% (753 of 766) (Table (Table2).2). The overlap of the 95% confidence levels for the wet and dry swabs listed in Table Table22 for both sensitivity and specificity indicates that the dry swabs were as accurate as the wet swabs.
This study demonstrated that a clinician-collected or self-administered vaginal swab detected more infections did the routine clinic diagnostic methods and was an accurate specimen collection method for the diagnosis of both chlamydia and gonorrhea infections. Other studies have reported that self-collected vaginal specimens can be used successfully to diagnose sexually transmitted infections, eliminating the need for a clinician and a pelvic examination for specimen collection (10-12, 18). These swabs have performed as well as or better than clinician-obtained endocervical swabs for diagnosis of either chlamydia or gonococcal infection by DNA amplification assays. Hook et al. reported a sensitivity of 91.8% for patient-obtained vaginal swabs, compared to 89.8% for clinician-obtained cervical swabs, for detection of chlamydia by LCR (11). For gonorrhea, the sensitivity for patient-obtained vaginal swabs was 100%, compared to 84.6% for clinician-obtained cervical swabs (12). Our study confirmed their results and indicated that vaginal swabs tested by LCR performed better than the currently used methods in the EDC clinic.
Additionally, we found that a vaginal swab that was transported to the laboratory in a dry state was as accurate as a wet swab shipped in the liquid transport medium recommended by the manufacturer. This was indicated by the overlap of the confidence intervals for the sensitivity and specificity of the wet and dry transported swabs. The ability to transport swabs in a dry state to the laboratory suggests that it might be possible to mail vaginal swabs in preaddressed mailing packets from remote sites to a central laboratory for testing.
Self-collected specimens could be obtained when it is desirable to test women who are not seeking health care but may have asymptomatic infections. Self-collected swabs should also be useful in large screening surveys of genital tract infections, when pelvic examinations are impractical or cost-prohibitive. For symptomatic women who cannot obtain a clinic appointment immediately, self-collected swabs could be obtained. This could be beneficial to military women when they are deployed to remote sites. The availability of a self-administered specimen collection system would give women in both military and civilian populations the opportunity to play a greater role in meeting their own health needs.
The need to develop acceptable, better, and more easily available techniques for diagnosing STD for all high-risk populations is significant. Five of the top 10 reportable diseases in the United States are STD, and sequelae include pelvic inflammatory disease, infertility, and cervical cancer (1). Additionally, STD are associated with increased risk for human immunodeficiency virus acquisition, and their costs are greater than $5 billion annually (5, 14). The population of military personnel is especially vulnerable, since they have been reported to have a higher prevalences of STD than comparable civilian populations (7, 21).
One objective of this study was to investigate whether a self-administered vaginal swab would be acceptable to women. We selected a military population for study because these young, sexually active women represented a high-risk group for STD acquisition, clinic facilities for diagnosing STD in military settings are often limited, and there is a strong incentive to simplify health care delivery in military settings. Although not formally evaluated, acceptability was suggested by the high volunteer rate among women who were approached in a clinic setting. However, self-administered swabs still need to be evaluated in other settings.
Of concern is the observation of a large number of false-positive gonorrhea results with the vaginal swab PCR, for which we used the research format of this assay. Many of our positive PCR tests could not be confirmed by extensive discrepant testing. These false-positives may be attributed to certain strains of Neisseria subflava and Neisseria cinerea that have been demonstrated to produce low-level positive optical density results (false-positive results) with the Amplicor N. gonorrhoeae PCR (6). Roche Molecular Systems has alleviated this problem in the test now cleared by the Food and Drug Administration by recommending in their package insert that a confirmatory repeat PCR test be used for positives, in addition to a higher cut-off optical density for positive results.
The accuracy of using a swab transported in a dry state has been demonstrated. Future studies to further explore the acceptability and usefulness of self-collected vaginal swabs should be undertaken to define their role in STD control. Additionally, comparison must be made of the results from transport of the dry swab specimens at both 4°C and ambient temperature, as elimination of a temperature requirement for transport would clearly enhance the utility of the intravaginal swabs.
This study was supported by grant DAMD17-1-6309 from the Department of the Army.