Context.—Data on high-risk human papillomavirus (hrHPV) DNA test results in vaginal, liquid-based cytology (LBC) specimens and corresponding cytologic and histopathologic correlation data are limited.
Objective.—To analyze follow-up correlation data associated with vaginal (after hysterectomy) low-grade squamous intraepithelial lesion (LSIL) LBC and hrHPV test results.
Design.—Hospital records were searched for vaginal LSIL LBC and hrHPV results between July 1, 2005, and July 30, 2009. Histopathologic and Papanicolaou test follow-up correlation data were analyzed.
Results.—During the study period, 2892 patients with test results from both posthysterectomy vaginal LBC and hrHPV were identified: 148 (5.1%) of the patients had vaginal Papanicolaou test results reported as LSIL, with hrHPV detected in 113 of the 148 patients (76.4%). Of 148 patients, 59 of those with vaginal LSIL including 48 (81.4%) with positive HPV testing and 11 (18.6%) with negative HPV testing results had a follow-up vaginal biopsy. Histopathologic vaginal intraepithelial neoplasia (VAIN) 2/3 was diagnosed in 7 of 59 patients (11.9%) with biopsies. In all 7 patients who were diagnosed with VAIN 2/3, hrHPV was detected in the LBC vial. No VAIN 2/3 diagnoses were documented in the biopsy specimens from the 11 patients with hrHPV− LSIL Papanicolaou test results. Histopathologic VAIN 2/3 was diagnosed from vaginal biopsies in 7 of the 48 patients (14.6%) with vaginal hrHPV+ LSIL test results.
Conclusions.—Sensitivity and specificity of hrHPV test results associated with histopathologic follow-up diagnoses of VAIN 2/3 in patients with vaginal LSIL results were 100% and 21.2%, respectively. The positive predictive value of a vaginal hrHPV+ LSIL result for a subsequent histopathologic VAIN 2/3 diagnosis was 14.6%. No cases of VAIN 2/3 were diagnosed in the 11 patients with vaginal hrHPV− LSIL results. Correlations of vaginal cytologic, histopathologic, and human papillomavirus findings were quite similar to correlation findings previously reported in older women with cervical LSIL test results.
Vaginal Papanicolaou (Pap) tests are primarily recommended for screening of women who have had hysterectomies and prior diagnoses of malignant or premalignant lower genital tract disease,1,2 and published literature has almost unanimously concluded that vaginal Pap screening for vaginal neoplasia after hysterectomies for benign disease is not cost-effective.3–6 When a hysterectomy has been performed because of a diagnosis of high-grade cervical intraepithelial neoplasia, up to 7.4% of patients have been reported to develop subsequent vaginal intraepithelial neoplasia (VAIN).7 Several possible hypotheses have been put forward8 to explain the pathogenesis of VAIN in this setting: (1) residual disease reflecting incomplete excision of a contiguous, preexisting intraepithelial lesion; (2) multicentric intraepithelial disease; (3) recurrent intraepithelial disease associated with compromised host defenses and occult-persistent human papillomavirus (HPV) infections; and (4) de novo development of intraepithelial lesions following an entirely new HPV infection. Therefore, follow-up of patients who have had hysterectomies for premalignant or malignant diagnoses is often recommended. Even though patients whose hysterectomies were performed for benign disease are theoretically at some risk for developing VAIN over time, the American College of Obstetricians and Gynecologists,1 the United States Preventive Screening Task Force,9 and the American Cancer Society10 all recommend against performing routine vaginal Pap testing on women who have had hysterectomies for benign conditions. Interestingly, data supporting vaginal screening recommendations are almost entirely from an era before widespread high-risk (hr) HPV DNA testing. There are scant published data on HPV test results in vaginal Pap specimens from patients either with or without detected cytologic abnormalities.10,11 Furthermore, the manufacturer of the most widely used hrHPV DNA test (Qiagen Corporation, Venlo, the Netherlands) specifically cautions in the package insert of its HPV test (Hybrid Capture 2 [HC2]) that the test has not been evaluated in vaginal specimens from patients who have had hysterectomies.12 Moreover there are currently no specific guidelines, to our knowledge, for management of patients with abnormal vaginal Pap test results and no specific recommendations for adjunctive hrHPV DNA testing on vaginal cytological specimens. Clinicians appear to generally adapt published guidelines and recommendations for cervical Pap tests.13
Vaginal Pap tests are performed in a unique patient population that predominantly comprises elderly women who have undergone total hysterectomies. Because of the unique characteristics of this population and the absence of substantial, published data from this population, we concluded that it would be useful to collect and analyze available data from our large, academic health system on the prevalence of HPV in vaginal specimens, the correlation of vaginal HPV test results with vaginal Pap test results, and the correlation of vaginal HPV and Pap test results with histopathologic follow-up. To our knowledge, this analysis of hrHPV DNA test results in vaginal specimens of women with low-grade squamous intraepithelial lesion (LSIL) Pap test results is the first of its kind. We also compared our findings on vaginal specimens to similar published data on cervical Pap tests.
MATERIALS AND METHODS
A retrospective, computer-based search was conducted in the CoPath database (Cerner Corporation, Kansas City, Missouri) of the laboratory information system of Magee-Womens Hospital of the University of Pittsburgh Medical Center for a period of more than 4 years from July 1, 2005, to July 30, 2009. All the vaginal Pap test results reported as abnormal (according to the Bethesda System; TBS 2001)14 with simultaneously available adjunctive hrHPV DNA test results were retrieved. For comparison all vaginal Pap test results reported as negative for intraepithelial lesion or malignancy were also retrieved.
All vaginal Pap tests were prepared using ThinPrep technology. ThinPrep Pap Tests (TPPT; Hologic, Bedford, Massachusetts) were prepared according to the manufacturer's specifications from PreserveCyt (Cytyc, Boxborough, Massachusetts) samples using an automated processor and staining performed on a Sakura Tissue Tek Automated slide stainer (Sakura Finetek USA, Torrance, California) according to a US Food and Drug Administration–approved manufacturer's protocol. Location-guided, computer-assisted screening of TPPT slides was accomplished using the ThinPrep Imaging System (Hologic).15 The ThinPrep Imaging System performed analysis of batches of up to 250 ThinPrep Pap test slides with specialized imaging software. For each slide, the locations were recorded from 22 microscopic fields that contained cells or cell clusters of interest. The imaged TPPT slides were placed on the cytotechnologists' review scopes, and the cytotechnologists reviewed the 22 fields in geographic order. If the cytotechnologists found no abnormalities in those 22 fields, the cytotechnologist could sign out the case as a negative result. In all cases in which any of the 22 fields contained any abnormality, reactive or reparative cellular changes, or microorganisms, the cytotechnologists manually rescreened the entire TPPT slide. All cases interpreted by cytotechnologists as abnormal or as showing reactive or reparative changes were referred to a pathologist for review.
Adjunctive hrHPV DNA testing was ordered by clinicians according to several ordering options as follows: reflex HPV testing following Pap test results of atypical squamous cell, HPV cotesting with Pap test results in women 30 years and older, or cotesting regardless of age or Pap test results. If hrHPV DNA was detected in negative Pap test results, the Pap test slides were routinely manually rescreened by the screening cytotechnologist, referred for further manual rescreening by a quality-assurance cytotechnologist, and reviewed by a pathologist.
The hrHPV DNA detection was performed with the commercially available US Food and Drug Administration–approved HC2 system,16 which tests for high-risk and intermediate-risk HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68. This enzyme-linked immunosorbent assay is based on a sandwich-capture molecular hybridization technique, followed by a nonradioactive alkaline phosphatase reaction with chemiluminescence in microplates. Cases were categorized as either HC2+ or HC2− based on results with a threshold of 1 pg/mL HPV DNA.
Histopathologic follow-up data on all patients with LSIL diagnoses from vaginal Pap tests were collected. Follow-up data included at least one vaginal biopsy after the vaginal Pap test was performed and time elapsed between the smear and biopsy procedure. In patients with more than one biopsy, the most severe histologic diagnosis was recorded. Hysterectomy records for all the patients were retrieved, and patients who had hysterectomy records in our system or who had repeated vaginal Pap tests with no transformation zone and an absence of cervical Pap tests after the index vaginal Pap test were included in this study. Among the patients for whom a hysterectomy record was available, it was determined whether the hysterectomy had been performed for a benign condition or for a premalignant or malignant lesion of the female genital tract. Patients with repeat vaginal Pap tests with an interpretation of LSIL were considered only as a single-study case. The hrHPV results were assessed in this study population and analyzed in the older and younger group with cutoff points (≤54 and ≥55 years).
Statistical analyses were performed with the χ2 test or the Fisher exact test for small-sample size using the SAS 9.1 system (SAS Institute, Cary, North Carolina). P values less than .05 were considered statistically significant; 95% confidence intervals for the frequency of hrHPV DNA detection rates were obtained by a Wald test.
During the 49-month study period, 2892 cases with both vaginal TPPT and adjunctive hrHPV DNA test results were identified. Of those, 1320 cases (45.6%) were interpreted as squamous cell abnormalities according to the Bethesda System 2001 (TBS 2001) classification, including 1125 cases (85%) of atypical squamous cells of undetermined significance; 36 cases (3%) of atypical squamous cells, for which a high-grade squamous intraepithelial lesion could not be excluded; 148 cases (11.2%) of LSIL; and 11 cases (1%) of high-grade squamous intraepithelial lesions. Of the 148 cases of LSIL, hrHPV DNA was detected in 113 (76.4%; 95% confidence intervals, 69.5%–83.2%) (Table 1). For comparison, cervical data from our institution are provided for the same period on women of all ages and on women older than 50 years. The average age of this LSIL study population was 56.9 years, ranging from 27 to 92 years. The rates of hrHPV DNA detection were not significantly different for women younger than 54 years or older than 55 years. A total of 89 cases were excluded from this study for the following reasons: (1) no hysterectomy procedure could be documented; (2) no histopathologic follow-up was available; (3) the patient was entirely lost to follow-up; or (4) repeat cases (details shown in the Figure).
Included in this study were 59 of the 148 cases (39.9%) with vaginal LSIL Pap test results (48 hrHPV+; 11 hrHPV−) and at least one vaginal biopsy after the index vaginal Pap test. The average age of these patients was 57 years (range, 27–92 years). The follow-up period was 0.2 to 43 months, with a median of 13 months. Histopathologic VAIN 2/3 was diagnosed in 7 of 59 patients (11.9%) with LSIL vaginal Pap test results who had vaginal biopsies, and all 7 patients with VAIN 2/3 diagnoses had hrHPV DNA detected in their vaginal cytologic specimens. The follow-up interval between the LSIL vaginal Pap test and histopathologic VAIN 2/3 detection ranged from 0.2 to 26 months, with a mean of 8.6 months. Among patients with hrHPV+ results, 14.6% were diagnosed with VAIN 2/3, but none of the 11 patients with hrHPV− results were diagnosed with VAIN 2/3; however, that difference did not reach statistical significance because of the low number of cases (P = .33) (Table 2). Histopathologic VAIN 1 was diagnosed in 41 of 59 patients (69.5%); 34 of these 41 patients (82.9%) tested positive for hrHPV DNA. The follow-up interval between an LSIL vaginal Pap test result and a histopathologic diagnosis of VAIN 1 ranged from 0.2 to 21 months, with a mean of 4.1 months. VAIN 1 was diagnosed in 70.8% of patients with hrHPV+ results and 63.6% of patients with hrHPV− results, and again, because of the low number of patients in the hrHPV− group, that difference did not reach statistical significance (P = .72) (Table 2). Overall, 41 of 59 cases (69.5%) of vaginal HPV+ LSIL and 6 of 11 cases (54.5%) of vaginal HPV− LSIL were subsequently diagnosed with some degree of vaginal dysplasia—VAIN 2/3 or VAIN 1. When those patients were segregated into 2 age groups (≤54 years and ≥55 years), there was no statistically significant difference in either VAIN 2/3 or VAIN 1 detection in both hrHPV+ and hrHPV− groups (Table 3).
Sensitivity and specificity of the hrHPV DNA detection in LSIL vaginal Pap tests for the subsequent histopathologic detection of high-grade vaginal dysplasia, VAIN 2/3, was 100% and 21.2%, respectively, with a positive predictive value of 14.6%. The negative predictive value of hrHPV− DNA test results for the detection of VAIN 2/3 was 100%. These results are quite similar to data from women with LSIL findings from cervical Pap tests results from our institution (Table 4).
In addition to the 59 women with histopathologic follow-up, another 37 had only cytologic follow-up data available. Among those 37 patients, 23 (62.2%) were reported to have LSIL follow-up Pap test results during a follow-up period of 3 to 22 months; 17 of the 23 (74%) tested hrHPV+, and 6 of 23 (26%) tested hrHPV−. None of the patients followed only with cytology had high-grade squamous intraepithelial lesion results reported.
Vaginal neoplasia after hysterectomy is rare and has been reported5,6,17–22 to develop in up to 7.4% of patients whose hysterectomies followed a premalignant cervical diagnosis, compared with 1.1% in patients who had hysterectomies performed for benign conditions. Therefore, health care organizations generally recommend limited vaginal Pap test screening after most hysterectomies.1,2,9 Because of an absence of specific guidelines on the management of patients with abnormal vaginal Pap test results, clinicians appear to be following recommendations that have been published for abnormal cervical Pap test results.13 The American Society of Colposcopy and Cervical Pathology guidelines for abnormal cervical Pap test results do not recommend routine, reflex HPV testing after LSIL findings from cervical Pap test results because of the high rate of associated HPV infection reported in many studies. However, as patients get older, both the incidence of transient HPV infection and histopathologic high-grade dysplasia decrease.23–27 Therefore, in the 2006 American Society of Colposcopy and Cervical Pathology management guidelines, the option of reflex HPV testing in postmenopausal women with LSIL diagnoses from cervical Pap test results was included as an acceptable management approach for postmenopausal patients.13 Little data are available on the detection rate of hrHPV DNA in vaginal Pap test specimens with abnormal results,10,11 nor are substantial histopathologic correlation data available on subsequent vaginal biopsy specimens.28,29 We describe here the first correlation study on the cytology, histopathology, and HPV findings for patients with LSIL diagnoses from vaginal, posthysterectomy Pap test results. We document a 76.4% detection rate for hrHPV DNA in patients with LSIL diagnoses from vaginal Pap test results. We also document histopathologic diagnosis of VAIN 2/3 in 11.9% of patients who had biopsies and positive LSIL findings from vaginal Pap tests. Among patients with HPV+ vaginal LSIL Pap results, VAIN 2/3 was diagnosed in 14.6%. Among patients with HPV− vaginal LSIL Pap results, no patient had a VAIN 2/3 diagnosis from histopathologic follow-up results.
Because of the lack of any study of a similar nature in the literature, we compared our results with published correlation data on cervical LSIL Pap results. We also compared our results with data on cervical LSIL findings on Pap tests from elderly women because our study population for vaginal Pap test results was similar in age (study population average age, 56.7 years). There are some inherent dissimilarities in the study populations of patients with vaginal versus cervical screenings because cervical Pap tests are screening tests predominantly performed on asymptomatic women, whereas vaginal Pap tests are performed on a more-selected, older than average, posthysterectomy population, with patients who may present with accompanying symptoms or with findings of a lesion on physical examination.
Reported hrHPV detection rates associated with LSIL findings from cervical Pap test results ranged from 58% to 85% in one prior meta-analysis.30 The average age of the subjects in all the studies in the meta-analysis was significantly younger than the age of patients in our current study. In an earlier Magee-Womens Hospital study on women older than 50 years, hrHPV was detected in 71% of patients with LSIL findings from cervical Pap tests (Table 1).31 By comparison, the hrHPV detection rate in our current study on LSIL findings from vaginal Pap tests was 76.4%, which documents that hrHPV detection rates in these 2 study populations with LSIL are quite similar, despite other inherent differences in those populations. The limited data presented here tend to validate the utility of vaginal HPV testing as a risk-stratification tool for assessing the presence of an underlying histopathologic, high-grade vaginal dysplasia in patients with abnormal vaginal-cytology results.
In this study, we also documented that 14.6% of patients with HPV+ LSIL results on vaginal Pap tests had subsequent diagnoses of VAIN 2/3, compared with no diagnoses of VAIN 2/3 in patients with HPV− LSIL diagnoses from vaginal Pap test results. We observed similar findings in Magee-Womens Hospital studies of cervical LSIL diagnoses from Pap tests results when patients were stratified by age (<50 versus ≥50 years).32 No women 50 years and older with diagnoses of cervical LSIL from Pap tests had histopathologic cervical intraepithelial neoplasia 2/3 diagnosed on follow-up. These data demonstrate a high, negative predictive value for histopathologic high-grade vaginal dysplasia in older women with hrHPV− LSIL findings from Pap tests from either the cervix or the vagina. These observations support the option of managing older women with LSIL diagnoses from Pap test results with hrHPV triage, referring only older patients with hrHPV+ LSIL findings to colposcopy.
Our study also notes a significant incidence of histopathologic VAIN 1 (64%) in patients with hrHPV− LSIL vaginal Pap tests. One possible explanation for this finding could be that the VAIN 1 lesions harbor low-risk HPV types that are not detected by the high-risk HC2 test. In one study,33 performed on vaginal biopsies, 35% of patients with VAIN 1 tested positive for low-risk HPV types. Another possible explanation could be an overinterpretation of nonneoplastic, mild vaginal squamous atypia as VAIN 1.
Despite this study being conducted for a period of 4 years in a large, integrated health system for women, the number of patients with vaginal LSIL Pap tests and histopathologic follow-up was limited for both hrHPV+ and hrHPV− findings. These few cases limit the measurable statistical significance of our findings. Nevertheless, our data on vaginal specimens are strikingly similar to results on cervical Pap test specimens from elderly patients. Larger multiinstitutional comparisons would be useful to further extend these observations.
From the Department of Pathology, Magee-Womens Hospital, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania.
The authors have no relevant financial interest in the products or companies described in this article.
Presented in part at the annual meeting of the United States and Canadian Academy of Pathology, Washington, DC, March 20–26, 2010.