Massad, L. Stewart MD; Xie, Xianhong PhD; Greenblatt, Ruth M. MD; Minkoff, Howard MD; Sanchez-Keeland, Lorraine PA-C; Watts, D. Heather; Wright, Rodney L. MD; D'Souza, Gypsyamber PhD; Merenstein, Daniel MD; Strickler, Howard MD, MPH
From the Washington University School of Medicine, St. Louis, Missouri; the Albert Einstein College of Medicine, Bronx, New York; the University of California, San Francisco, California; Maimonides Medical Center, State University of New York, Downstate, Brooklyn, New York; the Keck School of Medicine, University of Southern California, Los Angeles, California; the Eunice KennedyShriver National Institute of Child Health and Human Development, Bethesda, Maryland; the Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland; and the Georgetown University School of Medicine, Washington, DC.
The Women's Interagency HIV Study is funded by the National Institute of Allergy and Infectious Diseases (UO1-AI-35004, UO1-AI-31834, UO1-AI-34994, UO1-AI-34989, UO1-AI-34993, and UO1-AI-42590) and by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (UO1-HD-32632). The study is cofunded by the National Cancer Institute, the National Institute on Drug Abuse, and the National Institute on Deafness and Other Communication Disorders. Funding is also provided by the National Center for Research Resources (UCSF-CTSI grant number UL1 RR024131). Analysis was funded through R01-CA-085178.
The authors thank the Women's Interagency HIV Study (WIHS) Collaborative Study Group for data collection. Their centers (principal investigator) are located in New York City/Bronx Consortium (Kathryn Anastos); Brooklyn, NY (Howard Minkoff); Washington, DC, Metropolitan Consortium (Mary Young); The Connie Wofsy Study Consortium of Northern California (Ruth Greenblatt); Los Angeles County/Southern California Consortium (Alexandra Levine); Chicago Consortium (Mardge Cohen); and the Data Coordinating Center, Johns Hopkins Bloomberg School of Public Health (Stephen Gange).
Presented in part at the annual meeting of the International Papillomavirus Society, September 16–22, 2011, Berlin, Germany.
The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.
Corresponding author: L. Stewart Massad, MD, Division of Gynecologic Oncology, Washington University School of Medicine, 4911 Barnes-Jewish Hospital Plaza, St. Louis, MO 63110; e-mail: firstname.lastname@example.org.
Financial Disclosure Dr. D'Souza has been a consultant for and received research support from Merck. The other authors did not report any potential conflicts of interest.
Infection with carcinogenic types of human papillomavirus (HPV) and consequent cervical intraepithelial neoplasia (CIN) is common in women with human immunodeficiency virus (HIV).1,2 CIN may require hysterectomy to prevent progression to cancer and is the most common reason for hysterectomy in women with HIV, underlying 42% of all hysterectomies.3
The course of HPV-related disease after hysterectomy is unclear. Human papillomavirus can infect the vagina as well as the cervix, and oncogenic HPVs can cause vaginal intraepithelial neoplasia (VAIN) and vaginal cancer.4 We have found oncogenic HPV in a high percentage of cervicovaginal lavage specimens from women with HIV who had undergone hysterectomy.5 Although VAIN has been reported in 10% of women with HIV who were investigated in studies of global lower genital tract neoplasia, its presentation among HIV-seropositive women has not been explored in detail.6 We estimated the prevalence, incidence, and clearance of abnormal vaginal cytology and the risk of VAIN and vaginal cancer in HIV-seropositive women and comparison HIV-seronegative women in a long-term U.S. prospective cohort.
MATERIALS AND METHODS
The Women's Interagency HIV Study is an ongoing multicenter U.S. cohort study of HIV infection and related health conditions begun October 1, 1994, with 2,054 HIV-seropositive and 569 HIV-seronegative comparison women and expanded in 2001–2002 with 737 HIV-seropositive and 406 HIV-seronegative women. The protocols, recruitment processes, procedures, baseline results, expansion enrollment, and retention efforts of the Women's Interagency HIV Study have been described previously.7–9 Written informed consent for the study was obtained after approval from institutional review boards of the collaborating institutions and agencies. Follow-up continues, but this study includes information obtained through September 30, 2008.
Every 6 months, participants had histories taken by trained interviewers followed by a physical examination that included a gynecologic examination and conventional Pap test on a glass slide. Pap tests were read centrally and results were reported according to the 1991 Bethesda System,10 with high-grade squamous intraepithelial lesions (HSIL) subcategorized as consistent with either moderate or severe dysplasia or carcinoma in situ with modifications incorporating the 2001 Bethesda System in 2002. Colposcopy was prescribed by protocol for all women with abnormal cytology, defined as atypical squamous cells of undetermined significance (ASC-US) or a more severe lesion. Histology results were interpreted locally and were not centrally reviewed; women with VAIN that was not graded were included in analysis of “any VAIN” but not VAIN 2 or worse. Human immunodeficiency virus status was confirmed by Western blot. Women who seroconverted during follow-up were excluded. Highly active antiretroviral therapies (HAARTs) were defined according to Department of Health and Human Services/Kaiser guidelines.11
We focused on two end points. First, we identified all women with abnormal vaginal cytology after hysterectomy. For women without a self-reported history, hysterectomy was assumed if no cervix was reported at two consecutive examinations. Contingency tables of demographic characteristics were constructed using data at the first visit after hysterectomy. Characteristics of HIV-seropositive and HIV-seronegative women were compared using Pearson's χ2 tests or Fisher's exact tests.
The second end point was histologically confirmed VAIN regardless of prior hysterectomy. We included VAIN documented by biopsy at a Women's Interagency HIV Study visit or at another institution. Because VAIN 1 is considered to be a reflection of HPV infection of marginal oncogenic risk, we combined biopsies read as condyloma or koilocytosis with VAIN 1. For analyses of VAIN by grade, each woman contributed only once, classifying her according to the highest grade of VAIN she had identified during the study. Women were considered to have prevalent VAIN if it was diagnosed after one or more consecutive abnormal Pap test results beginning with enrollment. The prevalence of cytologic or histologic vaginal abnormality was compared between HIV-seropositive and HIV-seronegative women using the Pearson's χ2 test. Univariate and multivariable analyses that incorporated data from multiple visits were conducted using generalized estimating equation models for binary data, as previously described,12 which adjusted for repeated observations of cytologic or histologic abnormalities in the same women over time. Temporal trends in prevalence over time were examined using a generalized estimating equation model by incorporating visit number as a variable.
Incident cytologic abnormality was assessed in women who had undergone hysterectomy, because they received routine semiannual vaginal Pap testing. Incident VAIN was VAIN identified after one or more negative Pap test results regardless of hysterectomy. Each woman could contribute only a single incident cytologic or histologic abnormality analyzed as separate events. A χ2 test with one degree of freedom was used to compare differences in incidence rates between HIV-seropositive and HIV-seronegative women during a single year or all years pooled; the statistic was based on the normal approximation to the binomial distribution using the number of cases in the HIV-seropositive group as the binomial random variable.13
We also examined cumulative risk, including both prevalent and incident cases using the life table method.14 Comparison between HIV-seropositive and HIV-seronegative groups was conducted using the log-rank test. Multivariable Cox analysis was performed to examine time-to-event for incident abnormality (cytology or VAIN) with HIV status and CD4 count as our primary variables of interest with adjustment for age, smoking, and number of sexual partners in the past 6 months. The midinterval was used as the event time. Patients were time-censored at the last visit with available data if data were missing for two consecutive visits. Time to clearance of abnormality was also studied using life table and Cox models. All statistical tests in our study were two-sided.
After excluding 22 HIV seroconverters, 418 (11%; 95% confidence interval [CI] 10–12%) of 3,744 women in the Women's Interagency HIV Study had reported prior hysterectomy (343 HIV-seropositive, 75 HIV-seronegative), including 241 with hysterectomy before enrollment and 177 during follow-up. Of these, 410 (335 HIV-seropositive, 75 HIV-seronegative) contributed up to 28 vaginal Pap tests over a median of 5.6 years of follow-up after enrollment or hysterectomy for a total of 4,463 vaginal Pap test results (3,700 among HIV-seropositive and 763 from HIV-seronegative women). Table 1 presents the demographic and medical characteristics of these women; except for a higher rate of smoking in women without HIV, HIV-seropositive and HIV-seronegative women with hysterectomy were similar.
Pap test results were abnormal at 1,076 of 3,700 (29%; 95% CI 25–33%) visits among HIV-seropositive women, but only 31 of 763 (4%; 95% CI 2–8%) visits among HIV-seronegative women (P<.001), in analyses using generalized estimating equation logistic regression. The distribution of Pap test results by HIV serostatus is shown in Table 2; women with HIV had more Pap test result abnormalities, although most were ASC-US and low-grade squamous intraepithelial lesions (LSILs), suggesting poorly controlled HPV infection rather than precancer. Although HSIL Pap test results were uncommon, they were more frequent in HIV-seropositive (10 of 3,700 [0.3%]; 95% CI 0.1–0.6%) than in HIV-seronegative women (zero of 763; 95% CI 0–0%; P<.001). The frequency of vaginal Pap test result abnormality was similar for HIV-seropositive women with hysterectomy before (698 of 2,418 [29%]; 95% CI 27–31% tests) and after (378 of 1,282 [29%]; 95% CI 27–32% tests) study entry.
We previously have shown that visit-specific prevalence of cervical Pap test result abnormality among all Women's Interagency HIV Study participants fell across time, presumably as treatments decreased the burden of cervical disease.8 In contrast, Figure 1 shows no time trend in the prevalence of vaginal Pap test result abnormality in either HIV-seropositive or HIV-seronegative women (P=.11 for HIV-seropositive women and .50 for HIV-seronegative women).
As shown in Table 3, multivariable analyses found that prevalent detection of any abnormal vaginal Pap test result was more common in HIV-seropositive women with increasing immunosuppression and current smokers but less likely among women with multiple sexual partners. CD4 count was also significantly associated with HSIL Pap test results (P for trend <.001). As shown in Table 4, when analysis was repeated among only HIV-seropositive women, CD4 lymphocyte count, HIV RNA level, use of HAART, and smoking were independently associated with vaginal Pap test result abnormality; age, ethnicity, number of sexual partners in the prior 6 months, and prevalent compared with incident hysterectomy were not linked to a finding of abnormal Pap test results in these women. Similar results were obtained if analysis was restricted to any squamous intraepithelial lesions (excluding ASC-US) and to prevalent Pap test result abnormalities (not shown).
We also analyzed the incidence of abnormal vaginal Pap test results. Among the 410 hysterectomized women, 136 (121 HIV-seropositive, 15 HIV-seronegative) were excluded because of abnormal Pap test results at the time of hysterectomy (n=108) or lack of follow-up (n=28). Among the remaining 274 hysterectomized women, 114 (42%; 95% CI 38–46%) developed incident abnormal vaginal Pap test results, whereas 160 (58%; 95% CI 52–64%) did not. The incidence of abnormal vaginal Pap test results after hysterectomy was 14 per 100 person-years among HIV-seropositive and two per 100 person-years among HIV-seronegative women (P<.001).
The cumulative risk of an abnormal Pap test result, including baseline and follow-up, was high in both groups: after 12 years of observation, the risk of ever having abnormal vaginal cytology was 75% (95% CI 64–83%) in HIV-seropositive and 42% (95% CI 2–66%) in HIV-seronegative women (P=.13). The risk of ever having a Pap test read as HSIL or worse over 12 years of observation was 6.4% (95% CI 0.3–12.2%) for HIV-seropositive women and 0.0% (95% CI 0–0%) for HIV-seronegative women (P=.03).
Clearance of Pap test result abnormality was evaluated among all 193 women (177 HIV-seropositive, 16 HIV-seronegative) with prevalent or incident Pap test result abnormalities. After excluding 25 women without follow-up and four women who were treated (all excluded women were seropositive), the 193 women included 91 (84 HIV-seropositive and seven HIV-seronegative) women with prevalent Pap test result abnormalities and 102 (93 HIV-seropositive, nine HIV-seronegative) with incident abnormalities. Of these, 113 (64%; 95% CI 57–71%) of HIV-seropositive and 13 (81%; 95% CI 57–93%) of HIV-seronegative women cleared their abnormality without treatment (P=.16) across all visits. The 5-year clearance rate was 34 cases per 100 person-years for HIV-seropositive women and 116 per 100 person-years for HIV-seronegative women (P<.001). Clearance rates were higher for women with ASC-US (98 [72%]; 95% CI 63–78%) than LSIL (28 [50%]; 95% CI 37–63%; P=.004). In multivariable Cox analysis, clearance of any Pap test result abnormality was less likely among HIV-seropositive women with lower CD4 counts (hazard ratio 0.79, 95% CI 0.37–1.66 for CD4 greater than 500/microliter; 0.53, 95% CI 0.26–1.1 for CD4 200–500/microliter, and 0.36, 95% CI 0.16–0.80, for CD4 less than 200/microliter; P for trend=.002, compared with HIV-seronegative women) or with LSIL compared with ASC-US cytology (hazard ratio 0.52, 95% CI 0.32–0.87, P=.01). Those with more than one sexual partner in the 6 months before an abnormal Pap test result were more likely to clear Pap test result abnormalities (hazard ratio 2.34, 95% CI 1.19–4.59, P=.01 compared with no sexual partner). Age, ethnicity, smoking, parity, and incident compared with prevalent hysterectomy were not linked to clearance of cytologic abnormality. In a separate model limited to HIV-seropositive women, after adjusting for CD4 count, plasma HIV RNA level and HAART use were not associated with clearance.
We next assessed vaginal biopsy results, including women with and without prior hysterectomy. Biopsies were obtained from 269 women, including 255 of 2,791 (9%; 95% CI 8–10%) HIV-seropositive women and 14 of 953 (1%; 95% CI 1–2%) HIV-seronegative women (P<.001). Table 5 shows the highest grade VAIN for each woman, including results from women with multiple biopsies; although only HIV-seropositive women had VAIN 3, differences in the distribution of biopsy grade did not reach significance. Prevalent VAIN of any grade was found within 6 months of intake in 21 of 2,791 (1%; 95% CI 0.7–1.4%) HIV-seropositive women and no HIV-seronegative women (P=.01). Incident VAIN was found in 151 HIV-seropositive and seven HIV-seronegative women, and the incidence rate of VAIN was 0.8 per 100 person-years for HIV-seropositive and 0.1 per 100 person-years for HIV-seronegative women (P<.001).
We found no prevalent cases of VAIN 2 or worse in HIV-seropositive or HIV-seronegative women, whereas incident VAIN 2 or worse developed in 36 HIV-seropositive women and one HIV-seronegative woman. The incidence of VAIN 2 or worse was 0.2 per 100 person-years for HIV-seropositive women and 0.01 per 100 person-years for HIV-seronegative women (P=.001). Associations between VAIN and various risk factors are shown in Table 6. In multivariable Cox models, incident VAIN was associated with lower CD4 count, current smoking, and higher parity but not age, ethnicity, or number of recent sexual partners. In a separate model limited to HIV-seropositive women, HIV RNA level in blood was not associated with incident VAIN after controlling for CD4 count. In another model incorporating HIV-seropositive women that adjusted for multiple additional risk factors in addition to CD4 count and HIV RNA level, incident VAIN was linked to HAART use in the prior 6 months (hazard ratio 1.94, 95% CI 1.23–3.05, P=.004) and current smoking (hazard ratio 2.04, 95% CI 1.19–3.51, P=.01).
The incidence of VAIN remained higher among HIV-seropositive women when only the 418 women with hysterectomy were evaluated. Vaginal intraepithelial neoplasia was found within 6 months of enrollment in 13 women, all HIV-seropositive (4% of all seropositive women with hysterectomy, 95% CI 2–6%). In all, incident VAIN was found in 56 women (54 HIV-seropositive, two HIV-seronegative). The incidence of VAIN was 2.9 per 100 person-years for HIV-seropositive women and 0.4 per 100 person-years for HIV-seronegative women (P=.002). However, the incidence of VAIN 2 or worse in women after hysterectomy was only 0.8 per 100 person-years for HIV-seropositive and zero of 100 person-years in HIV-seronegative women (P=.05). Treatment for VAIN was undertaken for 41 women (38 HIV-seropositive, three HIV-seronegative). Only 12 women (11 HIV-seropositive, one HIV-seronegative) required treatment for VAIN detected after hysterectomy, but Pap test results remained abnormal after treatment in 10 (nine HIV-seropositive, one HIV-seronegative).
Two women, both HIV-seropositive, developed vaginal cancer after prior hysterectomy. The first was diagnosed with stage II squamous cell carcinoma in 2000 and was treated with radiotherapy; she died 10 months later from substance abuse complications but with persistent vaginal cancer. Prior Pap test results had shown ASC-US or LSIL, but three vaginal biopsies before diagnosis showed only condyloma. The second patient had undergone hysterectomy in 1997 for cervical carcinoma in situ and a Pap test read as atypical glandular cells. No cancer was found then and no subsequent vaginal biopsies were done. She was diagnosed with a stage II adenosquamous carcinoma of the vagina in 2004 and was treated with excision and radiotherapy and was free of disease 7 years after initial treatment.
Although HPV-related disease can occur throughout the lower genital tract, vaginal precancers and cancers are uncommon. Our findings demonstrate that although HIV-related immunosuppression raises the risk for abnormal vaginal Pap test results and VAIN, HIV-seropositive women remain at low absolute risk for vaginal precancer and cancer. The cumulative risk of abnormal Pap test results after up to 12 years of observation was 75% in HIV-seropositive women after hysterectomy, but the risk of HSIL cytology among these women was only 6.4% over years of observation. Most HIV-seropositive women with abnormal vaginal Pap test results cleared their abnormalities spontaneously, but more than one-third did not; their likelihood of clearance was lower than that among seronegative women and the most severely immunosuppressed HIV-seropositive women had the lowest likelihood of clearance. Treatments for VAIN after hysterectomy were relatively ineffective. Few abnormal Pap test results were associated with VAIN 2 or worse, and a lower proportion of HIV-seropositive than HIV-seronegative women with abnormal vaginal Pap test results had VAIN 2 or worse, probably a reflection of opportunistic expression of HPV infections that are controlled in HIV-seronegative women. In fact, among HIV-seropositive women after hysterectomy, which is commonly performed for CIN,3 the incidence of VAIN 2 or worse was only 0.8 per 100 person-years. Unfortunately, two cancers occurred in our cohort and were seen after only borderline cytology results.
Smoking, CD4 count, HIV RNA level, and HAART use were associated with higher risk for vaginal lesions. Future studies should explore whether smoking cessation and a rising CD4 cell count after HAART initiation lower risk for and speed clearance of abnormal vaginal Pap test results and VAIN. In the absence of prospective studies, these correlations may offer an additional incentive for smoking cessation and HAART adherence among women with these abnormalities.
Interpretation of our results is limited by several factors. We were unable to determine indications for prior hysterectomies in all cases, and the risk for abnormal Pap test results and VAIN may be lower among those undergoing hysterectomy for reasons other than CIN such as bleeding or pain. Among immunocompetent women, Pap testing is not indicated after hysterectomy for benign disease, whereas women with a history of CIN or cancer merit long-term annual screening.15 Future research should determine whether HIV-seropositive women with no history of prior CIN and women with multiple consecutive negative Pap test results benefit from continued Pap testing after hysterectomy.
Until those studies have been completed, all HIV-seropositive women who have had a hysterectomy should be advised to continue annual Pap testing. Women facing hysterectomy for cervical disease should be counseled that, although surgery may be lifesaving as a cervical cancer prevention measure, HPV is not eradicated and vaginal Pap test results often will be abnormal and must be followed up with colposcopy for all abnormalities, including ASC-US. Although HPV cannot be eradicated surgically and abnormal Pap test results often persist after therapy, VAIN 2 or worse should be treated, and intravaginal use of 5-fluorouracil may improve disease control. Vaginal intraepithelial neoplasia 1 can be treated or followed depending on symptoms and treatment risk. With such careful surveillance and targeted intervention, women living with HIV can be reassured that abnormal Pap test results and VAIN rarely presage vaginal cancer.
1. Palefsky JM, Minkoff H, Kalish LA, Levine A, Sacks HS, Garcia P, et al.. Cervicovaginal human papillomavirus infection in human immunodeficiency virus-1 (HIV)-positive and high-risk HIV-negative women. J Natl Cancer Inst 1999; 91: 226–36.
2. Massad LS, Seaberg EC, Wright RL, Darragh T, Lee YC, Colie C, et al.. Squamous cervical lesions in women with human immunodeficiency virus: long-term follow-up. Obstet Gynecol 2008; 111: 1388–93.
3. Massad LS, Evans C, Weber K, Cejtin HE, Golub ET, DiGilio K, et al.. Hysterectomy among women with HIV: indications and incidence. J Acquir Immune Defic Syndr 2007; 44: 566–8.
4. Insinga RP, Liaw KL, Johnson LG, Madeleine MM. A systematic review of the prevalence and attribution of human papillomavirus types among cervical, vaginal, and vulvar precancers in the United States. Cancer Epidemiol Biomarkers Prev 2008; 17: 1611–22.
5. D'Souza G, Burk RD, Zhong Y, Minkoff H, Massad LS, Xue X, et al.. Cervicovaginal HPV infection before and after hysterectomy: evidence of different tissue tropism for oncogenic and non-oncogenic HPV types in a cohort of HIV-positive and HIV-negative women. Int J Cancer 2011 [Epub ahead of print].
6. Dedes KJ, Beneder C, Samartzis N, Muller MD, Fink D, Fehr MK. Outcome of treated anogenital intraepithelial neoplasia among human immunodeficiency virus-infected women. J Reprod Med 2008; 53: 947–51.
7. Barkan SE, Melnick SL, Martin-Preston S, Weber K, Kalish LA, Miotti P, et al.. The Women's Interagency HIV Study. WIHS Collaborative Study Group. Epidemiology 1998; 9: 117–25.
8. Bacon M, von Wyl V, Alden C, Sharp G, Robison E, Hessol N, et al.. The Women's Interagency HIV Study: an observational cohort brings clinical sciences to the bench. Clin Diagn Lab Immunol 2005; 12: 1013–9.
9. Hessol NA, Weber KM, Holman S, Robison E, Goparaju L, Alden CB, et al.. Retention and attendance of women enrolled in a large prospective study of HIV-1 in the United States. J Womens Health (Larchmt) 2009; 18: 1627–37.
10. Kurman RJ, Solomon D. The Bethesda System for reporting cervical/vaginal cytologic diagnoses. New York (NY): Springer-Verlag; 1994.
12. Xue X, Gange SJ, Zhong Y, Burk RD, Minkoff H, Massad LS, et al.. Marginal and mixed-effects models in the analysis of human papillomavirus natural history data. Cancer Epidemiol Biomarkers Prev 2010; 19: 159–69.
13. Sahai H, Khurshid A. Statistics in epidemiology: methods, techniques, and applications. Boca Raton (FL): CRC Press; 1996.
14. Berkson J, Gage RP. Calculation of survival rates for cancer. Proc Staff Meet Mayo Clin 1950; 25: 270–86.
15. Cervical cytology screening. ACOG Practice Bulletin No. 109. American College of Obstetricians and Gynecologists. Obstet Gynecol 2009; 114: 1409–20.