Women with human immunodeficiency virus (HIV) are more likely than those without HIV to acquire genital human papillomavirus (HPV), an infection that persists longer in women with HIV than in women without HIV.1,2 As a consequence, women with HIV have higher rates of squamous intraepithelial lesions (SILs) on Pap tests and higher rates of cervical intraepithelial neoplasia.3,4 Their increased susceptibility is thought to be due, at least in part, to HIV-induced immune compromise. Because highly active antiretroviral therapy (HAART) lowers HIV viral load, slows the progression of HIV disease, and helps restore normal immune function,5 it has been hypothesized that HAART would enhance HPV clearance and the regression of cervical dysplasia. To this point, however, an effect of HAART on HPV infection and cervical dysplasia has not been demonstrated clearly. Three studies have failed to show an association between HAART and HPV persistence in women,6–8 and reports on the effect of HAART on HPV-associated cervical disease have been mixed. A few studies have reported that women on HAART are more likely to have regression of cervical lesions,7,9–11 whereas other studies have shown no effect of HAART.8,12,13 The purpose of the present study was to examine the association between HAART and HPV clearance and progression or regression of cervical cytologic abnormalities found on Pap test in a well-characterized cohort of women with HIV.
This analysis is part of the HIV Epidemiology Research (HER) Study, a prospective cohort study of women with HIV and at-risk women without HIV. Detailed descriptions of the HER Study have been reported previously.14,15 Briefly, beginning in April of 1993 and continuing until January of 1995, 871 women with HIV and 439 at-risk women without HIV were enrolled at four U.S. sites: Baltimore, MD, Detroit, MI, Bronx, NY, and Providence, RI. Each study site contributed approximately one fourth of the participants. Women with HIV without acquired immunodeficiency syndrome–defining conditions and women without HIV were selected to give an approximately 50:50 ratio between those with risk from injection drug use and those with increased risk due to sexual behavior but not injection drug use. Women were evaluated until March 2000. Semiannual research visits included core interviews on medical history, obstetric and gynecologic history, history of substance use, and sexual behavior, as well as a physical examination that included a pelvic examination. Institutional review boards approved the study protocol at each site and at the Centers for Disease Control and Prevention, and all participants gave written informed consent.
Antiretroviral use was ascertained by self-report at each visit. Regimen classifications were adapted from the U.S. Department of Health and Human Services’ guidelines for preferred or alternative combinations, issued in May of 1999.5 For this analysis, women were considered to be on HAART from the first study visit where they reported HAART use and every subsequent visit. Women who never reported HAART use but who may have been on antiretroviral therapy not defined as HAART were placed a priori in the never on HAART group.
For the first 10 study visits, cervical Pap tests were performed at every visit. After visit 10, Pap tests were performed at every other visit. All Pap tests were evaluated using predefined criteria by a single commercial cytologic laboratory (Kyto Diagnostics, LP, New York, NY). The predefined criteria were based on an expansion of the 1988 Bethesda scoring system for cervical cytology.16 Cytopathologists were blinded to participants’ HIV serostatus. For statistical analysis, the Pap tests were categorized as normal, atypical squamous cells of undetermined significance (ASC-US), low-grade SILs (LSIL), or high grade SILs (HSIL). Regression was defined as any lower grade Pap test result on a subsequent Pap test (HSIL followed by LSIL, LSIL followed by ASC-US, ASC-US followed by normal); Pap test progression was defined in the converse manner. Normal Pap test results at the index visit were excluded from Pap test regression analyses.
At the Bronx and Providence sites, women with abnormal Pap test results or with abnormalities noted on gynecologic examination were referred for colposcopy. At the Detroit site, all women received colposcopy at entry into the HER Study. After study entry, women with abnormal Pap test results or with abnormalities noted on physical examination were referred for colposcopy. At the Baltimore site, colposcopy was performed at each visit regardless of the Pap test result or gynecologic examination. Treatment for cervical dysplasia was identified by abstraction of the medical records at all study sites. Treatment was defined as the following procedures: excisional biopsy, cone biopsy, loop excision, knife conization, 5-fluorouracil, interferon, or cryotherapy.
Cervical vaginal lavage specimens were collected at each visit as a 10-mL saline wash, frozen, and tested for HPV using polymerase chain reaction (PCR) at a central laboratory as previously described.1 The HPV types that were identified by this method include the following: 6, 11, 16, 18, 26, 31, 33, 35, 39, 40, 42, 45, 51, 52, 53, 54, 55, 56, 58, 59, 66, 68, 73, 82, 83, and 84. Polymerase chain reaction amplification of a human Β-globin gene segment was used as an internal control for DNA quality. Samples that were Β-globin–negative were considered unsatisfactory and classified as having missing HPV data.
CD4+ T-lymphocyte counts were calculated at each visit and were grouped into three categories: less than 200, 200–499, and 500 or more cells/microliter. Viral load for HIV was determined at each visit using a third-generation branched DNA signal amplification assay (Chiron Corp, Emeryville, CA) with a lower quantification limit of 50 copies/mL.
This analysis includes a subset of 537 women with HIV who were HIV-positive before HER Study entry, had cervices, and had at least two study visits after April 30, 1996. The first protease inhibitor was commercially available in April of 1996, and women in the HER Study cohort began reporting HAART use after this date. Eighty-two percent of the women who started on HAART had protease inhibitors as part of their initial HAART regimen. Women who were never on antiretroviral therapy as well as those who were on antiretroviral therapy not determined to be HAART were placed in the never on HAART group.
Data were analyzed using SAS (SAS Institute Inc., Cary, NC). For the baseline demographic table, Pearson’s χ2 or Fisher exact tests were used to compare proportions, and Student’s t tests were used to compare means between groups.
To evaluate the effects of HAART, Cox proportional hazard models were used. Women who were never on HAART had time calculated from the first study visit after April 30, 1996, until their last study visit. For women on HAART, time was calculated from the first study visit where they reported HAART use until their last study visit. Women on HAART also were placed in the model for the time before they started HAART (pre-HAART). This time period was from the first study visit after April 30, 1996, until the visit before HAART use first was reported. The pre-HAART group was combined with the never on HAART group. Therefore, a woman who was never on HAART contributed one time period, whereas a woman who started HAART could have contributed two time periods in the model. Models were adjusted for repeated measures by using the sandwich variance–covariance estimator. Clearance models for HPV included women who were HPV-positive at the index visit. Women who were HPV-negative by PCR test were excluded from the clearance models. Clearance of HPV was defined as two negative PCR tests 6 months apart. A negative PCR test was defined as negative for any HPV type. Time to HPV clearance was calculated from the index visit until the date of the first of two consecutive HPV-negative results. Potential risk factors, identified a priori, were examined in multivariable models. Interaction between HAART and risk factors was included in a model if P<.2. The number needed to treat (NNT) at 2 years was calculated using the survival probability for the non-HAART group and the hazard ratio for HAART.17
Of the 537 women with HIV in this analysis, 285 (53.1%) never reported HAART use and 252 (46.9%) did report HAART use. Of the 285 women who never reported HAART use, 152 (53%) reported antiretroviral use at some point during the study, although their regimens were not classified as HAART. At enrollment in the HER Study, women who were never on HAART during the study were more likely to be in the risk cohort of injection drug use, African American, and current smokers and were more likely to use an injection drug and use crack in the 6 months before enrollment than women who started HAART during the study. The women never on HAART were less likely to be on antiretrovirals of any type and had higher CD4 counts and lower HIV viral loads than women on HAART. Women who were never on HAART did not differ significantly from women who were on HAART during the study in age and HPV presence at study enrollment (Table 1). Differences also were not seen in education, monthly income, number of live births, ever having had sex for drugs or money, condom use in the 6 months before enrollment, number of male sex partners in the 6 months before enrollment, or hormonal contraceptive use (data not shown).
For this analysis, the mean follow-up time for women on HAART (2.0±0.8 years) differed significantly from the mean follow-up time for women never on HAART (2.7±1.0 years) (P<.01). The results of the hazard regression models for HPV clearance are shown in Table 2. Because significant interaction between HAART and the index Pap test result was detected, hazard models for HPV clearance were calculated separately for women with normal, ASC-US, and SIL (LSIL and HSIL combined because of small sample size) index Pap test results. Study site was not shown to be a factor (P>.2) and was not included in the final HPV clearance models.
Among 168 women with normal Pap test results and HPV infection at the index visit, there were 196 time periods. Of these time periods, 78 were for women who were never on HAART, 63 were for women who were pre-HAART, and 55 were for women on HAART. In this model, HAART was not significantly associated with HPV clearance (hazard ratio 1.7, 95% confidence interval [CI] 0.9–3.1, NNT 6.5). Higher CD4 counts were significantly associated with HPV clearance (hazard ratio 1.1, 95% CI 1.0–1.2, P<.05). Age was not associated with HPV clearance (hazard ratio 0.8, 95% CI 0.4–1.4).
Among 122 women with ASC-US Pap test results and HPV infection at the index visit, there were 129 time periods. Of these time periods, 57 were for women who were never on HAART, 35 were for women who were pre-HAART, and 37 were for women on HAART. In this model, none of the following were significantly associated with HPV clearance: HAART (hazard ratio 1.0, 95% CI 0.4–2.5, NNT 174.0), CD4 counts (hazard ratio 1.1, 95% CI 1.0–1.3), and age (hazard ratio 0.5, 95% CI 0.2–1.0).
Among 144 women with SIL Pap test results and HPV infection at the index visit, there were 178 time periods. Of these time periods, 56 were for women who were never on HAART, 45 for were women who were pre-HAART, and 77 were for women on HAART. In the SIL Pap model, women on HAART were more likely to clear their HPV infections (hazard ratio 4.5, 95% CI 1.2–16.3, NNT 22.4). Women age 36 years and older were less likely to clear their HPV infections (hazard ratio 0.1, 95% CI 0.0–0.6). CD4 counts were not associated with HPV clearance (hazard ratio 1.2, 95% CI 1.0–1.4).
Because treatment of cervical disease may not clear HPV infection effectively, women included in the HPV clearance models were not censored at the time of treatment for cervical dysplasia in the primary analysis. However, as part of our secondary analyses, we did create additional models, censoring at time of cervical treatment. The point estimates for the risk factors in these models did not change significantly; however, the CIs did widen.
The hazard models for Pap test progression and Pap test regression are shown in Table 3. There were 700 time periods in the progression model: 285 for women who were never on HAART, 163 for women who were pre-HAART, and 252 for women on HAART. Women were censored at the time of treatment for cervical dysplasia. Study site was not shown to be a factor (P>.2) and was not included in the final progression or regression model. Use of HAART was not statistically associated with Pap test progression (hazard ratio 0.7, 95% CI 0.6–1.0, P=.06, NNT 12.8). Higher CD4 counts were less likely to exhibit Pap test progression (hazard ratio 0.9, 95% CI 0.9–1.0, P<.05). Infection with HPV was significantly associated with Pap test progression (hazard ratio 2.7, 95% CI 1.9–3.8). There were 328 time periods in the Pap test regression model: 123 for women who were never on HAART, 74 for women who were pre-HAART, and 131 for women on HAART. Use of HAART was not significantly associated with Pap test regression (hazard ratio 1.3, 95% CI 1.0–1.7, P=.05, NNT 10.9). Higher CD4 counts were significantly associated with Pap test regression (hazard ratio 1.1, 95% CI 1.0–1.2, P<.05). Infection with HPV and index Pap test result were not associated with Pap test regression.
In this large, prospective, cohort study of women with HIV, we found that HAART was associated with clearance of HPV infection in women with SIL cervical cytology but not in women with normal or ASC-US Pap test results. Other studies have not reported a decrease in HPV detection associated with HAART. For example, a 1998 French study found no difference in HPV prevalence in 26 women before and after the start of HAART.7 A 2001 Italian study of 163 women with HIV found that HAART was not associated with the persistence of high-risk HPV types.8 Another Italian study, published in 2004, evaluated 201 women with HIV and found no difference in HPV persistence among women who were on HAART compared with other antiretroviral therapies or undergoing no antiretroviral therapy.6 None of the above studies stratified the relationship between HPV persistence/clearance and HAART by initial Pap test, and this may explain why their results differ from ours—we found no significant effect of HAART on HPV clearance among women with normal or ASC-US Pap test results.
Our analysis suggested that women on HAART were 30% less likely to exhibit progression and 30% more likely to exhibit regression on their Pap tests than were women not on HAART. However, these differences did not reach statistical significance. In a previous analysis using the HER Study cohort, HAART was not correlated with the risk of incident SIL or regression of SIL measured 6 months later.18 The current analysis examined the long-term effect of HAART on cervical cytologic regression and progression in all Pap categories.
Several studies with large sample sizes have reported beneficial effects of HAART on cytologic findings. The Women’s Interagency HIV Study found that women on HAART were more likely to demonstrate regression in a 6-month pair analysis.11 A follow-up analysis found that women with incident SIL were more likely to regress if they were on HAART.9 The 2004 Italian study cited above found the rate of LSIL regression to be higher among women receiving HAART.6 A study of 168 women in France found a positive association between HAART and regression of cervical intraepithelial neoplasia,7 but a 2006 study from the same cohort found that HAART did not affect the risk of incident SIL.19 Other studies, with mostly smaller samples, did not find an effect of HAART on HPV-associated cervical disease regression.8,12,13
Our study has several limitations. First because of the relatively small sample, we were unable to examine HPV type-specific clearance. We also were unable to examine specific antiretroviral treatment regimens. Because many women were referred for colposcopy and histological information was not known for all participants, we used cytologic findings. Treatment for cervical dysplasia was defined by abstraction of the medical records, which may have led to an underestimation of cervical treatment for dysplasia. As in previous studies,9,11 we used an intent-to-treat analysis—once a woman reported HAART use, she was considered to be on HAART through her last study visit. This approach did not evaluate the effect of discontinuation, changes in regimens, or adherence to HAART. An earlier analysis using the HER Study cohort found that adherence to antiretroviral therapy was low and differed over time.20 The adherence rate for antiretroviral therapy was 64% during the first month of observation and dropped to 45% at the 6th month of observation. Low adherence, if present in the current study, likely would lead to an underestimation of effects associated with HAART.
Despite these limitations, our study had several strengths. This is one of the largest studies to examine the effect of HAART on HPV clearance and HPV-associated cervical disease. The women in the study are from a well-described cohort of women who were evaluated semiannually for up to 7 years. Testing for HPV by PCR was done for each semiannual visit. Our analysis considered long-term effects; this may explain the discrepancy of our results with other studies that looked at shorter time intervals.
In conclusion, our study found that HAART was associated with enhanced HPV clearance in women who already had cervical disease. Our findings, and those of others, suggest that antiretroviral therapy may ameliorate HPV-related disease. These effects do not occur rapidly and do not lead to complete resolution of cervical cytologic abnormalities in women with HIV; thus, close monitoring still is warranted.
1. Ahdieh L, Klein RS, Burk R, Cu-Uvin S, Schuman P, Duerr A, et al. Prevalence, incidence, and type-specific persistence of human papillomavirus in human immunodeficiency virus (HIV)-positive and HIV-negative women. J Infect Dis 2001;184:682–90.
2. Sun XW, Kuhn L, Ellerbrock TV, Chiasson MA, Bush TJ, Wright TC Jr. Human papillomavirus infection in women infected with the human immunodeficiency virus. N Engl J Med 1997;337:1343–9.
3. Ellerbrock TV, Chiasson MA, Bush TJ, Sun XW, Sawo D, Brudney K, et al. Incidence of cervical squamous intraepithelial lesions in HIV-infected women. JAMA 2000;283:1031–7.
4. Wright TC Jr, Ellerbrock TV, Chiasson MA, Van Devanter N, Sun XW. Cervical intraepithelial neoplasia in women infected with human immunodeficiency virus: prevalence, risk factors and validity of Papanicolaou smears. New York Cervical Disease study. Obstet Gynecol 1994;84:591–7.
6. Del Mistro A, Bertorelle R, Franzetti M, Cattelan A, Torrisi A, Giordani MT, et al. Antiretroviral therapy and the clinical evolution of human papillomavirus-associated genital lesions in HIV-positive women. Clin Infect Dis 2004;38:737–42.
7. Heard I, Schmitz V, Costagliola D, Orth G, Kazatchkine MD. Early regression of cervical lesions in HIV-seropositive women receiving highly active antiretroviral therapy. AIDS 1998;12:1459–64.
8. Lillo FB, Ferrari D, Veglia F, Origoni M, Grasso MA, Lodini S, et al. Human papillomavirus infection and associated cervical disease in human immunodeficiency virus-infected women: effect of highly active antiretroviral therapy. J Infect Dis 2001;184:547–51.
9. Ahdieh-Grant L, Li R, Levine AM, Massad LS, Strickler HD, Minkoff H, et al. Highly active antiretroviral therapy and cervical squamous intraepithelial lesions in human immunodeficiency virus-positive women. J Natl Cancer Inst 2004;96:1070–6.
10. Heard I, Tassie JM, Kazatchkine MD, Orth G. Highly active antiretroviral therapy enhances regression of cervical intraepithelial neoplasia in HIV-seropositive women. AIDS 2002;16:1799–802.
11. Minkoff H, Ahdieh L, Massad LS, Anastos K, Watts DH, Melnick S, et al. The effect of highly active antiretroviral therapy on cervical cytologic changes associated with oncogenic HPV among HIV-infected women. AIDS 2001;15:2157–64.
12. Moore AL, Sabin CA, Madge S, Mocroft A, Reid W, Johnson MA. Highly active antiretroviral therapy and cervical intraepithelial neoplasia. AIDS 2002;16:927–9.
13. Orlando G, Fasolo MM, Schiavini M, Signori R, Cargnel A. Role of highly active antiretroviral therapy in human papillomavirus-induced genital dysplasia in HIV-1-infected patients. AIDS 1999;13:424–5.
14. Smith DK, Warren DL, Vlahov D, Schuman P, Stein MD, Greenberg BL, et al. Design and baseline participant characteristics of the Human Immunodeficiency Virus Epidemiology Research (HER) Study: a prospective cohort study of human immunodeficiency virus infection in US women. Am J Epidemiol 1997;146:459–69.
15. Duerr A, Kieke B, Warren D, Shah K, Burk R, Peipert JF, et al. Human papillomavirus-associated cervical cytologic abnormalities among women with or at risk of infection with human immunodeficiency virus. Am J Obstet Gynecol 2001;184:584–90.
16. The 1988 Bethesda system for reporting cervical/vaginal cytological diagnoses. National Cancer Institute Workshop. JAMA 1989;262:931–4.
17. Altman DG, Andersen PK. Calculating the number needed to treat for trials where the outcome is time to an event. BMJ 1999;319:1492–5.
18. Schuman P, Ohmit SE, Klein RS, Duerr A, Cu-Uvin S, Jamieson DJ, et al. Longitudinal study of cervical squamous intraepithelial lesions in human immunodeficiency virus (HIV)-seropositive and at-risk HIV-seronegative women. J Infect Dis 2003;188:128–36.
19. Heard I, Potard V, Costagliola D. Limited impact of immunosuppression and HAART on the incidence of cervical squamous intraepithelial lesions in HIV-positive women. Antivir Ther 2006;11:1091–6.
20. Howard AA, Arnsten JH, Lo Y, Vlahov D, Rich JD, Schuman P, et al. A prospective study of adherence and viral load in a large multi-center cohort of HIV-infected women. AIDS 2002;16:2175–82.