Howard Minkoff; Linda Ahdieha; L. Stewart Massadb; Kathryn Anastosc; D. Heather Wattsd; Sandra Melnicke; Laila Muderspachf; Robert Burkg; Joel Palefskyh
 
From the departments of Obstetrics and Gynecology, Maimonides Medical Center and State University of New York Health Sciences at Brooklyn, New York, the aDivision of Epidemiology, Johns Hopkins School of Public Health, Baltimore, Maryland, bDepartment of Obstetrics and Gynecology, Cook County Hospital and Rush Medical College, Chicago, Illinois, cDepartment of Medicine, Montefiore Medical Center and Lincoln Medical and Mental Health Center, Bronx, New York, the dNational Institute of Child Health and Human Development, the eNational Cancer Institute, Bethesda, Maryland, the fDepartment of Obstetrics and Gynecology, University of Southern California, the gDepartment of Pediatrics, Microbiology & Immunology and Social Medicine, Albert Einstein College of Medicine, and the hDepartment of Medicine, University of California at San Francisco, San Francisco, California, USA.
 
AIDS 2001;15:2157-2164
 
Objective: Cervical intraepithelial neoplasia (CIN), a common condition among HIV-infected women, has been linked to HIV load and immune status. Highly active antiretroviral therapy (HAART) improves immunologic and virologic status. This study was undertaken to determine the relationship between HAART use and CIN.
 
Design: Cohort study. The Women's Interagency HIV Study (WIHS) in five cities in the USA (Bronx/Manhattan, New York; Brooklyn, New York; Chicago, Illinois; Los Angeles, California; San Francisco Bay area, California; Washington, District of Columbia).
 
Methods: HIV-infected women were followed every 6 months with Papanicolaou smears and cervicovaginal lavage for human papillomavirus (HPV) DNA testing. To characterize exposures that changed over time and to capture the dynamic nature of cytologic changes, Papanicolaou smear findings from each participant's consecutive visits were defined as a pair. We determined the proportion of all pairs that exhibited either regression or progression, according to HAART exposure, HPV results and Papanicolaou smear status. As participants could contribute multiple pairs, inferences were based on robust methods to adjust for correlated observations.
 
Results: Women with persistent HPV infection were more likely to have progression of their lesions. After adjustment for CD4 cell count and Papanicolaou smear status, women on HAART were 40% (95% confidence interval, 4­81%) more likely to demonstrate regression and less likely (odds ratio, 0.68; 95% confidence interval, 0.52­0.88) to demonstrate progression
 
Conclusions: HAART altered the course of HPV disease in HIV-infected women, reducing progression and increasing regression. As HPV disease is a common sex-specific manifestation of HIV disease this effect of HAART would be a major additional benefit from this modality of therapy.
 
editorial note: so this raises a question regarding when is the best time for a woman to begin HIV therapy, does it differ from men. The risk for HPV related disease appears to me to be a consideration. CD4 and viral load were important to improved outcome in this study.
 

---

Introduction
 
With the introduction of more effective antiretroviral therapy, the spectrum of disease in the AIDS epidemic has been shifting. It is projected that 20­40% of HIV-infected individuals will be diagnosed with a malignancy [1]. One malignancy that is unique to women is cervical cancer, the end stage of human papillomavirus (HPV)-associated cervical intraepithelial lesions. Numerous studies have shown that HIV-infected women, particularly those with advanced disease, have a greatly increased risk for the precursors to cervical cancer, both those measured by histology [cervical intraepithelial neoplasia (CIN)] and by cytology [squamous intraepithelial lesions (SIL)] [2­18]. Recent data have suggested that as many as one in five HIV-infected women with no evidence of cervical disease will develop SIL within 3 years. These data suggest a need to assess determinants of cytologic progression and to develop interventions. The effect of HAART on HIV-associated malignancies, particularly cervical lesions, has not been clearly delineated.
 
HAART has the potential to influence the relationship between HIV and cervical cancer precursors in two contrasting ways. First, by prolonging lives it may lengthen exposure to HPV, thereby allowing the accumulation of genetic somatic mutations that increase the likelihood of cervical disease. Conversely, because HAART can decrease HIV viral load, and can partially restore immune competence, HAART may mitigate the effect of HIV on the course of HPV disease. To date, these possibilities have been assessed only in small case series [23,24] and the implications of antiretroviral therapy for cervical neoplasia are unclear. Therefore, to determine if HAART can influence the course of HPV disease the following study was undertaken.
 
The Women's Interagency HIV Study (WIHS) is a prospective study of the natural and treated history of HIV infection in women that enrolled 2059 HIV-seroprevalent participants at six clinical consortia (Bronx, New York; Brooklyn, New York; Chicago; Los Angeles; San Francisco; Washington) between October 1994 and November 1995. The recruitment methods and data collection procedures for WIHS have been described previously [25]. In brief, participants undergo visits every 6 months that include an interview-administered questionnaire, a physical examination, and the collection of plasma, Papanicolaou smears, and cervicovaginal lavages for HPV testing.
 
The primary goal of our analysis was to identify whether exposure to HAART impacted rates of short-term (6 month) cytologic regression and progression.
 
The study population for our analysis of the impact of HAART on short-term cytological regression and progression analysis was composed of 1779 HIV-seropositive participants who had at least one study visit after 1 October 1995 (the approximate date at which HAART became available) and were tested for HPV infection during the 1 year following their enrollment. The study's primary analyses were limited to those individuals who had an oncognenic HPV infection (defined above). In addition, we compared short-term cytological progression and regression rates between those with and without an oncogenic HPV infection. Those with at least one oncogenic infection were further characterized as having 1/3, 2/3 or 3/3 oncogenic HPV infections.
 
Results
 
There were 741 women (41.6% of participants) with at least one oncogenic HPV at visits one through three. Among women who had HPV tests at all three baseline visits and who had at least one positive test (n = 507), oncogenic HPV was detected at one, two, and three visits in 189 (37.3%), 145 (28.6%), and 173 (34.2%) women respectively.
 
The group with oncogenic HPV was younger and more likely to smoke, to report HIV treatment, to have SIL at baseline, to be younger at first coitus, and to have more sexual partners. This group also had lower CD4 cell counts and higher HIV viral loads. None of these variables were found to be associated with cytologic progression or regression. In addition, ethnicity and report of a prior sexually transmitted disease were not associated with these outcomes.
 
Regression was more frequent (P < 0.001) and progression less common (P < 0.001) among women with no oncogenic HPV. Further, persistent HPV infection significantly decreased regression rates and increased progression rates. Among those with three measurements, regression rates among those with 1/3, 2/3, and 3/3 positive oncogenic HPV infections were 39.3%, 30.7% and 24.7% respectively; overall progression rates were 16.2%, 23.6% and 24.6% respectively (P < 0.001).
 
Cytological regression was also more likely among pairs characterized by HIV RNA levels < 4000 copies/ml (39.4%) compared to pairs characterized by RNA levels of 4000­20 000 copies/ml (27.7%), 20 000­100 000 copies/ml (29.9%), and > 100 000 copies/ml (22.7%) (P < 0.001).
 
Cytological regression was also more likely among pairs characterized by CD4 cell counts > 500 compared to pairs characterized by CD4 cell counts of 200­500 and < 200 (21.6%) (P < 0.001).
 
Cytological progression was less likely among pairs of measurements where the HIV RNA level at the index visit was < 4000 copies/ml (16.4%) compared to RNA levels of 4000­20 000 copies/ml (19.8%), 20 000­100 000 copies/ml (23.4%) and > 100 000/ml (27.5%) (P < 0.001). There was no association noted between CD4 cell count and progression; specifically, 17.9% of pairs with CD4 cell counts > 500 progressed, as compared to 21.8% of pairs with CD4 cell counts 200­500 ? and 21.2% of pairs with < 200.
 
Overall, pairs of measurements were 1.4 times [95% confidence interval (CI), 1.04­1.82] more likely to demonstrate regression on HAART. A beneficial effect of HAART was observed on progression. Pairs of visits on HAART demonstrated a 0.68 times (95% CI, 0.52­0.8) likelihood of progression of disease, i.e., pairs contributed off HAART were 1.5 times (95% CI, 1.3­1.9) more likely to progress as compared to pairs contributed on HAART.
 
In this large cohort of HIV-infected women with high-risk HPV DNA, we found that the use of HAART was associated with significantly increased cytological regression and decreased cytological progression, after controlling for stage of HIV disease and severity of cytological disease. We also confirmed previous reports that found high rates of HPV carriage, which were often persistent, among HIV-infected women. Ho reported that persistence of infection with HPV was related to the risk of persistent cervical dysplasia using similar methods [36]. We found that among HIV-infected women persistent carriage of HPV and high viral load were associated with cytologic progression and that low viral load and high CD4 cell count were associated with regression.
 
Relatively little attention has been paid to the relationship of HAART to malignancies in general or to cervical disease in particular. Since the introduction of HAART, some studies have shown stable levels of non-Hodgkin's lymphoma and other cancers, and declining rates of Kaposi's sarcoma and primary brain lymphoma among HIV-infected individuals [22,40]. Individual cases in which human herpes virus type 8 (HHV-8) was cleared from peripheral blood mononuclear cells following initiation of antiretroviral therapy have also been reported [41]. Conversely, at least one study suggested that few anal SIL regress in response to HAART [42].
 
In our cohort, the regression rate among women with high-grade and low-grade lesions who did not receive HAART were similar to, or greater than, those seen among women receiving protease inhibitors in the Heard study. However, when the rates of progression and regression among women with high-risk HPV who were on HAART were contrasted with changes among similar women not on HAAR T, the beneficial effect of antiretroviral therapy became clear. These results are particularly striking as women in both groups were comparably at risk in terms of the presence of oncogenic HPV. These results were also supported by multivariate analysis that included markers of HIV disease stage as well as other factors linked to changes in Papanicolaou smears over time.
 
The association between HIV disease and both HPV infections and SIL has been widely noted [6].Sun detected HPV at an initial examination in 56% of HIV-seropositive women and 31% of HIV-seronegative women [43]. Minkoff reported a prevalence of oncogenic HPV in 32.5% of HIV-infected and 17.0% of HIV-uninfected women [44]. Palefsky noted a correlation between HPV detection and low CD4 cell counts and high HIV viral loads in an earlier report of baseline data from this cohort [33]. Other authors have reported a similar inverse association between immune status and HPV detection rates [15,46­48].
 
Several investigators have demonstrated that in addition to an increased frequency of HPV infection, there is an increased frequency and severity of HPV-associated cervical lesions among HIV-infected women. Massad found that rates of SIL were five times greater among HIV-infected than uninfected women [6]. Mandelblatt, in a review of 15 studies comparing the association between HIV, HPV and CIN found that the odds ratio for CIN was eight times greater for HPV-infected women and that there was an interaction between HPV and HIV (P = 0.001) [5].
 
The clinical course of CIN has been noted to be marked by more progression and less regression among HIV-infected women [3,48]. Ellerbrock has recently reported that among 328 HIV-infected women without SIL at baseline 20% developed SIL during approximately 30 months of follow-up. That rate of progression was significantly greater than among uninfected women [16]. Massad, in an analysis of women in this cohort has shown that progression of SIL was linked to HIV serostatus and to viral load and CD4 cell count [49]. As noted, we have also found that progression and regression may be linked to immune status and viral load.
 
One possible explanation for the linkage of both serostatus and immune status to malignancies may be that HIV-associated products up-regulate cell proliferation [50]. Reduction of viral load and restoration of host immunity thus might retard the progression of these neoplasias.
 
It is impossible, however, to be certain that the association between the use of HAART and a less aggressive course of HPV disease is mediated simply by a reduced HIV viral load that results in less up-regulation of HPV. That HAART could have a direct antiviral effect on HPV has been suggested as a possibile explanation of the lack of correlation between decreases in HIV viral load and regression of cytologic abnormalities in women on HAART.
 
A similar mechanism has been posited as an explanation for the clearance of HHV-8 in a woman on HAART [41]. The resurgent immune system of the treated woman might also be better able to Œcontrol' HPV infections. Burk has shown a reduced rate of HPV as women age, even when controlling for the number of lifetime sexual partners [51], suggesting that there might be Œacquired immunity to HPV from past exposure' [52]. This mechanism has been proposed to explain the loss of cytomegalovirus (CMV) viremia following HAART therapy, even in the absence of specific anti-CMV therapy [53]. Similarly, Ahdieh, who reported that the likelihood of chronic carriage of HPV is linked to immune status, has suggested that Œthe reversal of immunosuppression possible in the current era of potent antiretroviral therapy may reduce HPV persistence and consequently the increased risk for cervical neoplasia' [54].
 
If these findings are confirmed, it would suggest that HAART may reduce the need for colposcopy, biopsy and therapeutic interventions among HIV-infected women, much as it has been reported by others to reduce the need for chemoprophylaxis against opportunistic infections [56]. Furthermore, it underscores the importance of full access to HAART by those for whom it is indicated.