Effect of Highly Active Antiretroviral Therapy on the Natural History of Anal Squamous Intraepithelial Lesions and Anal Human Papillomavirus Infection

The effect of highly active antiretroviral therapy (HAART) on the natural history of anal squamous intraepithelial lesions (ASIL)-the likely anal cancer precursor-and anal human papillomavirus (HPV) infection is unknown. ASIL severity and level of anal HPV DNA were evaluated among HIV-positive men who have sex with men (MSM) for at least 6 months before initiation of HAART. The results were compared with those from a 6-month period after initiation of HAART. Anal swabs for cytology and HPV studies were obtained, followed by high-resolution anoscopy and biopsy. Among men whose most severe pre-HAART diagnosis was atypical squamous cells of undetermined significance or low-grade ASIL, 18% (confidence interval [CI], 6-31%, 7 of 38) progressed and 21% (CI, 8-34%, 8 of 38) regressed 6 months after starting HAART. Seventeen percent (CI, 0-38%, 2 of 12) of study subjects who began with a normal diagnosis developed ASIL. Only 4% (CI, 0-10%, 1 of 28) of study subjects with high-grade ASIL regressed to normal. There was no reduction in the proportion of study subjects who tested positive for HPV DNA or HPV DNA levels after HAART initiation. The ASIL and HPV data were similar to those of the pre-HAART comparison period. These results indicate that HAART has little effect on either ASIL or HPV in the first 6 months after HAART initiation.

The relationship between cervical human papillomavirus (HPV) infection, cervical squamous intraepithelial lesions (CSILs), and invasive cervical cancer in women has been well established ^(1-6). In the last decade, an emerging body of evidence has shown that a parallel clinical phenomenon exists in the anal canal. Several studies have shown that men who have sex with men (MSM) are at high-risk for anal HPV infection ^(7-9) and anal squamous intraepithelial lesions (ASILs) ^(10-13). These studies also show that HIV-positive MSM are at even higher risk of anal HPV infection and ASILs than HIV-negative MSM, with a lower CD4 ⁺ lymphocyte count a finding that is associated with increased risk.

The clinical significance of these observations has not been clear because the relationship between ASILs and invasive anal cancer has not been adequately studied. Nevertheless, an indirect body of evidence suggests that ASILs represent a true anal cancer precursor lesion, particularly at the severely dysplastic end of the disease spectrum. Consistent with the high prevalence and incidence of ASIL in MSM, incidence of anal cancer was estimated to be as high as 35 in 100,000 before the onset of the HIV epidemic ⁽¹⁴⁾. This figure exceeds the current incidence of cervical cancer in the general population of women among whom it is estimated to be approximately 8 in 100,000. Moreover, incidence of anal cancer in HIV-positive MSM is estimated to be approximately twice that of HIV-negative MSM ⁽¹⁵⁾ and HIV-positive men were recently estimated to have a relative risk of 38 for developing anal cancer compared with findings in the general population ⁽¹⁶⁾. Using data from natural history studies of ASIL, anal cytology screening to detect and treat ASIL to prevent anal cancer has been projected to be cost effective for both HIV-positive and HIV-negative MSM ^(17,18).

This high incidence of ASIL and anal cancer is not restricted to MSM. HIV-positive women have been shown to be at increased risk of anal HPV infection ⁽¹⁹⁾ and ASIL ⁽²⁰⁾. As in HIV-positive men, a lower CD4 ⁺ lymphocyte count was associated with increased risk of both anal HPV infection and ASIL. Consistent with these data, the relative risk of anal cancer among HIV-positive women compared with findings in the general population was recently estimated to be 6.8 ⁽¹⁶⁾.

Given the relationship between HIV-associated immunosuppression and increased risk of anal HPV infection and ASIL, the introduction of highly active antiretroviral therapy (HAART) presents an interesting opportunity to examine this relationship, because HAART has been associated with lower HIV viral load, increased CD4 ⁺ lymphocyte count, and a lower incidence of opportunistic infections in many patients. The relationship between HAART and changes in the natural history of either cervical or anal HPV infection and squamous intraepithelial lesions is not adequately understood, and studies evaluating this relationship have yielded inconsistent results. Only one study has been published on the relationship between HAART and regression of CSIL ⁽²¹⁾. In that study, a statistically significant relationship was found between regression of CSIL and HAART. However, most CSIL lesions in that study did not regress, and 10 of 13 women with cervical high-grade squamous intraepithelial lesions (HSILs) showed persistent high-grade disease. There was no change in the HPV status of the women after initiation of HAART. No data on the relationship between HAART and anal HPV infection or ASIL have been reported. Therefore, the aim of this study was to characterize the relationship between treatment with HAART, anal HPV infection, and ASIL. Because we followed most of the MSM in our cohort studies for several years before HAART became available, we had the unique opportunity to compare the natural history of ASIL progression in the same study subjects before and after initiation of HAART.

METHODS

The study participants in this analysis were HIV-positive MSM who enrolled in two of our anal neoplasia cohort studies between November 1991 and March 1994 and who were treated with HAART continuously for at least 3 months while participating in these studies ^(13,22). To assess the effect of HAART on the natural history of ASIL, we studied anal HPV infection and ASIL in 98 HIV-positive MSM followed for at least 6 months before initiation of HAART, and the results were compared with data collected from the same study subjects observed during follow-up for 6 months after initiation of HAART. Clinical, laboratory, and interview methods are described in detail elsewhere ^(11-13).

Briefly, at baseline and every 6 months thereafter, participants in the prospective cohort study had a clinical examination that included anal cytology, a swab for anal HPV testing, and high-resolution anoscopy (HRA) with biopsy of visible anal lesions unless there were medical contraindications. All studies were performed with informed consent and the approval of the UCSF Committee on Human Research. Participants had blood tests to measure CD4 ⁺ lymphocyte counts and were interviewed to determine medications taken to treat HIV infection. The Bethesda System criteria to evaluate cervical cytology ⁽²³⁾ were used to classify anal cytology as normal, atypical squamous cells of undetermined significance (ASCUS), low-grade squamous intraepithelial lesions (LSIL), or HSIL. Anal histology was classified as normal, atypical, LSIL, or HSIL. Men with anal histologic or cytologic results that indicated LSIL or HSIL were observed more closely every 3 months with anal cytology, HRA, and a biopsy when clinically indicated. Participants with histologically confirmed HSIL were referred to physicians outside the study for treatment. Regardless of whether men received treatment of their HSIL, they were observed during follow-up for development of future lesions.

CD4 ⁺ lymphocyte counts, CD8 ⁺ lymphocyte counts, and the CD4 ⁺ /CD8 ⁺ ratio were determined by standardized two- or three-color fluorescence methods. Hybrid capture (HC) and polymerase chain reaction were used to detect HPV DNA in the anal swab as described previously ⁽⁹⁾. HC (Digene Corporation, Beltsville, MD, U.S.A.) was performed according to the manufacturer's recommendations. HC was performed on anal swab samples taken every 6 months, testing for two HPV groups using two mixtures of probes-a low-risk group composed of nononcogenic HPV types and a high-risk group composed of oncogenic HPV types. The results of HC were expressed as a relative light unit (RLU) ratio, determined by dividing the chemiluminescent signal of the test sample by that obtained with a control sample containing 10 pg/ml of HPV 11 DNA (low-risk group) or 10 pg/ml of HPV 16 DNA (high-risk group). The magnitude of the RLU ratio increases with increasing quantity of HPV in the specimen.

A participant was considered to be receiving HAART if he was taking a protease inhibitor (PI) in combination with one or more antiretroviral medications. Of these respondents, 88% used two antiretroviral drugs in combination with a PI and 12% used only one antiretroviral drug with a PI. Thirty percent of subjects used a combination of a PI, lamivudine (3TC), and zidovudine (ZDV), whereas 44% used a combination of a PI, 3TC, and stavudine (d4T). An additional 14% used a PI in combination with two of the following: ZDV, 3TC, didanosine (ddI), zalcitabine (ddC), d4T, delavirdine, nevirapine, or efavirenz (DMP266).

The date a study subject started HAART was defined as the date he began to use any PI continuously for at least 3 months. Study subjects were considered to be continuously taking PIs if they did not change drugs, or if they switched drugs, but there was no time lapse in the use of PIs. The most severe ASIL diagnosis on either cytology or histology during the year before starting HAART was used as the baseline. All data from the visit before the HAART start date and the visit closest to the date of the visit 12 months after the start date (between 6 months and 15 months) were used to determine this diagnosis. If there were no data between 6 and 15 months before starting HAART, the diagnosis was considered missing. The visit closest to 6 months (between 4 and 9 months) after starting HAART was used for the post-HAART diagnosis. Study subjects were included in this study if they had data from both the pre-HAART and post-HAART periods. Some of these study subjects were excluded from specific analyses as described later in this article.

Data collected earlier in the anal neoplasia studies from these 98 participants were used to provide a benchmark for progression and regression of anal lesions over a similar time period before the availability of HAART. For these data, the reference date was defined as 1 year before the actual date a participant started HAART. The prereference-date clinical diagnosis was determined from the visit closest to the 12 months before the reference date. If there were no data between 6 and 15 months before starting HAART, the clinical diagnosis was recorded as missing. The visit closest to 6 months (but between 4 and 9 months) after the reference date was used as the postreference diagnosis date. Results from the reference comparison period are provided to show the magnitude of changes found in a similar group of study subjects who were not receiving HAART. Because of the small number of participants involved and incomplete data, no formal statistical pre- and post-HAART comparisons of anal disease progression were done.

The effectiveness of HAART in individuals was examined by comparing differences in absolute CD4 ⁺ counts, CD8 ⁺ counts, and the CD4 ⁺ /CD8 ⁺ ratio before and after treatment with HAART. Cross-tabulations of anal lesion status before treatment with HAART versus anal lesion status after treatment with HAART were computed to evaluate the effect of HAART on anal disease. Analyses were done on a composite diagnosis defined as the most severe of the cytology and histology result when both were available. If no histologic results were available, the cytologic results were used. If the cytologic result was normal but anal lesions were seen but not subject to biopsy, the diagnosis was considered indeterminate, and the visit was excluded from the analysis. If a participant was treated for HSIL during the time interval involved in a particular cross-tabulation, his data was excluded from that comparison. For example, if he was treated for HSIL 3 months after starting HAART, his data would be included in cross-tabulations of anal disease during the reference period, but not cross-tabulations involving HAART. Analysis of variance was used to compare CD4 ⁺ lymphocyte counts stratified by diagnosis before treatment with HAART among those whose disease progressed, remained the same, or regressed. Confidence intervals (CIs) for proportions were computed from the binomial distribution for numbers that were ≤10 and from the normal approximation to the binomial distribution for numbers >10.

Samples taken 2 to 8 months before HAART and 4 to 10 months after HAART were used to evaluate the effect of HAART on HPV infection. A test result was considered positive for the HPV probe group if the RLU ratio was >0.65. Cross-tabulations of the probe groups detected before and after HAART were computed to evaluate the effect of HAART on the persistence of HPV infection. In addition, the paired t- test was used to compare the RLU levels before HAART with the RLU levels after HAART of the low- and high-risk group probes.

RESULTS

Of the 202 HIV-positive participants who were seen after January 1, 1995, 116 took PIs for at least 3 months. Four were excluded from this analysis because they took PIs intermittently over an extended period, 4 because they reported their first use of PIs as a part of a clinical trial so that it was unclear when they began taking them, 1 because he took PIs during several short, disjointed periods of time, and 9 because there were inadequate data during the analysis period. Of the 98 participants included in the analysis, 3 took PIs briefly (1-3 weeks) several months before starting continuous use for over 3 months, and 10 switched PIs 1 or more times after starting on PIs but took PIs continuously after starting the first drug.

The 98 men remaining in these analyses were 92% white, non-Hispanic, 4% Hispanic, and 4% other or unknown. Of these participants, 8% had 12 or fewer years of education, 47% reported 13 to 16 years and 44% had >16 years of education. Their mean age at HAART initiation was 46 years (range, 30-69) and they had been observed for an average of 48 months (range, 18-72) before treatment with HAART. All but 1 study subject tested positive for HPV using polymerase chain reaction (PCR) by the time they initiated HAART. By the time they started HAART, 64% (63 of 98) of study subjects had biopsy-confirmed HSIL, and 30% (29 of 98) had had their HSIL ablated with electrocautery during surgery at least once. In addition, 14 men were treated for HSIL within 1 year of initiating HAART, and 11 were treated during the pre-HAART comparison period.

CD4 ⁺ data were available before and after HAART for 81 participants, whereas CD8 ⁺ data were available before and after HAART for 81 participants. The participants' mean CD4 ⁺ lymphocyte count, CD8 ⁺ lymphocyte count, and CD4 ⁺ /CD8 ⁺ ratio increased in all strata after taking HAART (Table 1). The CD4 ⁺ lymphocyte count 6 months after starting HAART was higher than their last CD4 ⁺ count before starting HAART for 88% (78 of 89) of participants and 39% (35 of 89) of participants had an increase of more than 100/mm ³ . The average CD8 ⁺ lymphocyte count increased for those with initial CD4 ⁺ counts <200/mm ³ , remained the same for men with initial CD4 ⁺ counts between 200/mm ³ and 500/mm ³ , and decreased for those with initial CD4 ⁺ counts >500/mm ³ . The CD4 ⁺ /CD8 ⁺ ratio increased in 83% of study subjects and decreased in 17%. In contrast, during 6 months of follow-up during the pre-HAART comparison period, the CD4 ⁺ lymphocyte count decreased in 67% of the men, and the CD4 ⁺ /CD8 ⁺ ratio decreased in 64%.

Table 1

Data on anal lesion progression during the 6-month reference period in the year before initiation of HAART are shown in Table 2. Among these participants, 11 men were treated for HSIL during the period covered by this cross-tabulation, and 8 had a missing or indeterminate diagnosis. Among those whose most severe composite diagnosis during the year before HAART was ASCUS or LSIL, 27% (CI, 13%-40%; 11 of 41) progressed and 17% (CI, 6%-29%; 7 of 41) regressed at 6 months. Moreover, 38% (CI, 14-61%; 6 of 16) of men who began with a normal anal diagnosis progressed during the pre-HAART reference period, and 50% (CI, 29-71%; 11 of 22) of those with HSIL regressed.

Table 2

The most severe composite anal diagnosis during the year before starting HAART was compared with the composite anal diagnosis 6 months after starting HAART (Table 2). Of the 98 participants, 14 were excluded from this cross-tabulation because they were treated for HSIL during the period from 1 year before HAART and 6 months after HAART, and 6 men had a missing or indeterminate diagnosis. Among the 38 participants whose most severe composite diagnosis during the year before HAART was ASCUS or LSIL, 18% (CI, 6-31%; 7 of 38) progressed and 21% (CI, 8-34%; 8 of 38) regressed at 6 months after starting HAART, whereas 17% (CI, 0-38%; 2 of 12) who began with a normal anal diagnosis progressed at 6 months after starting HAART. Of those whose most severe diagnosis was HSIL in the pre-HAART period, 43% (CI, 25%-61%; 12 of 28) regressed and the remainder continued to have HSIL. Of the 28 participants who regressed, 27 continued to have ASCUS or LSIL, and only 1 regressed to normal. Similar analyses were performed using only those study subjects whose CD4 ⁺ /CD8 ⁺ ratio increased after taking HAART. Results were similar to those shown in Table 2. There was no association between anal lesion response to HAART and magnitude of increase in absolute CD4 ⁺ lymphocyte count (data not shown).

The relationship between the effect of mean CD4 ⁺ lymphocyte count at the time of HAART initiation and the effect of HAART on the most severe composite diagnosis is shown in Table 3. For each category of lesion before HAART initiation, there was a trend toward increased rates of disease regression after 6 months of HAART and higher mean CD4 ⁺ lymphocyte count at the time of HAART initiation. However, the differences were not statistically significant for any of these categories.

Table 3

After 6 months of HAART, anal HPV was detected by HC in 80% (CI, 35-100%; 4 of 5) of men with no detectable anal HPV infection at the pre-HAART HPV test. The low-risk HPV group was detected after HAART in 31% (CI, 13-49%; 8 of 26) of men with no detectable low-risk group HPV at the pre-HAART test, and the high-risk HPV group was detected after HAART in 50% (CI, 19%-81%; 5 of 10) of men in whom it was not detected at the pre-HAART test. Neither low-risk nor high-risk HPV was detected after HAART in 3% (CI, 0-6%; 2 of 79) of men with detectable anal HPV DNA at the pre-HAART test. The low-risk HPV group was not detected in 10% (CI, 3-18%; 6 of 58) of the men with low-risk group HPV at the pre-HAART test, and the high-risk HPV group was not detected in 9% (CI, 3%-16%; 7 of 74) of the men with high-risk HPV in the pre-HAART test.

Anal HPV was detected in 100% (CI, 65%-100%; 7 of 7) of men with no detectable anal HPV DNA at the test before the reference date (Table 4). Low-risk group HPV DNA was detected in 24% (CI, 9%-40%; 7 of 29) of men with no low-risk group HPV at the test before the reference date. High-risk group HPV DNA was detected in 75% (CI: 51%-100%; 9 of 12) of the men with no high-risk group HPV DNA at the earlier test. During the reference period, HPV was not detected in 5% (CI, 0-10%; 4 of 77) of the men with detectable anal HPV DNA at the test before the reference date. Low-risk group HPV DNA was not detected in 5% (CI: 0-12%; 3 of 55) of the men with low-risk group HPV DNA in the earlier test. High-risk group HPV DNA was not detected in 8% (CI, 2%-15%, 6 of 72) of the men with high-risk group HPV DNA in the earlier test.

Table 4

There was little change in the mean RLU ratio of the low-risk group HPV DNA after HAART in the men who were positive for the low-risk group HPV DNA in the test before HAART (Table 5). However, the RLU ratio for the high-risk group HPV DNA increased by 32 units (p = .005). During the comparison reference period, the RLU ratios for both the low-risk and high-risk HPV DNA groups rose slightly, but neither change was statistically significant.

Table 5

DISCUSSION

A clear relationship exists between HIV-related immunosuppression and risk of HPV-associated anogenital disease prevalence and progression ^{(2-4,10,12,20)}. However, the effect of HAART on the natural history of anogenital disease has not been well characterized. All of the existing data are from studies of cervical squamous intraepithelial lesions in HIV-positive women. In our work, we have reported the effect of HAART on the natural history of ASILs and anal HPV infection in men. In addition, this study is unique because the natural history of ASIL after HAART initiation was compared with a period before HAART initiation in the same population. As expected, our data show that HAART resulted in a considerable increase in CD4 ⁺ lymphocyte count. However, there was little difference in the natural history of ASIL 6 months after initiating HAART compared with a similar reference period before HAART initiation. Similar rates of disease progression were found and there was little regression to normal in either study period. These data indicate a limited effect of HAART on the natural history of ASIL. However, it should be noted that the comparison data differ from the HAART data because the participants were 1 year younger during the comparison period and had had their HIV infection for a shorter period of time. Their follow-up period for ASIL was 1 year shorter and they had slightly less severe anal disease. Consequently, our data may have underestimated the effect of HAART when compared with data from the comparison period. In addition, this study examined the effect of 6 months of HAART so that further follow-up is needed to determine whether longer courses of HAART might have a more beneficial effect on the natural history of ASIL.

Although the effect of HAART on morbidity and mortality of HIV-positive people is overwhelmingly beneficial, this study suggests that treatment with HAART may be of little value with respect to ASIL and also raises concerns about increased risk of anal cancer. If HAART has little or no impact on the natural history of SILs, then paradoxically the risk of anogenital cancer may increase in the HAART era. Progression from HSIL to invasive cancer may take many years, and before the availability of HAART, individuals with HSIL may have died of other HIV-associated complications before having sufficient time to progress to cancer. If HAART does not lead to significant ASIL regression, then the increased longevity may provide individuals with anal HSIL sufficient time to progress to cancer if left untreated. This is particularly worrisome in the anal canal because there is no routine screening and treatment program for anal HSIL.

Consistent with the effect of HAART on the natural history of ASIL, HAART had little effect on detection of anal HPV DNA using the HC method. There was little difference in the rate of detection of the low- and high-risk HPV DNA groups after initiation of HAART. In addition, HAART had no effect on the levels of HPV DNA as measured by the RLU ratio, and, interestingly, the quantity of HPV DNA in the oncogenic high-risk group increased 6 months after initiation of HAART.

One observation of interest was a trend toward a more beneficial effect of HAART among those with higher CD4 ⁺ lymphocyte counts at the time of initiation of HAART. These data were not statistically significant, in part because of the relatively small number of study participants. These men had relatively low CD4 ⁺ lymphocyte counts at the time of HAART initiation, and the effect of initiating HAART at higher CD4 ⁺ lymphocyte counts was not examined. However, if confirmed, our data suggest that HAART may have a more beneficial effect on the natural history of ASIL if the immune system is better preserved, as measured by CD4 ⁺ lymphocyte count, and that reconstitution of the immune response to HPV may be playing a role in this beneficial effect. If confirmed, our data also suggest that early institution of HAART may be beneficial in improving the course of ASIL, although the optimal CD4 ⁺ lymphocyte count at which HAART should be initiated is unknown.

In summary, our study shows that 6 months' treatment with HAART has relatively little impact on the natural history of ASIL and anal HPV infection. In the absence of screening and treatment of anal HSIL, the longer survival owing to HAART may increase the risk of progression to cancer because most of these lesions persist after HAART initiation. Confirmation of these observations is needed with a longer follow-up period on HAART as well as data on the effect of HAART failure.

Acknowledgments:

This research was financially supported by grant NCI R01CA54053 from the U.S. National Cancer Institute. These studies were carried out in the General Clinical Research Center, University of California, San Francisco with funds provided by the Division of Research Resources 5 M01-RR-00079, U.S. Public Health Service.