In data obtained before the HIV epidemic, one of the most important risk factors for anal cancer in men was a history of receptive anal intercourse (1,2). Consistent with this, the incidence of anal cancer in homosexual or bisexual men (hereafter referred to as homosexual men) was higher than that of heterosexual men. Further, the estimated incidence of anal cancer in homosexual men at approximately 35/100,000 (1) is higher than that of cervical cancer in women. Data on the relation between HIV positivity and the incidence of anal cancer in homosexual men are conflicting. When linked to HIV registry databases, data from the San Francisco Bay Area Surveillance, Epidemiology, and End Results (SEER) database indicate no substantial increased risk of anal cancer in HIV-positive men (3). In contrast, data from seven HIV registries around the United States suggest that the risk of anal cancer increases with increasing temporal proximity to an AIDS diagnosis (4).
Both the cervical and the anal epithelia may contain changes known as atypical squamous cells of undetermined significance (ASCUS) and squamous intraepithelial lesions (SIL). In the cervix, high-grade SIL (HSIL) is known to be the precursor lesion to invasive cervical cancer. Given the many similarities between cervical SIL (CSIL)/cervical cancer and anal SIL (ASIL)/anal cancer, it is likely that ASIL and anal HSIL, in particular, represent true anal cancer precursors. We showed that the prevalence of ASIL at baseline for this study was high in HIV-positive men (36%), whereas that of HIV-negative men was 7% (5).
Two of the most important risk factors for anal human papillomavirus (HPV) infection are HIV infection and receptive anal intercourse (5). Anal HPV infection has been shown in several studies to be more common in HIV-positive homosexual men than in HIV-negative homosexual men (6-10), as has ASIL (5,6,8,9,11). However, almost all these data were obtained from crosssectional samples. Consequently, little is known about the incidence of anal lesions or the natural history of anal ASCUS and LSIL, and risk factors for anal disease incidence and progression are poorly understood. The aims of this study were to describe the incidence and progression of anal ASCUS and ASIL in HIV-positive and HIV-negative homosexual men over a 2-year follow-up period and to characterize risk factors for disease progression in these groups.
This study was conducted with the approval of the University of California-San Francisco Committee on Human Research. A total of 346 HIV-positive and 262 HIV-negative men were enrolled at baseline and their demographic characteristics are described elsewhere (5). The subjects have been observed to date for an average of 3 years. For this analysis, we examined the first 2 years of follow-up to maximize the number of subjects available for analysis.
The criteria to interpret anal cytology and histology are described elsewhere (5), as are details of the clinical examination protocol, methods to test for HPV (12) and HIV, and measurement of CD4 levels. HIV-positive subjects with normal cytology and no lesions seen at baseline anoscopy underwent anal examinations every 6 months. HIV-negative subjects with normal cytology and no lesions seen at baseline anoscopy underwent anal examinations every 12 months to maximize subject retention. Both HIV-positive and HIV-negative subjects who had ASIL on cytology or histology at baseline were observed at 3-month intervals, as were subjects who had no lesions at baseline, but who had one or more lesions at any follow-up visit. All subjects with histologically confirmed HSIL were referred out of the study for treatment.
Subjects were interviewed at baseline and at interval visits about medical history, including substance use, sexually transmitted diseases, sexual practices, anal conditions (e.g., fissures, fistulas), and HIV-related symptoms, diseases, and treatments. Data collected from follow-up questionnaires were used to evaluate the association between progression of anal lesions and potential risk factors. A subject was considered to be positive for the attribute if he responded affirmatively on any of the follow-up questionnaires, and negative if he responded negatively on all of the follow-up questionnaires.
Subjects were considered to have relevant data at 2 years if they were seen within 18 to 30 months from their baseline visit. All data up to the subject's 2-year visit were included. The subject's status at 2 years was recorded as the highest grade of anal disease identified either by cytology or histology during the 2-year interval. For analysis, progression was defined as progression from normal to any grade of anal lesion or from a lower grade of anal lesion to a higher grade using both cytologic and histologic results. If biopsy results were available, the highest grade of lesion from cytology and histology was used as the composite diagnosis. If no lesions were present and cytology was normal, the composite diagnosis was considered to be normal. Cytology was not available for 1 subject at baseline, but inasmuch as he had no visible lesions, he was classified as normal. Examinations that resulted in a subject's having an indeterminate diagnosis as defined previously were excluded from analysis. The number of subjects was insufficient to analyze the risk of progression to ASCUS, LSIL, and HSIL separately, and therefore these were combined for purposes of analysis.
Risk factors examined only at baseline included HIV status, the presence of HPV, and the baseline level of anal disease. Both baseline CD4 level and the change in CD4 level over the 2-year period were included as risk factors in the HIV-positive subjects. Medical and behavioral factors examined were from the 2-year follow-up period.
Relative risks (RR) were estimated separately for HIV-positive and HIV-negative subjects and 95% confidence intervals (CI) were computed (13). The X2 test for homogeneity was used for comparisons between subjects observed for 2 years and those excluded from analysis. Significance levels reported for trend are from the Mantel-Haenszel X2 statistic for the test of linear trend. The SAS statistical package (SAS Institute, Inc., Cary, NC, U.S.A.) was used for all data analysis.
Of the 608 subjects who enrolled in the study, 333 (55%) were included in the 2-year analysis. Of the 275 subjects excluded from analysis, 17 had HSIL at baseline and were referred for treatment, 49 died or became too ill to participate by the 2-year point, 120 dropped out before 2 years, 2 seroconverted to HIV-positive by 2 years, 9 had no visit during the 18 to 30 month time period around 2 years, and 78 had an indeterminate diagnosis. Of the 333 remaining subjects, 169 (51%) were HIV-positive and 164 (49%) were HIV-negative.
The characteristics of the subjects who were analyzed after 2 years of follow-up were similar to those of subjects excluded from analysis except that the excluded subjects reported slightly fewer years of education; 30% of excluded subjects reported >16 years of education, compared with 43% of subjects observed for 2 years. A higher proportion of excluded subjects was HIV-positive (64%) than those observed for 2 years (51%), and in the HIV-positive subjects, a higher proportion of excluded subjects had lower CD4 counts at baseline than those who were observed for 2 years; 47% had CD4 counts <200/mm3 compared with 19% of men observed for 2 years. Subjects included in the 2-year analysis and those excluded were similar with respect to the proportion with ASIL at baseline.
Table 1 shows the anal disease status separately for the HIV-positive and HIV-negative subjects at baseline and the highest diagnosis determined within 2 years of follow-up. Men who were HIV-positive at baseline were more likely to progress to ASIL than HIV-negative men. This was true for each baseline diagnosis. Of HIV-positive men who were normal at baseline, more than one half progressed to ASIL within 2 years compared with 17% of HIV-negative men. An even higher proportion of HIV-positive and HIV-negative men progressed to ASIL when diagnosed with ASCUS at baseline (70% and 31%, respectively). A high proportion of those with LSIL at baseline progressed to HSIL (62% and 36%, respectively). Rates of lesion regression differed between HIV-positive and HIV-negative men (data not shown). Of 27 HIV-positive men diagnosed with ASCUS at baseline, 8 (30%) were normal at the visit closest to 2 years, compared with 8 of 13 (62%) HIV-negative men. Similarly, of 55 HIV-positive men diagnosed with LSIL at baseline, 3 (5%) were normal at the visit closest to 2 years, compared with 7 of 14 (50%) HIV-negative men.
During the 2-year follow-up period, two HIV-negative subjects seroconverted to HIV-positive, both after 12 months of follow-up and were excluded from analysis in this paper. One of these subjects had no anal disease at baseline and continued to be free of disease. The other subject had ASCUS at baseline. At his 1-year visit, he was HIV-positive and was diagnosed with LSIL. Nine months later, he was diagnosed with HSIL.
Table 2 shows the impact of baseline diagnosis, HIV status, and baseline CD4 level on risk of progressing to a higher stage of anal disease than that diagnosed at baseline. Baseline diagnosis of anal disease was not strongly predictive of progression (RR = 1.3). HIV-positive subjects had greater than a twofold risk of progression compared with HIV-negative subjects. When CD4 count at baseline was used as an indicator of the degree of severity of HIV-related immunosuppression, a significant trend was noted in the risk of anal lesion progression. HIV-positive subjects with CD4 counts <200/mm3 had a threefold risk of progression when compared with HIV-negative subjects, whereas HIV-positive men with CD4 counts >500/mm3 had a nearly twofold risk when compared with HIV-negative subjects. These RR showed little change after adjustment for HPV positivity using either polymerase chain reaction (PCR) or hybrid capture (HC; Digene Diagnostics, Inc., Silver Spring, MD, U.S.A.). We also examined the 2-year change in CD4 level as a predictor for anal lesion progression in HIV-positive men. Neither absolute change nor percentage change in CD4 count was associated with an elevated risk in our study (data not shown).
Table 3 shows the impact of HPV infection on risk of progression stratified by HIV status. Using PCR, we were able to detect 29 individual and one group of 10 HPV types. 69% of HIV-positive men and 27% of HIV-negative men who were HPV-positive had more than one HPV type. The frequency of infection with multiple HPV types was too high to permit analysis of the role of any single HPV type. Among both HIV-positive and HIV-negative subjects, the trend toward increased risk of anal lesion progression was significant as the number of HPV types increased when identified by PCR. Among the HIV-positive subjects, the risk was 1.6 for anal progression with presence of a single HPV type and 2.0 for multiple types when compared with HIV-positive men without HPV infection. These showed little change when adjusted for baseline CD4 count. Among HIV-negative subjects, the RR was 3.7 for anal lesion progression with presence of a single HPV type and 5.1 for multiple types when compared with HIV-negative men without HPV infection. These showed little change when adjusted for baseline anal lesion diagnosis.
HC positivity was predictive of anal disease progression. Among the HIV-positive subjects, a twofold RR was associated with group A HPV types and group A and B types combined. The quantity of HPV DNA as reflected by the HC group B relative light unit (RLU) ratio was examined as a risk factor for anal disease progression (see reference 5 for definition of RLU ratios). For HIV-positive men with CD4 counts >500/mm3, the risk for anal disease progression in those with a RLU ratio >10 (RR = 2.5; 95% CI, 1.3-4.8) was higher than for those with a RLU ratio between 0.65 and 10 (RR = 1.4; 95% CI, .62-3.2) when compared with men who were negative for group B DNA (p for trend = .003). Computation of the RR associated with higher RLU ratios in HIV-positive men with CD4 counts <500/mm3 was not meaningful because of the high proportion of the men who progressed in the reference group (RLU ratio, <0.65; 15 of 16; 93%).
Among the HIV-negative subjects, a twofold risk for anal disease progression was also associated with group A types whereas a greater than a threefold risk was associated with group B types and group A and B types combined (Table 3). Similar to HIV-positive men with CD4 counts >500/mm3, for HIV-negative men, the risk (RR = 3.6; 95% CI, 2.0-6.4) for anal disease progression in those with a group B RLU ratio >10 was higher than for those with a RLU ratio between 0.65 and 10 (RR = 2.5; 95% CI, 1.4-4.3) when compared with those who were negative for group B DNA (p for trend < .0005).
The relation between exposure during follow-up to risk factors and progression of anal disease was examined separately for HIV-positive and HIV-negative men. Among the HIV-positive men, a slight elevation in risk was seen for those who reported a new episode of genital warts since the baseline visit (RR = 1.3; 95% CI, 1.0-1.6). Among HIV-negative subjects, a new episode of genital warts (RR = 3.1; 95% CI, 1.9-5.2), blood in the stool (RR = 2.0; 95% CI, 1.1-3.5) and rectal bleeding after receptive anal intercourse (RR = 1.9; 95% CI, 1.0-3.6) were associated with an increased risk for anal disease progression. Multivariate analyses were not done because of small numbers of episodes for most of these risk factors. Other factors examined during the follow-up period were found not to be associated with increased risk in HIV-positive or HIV-negative men. These included alcoholic beverage consumption, new episodes of genital herpes, fissures or fistulas, rectal discharge, use of enemas or suppositories, douches or enemas before or after anal intercourse, and receptive anal intercourse.
It is unknown how frequently ASIL progresses to invasive anal cancer and whether treatment of ASIL will prevent development of cancer. However, using cervical disease as a model, anal cancer may be preventable with targeted screening and treatment of anal HSIL. Thus, an understanding of the natural history of ASIL and risk factors for progression in those at risk is important.
The natural history of ASCUS and LSIL in the anus have not been reported previously. This study differs from earlier work by characterizing the natural history of these lesions in HIV-positive and HIV-negative men. This study also differs from a previous report (14) in its use of more advanced and sensitive HPV detection technology (PCR) to detect 29 individual HPV types and one group of 10 types, which most likely accounts for the higher prevalence of anal HPV infection in the present report. We also used HC to detect HPV infection, permitting semiquantitative analyses of the amount of HPV DNA in the anal specimen and characterization of the relation between HC RLU levels and anal lesion progression. Finally, inasmuch as the degree of anal cytologic abnormality does not always correlate with the degree of histologic abnormality (15), the use of anal histologic examination to classify the stage of anal lesions in many cases resulted in improved diagnostic accuracy.
In this study, high rates of incident ASIL were seen within 2 years in both HIV-positive men (52%) and HIV-negative men (17%) who had no evidence of lesions at baseline. In total, 20% of the HIV-positive men who had no evidence of lesions at baseline developed HSIL, similar to the 15% incidence of HSIL reported in another study (16) and to the 16% incidence of HSIL in an earlier study after an average of 17 months of follow-up (17). Among the HIV-negative men with no anal lesions at baseline, 8% developed HSIL, a number similar to the 5% reported earlier (16).
ASCUS on anal cytology was associated with a high rate of detection of ASIL within 2 years (70% in HIV-positive men and 31% in HIV-negative men), indicating that detection of ASCUS on cytology is a strong indicator of need for follow-up. Finally, more than one half of the HIV-positive subjects with anal LSIL at baseline progressed to HSIL within 2 years, and there was little regression in those who did not. Regression rates were higher in the 14 HIV-negative men with LSIL at baseline (50%), but more than one third progressed to HSIL. These data demonstrate the high incidence rate of ASIL, including HSIL, and the potential for LSIL to progress to HSIL in the anal canal, particularly in the HIV-positive subjects. These data further highlight the importance of LSIL. Although LSIL does not require immediate therapy because it is unlikely to progress to invasive cancer, it warrants close attention because it may progress to HSIL, which should be treated as a possible anal cancer precursor.
Despite the high documented incidence of ASIL in this study, the true incidence in the HIV-positive men may have been underestimated, because subjects with the most advanced immunosuppression were less likely to have been included in the 2-year analysis. Likewise, the incidence in HIV-negative men with no lesions at baseline may have been slightly underestimated relative to the HIV-positive men, because they were observed at longer time intervals than the HIV-positive men, so that lesion development and regression during this longer time interval may have been missed. Differences in lesion progression in HIV-negative men with a lesion detected at any time would not have been underestimated relative to HIV-positive men because all men with anal lesions were observed at the same intervals regardless of HIV status. For this reason and because of possible biologic differences between HIV-positive and HIV-negative men in the cohort, most analyses of risk factors for anal lesion progression were conducted separately for the two groups.
Univariate analysis indicated that the most important risk factors for anal lesion progression were characteristics at baseline including HIV positivity, lower CD4 count, and HPV infection as reflected by detection of more than one HPV type using PCR, detection of group A or group B HPV types using HC, and higher baseline levels of group B HPV DNA as measured by RLU ratios. These factors are the same as those associated with prevalent anal lesions (5), and our data indicate that the factors most strongly associated with detection of an anal lesion are associated with disease progression.
Exposure during the follow-up period to most of the behavioral and medical risk factors examined was not associated with increased RR for progression to a higher level of anal disease in either HIV-positive or HIV-negative men. Not surprisingly, new episodes of warts since baseline examination were associated with progression in both HIV-positive and HIV-negative men. Reports of rectal bleeding after sex and blood in the stool in HIV-negative men may have reflected the development of these new lesions. These data raise the concern that ASIL may be clinically important for its ability to potentiate HIV transmission because of lesion friability and trauma-induced bleeding.
A surprising amount of new cases of HSIL was detected during the 2 years of follow-up, particularly when compared with the prevalence of HSIL at baseline (5). Some of the higher-than-expected incidence of HSIL may be explained by improvement in diagnostic skills during the study. This would result in underestimation of HSIL prevalence at baseline and overestimation of HSIL incidence. However, because our subjects were relatively uniform in age, the higher-than-expected incidence may also represent a cohort effect. Comparison of cytology results during the 2-year follow-up period, which are not as subject to changes in diagnostic skills as anoscopically directed biopsies, were consistent with a true increase in lesion prevalence (data not shown).
In summary, after 2 years of follow-up, high rates of anal lesion incidence and progression, including progression to HSIL, were seen in HIV-positive men, and to a lesser extent in HIV-negative men. Because our data include only 2 years of follow-up, this incidence may increase with longer follow-up. HIV infection was an important risk factor, and in the HIV-positive men, lower CD4 counts also contributed to risk of anal disease progression. Among both HIV-positive and HIV-negative men, HPV infection, particularly with multiple types, and HC group B infection with a high RLU ratio, were important risk factors. LSIL and ASCUS progressed to HSIL in a large proportion of subjects, and men with LSIL should be monitored carefully for development of HSIL. Because progression of HSIL to anal cancer may require several years, HIV-negative men with HSIL may be at particularly high risk given their potential for a normal life span. Paradoxically, HIV-positive individuals with HSIL on highly active antiretroviral therapy (HAART) may be at increased risk of progression to anal cancer if HAART prolongs their life span but does not lead to HSIL regression. Data are needed on the natural history of ASIL in individuals on HAART.
Acknowledgments: The authors are grateful to Jennifer Kristianson, Pam Elmore, and Rachel Remler for their expert assistance. Supported by grant number R01 CA54053 from the National Cancer Institute. These studies were performed 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.
1. Daling JR, Weiss NS, Hislop TG, et al. Sexual practices, sexually transmitted diseases, and the incidence of anal cancer
. N Engl J Med
2. Holly EA, Whittemore AS, Aston DA, Ahn DK, Nickoloff BJ, Kristiansen JJ. Anal cancer
incidence: genital warts, anal fissure or fistula, hemorrhoids, and smoking. J Natl Cancer Inst
3. Rabkin CS, Yellin F. Cancer incidence in a population with a high prevalence of infection with human immunodeficiency virus type 1. J Natl Cancer Inst
4. Melbye M, Cote TR, Kessler L, Gail M, Biggar RJ. High incidence of anal cancer
among AIDS patients: the AIDS/Cancer Working Group. Lancet
5. Palefsky JM, Holly EA, Ralston ML, et al. Anal squamous intraepithelial lesions among HIV
-positive and HIV
-negative homosexual and bisexual men: prevalence and risk factors. J Acquir Immune Defic Syndr Hum Retrovirol
6. Caussy D, Goedert JJ, Palefsky J, et al. Interaction of human immunodeficiency and papilloma viruses: association with anal epithelial abnormality in homosexual men. Int J Cancer
7. Critchlow CW, Holmes KK, Wood R, et al. Association of human immunodeficiency virus and anal human papillomavirus
infection among homosexual men. Arch Intern Med
8. Palefsky JM, Shiboski S, Moss A. Risk factors for anal human papillomavirus
infection and anal cytologic abnormalities in HIV
-positive and HIV
-negative homosexual men. J Acquir Immune Defic Syndr Hum Retrovirol
9. Kiviat N, Critchlow C, Holmes K, et al. Association of anal dysplasia
and human papillomavirus
with immunosuppression and HIV
infection among homosexual men. AIDS
10. Breese PL, Judson FN, Penley KA, Douglas J Jr. Anal human papillomavirus
infection among homosexual and bisexual men: prevalence of type-specific infection and association with human immunodeficiency virus. Sex Transm Dis
11. Melbye M, Palefsky J, Gonzales J, et al. Immune status as a determinant of human papillomavirus
detection and its association with anal epithelial abnormalities. Int J Cancer
12. Palefsky JM, Holly EA, Ralston ML, Jay N. Prevalence and risk factors for human papillomavirus
infection of the anal canal in HIV
-positive and HIV
-negative homosexual men J Infect Dis
13. Breslow NE, Day NE. The design and analysis of cohort studies. Statistical methods in cancer research,
vol. 2, Scientific Publication No. 82. Lyons, France: International Agency for Research on Cancer, 1987.
14. Palefsky JM, Gonzales J, Greenblatt RM, Ahn DK, Hollander H. Anal intraepithelial neoplasia and anal papillomavirus infection among homosexual males with group IV HIV
15. Palefsky JM, Holly EA, Hogeboom CJ, Jay N, Berry M, Darragh TM. Anal cytology
as a screening tool for anal squamous intraepithelial lesions. J Acquir Immune Defic Syndr Hum Retrovirol
16. Critchlow CW, Surawicz CM, Holmes KK, et al. Prospective study of high grade anal squamous intraepithelial neoplasia in a cohort of homosexual men: influence of HIV
infection, immunosuppression and human papillomavirus
17. Palefsky JM, Holly EA, Gonzales J, Lamborn K, Hollander H. Natural history of anal cytologic abnormalities and papillomavirus infection among homosexual men with group IV HIV
disease. J Acquir Immune Defic Syndr Hum Retrovirol