The incidence of anal cancer prior to the onset of the HIV epidemic was estimated to be as high as 35/100 000 among men with a history of receptive anal intercourse [1,2]. The incidence of anal cancer among HIV-positive men who have sex with men (MSM) was estimated to be twice that of HIV-negative MSM  and in another study the risk of anal cancer was 37-fold higher among HIV-positive men compared with the general population . Since the introduction of highly active antiretroviral therapy (HAART) in 1996, the incidence of anal cancer has continued to increase. In San Francisco County the incidence of anal cancer quadrupled between the 1973–1979 period and the 1996–1999 period among all Caucasian and Hispanic men aged 40–64  and the incidence of anal cancer in HIV-positive MSM is likely to be considerably higher than the incidence of cervical cancer in the general population of women (8 to 10/100 000) .
Cervical cancer is preceded by cervical intraepithelial neoplasia (CIN) grade 2 or 3, primarily CIN 3. Identification of CIN 2 and 3 through cervical cytology screening and treatment of these precursors to cervical cancer has contributed to a decline in the incidence of cervical cancer in countries where screening is practised routinely. Likewise, anal cancer may be preceded by anal intraepithelial neoplasia (AIN). Several prospective studies performed before HAART was introduced showed a high prevalence  and incidence of AIN 2 and 3 among both HIV-negative and HIV-positive MSM [8–10] consistent with the observed increase in incidence of anal cancer. Among HIV-positive MSM, anal human papillomavirus (HPV) infection and more advanced immunosuppression as measured by low CD4+ cell counts were the most important risk factors for the development of AIN [8–10].
HAART has led to the reduction in complications of several HIV-associated viral opportunistic infections including cytomegalovirus-associated retinitis and human herpes virus-8-associated Kaposi's sarcoma [11–13]. Data are limited on the effect of HAART on HPV-associated AIN. In one study, 6 months of HAART had little effect on the natural history of AIN in MSM followed before and after HAART initiation . Understanding the effect of HAART on AIN is important because analogous to progression from CIN to invasive cervical cancer, AIN most likely takes many years to progress to anal cancer. If individuals on HAART are living longer, but HAART does not reduce the incidence or progression of AIN, then the incidence of anal cancer may continue to increase among those on HAART if AIN remains untreated.
The purpose of this study was to determine the factors associated with and prevalence of AIN in a cohort of HIV-positive MSM and to determine whether HAART use was associated with lower prevalence of AIN.
All procedures were performed after obtaining informed consent and with the approval of the Committee on Human Research of the University of California, San Francisco (UCSF). MSM were recruited into this cohort study between February 1998 and January 2000 through newspaper advertisements and other community outreach at UCSF. A total of 357 HIV-positive MSM were enrolled. Interviews were conducted in person at baseline and included medical history, sexual practices and substance use. Participants had a clinical examination beginning with an anal swab for cytology followed by a second swab for anal HPV testing. High-resolution anoscopy (HRA) was completed with a biopsy of visible anal lesions unless there were medical contraindications. Anal cytology (conventional, direct smears) was performed and interpreted using the Bethesda criteria as described previously [7,9,10]. HRA and anal biopsy were performed as described previously [9,10,15]. Anal cytology was classified as normal, atypical squamous cells of undetermined significance (ASCUS), AIN 1, AIN 2 or AIN 3. Anal biopsies were classified as normal, atypical, AIN 1, AIN 2 or AIN 3. Blood was obtained for CD4+ lymphocyte counts and HIV viral load tests.
CD4+ lymphocyte counts were determined by standardized two- or three-color fluorescence methods. HIV viral load was measured using the branched-chain Chiron assay (Chiron, Emeryville, California, USA). When the study was initiated, the lower limit of sensitivity was 500 copies/ml; midway through the study, the sensitivity of the test improved to 50 copies/ml.
HPV testing was performed as described previously using the polymerase chain reaction (PCR) with L1 consensus primers and probes specific for 29 individual HPV types and a mixture of 10 other types [16,17]. Beta-globin negative samples were excluded from analysis.
Participants were interviewed in person regarding medical history and behavioral factors. Men were queried about lifetime history of medical conditions including hemorrhoids, anal fissures or fistulas, history of AIN, sexually transmitted infections, enemas and suppositories. Behavioral factors included lifetime and recent receptive and insertive anal intercourse, rimming and use of objects in the anus, lifetime history of cigarette use, alcohol consumption, and use of recreational drugs. The men were queried about lifetime history of common HIV-related conditions and finally, participants were queried in detail about their lifetime use of antiretroviral medications.
A composite anal diagnosis based on cytology and histology was used in analysis, as described previously [9,10]. If both cytological and histological results were available, the diagnosis used for analyses was the more severe result. Participants were classified as normal if they had normal cytology and no lesions seen by anoscopy, or if both the cytology and histology results were normal. If a lesion was seen by anoscopy but not biopsied, the cytology result was used for diagnosis unless the cytology was normal. In that case, the diagnosis was considered indeterminate and the data were excluded from analysis.
Whenever possible, study variables were dichotomized into presence or absence of the anal lesion, condition, treatment or behavior to evaluate their association with the anal lesion. For many of the behavioral factors almost all participants engaged in the particular activity at least occasionally. For these factors summary variables were created to characterize the level of lifetime exposure . The antiretroviral drug regimens were classified as HAART and non-HAART regimens. HAART was defined as the use of three or more antiretroviral drugs taken in combination and consisting of nucleoside reverse transcriptase inhibitors (NRTI) and one or more protease inhibitor (PI) or non-nucleoside reverse transcriptase inhibitor (NNRTI). Both the HAART and non-HAART regimens were further subdivided into PI-containing, NRTI-containing and PI plus NNRTI-containing regimens. The types of drug regimen and length of use during the past 3 years was determined. The length of use of a particular regimen was adjusted for the time spent not taking their medications. The time on current regimen and time on the regimens in the past 3 years were divided into categories of up to 1–12 months, 13–24 months, and 25 or more months. The use of Combivir (lamivudine and zidovudine) was counted as the use of two NRTIs to determine the type of drug regimen. CD4+ cell counts were stratified into three groups: < 200, 200–500 and > 500 × 106 cells/l as published previously [10,17]. HIV viral load data were stratified into four groups: < 500, 500–4000, 4001–20 000 and > 20 000 copies/ml.
Odds ratios (ORs) as the measure of association were computed using unconditional logistic regression. Normal theory approximation was used to determine the 95% confidence intervals (CI) for the OR estimates (hereafter called ‘risk’). Significance testing of regression coefficients was based on Wald's statistic. Tests for trend were assessed by assigning ordinal dose-specific categories, fitting this variable as a continuous linear term and testing the significance of the regression coefficient using the likelihood ratio test. Fisher's exact test was used to test for association in 2 × r tables. One-way analysis of variance (ANOVA) was used to compare means. Post-hoc pair-wise multiple comparisons were evaluated using the Ryan–Einot–Gabriel–Welsch multiple range test. All P-values were based on two-sided tests. Results were considered significant at the P < 0.05 level. Statistical analyses were performed using SAS software (SAS Institute, Cary, North Carolina, USA).
A total of 357 HIV-positive MSM were enrolled in the study. The age, education level, race and ethnicity, and years of HIV-positivity stratified by use or non-use of HAART are shown in Table 1. There were no significant differences in these characteristics among men who did not use antiretrovirals and men who did use HAART, including mean time of HIV positivity, although men using a non-HAART regimen were older than men not using an antiretroviral regimen (P < 0.05).
The most common HAART regimens are shown in Table 2. For all men the median time since their first HIV-positive blood sample was 130 months (range, 0–259 months). Only 26 (7%) never had taken any antiretroviral medication and most (307, 86%) were taking at least one antiretroviral drug at the time of study entry (Table 2). A total of 137 different combinations of medications were in current use by these men. For 267 (75%) men, the regimen was considered to be HAART. Of the remaining 40 men whose regimen was not HAART, 31 were taking NRTIs only and nine were taking a PI or NNRTI but using fewer than three drugs.
Of those taking antiretrovirals, 57 (19%) skipped some medication in the past 3 days. A drug holiday of at least 2 days from some or all of the drugs in their current regimen was taken by 47 men. The medication holidays were 1 week or less for 31 (66%) of the men, 1 week to 1 month for 10 (21%) men, and 1–6 months for six (13%) men. Most men reported using the drugs as prescribed, but 20 (7%) reported taking one or more medications in different amounts or at different times than prescribed.
Relationship between anal lesions and HIV status, CD4+ cell counts, HIV viral load and anal HPV infection
The relationship between anal lesions and HIV status, CD4+ cell counts, HIV viral load and anal HPV infection is shown in Table 3. AIN was found in 81% of the men with available CD4+ cell counts at the baseline visit. AIN 1 or condyloma was found in 29% of the men, 39% had AIN 2 and 13% had AIN 3. There was a trend toward an association between increased ORs for AIN of any grade, or AIN 2 or 3, and decreasing CD4+ cell counts but these results were not statistically significant (P < 0.07). There was no association between AIN of any grade or AIN 2 or 3 with higher HIV viral load.
Of the 354 HIV-positive men with an anal swab available for HPV testing, 31 (8.8%) tested negative for beta-globin by PCR. Of the 323 evaluable HIV-positive MSM, 308 (95%) had anal HPV infection. The men with anal HPV infection had a 15-fold increased estimated risk (CI = 4.0–57) for AIN of any grade and a 19-fold increased estimated risk (CI = 3.7–101) for AIN 2 or 3 compared with HPV-negative men. The increased estimated risk for having AIN with increasing number of specific HPV types was statistically significant (P-for-trend < 0.001). Of the 308 men who had anal HPV infection, 270 (88%) had more than one specific HPV type and 128 (42%) had six or more HPV types. Furthermore, 120 (94%) of the 128 men with six or more types had AIN and 85 (66%) had AIN 2 or 3. Among the 163 men diagnosed with AIN 2 or 3, HPV was detected in 160 (98%).
Antiretroviral drug use and other risk factors for AIN
Men who were taking antiretroviral medications at baseline had an increased estimated risk for any grade of AIN and of AIN 2 or 3 compared with men who were not currently taking medications (Table 4). When the drug regimens were separated into HAART and non-HAART regimens the risks were higher for those on HAART. Risks for any anal grade of AIN and AIN 2 or 3 increased with increasing time on HAART within the prior 3 years. Similarly, the risks increased with increase in the total number of different antiretroviral drugs ever used (P < 0.001 for each trend), and with the number of antiretroviral drugs currently being taken (data not shown).
Lifetime use of zidovudine, lamivudine, didanosine, indinavir and ritonavir and current use of stavudine and efavirenz each were significantly associated with increased risks for AIN and AIN 2 or 3 (ORs ranged between 2.4 and 5.0). Those with a history of acyclovir use, Pneumocystis carinii, oral or esophageal Candidiasis, shingles and Molluscum contagiousum also showed increased risks for AIN and AIN 2 or 3. The ORs showed little change when adjusted for CD4+ cell counts (data not shown).
There was no association between any grade of AIN or AIN 2 or 3 and race, education or time since HIV diagnosis. There also was no association with sexual activities, use of alcohol, tobacco, recreational drugs, or other medical conditions or therapeutic drugs. The number of changes in medications in the past 3 years, the use of drug holidays, compliance with drug regimen, and time on current regimen were not associated with AIN.
Separate multivariate models for any grade of AIN and for AIN 2 or 3, were developed. Factors included in the model were number of HPV types, CD4+ cell count, HIV viral load and current use of HAART (Table 5). After adjustment for length of time HIV positive, CD4+ cell count and HIV viral load, both the number of specific HPV types and current use of HAART were associated with increased risk for any grade of AIN and for AIN 2 or 3. Models using other methods of measurement of anal HPV infection such as positive versus negative provided results similar to models using the number of specific HPV types.
Understanding the effect of HAART on the natural history of AIN is critical to estimate the future direction of this disease as the HIV epidemic proceeds into its third decade. Our data indicate that HAART does not reduce the prevalence of AIN among HIV-positive MSM. In multivariate analysis, HAART use was positively associated with detection of AIN, including AIN 2 or 3. Another important factor in multivariate analysis was anal HPV infection and the number of specific HPV types detected.
A similar positive association between use of HAART and an increased risk for oral warts has been previously reported . Studies of the effects of HAART on the natural history of CIN have shown mixed results, with some showing a reduction in CIN progression and others showing little effect [19–21]. However, it is clear that any positive effect of HAART on CIN is modest [22,23]. In an earlier prospective report, 6 months of use of HAART had no impact on the natural history of AIN . Consistent with this observation, in the present analysis, longer duration of HAART use prior to baseline had no impact on the prevalence of AIN.
One limitation of this study is that it included a high percentage of men who had been HIV positive for many years and a high proportion who were taking HAART. We believe that our study group is representative of HIV-positive men in the reported age range, but our results may not be generalizable to all HIV-positive MSM, particularly younger HIV-positive men who may have been HIV infected for fewer years.
In our study the magnitude of the risk for AIN associated with use of HAART was substantial. Although it is likely that men on HAART have more advanced HIV disease than those not on HAART, and hence might have a higher risk for AIN a priori, our multivariate model included length of time of HIV positivity, CD4+ cell count and HIV viral load. Nevertheless men on HAART are different from those not on HAART in several ways and it is possible that the increased risk for AIN in this study reflected unmeasured differences between the two groups. Adjusting for the variables in our model does not exclude a role for lifetime cumulative immunosuppression. Evidence that cumulative immunosuppression may play a role is suggested by the association between increased risk for AIN and lifetime usage of antiretrovirals, history of pneumocystis and candidiasis, and the total number of different antiretroviral drugs ever used.
Lower CD4+ cell counts were associated with an increased OR for AIN in our study as in earlier reports, but this association was weaker than in the earlier studies conducted prior to the introduction of HAART [7–10]. In a recent report the prevalence of AIN among 45 MSM was similar in those exhibiting a significant increase in CD4+ cell count after HAART initiation compared with those who did not . These results are not surprising since HAART leads to increased CD4+ cell counts in many men, and CD4+ cell counts may not be as reliable a measure of immune suppression as they were in the pre-HAART era. Another reason for the lack of clear association between CD4+ cell count and AIN in the HAART era is that CD4 cell increases due to HAART may not correspond to an increase in HPV-specific immune response. Finally immune restoration due to HAART may play a limited role in clearing AIN, regardless of how it is measured.
Data from the pre-HAART era are informative with respect to the role of the immune response in controlling AIN and anal cancer. During that time neither cervical nor anal cancer were associated with lower CD4+ cell counts, unlike other HIV-associated malignancies such as Kaposi's sarcoma and non-Hodgkin lymphoma . At the same time, however, progression to AIN 2 or 3 pre-HAART clearly was associated with lower CD4+ cell counts [10,26]. Together these data are consistent with a model wherein immune suppression may play a more prominent role early in the natural history of AIN, but other factors such as host chromosomal changes may play a more important role in persistence of high-grade AIN and progression to cancer . If true, then institution of HAART would be unlikely to lead to regression of AIN if instituted when the individual already has AIN 2 or AIN 3. It is not yet known whether the use of HAART at an earlier stage of HIV disease, namely, at high CD4+ cell counts, might reduce the incidence of AIN since guidelines for the institution of HAART typically recommend delaying therapy until the CD4+ cell count is below 350 × 106 cells/l. Data from treatment of CIN in HIV-positive women suggest a possible benefit of HAART in decreasing recurrence rates after ablative therapy . Therefore it might be useful to evaluate the relationship of HAART use and response to treatment for AIN.
Comparisons between data collected prior to the introduction of HAART and the data in the present cohort also suggest that there has been a marked change in the epidemiology of AIN in the past 10 years. In a point-prevalence analysis of a cohort enrolled between 1991 and 1994, the prevalence of AIN among HIV-positive MSM with a mean age of 42 years was 36% . However, in the follow-up of that cohort, more than 50% of HIV-positive MSM with a CD4+ cell count of less than 500 × 106 cells/l developed AIN 2 or 3 within 4 years. Consistent with these data, the prevalence of AIN (81%) was markedly higher at baseline in the present cohort, which had a similar mean age when enrolled between 1998 and 2000. These data indicate that the prevalence of AIN has rapidly increased among HIV-positive MSM in the San Francisco Bay Area since 1991–1994. A higher prevalence of AIN among HIV-negative MSM in the San Francisco Bay Area enrolled between 1998 and 2000 has also been noted in comparison with a similarly-aged cohort of HIV-negative MSM studied between 1991 and 1994 (J. Palefsky, unpublished data). It is unknown whether the prevalence of anal HPV infection and AIN is equally high among MSM at other geographic locations but a recent study by Chin-Hong et al. of 1409 HIV-negative MSM in Denver, Boston, New York and San Francisco showed a similar prevalence of anal HPV infection among men in these cities . The mechanisms underlying the increase in AIN among MSM in the San Francisco Bay Area are unclear. Although we cannot exclude the possibility that some of the increased disease detection may have reflected improved diagnostic acumen by the study investigators over time, the data are most consistent with a cohort effect reflecting higher levels of sexual activity that began in the 1970s.
Similar to cervical cancer screening, an anal cancer screening program that uses anal cytology, followed by referral of those with abnormal results to high resolution anoscopy, and that provides treatment of biopsy-proven high-grade AIN, may prevent the development of anal cancer. Cranston et al. recently showed that anal cytology sampling can be effectively achieved by patients themselves with a sensitivity similar to that of clinician-performed sampling, suggesting that patient-based screening may be possible . Using models based on pre-HAART data, Goldie et al. projected screening with anal cytology to be cost-effective to prevent anal cancer for both HIV-positive and HIV-negative MSM [31,32]. Under the assumptions of those models, the cost-effectiveness of screening HIV-positive MSM on HAART may be even higher given the data reported here.
Taken together, the results of this study indicate that AIN, including AIN 2 or 3, continues to be detected in a high number of HIV-positive MSM, regardless of whether they are on HAART. While our data do not suggest that use of HAART directly increases the risk of AIN, at a minimum they do support the conclusion that HAART use does not reduce the risk for AIN. Unlike most other HIV-related malignancies, the incidence of anal cancer has been increasing since the introduction of HAART. Studies to document the utility of treatment of AIN 2 or 3 to prevent anal cancer are urgently needed.
Sponsorship: These studies were conducted in the General Clinical Research Center, University of California, San Francisco with funds provided by the Division of Research Resources 5 M01-RR-00079, US Public Health Service. The study was supported by NCI grant number R01CA54053.
1. Daling JR, Weiss NS, Hislop TG, Maden C, Coates RJ, Sherman KJ, et al
. Sexual practices, sexually transmitted diseases, and the incidence of anal cancer. N Engl J Med 1987; 317:973–977.
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 1989; 81:1726–1731.
3. Goedert JJ, Cote TR, Virgo P, Scoppa SM, Kingma DW, Gail MH, et al
. Spectrum of AIDS-associated malignant disorders. Lancet 1998; 351:1833–1839.
4. Frisch M, Biggar RJ, Goedert JJ. Human papillomavirus-associated cancers in patients with human immunodeficiency virus infection and acquired immunodeficiency syndrome. J Natl Cancer Inst 2000; 92:1500–1510.
5. Cress RD, Holly EA. Incidence of anal cancer in California: increased incidence among men in San Francisco, 1973–1999. Prev Med 2003; 36:555–560.
6. Qualters JR, Lee NC, Smith RA, Aubert RE. Breast and cervical cancer surveillance, United States, 1973–1987. MMWR CDC Surveill Summ 1992; 41:1–7.
7. Palefsky JM, Holly EA, Ralston ML, Arthur SP, Jay N, Berry JM, et al
. Anal squamous intraepithelial lesions in HIV-positive and HIV-negative homosexual and bisexual men: prevalence and risk factors. J Acquir Immune Defic Syndr Hum Retrovirol 1998; 17:320–326.
8. Critchlow CW, Surawicz CM, Holmes KK, Kuypers J, Daling JR, Hawes SE, 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 infection. AIDS 1995; 9:1255–1262.
9. Palefsky JM, Holly EA, Hogeboom CJ, Ralston ML, DaCosta MM, Botts R, et al
. Virologic, immunologic, and clinical parameters in the incidence and progression of anal squamous intraepithelial lesions in HIV-positive and HIV-negative homosexual men. J Acquir Immune Defic Syndr Hum Retrovirol 1998; 17:314–319.
10. Palefsky JM, Holly EA, Ralston ML, Jay N, Berry JM, Darragh TM. High incidence of anal high-grade squamous intra-epithelial lesions among HIV-positive and HIV-negative homosexual and bisexual men. AIDS 1998; 12:495–503.
11. Whitcup SM, Fortin E, Lindblad AS, Griffiths P, Metcalf JA, Robinson MR, et al
. Discontinuation of anticytomegalovirus therapy in patients with HIV infection and cytomegalovirus retinitis. JAMA 1999; 282:1633–1637.
12. Deayton JR, Wilson P, Sabin CA, Davey CC, Johnson MA, Emery VC, et al
. Changes in the natural history of cytomegalovirus retinitis following the introduction of highly active antiretroviral therapy. AIDS 2000; 14:1163–1170.
13. Appleby P, Beral V, Newton R, Reeves G. Highly active antiretroviral therapy and incidence of cancer in human immunodeficiency virus-infected adults. J Natl Cancer Inst 2000; 92:1823–1830.
14. Palefsky JM, Holly EA, Ralston ML, Da Costa MM, Bonner H, Jay N, et al
. The effect of highly active antiretroviral therapy on the natural history of anal squamous intraepithelial lesions and anal human papillomavirus infection. J Acquir Immun Defic Syndr 2001; 28:422–428.
15. Jay N, Berry JM, Hogeboom CJ, Holly EA, Darragh TM, Palefsky JM. Colposcopic correlates of anal squamous intraepithelial lesions. Dis Col Rectum 1997; 40:919–928.
16. Palefsky JM, Holly EA, Ralston ML, Jay N. Prevalence and risk factors for human papillomavirus infection of the anal canal in human immunodeficiency virus (HIV)-positive and HIV-negative homosexual men. J Infect Dis 1998; 177:361–367.
17. 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; 91:226–236.
18. Greenspan D, Canchola A, MacPhail L, Cheikh B, Greenspan J. Effect of highly active antiretroviral therapy on frequency of oral warts. Lancet 2001; 357:1411–1412.
19. Heard I, Schmitz V, Costagliola D, Orth G, Kazatchkine M. Early regression of cervical lesions in HIV-seropositive women receiving highly active antiretroviral therapy. AIDS 1998; 12:1459–1464.
20. 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; 15:2157–2164.
21. Lillo F, Ferrari D, Veglia F, Uberti-Foppa C. Human papillomavirus infection and associated cervical disease in human immunodeficiency virus-infected women: effect of highly active antiretroviral therapy. J Infect Dis 2001; 184:547–551.
22. Heard I, Tassié J, Kazatchkine M, Orth G. Increased regression of CIN in HIV-seropositive women under HAART. AIDS 2002; 16:1799–1802.
23. Heard I, Palefsky JM, Kazatchkine MD. The impact of HIV antiviral therapy on human papillomavirus (HPV) infections and HPV-related diseases. Antivir Ther 2004; 9:13–22.
24. Piketty C, Darragh TM, Heard I, Da Costa M, Bruneval P, Kazatchkine MD, et al
. High prevalence of anal squamous intraepithelial lesions in HIV-positive men despite the use of highly active antiretroviral therapy. Sex Transm Dis 2004; 31:96–99.
25. Rabkin CS. Association of non-acquired immunodeficiency syndrome-defining cancers with human immunodeficiency virus infection. J Natl Cancer Inst Monogr 1998; 23:23–25.
26. Critchlow CW, Surawicz CM, Holmes KK, Kuypers J, Daling JR, Hawes SE, 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 infection. AIDS 1995; 9:1255–1262.
27. Palefsky JM, Holly EA. Immunosuppression and co-infection with HIV. J Natl Cancer Inst Monogr 2003; 31:41–46.
28. Robinson WR, Hamilton CA, Michaels SH, Kissinger P. Effect of excisional therapy and highly active antiretroviral therapy on cervical intraepithelial neoplasia in women infected with human immunodeficiency virus. Am J Obstet Gynecol 2001; 184:538–543.
29. Chin-Hong PV, Vittinghoff E, Buchbinder S, Cohen D, Colfax G, Da Costa M, et al
. Age-related prevalence of anal human papillomavirus infection among sexually active men who have sex with men.J Infec Dis
30. Cranston RD, Darragh TM, Holly EA, Jay N, Berry JM, Costa M, et al
. Self-collected versus clinician-collected anal cytology specimens to diagnose anal intraepithelial neoplasia in HIV-positive men. J Acquir Immune Defic Syndr 2004; 36:915–920.
31. Goldie SJ, Kuntz KM, Weinstein MC, Freedberg KA, Welton ML, Palefsky JM. The clinical-effectiveness and cost-effectiveness of screening for anal squamous intraepithelial lesions in homosexual and bisexual HIV-positive men. JAMA 1999; 281:1822–1829.
32. Goldie SJ, Kuntz KM, Weinstein MC, Freedberg KA, Palefsky JM. The clinical benefits and cost-effectiveness of screening for anal squamous intraepithelial lesions and anal squamous cell cancer in homosexual and bisexual men. Am J Med 2000; 108:634–641.