Human papillomavirus infection and abnormal cytology of the anus in HIV-infected and uninfected adolescents

Abstract

Objective: 

To examine the prevalence of and risk for anal human papillomavirus (HPV) infection and abnormal anal cytology in sexually active adolescents.

Design: 

Prevalence data from adolescents aged 13–18 years with and without HIV infection and with a history of high-risk sexual behavior.

Methods: 

HPV DNA was detected using amplification techniques. Abnormal anal cytology was defined as atypical squamous cell of undetermined significance or worse.

Results: 

Prevalence of anal HPV infection was similar in HIV-infected [28/58 (48%)] and uninfected [9/25 (36%)] boys (P = 0.3). but greater in HIV-infected [59/183 (59%)] than in uninfected [11/82 (13%)] girls (P < 0.001). Perianal warts were a risk for anal HPV in both boys [odds ratio (OR), 15.5; 95% confidence interval (CI), 1.6–149] and girls (OR, 9.9; 95% CI, 1.9–51.3). In subjects without anal warts, HIV infection was significant for girls (OR, 2.3; 95% CI, 1.1–4.9) and homosexual/bisexual orientation was significant for boys (OR, 5.2; 95% CI, 1.3–20.6). Abnormal anal cytology was more common among boys [32/77 (41.6%)] than girls [38/230 (16.5%)] (P < 0.001) and in addition to anal HPV, independent risk factors were positive HIV status in boys (OR, 6.5; 95% CI, 1.5–11.9) and number of partners within the past 3 months in girls (OR, 4.2; 95% CI, 1.5–11.9).

Conclusions: 

Strong risk factors for abnormal anal cytology were HIV infection and anal HPV in boys and anal HPV and higher number of sexual partners for girls. The results suggest that anal cytology screening should be considered in HIV infected homosexual/bisexual males.

Introduction

Although anal cancers are rare, there is evidence that these cancers are increasing in HIV- infected persons [^1,2]. As with cervical cancers, human papillomavirus (HPV) has been strongly associated with the development of anal cancer. In addition, progression to anal cancer is thought to reflect the natural history of cervical cancer, with the development of pre-invasive, squamous intraepithelial lesions (SIL) prior to invasion. Many recent studies have demonstrated the high rate of HPV infection and SIL in HIV-infected men and women, underscoring the potential magnitude of this emerging health problem in persons living with HIV [^3–7].

Several investigators now promote anal cytology screening in HIV-infected persons in order to detect and triage these precancers for preventative treatment [⁸]. Although several epidemiological studies have shown that cervical HPV infections and SIL are more common in adolescents than in older women [^9–11], no studies to date have determined the magnitude of anal HPV infection and disease in this age group.

The higher rate of HPV infections compared with SIL and cancer in general populations suggest that cofactors are necessary for the development of these lesions [^5–12]. The higher rates of anal SIL and cancer reported in HIV-infected men and women compared with uninfected persons suggest that one of the important risks is immune dysfunction [^5–7]. Studies of anal HPV infection to date have primarily involved homosexual men, who most likely acquired HPV many years ago with the onset of sexual activity [^5–7]. The few studies reported in women include adults who also have longer histories of sexual activity than most adolescents [^13,14]. Risks for anal HPV infection in these studies suggest that risks for acquisition differ between men and women [^5–13]. Studies of HPV in HIV-infected adolescents may provide a unique opportunity to examine the associations between HPV and HIV early in their natural histories, since most adolescents have limited sexual experience compared with adults.

The present study examines the association between prevalence of anal HPV infection, abnormal anal cytology, and HIV status in a population of high-risk adolescents and risk factors for anal HPV and abnormal anal cytology in both boys and girls.

Methods

The REACH Project (Reaching for Excellence in Adolescent Care and Health) of the Adolescent Medicine HIV/AIDS Research Network recruited HIV-infected and high-risk HIV-uninfected adolescents ages 13 to 18 years into a study of biomedical and behavioral characteristics of HIV infection as seen in primary care across 15 different clinical sites in 13 cities in the United States. Characteristics of the cohort, recruitment and eligibility criteria, and study design are reported in detail elsewhere [¹⁵]. Briefly, the study enrolled eligible adolescents who were in care and by history had acquired HIV through sexual activity or injection drug use, of which the latter was rarely reported in our cohort (n = 2). The study also recruited HIV-negative girls and boys of comparable age and high-risk behaviors (i.e., sexual activity, substance use) in a preplanned ratio of 2:1 (HIV infected:uninfected). All subjects consented to participate in this study according to the approval of Institutional Review Boards of the respective institutions.

Data on antiretroviral therapy was obtained through interview and chart review for current prescriptions, and compliance data was obtained through interview. Demographic information and contraceptive behavior was collected in the face-to-face interview. Audio-Computer Assisted Self Interview (ACASI) was utilized for collection of all sensitive and sexual and illicit drug use behaviors [¹⁵].

Examinations for girls and boys included a visual inspection of the external anogenital area for genital warts, anal samples for cytology and HPV DNA testing using a single dacron swab, and a second anal sample for Chlamydia trachomatis testing (LCX STD system, Abbott Laboratories, Abbott Park, Illinois, USA). For girls, additional samples were collected: vaginal samples for Gram-stain evidence of bacterial vaginosis and culture for Trichinoma vaginalis infection, a 10 ml cervical vaginal lavage for HPV testing, an endocervical and exocervical sample for standard Papanicolaou smear, and an endocervical sample for C. trachomatis and Neisseria gonorrhoeae testing (LCX STD system, Abbott Laboratories) [¹⁰]. For boys, a first void urine sample for C. trachomatis and N. gonorrhoeae were also obtained. (LCX STD system, Abbott Laboratories).

Quantitative immunophenotyping of CD4 T cell lymphocytes counts were determined at the individual clinical sites in certified laboratories using AIDS Clinical Trials Group (ACTG) standardized flow cytometry protocols. Quantitative HIV-1 RNA viral load in plasma was measured in a central laboratory on frozen specimens, using either nucleic acid sequence-based amplification (NASBA) or NucliSens assays (Organon Teknika, Durham, North Carolina, USA). The respective lower limits of detection for the NASBA and Nuclisens assays were 400 and 80 copies/ml, respectively.

Clinicians were instructed to insert a moistened Dacron swab 2–4 cm into the canal and rotate the swab in a circular motion with firm pressure on the anal wall. Anal samples were immediately placed into PreservCyt (Cytyc Corp, Boxborough, Massachusetts, USA) media. Samples were shipped to a central laboratory; after mixing with a vortex, 7 ml was removed for HPV DNA testing. The remaining 13 ml was sent for routine processing using ThinPrep 2000 Processor (Cytyc Corp). All anal and cervical cytology was reviewed by a single pathologist (Dr Darragh). Diagnosis of atypical cells of undetermined significance (ASCUS), low-grade squamous intraepithelial lesions (LSIL) and high-grade squamous intraepithelial lesions (HSIL) was made according to the Bethesda system for rating cytology [¹⁶] for both cervical and anal samples. For analysis, abnormal cytology was defined as presence of ASCUS, LSIL, or HSIL [¹⁷]. Histology was not available since referral for anoscopy was not available at most sites.

Testing for human papillomavirus

Pellets from centrifuged anal PreservCyt samples were resuspended in 100 μl 50 mmol/l Tris (pH 8.3), 1 mmol/l ethylenediaminetetraacetic acid (pH 8.5), Protease K 200 μg/ml and 0.1% of Laureth 12 and then incubated at 56°C for 2 h and at 95°C for 10 min. A 5 μl sample of the incubation mix was added to a solution containing 500 mmol/l of each of HPV primers (MY09, MY11 and HMB01), 2.5 mmol/l dNTP, 2.5 units Taq polymerase, 500 mmol/l KCl, 40 mmol/l MgCl₂, and 100 mmol/l Tris pH 8.5. The solution also contained primers for the amplification of betaglobin (GH20 and PC04) [¹⁰]. The solution underwent 40 cycles of amplification, and the product was assayed by dot blot analysis for betaglobin and for HPV using a ‘generic probe’ mix that identified over 25 different HPV types and type-specific probes for HPV 6, 11, 42, and 44 (low risk for cancer HPV types), and HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, and 58 (high risk for cancer HPV types). Cervical vaginal lavages samples were processed at the same central laboratory using polymerase chain reaction amplification technique as described in detail previously [¹⁰].

Samples negative for betaglobin were considered inadequate and excluded from analysis. For analysis, HPV was defined by the presence of any HPV (generic probe positive), positive for any of the high-risk types, positive for any low-risk types only and HPV type ‘unknown’ (positive for the generic probe but type not identified).

Statistical methods

Anal HPV infection and abnormal anal cytology are the two outcomes in this analysis. All analysis were performed separately for boys and girls since risk factors were likely to be different for the groups. Risk factors associated with the outcomes were analyzed by both univariate and multivariate methods. Risk factors included HIV status, age at visit, living arrangements, currently smoking, positive chlamydial or gonorrheal test from the urine (boys) or cervix (girls), anal chlamydial infection, cervical HPV infection or SIL (girls only), number of partners in the last 3 months, ever engaged in anal receptive sex, abstinence over the past 3 months, current diagnosis of external genital warts, current diagnosis of perianal warts, sexual orientation (history of same sex behavior), number of sexual lifetime partners (included either gender and oral, anal, and vaginal sex), time since first reported anal or vaginal sex, age of first reported anal or vaginal sex, and age of first reported vaginal intercourse. For those reporting anal intercourse, the following covariates were examined: recent (past 3 months) anal intercourse, number of protected and unprotected anal intercourse events in the past 3 months with the three most recent partners. For the HIV-positive groups, HIV plasma viral load, currently on antiretroviral therapy and CD4 cell counts were also examined.

Potential risk factors were first examined by cross-tabulation. Continuous variables were categorized to avoid extreme values and linear assumptions in the logistic regressive models. Three different tests were applied according to the characteristics of the tables: Pearson's chi-square test, Fisher's exact test, and Cochran–Armitage Trend test [¹⁸]. The multivariable logistic regression models used a backwards stepwise elimination process. All covariates that had a P value of ≤ 0.10 in the univariate analysis were entered into the initial model. The final model included only those covariates with P ≤ 0.05. As a final step, two-way interactions between each pair of significant covariates were added to the model. Since none of the interactions were significant for any of our four analyses, they were removed from the models. Potential interactions between categorical covariates were examined by Zelen's exact homogeneity test [¹⁹].

Results

Characteristics of the cohort

At the time of analysis, 443 adolescents were enrolled in the REACH study and 95 subjects refused anal sampling or had betaglobin-negative samples, resulting in 348 subjects available for the HPV analysis. HIV infection occurred in 58 of the 83 males and 183 of the 265 females. The population was predominantly African-American (71%) or other minority groups (20%) and 81% of the boys and 72% of the girls were ≥ 17 years. The excluded group was more likely to be HIV uninfected (44%) and younger (57% were ≥ 17 years). No differences were found for race/ethnicity, sexual orientation, and gender. Of these 348, 41 had inadequate samples for cytological review, resulting in 307 subjects available for the abnormal anal cytology analysis.

The population of HIV-infected youth was relatively healthy at baseline, with 43% having a CD4 cell count ≥ 500 × 10⁶ cells/l, 44% having a CD4 cell count of 200–499 × 10⁶ cells/l, and only 13% having a CD4 cell count < 200 × 10⁶ cells/l. Girls were more likely to have higher CD4 cell counts than boys: 86 (48.6%) versus 17 (29.3%), respectively, had CD4 cell counts ≥ 500 × 10⁶ cells/l (P = 0.04). Plasma HIV RNA levels < 10 000 copies/ml were found in 63%; girls were more likely to have plasma HIV RNA < 10 000 copies/ml than boys [122 (73%) versus 31 (55.4%), respectively; P = 0.04]. There were no gender differences in antiretroviral therapy, which was being taken by 49%.

Prevalence of anal human papillomavirus infection

Anal HPV was detected in 107 (31%) subjects with adequate anal HPV samples. Gender was significantly associated with anal HPV infection (P = 0.002): 37 (44.6%) of the boys and 70 (26.4%) of the girls were infected. Table 1 gives the distribution of specific HPV types by gender.

Table 1: Human papillomavirus (HPV) types by gender.

There were no differences in specific HPV types or number of HPV types and HIV status. Among the boys with HPV, 14 (38%) had a low-risk type only, 17 (46%) had a high-risk HPV type, of whom nine had mixed high/low-risk types, and six (16%) were HPV type unknown. Among the girls, 16 (23%) had a low-risk HPV type only, 39 (56%) had a high-risk HPV type (of whom only six had a mixed low/high-risk type infection), and 15 (21%) had HPV type unknown. Among the HPV-infected subjects, boys were more likely to have a low-risk HPV type infection than girls, and girls were more likely to have HPV 58. Of the boys with HPV, 23 (62%) had a low-risk HPV infection compared with 22 (31%) of the girls with HPV (P = 0.002); of the girls with HPV, 15 (21%) had HPV 58 compared with one (3%) boy (P = 0.01).

Anal HPV infection was significantly associated with HIV in girls but not in boys. The prevalence of anal HPV infection was 48% (28/58) in HIV-infected boys and 36% (9/25) in HIV-negative boys (P = 0.3). HPV prevalence was 32% (59/183) in HIV-infected girls and 13% (11/82) in HIV-negative girls (P < 0.001).

Risk factors for anal human papillomavirus infection

The risk factors that were statistically significant at the P < 0.1 level (Tables 2 and 3) were included in the multivariate analysis. The multivariate analysis indicated that current diagnosis of anal warts and homosexual or bisexual orientation were independently associated with HPV infection for boys (see Table 4). The HPV type distribution for those with anal warts showed that all eight of the boys with anal warts had low-risk HPV infections; two of the eight had mixed low/high-risk HPV infections. In comparison, of the 29 boys with anal HPV but without anal warts, 28% had low-risk HPV only, 28% had high-risk types only, 24% had mixed low/high-risk types, and 21% had type unknown (Fisher's exact, P = 0.06).

Table 2: Univariate analysis of potential risk factors for human papillomavirus (HPV) infection in males.

Table 3: Univariate analysis of potential risk factors for human papillomavirus (HPV) infection in females.

Table 4: Multivariate analysis for risk factors for anal human papillomavirus (HPV) infection.

For girls in the multivariate analysis, the independent risk factors associated with anal HPV included current diagnosis of anal warts, current diagnosis of external genital warts and cervical HPV infection (see Table 4). Age of first sex, age of first vaginal sex, and cervical squamous cell abnormality were not included in the initial logistic regression model because of the number of missing observations.

The HPV type distribution for girls with external genital warts showed that among the 22 girls with anal or external warts, 12 (55%) had low-risk HPV only, seven (32%) had high-risk HPV only, one (5%) had mixed low/high-risk types and two (9%) had HPV type unknown. In comparison, of the 48 girls with HPV but without anal or external warts, 8% had low-risk HPV only, 54% had high-risk HPV only, 10% had mixed low/high-risk HPV, and 27% had HPV type unknown (Fisher's exact, P < 0.001).

Because of the concern that the presence of anal HPV may be a result of contamination from anal or external genital warts, and not from actual intra-anal infection, the factors for anal HPV infection were re-examined in subjects without visible warts. Again, univariate and multivariate logistic regression models were run on this subpopulation to determine the independently associated risks for HPV, separately for boys and girls. In the multivariate analysis, the significant risk factors for boys without anal warts continued to be homosexual or bisexual orientation. For girls without external or anal warts, positive HIV status was now significant [odds ratio (OR), 2.3; 95% confidence interval (CI), 1.1–4.9; P = 0.03]. Because greater than 10% of the observations had missing data for age of first sex and cervical squamous cytology, these covariates were not included in the model for girls.

In order to assess the association of CD4 cell count, viral load and antiretroviral therapy use with anal HPV infection, a separate analysis was conducted on the HIV-infected girls. The multivariate model indicated that only current diagnosis of genital warts (OR, 3.9; 95% CI, 1.3–11.9; P = 0.02), current diagnosis of anal warts (OR, 23.0; 95% CI, 2.6–202; P = 0.005), and CD4 cell count (CD4 cell count < 200 versus ≥ 500 × 10⁶ cells/l: OR, 9.4; 95% CI, 2.8–31.3; P < 0.001; for CD4 cell count 200–499 versus ≥ 500 × 10⁶ cells/l: OR, 3.3; 95% CI, 1.5–7.3; P = 0.003) were independently associated with anal HPV infection in this population. This analysis was not repeated for boys since HIV status was not associated with anal HPV for boys in any of the models.

Prevalence of abnormal anal cytology

Of the 307 subjects with adequate anal cytology and anal HPV DNA results available, abnormal anal cytology was diagnosed in 70 (22.8%) subjects. Gender was significantly associated with abnormal anal cytology, with 41.6% of the boys and 16.5% of the girls diagnosed with abnormal anal cytology (P = 0.001). However, the distribution of the grade of abnormal anal cytology did not differ significantly by gender. Among the 32 boys with abnormal anal cytology, seven (21.9%) had ASCUS, 20 (62.5%) had LSIL, and five (15.6%) reported HSIL. Among the 38 girls with abnormal anal cytology, 18 (47.4%) had ASCUS, 16 (42.1%) had LSIL, and four (10.5%) had HSIL.

Boys with only low-risk HPV type infections or with mixed low/high-risk HPV infections had similar rates of abnormal anal cytology (79% and 78%, respectively). Abnormal anal cytology was present in 50% of the boys with only high-risk HPV type infections. The abnormal anal cytology rate was only 17% in boys without HPV infection.

Girls had similar abnormal anal cytology rates (80%) for low-risk only HPV infections, as was seen in the boys. Subjects with mixed low/high-risk HPV types or only high-risk HPV type infections had similar rates of abnormal anal cytology (50% and 45%, respectively). Girls without HPV infection had an abnormal anal cytology rate of only 5%. Figure 1 shows the association between anal HPV infection and cytological diagnosis for all subjects.

Fig. 1.:  Distribution of human papillomavirus (HPV) by cytology. NI, normal cytology; ASCUS, atypical cells of undetermined significance; LSIL, low-grade squamous intraepithelial lesions; HSIL, high-grade squamous intraepithelial lesions.

Risk factors for abnormal anal cytology

Abnormal anal cytology was significantly associated with HIV status in the univariate analysis for both boys and girls: 28 (52.5%) of the HIV-infected boys had abnormal cytology compared with four (16.7%) of uninfected boys (P = 0.003). Abnormal anal cytology was seen in 34 (21.3%) of the HIV-infected girls and four (5.7%) of the uninfected girls. Although the prevalence of abnormal anal cytology was higher in the former group, the distribution of grade of abnormal anal cytology was not significantly altered by HIV status.

The multivariate analysis using logistic regression indicated that anal HPV infection and HIV status were independent predictors for abnormal anal cytology in boys (see Table 5). Although HIV status was significant in the multivariate model, univariate findings indicated that the covariates related to HIV status, such as CD4 cell count, viral load, and antiretroviral therapy use, were not associated with abnormal anal cytology; therefore, a multivariate subanalysis for HIV positive boys was not performed.

Table 5: Multivariate analysis for risk factors for abnormal anal cytology.

For girls, the independent risk factors in the multivariate analysis associated with abnormal anal cytology included anal HPV infection and number of partners in the past 3 months (see Table 5). As in the HPV analysis for girls, cervical squamous cell abnormalities and number of unprotected vaginal sex events were not included in the initial logistic regression model because of the number of missing observations.

Since anal warts were found to be associated with abnormal anal cytology in the univariate analysis among girls, the analysis was repeated for those girls without warts. The regression model indicated the same covariates were independently associated with abnormal anal cytology in girls without warts as in the total female group. The analysis was not repeated for boys without warts since no univariate associations were found between anal warts and abnormal anal cytology in the boys.

Discussion

Several studies have indicated that anal HPV infection, anal receptive sex, and HIV status are significant risks for anal cancer; these observations support the initiation of anal cytology screening programs for cancer and precancer detection in homosexual men [^4,8,20]. Our findings demonstrate that sexually active high-risk adolescents are at significant risk for anal HPV infections and abnormal anal cytology, suggesting that HIV-infected adolescents should be included in screening programs. In our study of HIV-infected and uninfected high-risk youth, we found that almost half of the boys and a quarter of girls had anal HPV infection. Even more alarming, almost one third of the boys and one tenth of the girls had abnormal anal cytology. As previously found in adults, anal HPV infection and abnormal anal cytology were highest among boys with same sex behaviors. In comparison, the lowest rates for both were found in HIV-uninfected adolescent girls.

Although adolescents have the highest rates of many sexually transmitted infections [²¹], high rates of anal infections have not been reported, partly because they have not been studied. Studies of cervical HPV show that the highest rates of HPV are seen in adolescents. In addition, these studies of cervical HPV infection have consistently shown that sexual intercourse is the primary risk associated with infection [⁹]. The association with homosexual/bisexual orientation in the boys supports these findings for anal infection in boys. In contrast, anal sex was not a significant risk in girls. The lack of significance with anal intercourse may have been because of the lack of power in the analysis since few girls reported anal intercourse. However, the data described here and in other studies [¹³] suggest that anal infection in women can occur without direct sexual transmission. The association between cervical HPV infection and anal HPV indicates that HPV shed in vaginal discharge may be a risk for transmission to the anus owing to the close proximity of the two. The association with HIV infection and low CD4 cell count in girls without warts underscores the importance of these two factors in acquisition of anal HPV in girls; by comparison, the lack of association of these factors in boys underscores the strong risk of homosexual behavior. This lack of association with HIV infection and low CD4 cell count in the boys contrasts with adult studies [^6,22–24].

Palefsky et al. [²³] found that over 90% of HIV-infected homosexual adult men, and over 60% of the uninfected homosexual men, had anal HPV; however, essentially all of the HIV-infected men with low CD4 cell counts had anal HPV. The small number of HIV-infected boys in our group precludes any generalization but the association with behavior in this group was strong.

In comparison with homosexual men, little information is available for rates of anal HPV infection in women and heterosexual men. Van Doornum et al. [²⁵] examined non-HIV-infected heterosexual partners from a clinic for sexually transmitted diseases, who were recruited as part of an HIV transmission study, and found that only 1.2% of the men and 2.5% of the women had an anal HPV infection. Our higher rate may be related again to young age as found for cervical infection, or it may be because of the reluctance of our boys and girls to reveal anal intercourse behavior. Too few heterosexual boys were HPV infected for further analysis.

The association with perianal warts and HPV in both genders was not unexpected since HPV is the cause of perianal warts. However, the question arises whether the intra-anal samples reflect contamination during specimen collection or actual intra-anal infection. Most studies have found that anogenital warts are a significant risk for intra-anal precancers, as seen in our analysis for the girls [²⁴]. Whether the detection of HPV in our study was contamination or evidence of an intra-anal infection remains to be determined.

The rate of abnormal anal cytology, particularly among the boys, was also surprisingly high given the relatively healthy nature of the HIV-infected adolescents and the recent history of onset of sexual activity. Not surprisingly, anal HPV was the most important risk for abnormal anal cytology in both boys and girls. The strong association between low-risk HPV type infections and abnormal cytology group in the girls appeared to be driven by the presence of perianal warts. However, the rates of low-risk HPV types among the boys with abnormal anal cytology were significantly high even among those without perianal warts. This finding is in contrast to adult studies, where over 90% of men with abnormal anal cytology have high-risk HPV types [^5,6,20]. Our findings may suggest that early on in HIV and HPV infection, low-risk HPV types are common and just as likely as high-risk HPV types to cause abnormal cytology in the anal canal. Since low-risk HPV types are likely to result in the benign changes of ASCUS and LSIL, this finding is also consistent with the relatively large number of ASCUS and LSIL cytological diagnoses found in our study. Unfortunately, histological verification was not available for the majority of adolescents because of poor compliance and the lack of availability for anoscopy.

Unlike anal HPV, HIV status was an important risk for abnormal anal cytology in boys. The lack of association with CD4 cell count among boys with HIV was most likely a consequence of the small number of subjects with a CD4 cell count < 200 × 10⁶ cells/l. However, we recently reported a similar lack of association with low CD4 cell count in the REACH cohort for cervical SIL in girls [¹⁰]. These studies suggest that, early on in HIV infection, HIV may interact with HPV in ways independent of CD4 cell immunosuppression [^26–28]. In contrast, HIV status was not an important risk in girls in the final multivariate model, suggesting that the risk for abnormal anal cytology in girls was directly related to the presence of anal HPV; this, in turn, was related to HIV status or presence of HPV infections at other sites (i.e., external genital warts or cervical infection). Similar rates of abnormal cytology have been reported in adult women. A recent study by Holly et al. [²⁹]. reported abnormal cytology (including ASCUS) of the anus in 26% of HIV-infected women and 8% of high-risk but uninfected controls.

The clinical significance of abnormal anal cytology in adolescents and adults remains unknown. The rate of anal cancer is substantially lower than that of cervical cancer. In the few studies available for comparisons of abnormal anal cytology in HIV-uninfected groups, the low rates found in our study are similar to those reported in other sexually active groups [^12,13]. However, the reported escalation of anal cancers in HIV-infected men [^5,23,24] strongly suggest that finding of SIL in the anus of HIV-infected persons is clinically significant. In a recent study, 62% of LSIL was found to progress to HSIL in HIV-infected homosexual men. Although one of the limitations of our study was the lack of anoscopy and histology, the specificity of abnormal cytology is expected to be high [¹⁷]. Hence, our study most likely underdiagnosed anal disease rather than overdiagnosed. Continued observation of this cohort and histological verification of abnormal cytology will be essential. If the natural history of anal HPV infections were similar to the cervix, then it is apparent that screening practices for anal cancers and precancers should be based on similar models to those practiced for cervical screening (i.e., cytology screening for precancer lesions). Recent cost-effective analysis showed that screening for anal cancer in homosexual HIV-infected men is worthwhile [⁸]. Our data support this screening recommendation for adolescent males. The recommendation for women is less clear.

In summary, our data show that adolescents early in their course of HIV and HPV infections are at significant risk of anal HPV and abnormal anal cytology. Risks for anal HPV and abnormal anal cytology differ between boys and girls, underscoring the importance of examining these groups individually. The high rates of abnormal anal cytology in our study suggest anal screening should be considered among HIV-infected homosexual or bisexual adolescent boys.

Acknowledgements

The following investigators, listed in order of the numbers of subjects enrolled, are participating in this study. University of Miami: L. Friedman, L. Pall, D. Maturo, A. Pasquale; Montefiore Medical Center: D. Futterman, D. Monte, M. Alovera-DeBellis, N. Hoffman, S. Jackson; University of Pennsylvania and the Children's Hospital of Philadelphia: D. Schwarz, B. Rudy; Children's Hospital of Philadelphia: M. Tanney, A. Feldman; Children's Hospital of Los Angeles: M. Belzer, D. Tucker, C. Hosmer, K. Chung; Tulane Medical Center: S. E. Abdalian, L. Green, C. McKendall, L. Wenthold; Children's National Medical Center: L. J. D'Angelo, C. Trexler, C. Townsend-Akpan, R. Hagler, J. A. Morrissy; University of Maryland: L. Peralta, C. Ryder, S. Miller, S. Calianno; Cook County Hospital/University of Chicago: L. Henry-Reid, R. Camacho; Children's Hospital, Birmingham: M. Sturdevant, A. Howell, J. E. Johnson; Children's Diagnostic and Treatment Center: A. Puga, D. Cruz, P. McLendon; Emory University: M. Sawyer, J. Tigner, A. Simmonds; St. Jude Children's Research Hospital: P. Flynn, K. Lett, J. Dewey, S. Discenza; Mt. Sinai Medical Center: L. Levin, M. Geiger; University of Medicine and Dentistry of New Jersey: P. Stanford, F. Briggs; SUNY Health Science Center at Brooklyn: J. Birnbaum, M. Ramnarine, V. Guarino.

The following investigators have been responsible for the basic science agenda: C. Holland (Center for Virology, Immunology, and Infectious Disease, Children's Research Institute, Children's National Medical Center), A. B. Moscicki (University of California at San Francisco), D. A. Murphy (University of California at Los Angeles), S. H. Vermund (University of Alabama at Birmingham), P. Crowley-Nowick (The Fearing Laboratory, Brigham and Women's Hospital, Harvard Medical School, Boston), S. D. Douglas (University of Pennsylvania and the Children's Hospital of Philadelphia).

Network operations and analytic support are provided by C. M. Wilson, C. Partlow (University of Alabama at Birmingham), S. J. Durako, J. H. Ellenberg, B. Hobbs, A. Bennett, M. Camarca, K. Clingan, J. Houser, V. Junankar, O. Leytush, L. Muenz, Y. Ma, R. Mitchell, T. Myint, P. Ohan, L. Paolinelli, M. Rakheja, M. Sarr, A. Soloviov (Westat, Inc.).

Staff from sponsoring agencies include A. Rogers, A. Willoughby, (National Institute of Child Health and Human Development), K. Davenny, V. Smeriglio (National Institutes on Drug Abuse, Allergy and Infectious Diseases), E. Matzen, (National Institutes on AIDS), B. Vitiello (National Institute of Mental Health).

The investigators are grateful to the members of the Community Advisory Board for their insight and counsel and particularly indebted to the youth who made this study happen.

Sponsorship: The Adolescent Medicine HIV/AIDS Research Network is supported by the National Institute of Child Health and Human Development, with supplemental funding from the National Institutes on Drug Abuse, Allergy and Infectious Diseases, and Mental Health and the Health Resources and Services Administration. This work was also supported in part by NCI R01 CA 51323.We would also like to thank Cytyc Corporation for providing supplies for the ThinPrep.