Abstract

Little is known about the epidemiology of anal human papillomavirus (HPV) infection in women. We studied 251 human immunodeficiency virus (HIV)—positive and 68 HIV-negative women for the presence of anal HPV by use of polymerase chain reaction (PCR) and hybrid capture. Medical and behavioral risk factors were evaluated; 76% of HIV-positive and 42% of HIV-negative women were found to have anal HPV DNA via analysis by PCR (relative risk [RR], 1.8; 95% confidence interval [CI], 1.3–2.5). Among 200 women for whom there were concurrent anal and cervical HPV data, anal HPV was more common than cervical HPV in both HIV-positive (79% vs. 53%) and HIV-negative women (43% vs. 24%). By multivariate analysis of HIV-positive women, CD4+ cell counts ⩽200 cells/mm3, compared with counts >500 cells/mm3 (RR, 1.4; 95% CI, 1.1–1.5), and cervical HPV infection (RR, 1.3; 95% CI, 1.1–1.4) were associated with anal HPV infection. Women >45 years old had reduced risk, compared with women <36 years old (RR, 0.80; 95% CI, 0.50–0.99), as did African American women (RR, 0.86; 95% CI, 0.72–1.0), compared with white women. Anal HPV infection is underrecognized in HIV-positive and high-risk HIV-negative women.

Substantial evidence has accumulated that links infection with human papillomavirus (HPV) to the development of cervical cancer. HPV 16, 18, 31, 33, and 45 are most strongly associated with invasive squamous cell cancer of the cervix, and HPV 16 is found to be the most common of these in most studies [1, 2]. Several studies also have shown a strong association between anal cancer and HPV of the same types as those associated with cervical cancer [3–6]. Likewise, there is a strong relationship between HPV infection and anal squamous intraepithelial lesions (ASILs), the putative anal cancer precursor, again with the same spectrum of HPV types as those found in the cervix [3, 7–12].

The incidence of anal cancer is increasing in the United States among both men and women, but the condition is about twice as common in women as it is among men [13, 14]. However, this ratio may be changing with the high incidence of anal cancer among men who have sex with men (MSM), particularly those MSM who are infected with human immunodeficiency virus (HIV). Since 1960, the incidence of anal cancer in Connecticut increased ∼2-fold among men and 2.3-fold among women. The incidence was lowest among white men (1973–1989 average, 0.41 per 100,000 population) and highest among African American women (1973–1989 average, 0.74 per 100,000 population) [15]. Anal cancer is most common among men with a history of receptive anal intercourse, with an estimated incidence as high as 36 per 100,000 population. Thus the incidence of anal cancer in this group of men is ∼5 times higher than the incidence of cervical cancer among women in the United States [16]. Analysis of data suggests that the incidence of anal cancer among HIV-positive MSM may be even higher than that among HIV-negative MSM. During the early post-AIDS period of 1985–1989, the incidence of anal cancer among people with AIDS was reported to be approximately twice that of people of the same age, race, and sex distribution in the early pre-AIDS period 1975–1979 [17]. There are no data yet on the incidence of anal cancer among HIV-positive women or HIV-negative women with a history of high-risk sexual behaviors.

Several studies have shown that HIV-positive women are at increased risk of cervical HPV infection, compared with HIV-negative women [18–24]. Consistent with increased prevalence of cervical HPV infection in HIV-positive women, multiple studies have documented increased prevalence of cervical squamous intraepithelial lesions (CSILs) in HIV-positive women, compared with HIV-negative women [18, 25–29]. In a meta-analysis of results published between 1986 and 1990, the relative risk of CSILs in HIV-positive women was ∼5 times that of HIV-negative women [30]. However, the computation of these risks was not adjusted for HPV infection. In the Women's Interagency HIV Study (WIHS) study, a natural history cohort study of HIV-positive and comparable HIV-negative adult women, cervical cytology was abnormal for 38% of HIV-positive women and for 16% of the HIV-negative women [31]. Similarly, the prevalence of abnormal cervical cytology was recently shown to be higher among HIV-positive adolescent girls (61%) than among HIV-negative adolescent girls (26%) [32]. The risk factors for abnormal cervical cytology identified in the WIHS study by multivariate analysis included HIV infection, lower CD4+ cell count, higher plasma HIV RNA level, the presence of HPV DNA in cervicovaginal lavage specimens, a history of abnormal cytology, and number of male sex partners within 6 months of enrollment.

Compared with cervical HPV infection and CSILs, little is known about anal HPV infection and ASILs in HIV-positive and high-risk HIV-negative women. In one report, anal HPV infection was associated with HIV infection (odds ratio [OR], 2.6; 95% confidence interval [CI], 1.03–6.8) and was twice as frequent as cervical HPV infection [18]. Fifteen (14%) of 109 women in that study had abnormal anal cytology. The higher prevalence of anal versus cervical HPV infection in HIV-positive women was subsequently confirmed in a second study of 151 women [33]. To our knowledge, to date, there have been no studies documenting the prevalence of specific HPV types in HIV-positive or HIV-negative women and how these types may differ from those found in the cervix. Further, to our knowledge, no data have been published on risk factors for anal HPV infection in women. The aim of this study was to perform a detailed analysis of anal HPV infection in HIV-positive and high-risk HIV-negative women and to characterize risk factors for HPV infection in these groups.

Subjects and Methods

Participants in this study were recruited between November 1995 and January 1997 from women who were participating at the San Francisco Bay Area site of the multicenter WIHS, a natural-history cohort study of HIV disease in women. The methods for this cohort study have been summarized in greater detail elsewhere [34]. A total of 251 HIV-positive and 68 HIV-negative women enrolled in the study of anal HPV. The HIV-negative women were recruited to the WIHS study on the basis of characteristics (age, HIV risk factors, and other demographics) that would render them similar to the HIV-positive women in the WIHS study population. On the basis of prevalence of injection drug use and sexual risk behaviors, the HIV-negative WIHS participants therefore constituted a group that shared high risk characteristics for acquiring HIV. During each WIHS visit, patients were interviewed by trained staff to obtain information on behavioral and medical history. Cervicovaginal lavage samples (using 10 mL of sterile saline) for HPV testing were collected; blood was collected to assess HIV serology, plasma HIV RNA level, and lymphocyte subsets. HIV status was determined for all subjects with ELISA. All positive results of ELISA were confirmed with a Western blot assay. Lymphocyte subsets were measured with standardized 2- or 3-color fluorescence methods, and HIV virus load was measured with a nucleic acid sequence-based amplification assay (NASBA; Organon Teknika) in laboratories participating in National Institute of Allergy and Infectious Diseases—sponsored quality assurance programs [34].

During regular WIHS visits, women who consented to participate in the study of anal HPV had samples taken for cytology and HPV tests. A brief additional questionnaire was administered that covered medical history and sexual practices specific to the study. To obtain samples for cytology and testing for HPV, 2 consecutive Dacron swabs moistened with tap water were inserted into the anal canal. The first was used for cytology, as described elsewhere [35]. The second swab for HPV testing was inserted into a 1-mL vial of sample transport medium (Digene Diagnostics) and was frozen at −70°C until analysis.

Because of its high sensitivity, polymerase chain reaction (PCR) was used in this study to detect low-level HPV infection, as described elsewhere [9]. Because a positive PCR result does not discriminate between low-level and high-level infection, hybrid capture (HC; Hybrid Capture II, Digene Diagnostics), a non—amplification-based test, was used to indicate quantity of HPV DNA present in specimens.

Two hundred microliters of specimen in sample transport medium was used for PCR, and the remainder was used for HC. PCR was performed with MY09/MY11 consensus HPV L1 primers, as well as primers for amplification of the human β-globin as an indicator of specimen adequacy, as described elsewhere [12]. After 30 amplification cycles, specimens were probed with a biotin-labeled HPV L1 consensus probe mixture. A separate membrane was probed with a biotin-labeled probe to the human β-globin gene. Specimens were then studied to determine the specific HPV type by probing with probes to 29 different HPV types (6, 11, 16, 18, 26, 31, 32, 33, 35, 39, 40, 45, 51, 52, 53, 54, 55, 56, 58, 59, 61, 66, 68, 69, 70, 73, AE2, Pap 155, and Pap 291, as well as the following 10 types together in a probe mixture: HPV 2, 13, 34, 42, 57, 62, 64, 67, 72, and W13B). AE2, Pap 155, Pap 291, and W13B are recently recognized HPV types, in which a new HPV type is defined as one in which the L1 open reading frame of the putative new type has <90% homology to its closest related type [36]. Specimens negative for β-globin gene amplification were excluded from analysis. The results of PCR were recorded on a 0–5 scale, rather than as positive or negative. The strength of the PCR signal was recorded on a scale from 0 (negative) to 5, on the basis of the intensity of the signal on the dot blots, as described elsewhere [37]. Cervical HPV infection was evaluated, as described elsewhere, with PCR analysis of a cervicovaginal lavage sample obtained at the same visit as the anal specimen [23].

HC was conducted on the anal specimens, according to the manufacturer's recommendations. 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 by averaging control samples containing 1 pg/mL of HPV 16 DNA and 1 pg/mL of HPV 11 DNA. The magnitude of the rlu ratio increases with increasing quantity of HPV DNA in the specimen. The rlu ratio was computed and classified as negative (<1.0) or positive (⩾1.0); for some analyses, positive specimens were further categorized as 1.0 to <10, 10–100, and >100.

Data analysis

SAS software (version 6.12; SAS) was used for data analysis. In analyses of the presence or absence of HPV, all PCR test results were regarded as negative if the level was equal to 0. The test result for the consensus probe was considered to be positive if the level was ⩾1. Test results for the individual types were considered to be equivocal if the level was 1 and positive if the level was ⩾2. Unless otherwise stated, in analyses evaluating the presence versus absence of HPV infection, equivocal results were treated as negative. The HC test for HPV infection was considered to be positive if the rlu ratio was ⩾1.0 and negative if it was <1.0. For some analyses, HIV-positive subjects were stratified by CD4+ cell counts into 3 groups (>500, >200–500, and ⩽200 cells/mm3) and by HIV RNA level into 4 groups (<4000, 4000–20,000, >20,000–100,000, and >100,000 copies/mL). These groups were identical to those used in analysis of cervical HPV HIV RNA infection in previous analyses of findings from the entire WIHS cohort [23]. The threshold for HIV RNA quantitation was 4000 copies/mL for the NASBA assay used to generate WIHS data. The other 2 RNA cutpoints were chosen a priori to approximate the median and 75th percentiles (20,000 and 100,000 copies/mL). Relative risks (RRs) and 95% CIs were used to evaluate potential risk factors for HPV infection. The Mantel-Haenszel χ2 test was used to test for trends in multilevel risk factors. The Mantel-Haenszel procedure was used to adjust RRs for 1 or 2 other risk factors. In addition, logistic regression was used to evaluate models involving >3 risk factors simultaneously. The Zhang and Yu correction was applied to the adjusted ORs from logistic regression to approximate the adjusted RRs [38].

In this evaluation, HIV-negative subjects were compared with HIV-positive subjects as a group and with HIV-positive subjects stratified by HIV RNA level and by CD4+ cell count. The interviews, including those administered as part of the WIHS study, were used to determine exposure to potential medical and behavioral risk factors for HIV-positive women only. For these risk factors, HPV infection in women who reported exposure to the risk factor was compared with that of those who did not report exposure.

The level of consensus probe, number of specific types detected by PCR, and the HC rlu ratio were used as measures of the amount of HPV detected in the anal canal. The t test allowing for unequal variance was used to compare the amount of HPV detected in HIV-positive and HIV-negative women. Analysis of variance and the least significant difference were used to compare the amount of HPV detected in the anal samples of HIV-positive women stratified by CD4+ cell counts and stratified by HIV virus load. The rlu ratio was logarithmically transformed before analysis, and the means were transformed back to their original scale after analysis.

PCR test results for cervical HPV were available from the same visit for women who enrolled in the study of anal HPV at their 6-or 12-month WIHS visit. The χ2 test was used to test for associations between HPV in the anal and cervical samples. The Cochran-Mantel-Haenszel statistic was used to control for CD4+ cell counts in HIV-positive women. The similarity between measures of amount of HPV detected in the anal and cervical samples was examined with Spearman correlations. Similarity of specific types found in the anal and cervical samples was examined in women who were positive for HPV in both samples.

Results

At the San Francisco Bay Area site of the WIHS cohort, 319 of 422 women agreed to participate in this study. The mean age of the women at the time of entry into the anal study was 40 years (range, 20–61 years); 31% had <12 years of education, 40% had a high-school diploma or general equivalency degree, and 29% had at least some college education. Sixty percent of the women were African American, 25% were non-Hispanic white, 12% were Hispanic, and 4% were of other or mixed races.

The distributions of socioeconomic and demographic characteristics were similar to those reported for the WIHS study [34], except for older age and fewer women of Hispanic origin among the study participants. However, the subjects in this study reflect the demographics of the San Francisco Bay Area site of the WIHS cohort study with respect to these characteristics. The median age for the women in all the geographic sites at the WIHS baseline visit was 36 years for HIV-positive WIHS participants and 34 years for HIV-negative WIHS participants [34]. In contrast, the median age at the San Francisco WIHS baseline visit (6–18 months before the baseline visit for the study of anal HPV) was 38 years for all women in the San Francisco Bay Area cohort and 39 years for the participants in the study of anal HPV. Overall, 24% of all WIHS participants were Hispanic, but 13% of the San Francisco Bay Area cohort and 12% of the participants in this study were Hispanic.

Anal HPV infection

An anal sample was available for PCR analysis for 306 of 319 women who enrolled in the study. Of the 306 samples tested, 26 (8.5%) were negative for β-globin and were excluded from analysis. β-Globin negativity was somewhat more common among HIV-negative women (14%) than among HIV-positive women (7%; P = .09). HPV DNA, as determined by consensus probes, was detected in the anal samples of 170 (76%) of 223 HIV-positive women and 24 (42%) of 57 HIV-negative women (P<.001). In addition, >1 of the specific HPV types were found in 5 of the HIV-positive and 2 of the HIV-negative women who had negative test results via the consensus probe.

Table 1 shows the distributions of specific HPV types found in HIV-positive women and in HIV-negative women. Of the specific types found in HIV-positive women, the most frequently detected was HPV 16 (15% of HPV-positive samples). This was followed by types 58, 53, 61, 70, Pap 155, 18, and Pap 291, which were found in ⩾5% of the HPV-positive samples. In addition, 59 (35%) of 170 HIV-positive women who had detectable anal HPV DNA had an unknown HPV type. Forty-nine women had negative test results for all specific types, and 10 had equivocal results for specific types. Only HPV types 18, 53, 54, and 70 were found in ⩾2 HIV-negative women. Seventeen (71%) of 24 HIV-negative women with detectable anal HPV had an unknown HPV type, 15 were negative for all specific types, and 2 had equivocal results. Multiple HPV types were detected in 47 (28%) of 170 HIV-positive women and 2 (8%) of 24 HIV-negative women with detectable anal HPV (P = .04). By use of HC, HPV was detected in 182 (75%) of 242 anal samples available from HIV-positive women and in 20 (30%) of 67 samples from HIV-negative women (P< .001). Among HIV-positive women for whom both PCR and HC results were available for analysis, HPV was detected in 22 samples by PCR only and in 22 samples by HC only. Among HIV-negative women, HPV was detected in 8 samples by PCR only and in 2 samples by HC only.

Table 1

Anal human papilloma virus (HPV) infection and distribution of HPV types in human immunodeficiency virus-positive (HIV+) and HIV-negative (HIV) women with anal HPV infection, as measured by polymerase chain reaction (PCR).

Table 1

Anal human papilloma virus (HPV) infection and distribution of HPV types in human immunodeficiency virus-positive (HIV+) and HIV-negative (HIV) women with anal HPV infection, as measured by polymerase chain reaction (PCR).

HIV infection was associated with an increased risk of anal HPV infection as detected by PCR (RR, 1.8) or by HC (RR, 2.5; table 2). Detection of HPV was inversely associated with CD4+ cell counts for either method of HPV detection. However, there was no association between plasma HIV RNA level and detection of anal HPV.

Table 2

Anal human papilloma virus (HPV) infection detected by polymerase chain reaction (PCR) and by hybrid capture (HC), stratified by human immunodeficiency virus (HIV) status, CD4+ cell count, and HIV virus load (VL).

Table 2

Anal human papilloma virus (HPV) infection detected by polymerase chain reaction (PCR) and by hybrid capture (HC), stratified by human immunodeficiency virus (HIV) status, CD4+ cell count, and HIV virus load (VL).

The relationships between the quantity of HPV DNA found among HPV-positive women and HIV status, CD4+ cell counts, and HIV RNA level were also examined (table 3). The mean level of the consensus probe intensity detected by PCR, mean number of specific HPV types detected by PCR, mean sum of the levels of the PCR tests for specific types, and mean HC rlu ratio were higher in HIV-positive women than in HIV-negative women. Among HIV-positive women, there was a trend toward increases in the number of specific types and HC rlu ratio as CD4+ cell counts decreased, but this trend was not observed in the amount of HPV DNA detected with the PCR consensus probe. There was no association between HIV RNA level and any of the measures of the amount of HPV DNA.

Table 3

Comparisons of mean level of consensus probe by polymerase chain reaction (PCR), mean number of specific human papilloma virus (HPV) types by PCR, and mean hybrid capture (HC) relative light unit (rlu) ratio level, stratified by human immunodeficiency virus (HIV) status, CD4+ cell count, and HIV RNA level in women who were positive for HPV by either PCR or HC.

Table 3

Comparisons of mean level of consensus probe by polymerase chain reaction (PCR), mean number of specific human papilloma virus (HPV) types by PCR, and mean hybrid capture (HC) relative light unit (rlu) ratio level, stratified by human immunodeficiency virus (HIV) status, CD4+ cell count, and HIV RNA level in women who were positive for HPV by either PCR or HC.

Risk factors for anal HPV infection among HIV-positive women

Lifetime and recent exposure to potential risk factors for detection of anal HPV DNA by HC were evaluated in HIV-positive women (table 4). Potential risk factors were not evaluated for HIV-negative women because of the small number of subjects in this group. In univariate analysis, age, race, lifetime history of some AIDS-related diseases (herpes zoster episodes, Pneumocystis carinii pneumonia, and being diagnosed with AIDS), lifetime use of several medications (zidovudine, didanosine, zalcitabine, stavudine, Bactrim/Septra, ketoconazole, fluconazole, and lamivudine), and use in the last 6 months of several medications (zidovudine, Bactrim/Septra, dapsone, fluconazole, and lamivudine) showed increased risk for anal HPV detection. However, after adjustment for CD4+ cell count, only the lifetime use of zidovudine, being white, and age <36 years remained significantly associated with detection of HPV DNA in anal specimens.

Table 4

Univariate and adjusted relative risk (RR) of anal human papilloma virus (HPV) infection detected by hybrid capture for various risk factors in human immunodeficiency virus (HIV)—positive women.

Table 4

Univariate and adjusted relative risk (RR) of anal human papilloma virus (HPV) infection detected by hybrid capture for various risk factors in human immunodeficiency virus (HIV)—positive women.

Other potential risk factors were evaluated in univariate analysis and were found not to be significantly associated with detection of anal HPV among HIV-positive women. These included the following: history of genital herpes; gonorrhea; syphilis; chlamydial infection; trichomonas; bacterial vaginosis or vaginal candidiasis; engaging in anal intercourse; insertion of objects into the anus; oral sexual activities; cigarette smoking; use of alcohol, marijuana, amphetamines, cocaine, or injection drugs; history of hemorrhoids; anal fissures or fistulas; blood in stool; cancer of any type; pregnancy; use of oral contraceptives; history of oral candida; skin rashes; pneumonia other than P. carinii pneumonia; wasting syndrome; weight loss or night sweats; lifetime therapeutic use of stavudine, dapsone, or pentamidine; current therapeutic use of didanosine, zalcitabine, stavudine, ketoconazole, or pentamidine; marital status; or sex of sex partners.

To study the independent relationships of detection of anal HPV by PCR with other factors, we developed a multivariate model of risk factors that included CD4+ cell count, race, age, and presence of cervical HPV (table 5). These models demonstrated that having a CD4+ count ⩽200 cells/mm3 and the presence of cervical HPV were independent risk factors for the detection of anal HPV. Risk for anal HPV infection was significantly lower for races other than white and the oldest age group.

Table 5

Adjusted relative risks (RRs) in a multivariate model of anal human papilloma virus (HPV) infection, detected by hybrid capture in human immunodeficiency virus (HIV)—positive women.

Table 5

Adjusted relative risks (RRs) in a multivariate model of anal human papilloma virus (HPV) infection, detected by hybrid capture in human immunodeficiency virus (HIV)—positive women.

Comparisons with cervical HPV infection

Cervical samples that were obtained at the same visit as the baseline anal visit were available for HPV testing with PCR. Among the 200 HIV-positive women for whom simultaneous anal and cervical specimens were analyzed, 79% had anal HPV infection, compared with 53% who had cervical HPV infection (table 6). Among the 49 HIV-negative women for whom simultaneous anal and cervical specimens were analyzed, 43% had anal HPV infection, compared with 24% who had cervical HPV infection. Detection of HPV in one sample was associated with detection in the other sample for HIV-positive women, even after controlling for CD4+ cell counts (P<.0001). HIV-negative women also showed an association between HPV infection in the 2 samples, but the number of subjects in this analysis was small (P = .06).

Table 6

Any type of human papilloma virus (HPV) detected by polymerase chain reaction (PCR) in anal samples versus cervical samples for human immunodeficiency virus (HIV)—positive and HIV-negative women.

Table 6

Any type of human papilloma virus (HPV) detected by polymerase chain reaction (PCR) in anal samples versus cervical samples for human immunodeficiency virus (HIV)—positive and HIV-negative women.

Data from the HIV-positive and HIV-negative women were combined to compare specific HPV types in cervical and anal samples in individual patients. Overall, the spectrum of HPV types found in the anus and cervix was similar, except for HPV type 56, which was found in 5% of the HPV-positive cervical samples but in none of the anal samples. The correlation between the number of specific types detected in the anal and cervical samples was small (r = .15) but different from 0 (P = .03). In 7 (7%) of 106 women for whom HPV was detected in both samples, no specific types were detected in either—that is, both specimens were positive with the consensus HPV probes but negative for all specific HPV types. Exactly the same specific types were found in both samples for 7 (7%) women. None of the same specific types were found in the anal and cervical samples for 63 (64%) of 99 women for whom any specific type was detected. A single identical HPV type was found in both samples in 26 (26%) women. Two identical specific types were found in both samples in 8 (8%) women, and 3 identical types were found in 2 (2%) women.

Discussion

This study of anal HPV infection in women is the largest reported to date. The study also provides the first report of direct comparisons between anal and cervical HPV infection. Overall, 76% of HIV-positive and 42% of high-risk HIV-negative women had anal HPV infection, and analysis of our data confirms the findings of 2 smaller studies that anal HPV infection is more common than cervical HPV infection in both HIV-infected and high-risk HIV-uninfected women [18, 33].

Via univariate analysis, risk factors for anal HPV infection among HIV-positive women included lower CD4+ cell count, the presence of cervical HPV, younger age, and being non-Hispanic white, compared with being African American or of other racial background. History of receptive anal intercourse was not associated with anal HPV infection, a finding similar to that of Williams et al. [18]. The absence of a significant risk associated with a history of anal intercourse suggests that other means of acquisition may occur. In multivariate analysis, having a lower CD4+ cell count, the presence of cervical HPV, younger age, and being white persisted as independent risk factors for anal HPV infection.

In a study of risk factors for cervical HPV infection in the WIHS cohort, both lower CD4+ cell count and higher HIV RNA level were associated with cervical HPV infection [23]. In that study, other risk factors for cervical HPV infection included being African American, compared with being non-Hispanic white, current smoking status, and younger age. In this study of anal HPV infection, younger age was also a risk factor for anal HPV infection, but current smoking status was not associated with anal HPV infection, and being nonwhite was protective. Differences in risk factors may reflect differences in the biology of HPV infection in the anal canal versus the cervix but may also reflect the smaller sample size of the study of anal HPV, compared with the study of cervical HPV.

HIV positivity and lower CD4+ cell count, but not HIV virus load, were associated with increased risk of anal HPV infection, higher number of detectable HPV types with PCR, and higher level of HPV DNA, as determined by the HC rlu ratio. A higher proportion of the HPV infections detected with PCR were positive at the level of HC in the HIV-positive women, compared with HIV-negative women. These data are consistent with more active HPV replication among HIV-positive women, a larger number of HPV types, or both. As in the case of cervical HPV infection, analysis of these data suggests an important role for the immune response in controlling HPV infection in the anal canal. Moreover, the higher number of HPV types detected in women with lower CD4+ cell counts may reflect either persistence or activation of preexisting HPV infections with declining immunity rather than recent acquisition, given the lack of association with specific sexual practices. It is not known if or how the presence of multiple HPV types potentiates the pathogenesis of squamous intraepithelial lesions. However, in studies of anal HPV infection in HIV-positive and HIV-negative men, detection of multiple HPV types was associated with prevalence of ASILs [10], progression of anal lesions to a higher grade over a 2-year follow-up period [11], and development of high-grade squamous intraepithelial lesions over a 4-year follow-up period [39].

Despite some differences in risk factors for anal and cervical HPV infection, this study demonstrates a clear relationship between HPV infection at the 2 sites. Having cervical HPV infection was associated with having anal HPV infection. This association may implicate unmeasured common sexual practices other than intercourse that involve both the cervix and the anus. Although the overall spectrum of HPV types found in the anus was similar to that found in the cervix, it was striking that, in most of the women in our study, the anus and cervix had different HPV types detected in samples collected at the same visit. It is possible that cervical and anal HPV infections may have originated from separate sexual exposures. However, because a high proportion of specimens contained unidentified HPV types, it is likely that we underestimated the degree of concordance because we would not have known whether the same unidentified types were present in both the anus and cervix. Moreover, differences between HPV types found in the cervix and anus may reflect variation in the detection of specific HPV types at each site at any one sampling, because variation has been shown to occur even during short testing intervals [22, 40, 41]. Given the potential for movement of HPV between the cervix and the anus, it is possible that HPV infection in the cervix or anus modulates the course of HPV infection and squamous intraepithelial lesions at the other site. However, these interactions, if any, are not yet understood.

The proportion of samples with unidentified HPV types was higher in those taken from the anal canal rather than the cervix in this study, as well as in cervical samples taken at baseline from the entire WIHS cohort [23]. As in that study of cervical HPV infection in the entire WIHS cohort, in this study, there was a wide distribution of anal HPV types, and no single type was predominant. HPV 16 was the most common type found in the anal canals of HIV-positive women, followed by HPV 53 and 58. These types were also among the most common in the cervix [23].

It is also of interest to compare anal HPV infection in HIV-positive and HIV-negative women with that of HIV-positive and HIV-negative MSM. Because of their generally higher number of acts of receptive anal intercourse, it might have been assumed a priori that MSM would be at higher risk of anal HPV infection than women. However, when comparing HIV-positive MSM to HIV-positive women, the differences in prevalence of anal HPV infection were relatively small when comparing groups with the same CD4+ cell counts [12]. Similarly, the difference in prevalence of anal HPV infection between HIV-negative MSM and high-risk HIV-negative women was surprisingly small. However, the mean number of specific anal HPV types was higher among the HIV-positive MSM than among the HIV-positive women [12]. The mean number of anal HPV types was also higher among HIV-negative MSM than among the HIV-negative women in this study.

The data in this study should be interpreted with caution because we were unable to assess the impact of highly active antiretroviral therapy on HPV detection. Virtually none of the data in this study were collected before the widespread use of protease inhibitors among cohort participants. In addition, there were a large number of statistical tests in this analysis that increased the probability of type I error. Finally, our study does not permit us to exclude the possibility that the anal HPV detected represents contamination or colonization in the anal canal. However, several pieces of evidence suggest that we are detecting bona fide HPV infection. First, the high frequency with which different HPV types were detected in the cervix and anus argues against contamination. Second, most women whose HPV infections were detected via PCR were also found to be positive via HC, which suggests that relatively high levels of HPV DNA were present. Third, anal HPV was not likely to be merely colonizing, because detection of anal HPV infection via HC was a significant risk factor for ASILs in multi-variate analysis, with higher rlu ratios associated with increased risk of ASILs (data not shown).

In summary, anal HPV infection is clearly underrecognized among HIV-positive and high-risk HIV-negative women. Anal HPV infection is associated with the detection of ASILs in women (J.M.P., data not shown). Because ASILs are likely to be an anal cancer precursor, these findings may portend increased risk of development of anal cancer in these women. Analysis of our data suggests that HIV-positive women with lower CD4+ cell counts may be at particularly high risk of anal HPV infection, although high rates of infection were found with all CD4+ cell counts. Further work on the prevalence, incidence, and natural history of anal HPV infection and ASILs will be needed to determine whether these women, as well as high-risk HIV-negative women, would benefit from screening for ASILs. Detection of cervical HPV infection is associated with the detection of anal HPV infection, but the high proportion of discordant HPV results at the 2 sites suggests separate exposures. However, our data do not exclude a role for cervical HPV infection as a reservoir and source of anal HPV infection or vice versa. Further work is required to define the mode of acquisition of anal HPV infection and its relationship to the natural history of cervical HPV infection and CSILs.

Acknowledgments

We thank Rosana Botts and Michelle Moghadassi (Departments of Laboratory Medicine and Stomatology, University of California, San Francisco) for their technical assistance.

References

1.
Bosch
FX
Manos
MM
Munoz
N
, et al.  . 
Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International Biological Study on Cervical Cancer (IBSCC) Study Group
J Natl Cancer Inst
 , 
1995
, vol. 
87
 (pg. 
796
-
802
)
2.
Walboomers
JM
Jacobs
MV
Manos
MM
, et al.  . 
Human papillomavirus is a necessary cause of invasive cervical cancer worldwide
J Pathol
 , 
1999
, vol. 
189
 (pg. 
12
-
9
)
3.
Palefsky
JM
Holly
EA
Gonzales
J
Berline
J
Ahn
DK
Greenspan
JS
Detection of human papillomavirus DNA in anal intraepithelial neoplasia and anal cancer
Cancer Res
 , 
1991
, vol. 
51
 (pg. 
1014
-
9
)
4.
Zaki
SR
Judd
R
Coffield
LM
Greer
P
Rolston
F
Evatt
BL
Human papillomavirus infection and anal carcinoma: retrospective analysis by in situ hybridization and the polymerase chain reaction
Am J Pathol
 , 
1992
, vol. 
140
 (pg. 
1345
-
55
)
5.
Frisch
M
Glimelius
B
van den Brule
AJ
, et al.  . 
Sexually transmitted infection as a cause of anal cancer
N Engl J Med
 , 
1997
, vol. 
337
 (pg. 
1350
-
8
)
6.
Frisch
M
Fenger
C
van den Brule
AJ
, et al.  . 
Variants of squamous cell carcinoma of the anal canal and perianal skin and their relation to human papillomaviruses
Cancer Res
 , 
1999
, vol. 
59
 (pg. 
753
-
7
)
7.
Kiviat
NB
Critchlow
CW
Holmes
KK
, et al.  . 
Association of anal dysplasia and human papillomavirus with immunosuppression and HIV infection among homosexual men
AIDS
 , 
1993
, vol. 
7
 (pg. 
43
-
9
)
8.
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 infection
AIDS
 , 
1995
, vol. 
9
 (pg. 
1255
-
62
)
9.
Palefsky
JM
Holly
EA
Ralston
ML
Jay
N
Berry
JM
Darragh
TM
High incidence of anal high-grade squamous intraepithelial lesions among HIV-positive and HIV-negative homosexual/bisexual men
AIDS
 , 
1998
, vol. 
12
 (pg. 
495
-
503
)
10.
Palefsky
JM
Holly
EA
Ralston
ML
, 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
, vol. 
17
 (pg. 
320
-
6
)
11.
Palefsky
JM
Holly
EA
Hogeboom
CJ
, 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
, vol. 
17
 (pg. 
314
-
9
)
12.
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
, vol. 
177
 (pg. 
361
-
7
)
13.
Holmes
F
Borek
D
Owen-Kummer
M
, et al.  . 
Anal cancer in women
Gastroenterology
 , 
1988
, vol. 
95
 (pg. 
107
-
11
)
14.
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
, vol. 
81
 (pg. 
1726
-
31
)
15.
Melbye
M
Rabkin
C
Frisch
M
Biggar
RJ
Changing patterns of anal cancer incidence in the United States, 1940–1989
Am J Epidemiol
 , 
1994
, vol. 
139
 (pg. 
772
-
80
)
16.
Daling
JR
Weiss
NS
Hislop
TG
, et al.  . 
Sexual practices, sexually transmitted diseases, and the incidence of anal cancer
N Engl J Med
 , 
1987
, vol. 
317
 (pg. 
973
-
7
)
17.
Goedert
JJ
Cote
TR
Virgo
P
, et al.  . 
Spectrum of AIDS-associated malignant disorders
Lancet
 , 
1998
, vol. 
351
 (pg. 
1833
-
9
)
18.
Williams
AB
Darragh
TM
Vranizan
K
Ochia
C
Moss
AR
Palefsky
JM
Anal and cervical human papillomavirus infection and risk of anal and cervical epithelial abnormalities in human immunodeficiency virus-infected women
Obstet Gynecol
 , 
1994
, vol. 
83
 (pg. 
205
-
11
)
19.
Sun
XW
Ellerbrock
TV
Lungu
O
Chiasson
MA
Bush
TJ
Wright
TC
Jr
Human papillomavirus infection in human immunodeficiency virus—seropositive women
Obstet Gynecol
 , 
1995
, vol. 
85
 (pg. 
680
-
6
)
20.
Vernon
SD
Reeves
WC
Clancy
KA
, et al.  . 
A longitudinal study of human papillomavirus DNA detection in human immunodeficiency virus type 1—seropositive and —seronegative women
J Infect Dis
 , 
1994
, vol. 
169
 (pg. 
1108
-
12
)
21.
Gemignani
M
Maiman
M
Fruchter
RG
Arrastia
CD
Gibbon
D
Ellison
T
CD4 lymphocytes in women with invasive and preinvasive cervical neoplasia
Gynecol Oncol
 , 
1995
, vol. 
59
 (pg. 
364
-
9
)
22.
Sun
XW
Kuhn
L
Ellerbrock
TV
Chiasson
MA
Bush
TJ
Wright
TC
Jr
Human papillomavirus infection in women infected with the human immunodeficiency virus
N Engl J Med
 , 
1997
, vol. 
337
 (pg. 
1343
-
9
)
23.
Palefsky
JM
Minkoff
H
Kalish
LA
, et al.  . 
Cervicovaginal human papillomavirus infection in HIV-positive and high risk HIV-negative women
J Natl Cancer Inst
 , 
1999
, vol. 
91
 (pg. 
226
-
35
)
24.
Hankins
C
Coutlee
F
Lapointe
N
, et al.  . 
Prevalence of risk factors associated with human papillomavirus infection in women living with HIV
CMAJ
 , 
1999
, vol. 
160
 (pg. 
185
-
91
)
25.
Sillman
FH
Sedlis
A
Anogenital papillomavirus infection and neoplasia in immunodeficient women
Obstet Gynecol Clin North Am
 , 
1987
, vol. 
14
 (pg. 
537
-
58
)
26.
Maiman
M
Fruchter
RG
Serur
E
Remy
JC
Feuer
G
Boyce
J
Human immunodeficiency virus infection and cervical neoplasia
Gynecol Oncol
 , 
1990
, vol. 
38
 (pg. 
377
-
82
)
27.
Vermund
SH
Kelley
KF
Klein
RS
, et al.  . 
High risk of human papillomavirus infection and cervical squamous intraepithelial lesions among women with symptomatic human immunodeficiency virus infection
Am J Obstet Gynecol
 , 
1991
, vol. 
165
 (pg. 
392
-
400
)
28.
Klein
RS
Ho
GY
Vermund
SH
Fleming
I
Burk
RD
Risk factors for squamous intraepithelial lesions on Pap smear in women at risk for human immunodeficiency virus infection
J Infect Dis
 , 
1994
, vol. 
170
 (pg. 
1404
-
9
)
29.
Wright
TC
Jr
Sun XW Anogenital papillomavirus infection and neoplasia in immunodeficient women
Obstet Gynecol Clin North Am
 , 
1996
, vol. 
23
 (pg. 
861
-
93
)
30.
Mandelblatt
JS
Fahs
M
Garibaldi
K
Senie
RT
Peterson
HB
Association between HIV infection and cervical neoplasia: implications for clinical care of women at risk for both conditions
AIDS
 , 
1992
, vol. 
6
 (pg. 
173
-
8
)
31.
Massad
LS
Riester
KA
Anastos
KM
, et al.  . 
Prevalence and predictors of squamous cell abnormalities in Papanicolaou smears from women infected with HIV-1. Women's Interagency HIV Study Group
J Acquir Immune Defic Syndr
 , 
1999
, vol. 
21
 (pg. 
33
-
41
)
32.
Moscicki
AB
Ellenberg
JH
Vermund
SH
, et al.  . 
Prevalence and risks factors for cervical human papillomavirus infection and squamous intraepithelial lesions in adolescent girls: impact of infection with human immunodeficiency virus
Arch Pediatr Adolesc Med
 , 
2000
, vol. 
154
 (pg. 
127
-
34
)
33.
Melbye
M
Smith
E
Wohlfahrt
J
, et al.  . 
Anal and cervical abnormality in women: prediction by human papillomavirus tests
Int J Cancer
 , 
1996
, vol. 
68
 (pg. 
559
-
64
)
34.
Barkan
SE
Melnick
SL
Preston-Martin
S
, et al.  . 
The Women's Interagency HIV Study
Epidemiology
 , 
1998
, vol. 
9
 (pg. 
117
-
25
)
35.
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
 , 
1997
, vol. 
14
 (pg. 
415
-
22
)
36.
de Villiers
EM
Papillomavirus and HPV typing
Clin Dermatol
 , 
1997
, vol. 
15
 (pg. 
199
-
206
)
37.
Morrison
EA
Goldberg
GL
Kadish
AS
Burk
RD
Polymerase chain reaction detection of human papillomavirus: quantitation may improve clinical utility
J Clin Microbiol
 , 
1992
, vol. 
30
 (pg. 
2539
-
43
)
38.
Zhang
J
Yu
KF
What's the relative risk?
JAMA
 , 
1998
, vol. 
280
 (pg. 
1690
-
1
)
39.
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
, vol. 
12
 (pg. 
495
-
503
)
40.
Moscicki
AB
Palefsky
J
Smith
G
Siboshski
S
Schoolnik
G
Variability of human papillomavirus DNA testing in a longitudinal cohort of young women
Obstet Gynecol
 , 
1993
, vol. 
82
 (pg. 
578
-
85
)
41.
Hildesheim
A
Schiffman
MH
Gravitt
PE
, et al.  . 
Persistence of type-specific human papillomavirus infection among cytologically normal women
J Infect Dis
 , 
1994
, vol. 
169
 (pg. 
235
-
40
)
Presented in part: 17th International Papillomavirus Conference, Charleston, South Carolina, January 1999 (abstract EPI 30).
Informed consent was obtained from the participants of this study, with procedures and consent materials that were reviewed and approved by the committee on human experimentation at the University of California, San Francisco.
Financial support: National Cancer Institute (grants R01CA 63933 and UO1-AI-34989). Data in this article were derived in part from studies carried out in the General Clinical Research Centers, University of California, San Francisco, with funds provided by the Division of Research Resources, US Public Health Service (grants 5-M01-RR-00079 and 5-MO1-RR-00083).

Author notes

a
J.M.P. holds stock in Digene Diagnostics, maker of the hybrid capture test.