Anal intraepithelial neoplasia in a multisite study of HIV-infected and high-risk HIV-uninfected women: "Our findings show that the prevalence of AIN among HIV-infected women is high and significantly increased above a comparison group of HIV-uninfected women."
AIDS:Volume 23(1)2 January 2009p 59-70
Hessol, Nancy Aa; Holly, Elizabeth Aa; Efird, Jimmy Tb; Minkoff, Howardc; Schowalter, Karlened; Darragh, Teresa Ma; Burk, Robert De; Strickler, Howard De; Greenblatt, Ruth Ma; Palefsky, Joel Ma
aUniversity of California, San Francisco, California, USA
bCincinnati Children's Hospital Medical Center, Cincinnati, Ohio, USA
cHealth Science Center, State University of New York, Brooklyn, New York, USA
dThe Core Center, Cook County Hospital, Chicago, Illinois, USA
eAlbert Einstein College of Medicine, Bronx, New York, USA.
"70% were African-American. HIV-infected women were older, more likely to be widowed, divorced, or separated, reported lower household incomes, were less likely to be employed, drank less alcohol, and were more likely to have injected drugs than HIV-uninfected women. The proportion of women reporting anal intercourse was high among both HIV-infected (47%) and HIV-uninfected women (46%) and was not statistically different between the two groups. However, women with HIV infection were significantly more likely to have abnormal anal cytology/histology (31%) compared with the HIV-uninfected women (9%) (P < 0.001; Table 1). Similarly HIV-infected women were more likely to have abnormal cervical cytology/histology (34%) compared with the HIV-uninfected women (19%) (P = 0.003). HIV-infected women were more likely than HIV-uninfected women to have anal, cervical, and both anal and cervical HPV infection (P < 0.001)......When cigarette smoking was evaluated by grade of anal disease among all women, current cigarette smokers were more likely to have higher grade anal disease.....In multivariable analysis, women with a history of receptive anal intercourse were 3.8 times more likely (95% CI = 1.8-7.8).... to have LGAIN (low-grade anal intraepithelial neoplasia)"
Objectives: To study anal intraepithelial neoplasia and its associations with anal and cervical human papillomavirus (HPV), cervical neoplasia, host immune status, and demographic and behavioral risk factors in women with and at risk for HIV infection.
Design: Point-prevalence analysis nested within a prospective study of women seen at three clinical centers of the Women's Interagency HIV Study.
Methods: In 2001-2003 participants were interviewed, received a gynecological examination, anal and cervical cytology testing and, if abnormal, colposcopy-guided or anoscopy-guided biopsy of visible lesions. Exfoliated cervical and anal specimens were assessed for HPV using PCR and type-specific HPV probing. Logistic regression analyses were performed, and odds ratios (ORs) estimated risks for anal intraepithelial neoplasia.
Results: Four hundred and seventy HIV-infected and 185 HIV-uninfected women were enrolled. Low-grade anal intraepithelial neoplasia was present in 12% of HIV-infected and 5% of HIV-uninfected women. High-grade anal intraepithelial neoplasia was present in 9% of HIV-infected and 1% of HIV-uninfected women. In adjusted analyses among HIV-infected women, the risk factors for low-grade anal intraepithelial neoplasia were younger age [OR = 0.59, 95% confidence interval (CI) = 0.36-0.97], history of receptive anal intercourse (OR = 3.2, 95% CI = 1.5-6.8), anal HPV (oncogenic types only OR = 11, 95% CI = 1.2-103; oncogenic and nononcogenic types OR = 11, 95% CI = 1.3-96), and cervical HPV (oncogenic and nononcogenic types OR = 3.5, 95% CI = 1.1-11). In multivariable analyses among HIV-infected women, the only significant risk factor for high-grade anal intraepithelial neoplasia was anal HPV infection (oncogenic and nononcogenic types OR = 7.6, 95% CI = 1.5-38).
Conclusion: Even in the era of highly active antiviral therapy, the prevalence of anal intraepithelial neoplasia was 16% in HIV-infected women. After controlling for potential confounders, several risk factors for low-grade anal intraepithelial neoplasia differed from risk factors for high-grade anal intraepithelial neoplasia.
The incidence of anal cancer is increasing among several important subgroups of the US population, in particular, among people with HIV/AIDS. The incidence of anal cancer is 10-50 times greater in men who have sex with men (MSM) and adults with AIDS than in the general population [1-8]. Among women with HIV/AIDS, the incidence of invasive anal cancer is 7-28 times greater than in the general population [9,10]. Risk factors for the development of anal cancer include sexual practices that increase exposure to anal human papillomavirus (HPV) such as receptive anal intercourse, history of anal warts, and cigarette smoking [11-13].
Most invasive anal cancers, like cervical cancers, are squamous cell carcinomas and are associated with HPV infection. The natural history of squamous cell carcinomas of the anus is similar to other neoplasms in which HPV plays an etiological role . A small percentage of high-grade anal intraepithelial neoplasias (HGAIN) will become malignant, with invasion into the subjacent stroma. However, the natural history of anal intraepithelial neoplasia (AIN) is not as well defined as that of cervical intraepithelial neoplasia (CIN), and most of the research on AIN has been done in men, particularly MSM.
The prevalence of anal HPV infection is higher in HIV-infected MSM (85-95%) [15-17] than in HIV-uninfected MSM (66%)  but is high in both groups compared with heterosexual men (24%) . HIV-infected MSMs are more likely to have multiple HPV types (61%) and oncogenic HPV types (65%) than HIV-infected injection drug users or heterosexual men (26 and 44%, respectively) [16,18]. The risk factors for the detection of anal HPV in MSM include HIV infection, low CD4+ cell counts, presence of anal warts, and history of receptive anal intercourse . The established risk factors for developing AIN are HIV infection, low CD4+ cell counts, persistent anal HPV infection, and infection with oncogenic HPV genotypes . Less is known about the risk factors for anal HPV infection and AIN in women, either with or without HIV infection, although the sparse existing data indicate that the risk factors are similar to those in men [20,21].
Anal cancer remains one of the few carcinomas of the gastrointestinal tract that can be successfully treated with a combination of chemotherapy and radiation therapy, with or without surgery . However, such treatment is often accompanied by substantial morbidity due to the radiation therapy, and treatment failure is more likely with advanced disease. An understanding of the risk factors for AIN is important as detection and treatment of precancerous lesions such as HGAIN may prevent squamous cell carcinomas of the anus, similar to treatment of high-grade CIN to prevent cervical cancer.
In women, anogenital HPV infection is likely to be multicentric and cervical HPV infection may serve as a reservoir and source of anal HPV infection or vice versa. Studies in HIV-infected women [20,23,24] have found that the prevalence of HPV infection and disease was higher in the anus than in the cervix. Because AIN may progress to invasive cancer, an understanding of the biology of AIN in high-risk groups of women is important to the design of anal cancer prevention efforts in these women. The goal of this investigation was to determine the prevalence of and risk factors for AIN in a representative cohort study of US women with HIV and at risk for HIV infection.
A total of 655 women, 470 HIV-infected and 185 HIV-uninfected, were enrolled in this study (Table 1). Because of missing data for some study variables, the sample size varied for each analysis. HIV-infected women were older, more likely to be widowed, divorced, or separated, reported lower household incomes, were less likely to be employed, drank less alcohol, and were more likely to have injected drugs than HIV-uninfected women. (from Jules: among the HIV+ women, 70% were African-American, 13% white, 5% white/Hispanic, 3% Black/Hispanic, 5% Hispanic, with similar percentages in the HIV-neg group).
The proportion of women reporting anal intercourse was high among both HIV-infected (47%) and HIV-uninfected women (46%) and was not statistically different between the two groups. However, women with HIV infection were significantly more likely to have abnormal anal cytology/histology (31%) compared with the HIV-uninfected women (9%) (P < 0.001; Table 1). Similarly HIV-infected women were more likely to have abnormal cervical cytology/histology (34%) compared with the HIV-uninfected women (19%) (P = 0.003). HIV-infected women were more likely than HIV-uninfected women to have anal, cervical, and both anal and cervical HPV infection (P < 0.001).
Among the women who had both anal cytologic and histologic evaluation, 20 women had atypia on anal cytology and eight (40%) were upgraded to more severe disease based on histological evaluation and 28 women had LSIL on anal cytology and five (18%) were upgraded to more severe disease based on histological evaluation. Among the women who had both cervical cytologic and histologic evaluation, 63 women had atypia on cervical cytology and 46 (73%) were upgraded to more severe disease based on histological evaluation and 25 women had LSIL on cervical cytology and eight (32%) were upgraded to more severe disease based on histological evaluation (not shown in table).
When cigarette smoking was evaluated by grade of anal disease among all women, current cigarette smokers were more likely to have higher grade anal disease: 52% of women with normal cytology/histology were smokers, 57% of women with atypia, 63% of women with low-grade anal intraepithelial neoplasia (LGAIN), and 64% of women with HGAIN, one-sided trend test P value 0.02 (not shown in tables). Similar results were seen when these analyses were restricted only to the HIV-infected women: 51% of women with normal cytology/histology were smokers, 58% of women with atypia, 65% of women with LGAIN, and 65% of women with HGAIN, one-sided trend test P value 0.02.
The unadjusted and adjusted analyses for risk of LGAIN among all women are shown in Table 2. In multivariable analysis, women with a history of receptive anal intercourse were 3.8 times more likely (95% CI = 1.8-7.8), and women with anal HPV infection were 12-17 times more likely (95% CI = 1.5-143) to have LGAIN than women with no history of receptive anal intercourse or anal HPV infection. Women with nononcogenic cervical HPV infection or both oncogenic and nononcogenic cervical HPV infection were than 3.3-4.5 times more likely (95% CI = 1.5-13) to have LGAIN than women with no cervical HPV infection.
When restricted to women with HIV infection and controlling for age, smoking status, HAART use, CD4+ cell count and viral load, the risk factors for LGAIN were younger age (OR = 0.59, 95% CI = 0.36-0.97), history of receptive anal intercourse (OR = 3.2, 95% CI = 1.5-6.8), anal HPV (oncogenic types only OR = 11, 95% CI = 1.2-103; oncogenic and nononcogenic types OR = 11, 95% CI = 1.3-96), and cervical HPV (oncogenic and nononcogenic types OR = 3.5, 95% CI = 1.1-11) (Table 3).
The unadjusted and adjusted analyses for risk of HGAIN among all women are shown in Table 4. Controlling for HIV infection, age and smoking status, women with oncogenic anal HPV infection only or both oncogenic and nononcogenic anal HPV infection were 7-10 times more likely (95% CI = 1.3-50) to have HGAIN than women with no anal HPV infection.
When restricted to women with HIV infection and controlling for age, smoking status, HAART use, CD4+ cell count and viral load, women with concomitant oncogenic and nononcogenic anal HPV infection were 7.6 times more likely (95% CI = 1.5-38) to have HGAIN than women with no anal HPV infection (Table 5).
To determine whether the results of the logistic regression analyses varied if our outcome was based on cytology data alone, we also performed all of our logistic regression models using only the cytology data. We found little to no difference in our estimated ORs and 95% CIs (data not shown) compared with data using the combined cytology/histology results.
Only a few studies have assessed the prevalence of AIN in HIV-infected women and even fewer had a comparison group of high-risk HIV-uninfected women. Our findings show that the prevalence of AIN among HIV-infected women is high and significantly increased above a comparison group of HIV-uninfected women.
Investigators from a smaller study of HIV-infected women in France  reported the prevalence of HPV-related lesions to be 11.3%, with 4.6% for anal condyloma and 6.7% for AIN. In an earlier study of the San Francisco WIHS participants , the prevalence among the HIV-infected women was 15% for anal atypia, 10% for LGAIN, and 1% for HGAIN. This same study found that these rates were substantively higher than those of the HIV-uninfected women, 7, 2, and 0%, respectively. We now report follow-up from that study using a larger sample size and a more recent time point. We found that the rates of atypia were lower than the prior rates but rates of HGAIN were higher than the prior rates. This was true for both the HIV-infected and HIV-uninfected women and may be the result of progression of dysplasia over time or due to the addition of histological evaluation (anal biopsies) that may have increased detection of higher grade disease.
Consistent with our prior results , we also found a high prevalence of anal HPV infection among HIV-infected women (80%) and that the prevalence of anal HPV infection exceeded that of cervical HPV infection (45%), suggesting that HPV infection is either more common in the anus or is present at higher levels in the anus and is thus easier to detect.
We demonstrated that several of the risk factors for LGAIN and HGAIN in women differ. Most notably, having either oncogenic or nononcogenic HPV types was strongly associated with LGAIN, whereas having oncogenic HPV types was associated with HGAIN. These results are similar to what has been shown in the cervix with CIN [37,38] and in the anus with AIN [39,40]. Additionally, we found that history of receptive anal intercourse was a risk factor for LGAIN but not for HGAIN. This was true when analyzed as a dichotomous variable (ever versus never) or as a continuous variable (lifetime number of receptive anal intercourse acts). However, receptive anal intercourse is not necessary to acquire anal HPV infection or AIN. One study of anal HPV infection and AIN in HIV-infected men  (50 heterosexual injection drug users and 67 MSM) found that anal HPV infection and AIN may be acquired in the absence of anal intercourse. In women, HPV autoinoculation (e.g., HPV transmission from the cervicovaginal compartment to the anal mucosa) is certainly possible [20,41]. We also found that low CD4+ cell count, less than 200, had an increased (but not significant) OR for LGAIN but not HGAIN. Although few studies have analyzed LGAIN separately from HGAIN, our results are consistent with other studies in HIV-infected adults [17,36,40] and may explain why HIV-infected individuals with invasive anogenital disease (anal or cervical) have only moderate immunodeficiency at the time of diagnosis [9,42,43].
Among HIV-infected women, in models adjusted for CD4+ cell count, neither HAART use nor plasma HIV RNA had independent effects on the risk of AIN. However, by adjusting for current CD4+ cell count we may be incorporating the beneficial effects of HAART. Longitudinal analyses are needed to adequately incorporate time-varying variables such as CD4+ cell count, HIV RNA, and HAART use and the risk of persistent AIN.
In trend test analyses, we found that current smoking was associated with a higher grade of anal disease. However, in adjusted analyses we found no association with number of years of cigarette smoking and the risk of anal disease; this was true when we used current smoking status (data not shown). A similar study of AIN in HIV-infected MSM  also failed to find an association between smoking and any grade of AIN. Our results may reflect our sample size or the different impact that smoking has on the anal mucosa versus the cervix and the risk of intraepithelial neoplasia. Although several studies have reported an association between cigarette smoking and the risk of anal cancer [13,44,45], the results are not consistent  and in that regard differ from what has been reported in the cervix .
Although HIV infection was significant in unadjusted models for LGAIN and HGAIN, in adjusted analyses that included anal HPV infection, the OR for HIV infection was increased but the 95% CI spanned unity. This suggests that HIV infection may impact risk of neoplasia through its effect on immune suppression and persistence of anal HPV infection. Age also was included in all of our multivariable analyses but was found to be a significant risk factor for LGAIN only among HIV-infected women. Consistent with our findings, younger age has been found to be associated with anal HPV infection  and for AIN  among HIV-infected women.
In unadjusted analyses, both cervical HPV infection and cervical disease were associated with an increased risk of LGAIN and HGAIN. However, in analyses that adjusted for anal HPV infection, cervical HPV infection contributed only to LGAIN. The most likely explanation for this finding is that most women with cervical HPV infection or disease also have anal HPV infection. This finding warrants further investigation and may impact screening for anal HPV infection and disease in HIV-infected women.
The following limitations should be noted. Not all women with abnormal anal and cervical cytology had corresponding biopsy-based histological results. In addition, anal and cervical biopsies were performed only on participants with cytological abnormalities. Therefore, the true prevalence of AIN and CIN is likely to be even higher than what we observed because some of the women with normal cytological findings may have had a false-negative cytology result . Although this represents the largest group of HIV-infected women to be studied to date for AIN, in some of the subgroup analyses the samples sizes were small.
In summary, the high prevalence of anal and cervical HPV-related disease and anal HPV infection among the HIV-infected women is essential information that will help guide screening and treatment practices. Among all women, detection of anal oncogenic HPV types, either alone or in combination with nononcogenic types, were the only significant risk factors associated with HGAIN. In contrast, LGAIN was significantly associated with both oncogenic and nononcogenic HPV types, as is observed in studies of CIN. Other significant risk factors for LGAIN included history of receptive anal intercourse, cervical HPV and, among HIV-infected women, younger age.
These results are from a study nested within the Women's Interagency HIV Study (WIHS), a prospective study of HIV-1 infection in women, conducted in five locations within the United States. The WIHS methods and baseline cohort characteristics have been described previously . Briefly, between October 1994 and November 1995, 2056 HIV-1 infected and 569 uninfected women were enrolled. A second enrollment, between October 2001 and September 2002, added 737 HIV-infected and 406 HIV-uninfected women. The HIV-infected women in the WIHS cohort reflect the ethnicity, exposure status, and ages of the national AIDS cases among US women .
Every 6 months, WIHS participants were interviewed and received a physical examination. Multiple gynecologic and blood specimens were collected at each visit. Interviewers assessed self-reported antiretroviral therapy (ART) use during the period prior to the study visit. ART medications were identified by photomedication cards and recorded by both brand and generic drug names. The physical examinations included a pelvic examination with cervical Pap testing and cervicovaginal lavage (CVL) fluid collection.
For this WIHS anal substudy, women were recruited in years 2001-2003 from the Brooklyn, Chicago, and San Francisco Bay Area WIHS sites. Women were eligible if they had no history of HGAIN or anal cancer. In addition to the WIHS interview, physical examination and specimen collection, women enrolled in the anal substudy were administered a short questionnaire containing questions specific to anal disease. Participants also received an external anal examination for the presence of visible warts, hemorrhoids, discharge, or ulcerations. Anal swab specimens were collected for anal cytology (ThinPrep; Cytyc Corporation, Marlborough, Massachusetts, USA) and anal HPV testing. Among HIV-infected women, blood collected at the baseline anal study visit was tested for CD4+ lymphocytes and HIV RNA. Women with abnormal anal cytology were referred for high-resolution anoscopy (HRA) and biopsy of visible disease at a separate visit, as described previously . Women enrolled in this substudy were followed every 6 months through April 2006. The data analyzed in this report are from the baseline visit of the anal substudy. Study protocols were reviewed and approved by the institutional review boards and informed consent was obtained from the participants.
Exfoliated cells for HPV DNA testing were obtained using CVL and by anal swabs, as previously described [27,28]. HPV DNA was detected with L1 consensus primer MY09/MY11/HMB01 PCR assays. Primer set PC04/GH20 that amplifies a 268 base pair cellular _-globin DNA fragment was included in each assay as an internal control to assess the adequacy of amplification. Specimens that were negative for _-globulin amplification were considered inadequate, and their HPV DNA data were considered as missing. Details of these laboratory methods have been published previously , and the results were shown to have high reproducibility, sensitivity, and specificity [28,30,31]. Amplification products were probed for the presence of any HPV DNA with a generic probe mixture and probed for HPV DNA with filters individually hybridized with type-specific biotinylated oligonucleotide probes for oncogenic HPV (types 16, 18, 31, 33, 39, 45, 51, 52, 56, 58, 59, and 66) and nononcogenic HPV (types 6, 11, 13, 26, 32, 34, 40, 42, 53, 54, 55, 57, 61, 62, 64, 67, 68, 69, 70, 71 (AE8), 72, 73 (PAP238A), 81 (AE7), 82 (W13B and AE2), 83 (PAP291), 84 (PAP155), 85 (AE5), 89 (AE6), AE9, and AE10) [29,30].
Both anal and cervical cytology results were categorized according to the Bethesda system: normal; atypical squamous cells (atypia); low-grade squamous intraepithelial lesions (LSIL); and high-grade squamous intraepithelial lesions (HSIL). Women with unsatisfactory cervical or anal cytology due to insufficient cellularity were excluded from cytology classification. Anal and cervical histological results were categorized as benign, atypia, condyloma, AIN-1/CIN-1, AIN-2/CIN-2, and AIN-3/CIN-3. For this analysis, we defined the grade of the lesion as the more advanced diagnosis on cytology or histology when both were available. In the absence of histological data, grade was based on cytology alone. We hereafter refer to this combined categorization as 'cytology/histology'.
After the results from the anal and cervical cytology testing were available, participants with abnormal results were contacted and asked to return for anoscopy and/or colposcopy. Among the 148 women with abnormal anal cytology results, 50% had anoscopy and 41% had a biopsy for histological evaluation. Among the 166 women with abnormal cervical cytology results, 72% had colposcopy and 55% had a biopsy for histological evaluation. The main reason for lack of histological confirmation of the abnormal cytology results was failure of participants to return for anoscopy and colposcopy within 6 months after their baseline visit (74 women with abnormal anal cytology and 43 women with abnormal cervical cytology). The other reason was that no lesions were found on visual inspection (13 women with anoscopy and 32 women with colposcopy).
HPV genotype for our analysis was grouped into oncogenic HPV types (16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 66) and nononcogenic types (6, 11, 26, 32, 40, 53, 54, 55, 61, 68, 70, 73, 82, 83, and 84 plus all those HPV types that hybridized only with the consensus probe) . In addition, anal and cervical HPV infection was categorized into the following mutually exclusive groups: only oncogenic HPV types; only nononcogenic HPV types; both oncogenic and nononcogenic HPV types; or no HPV infection (the reference group).
Contingency table analyses were conducted to compare the distribution of participant characteristics by HIV serostatus. Cochran-Armitage Trend test analyses were conducted to assess the association between cigarette smoking and grade of anal disease. Logistic regression analyses were done to compute unadjusted odds ratios (ORs, hereafter referred to as risk), 95% confidence intervals (95% CI), and P values to estimate the relationship between each factor and risk of anal abnormalities. For the regression analyses, women with normal anal cytology/histology were the reference group, and they were compared with those with low-grade AIN (LSIL, condyloma, AIN-1) or high-grade AIN (HSIL, AIN-2, and AIN-3). We then performed multivariable logistic regression analysis, including those characteristics that were significantly associated with anal disease in the unadjusted analyses. As abnormal cervical cytology/histology and cervical HPV infection were highly correlated with each other, only cervical HPV infection (the stronger predictor variable) was included in the multivariable models. All multivariable models also included those variables that have been found to be associated with AIN, regardless of their statistical significance in our data set. These included HIV status, age, cigarette smoking status, HAART use, HIV RNA level, and CD4+ cell count. In the case of quasicomplete separation of data points for the anal HPV dummy variables, we employed a dispersion matrix algorithm to identify null response patterns in our logistic regression models to obtain convergent maximum likelihood model estimates [33,34]. P values 0.05 or less were considered to be statistically significant. All analyses were performed using SAS/STAT software version 9.1 (SAS Institute Inc., Cary, North Carolina, USA) .