Human papillomavirus type distribution in anal cancer and anal intraepithelial lesions

Authors


  • Conflicts of interest: J.S.S. has received research grants, honoraria, and consulting fees during the last four years from GlaxoSmithKline (GSK) and Merck Corporation. J.M. Pimenta is a full-time employee of GSK.

Abstract

A systematic review was conducted of HPV type distribution in anal cancer and anal high-grade and low-grade squamous intraepithelial lesions (HSIL and LSIL). A Medline search of studies using PCR or hybrid capture for HPV DNA detection was completed. A total of 1,824 cases were included: 992 invasive anal cancers, 472 HSIL cases and 360 LSIL cases. Crude HPV prevalence in anal cancer, HSIL, and LSIL was 71, 91 and 88%, respectively. HPV16/18 prevalence was 72% in invasive anal cancer, 69% in HSIL and 27% in LSIL. The HPV 16 and/or 18 prevalence in invasive anal cancer cases was similar to that reported in invasive cervical cancer. If ongoing clinical trials show efficacy in preventing anal HPV infection and associated anal lesions, prophylactic HPV vaccines may play an important role for the primary prevention of these cancers in both genders. © 2008 Wiley-Liss, Inc.

Anal cancer is a slowly progressing disease that begins as a superficial mass and may spread locally, involve regional lymph nodes or metastasize to distant organs.1 Globally, annual incidence rates of invasive anal cancer range from 0.1 to 2.8 cases per 100,000 among men and 0.0 to 2.2 per 100,000 among women (specific rates as defined by C21 ICD-9 code).2 Although the incidence of anal cancer in the United States (0.9 cases per 100,000 population) is relatively low compared with cervical cancer (8 cases per 100,000 women), anal cancer rates among men who have sex with men (MSM) are notably higher. Among HIV-seronegative MSM, the estimated incidence of anal cancer is 35 cases per 100,000, and this rate doubles to ∼70 cases per 100,000 among HIV-seropositive MSM.3 With the introduction of highly active antiretroviral therapy, anal cancer rates have continued to increase, possibly reflecting the longer life-expectancy in a generally unscreened HIV-seropositive population.4

It is known that invasive cervical cancer among women is preceded by cervical intraepithelial neoplasia (CIN) 2–3. Similarly, invasive anal cancer has well-documented precursors, known as anal intraepithelial neoplasia (AIN) 2–3 (histology) or high-grade squamous intraepithelial lesions (HSIL) (cytology).5 AIN 1 and cytological low-grade squamous intraepithelial lesions (LSIL) are not considered direct precursors of invasive anal cancer, but may precede the later development of AIN2/3 or HSIL.

Invasive anal cancer, like invasive cervical cancer, has been causally linked to high-risk human papillomavirus (HPV) infection.6–9 Carcinogenic types of human papillomavirus (HPV) infection have been detected in more than 99% percent of invasive cervical cancer cases using sensitive polymerase chain reaction (PCR) assays.10, 11 To date, no literature review has been conducted to quantify overall and type-specific prevalence of HPV among invasive anal cancer and preinvasive anal lesions. Data on the proportion of anal cancer cases attributable to specific types of HPV infection are important to predict the potential future impact of HPV-based screening tests and prophylactic HPV16/18 vaccines in women and in men (assuming demonstrated efficacy in men).12

Here, we present results of the first review of available published literature on HPV type distribution in invasive anal cancer and its precursors, high-grade and low-grade anal squamous intraepithelial lesions (HSIL and LSIL).

Abbreviations:

ADC, adenocarcinoma; AIN, anal intraepithelial neoplasia; CI, confidence interval; CIN, cervical intraepithelial neoplasia; DNA, deoxyribonucleic acid; HIV, human immunodeficiency virus; HPV, human papillomavirus; HSIL, high-grade squamous intraepithelial lesion; LSIL, low-grade squamous intraepithelial lesion; MSM, men who have sex with men; PCR, polymerase chain reaction; SCC, squamous cell carcinoma.

Material and methods

A systematic review was conducted in Medline by one investigator for all peer-reviewed studies published until July 2007 with no specified start date using the following MeSH search terms recommended by an information services librarian: “anal cancer” or “anal intraepithelial neoplasia,” and “HPV” or “human papilloma virus.” Articles written in languages other than English were translated and reviewed for inclusion criteria. Abstracts and unpublished studies were not included. Studies were included if HPV DNA was detected using polymerase chain reaction (PCR)-based technology or Hybrid Capture, and more than one case had HPV data available. References cited in included articles were also reviewed for inclusion criteria. Studies using other HPV detection assays (e.g., in situ or dot blot hybridization) were excluded due to limited sensitivity for HPV DNA detection. A flow diagram of inclusion and exclusion of the identified studies is presented in Figure 1.

Figure 1.

Flow diagram of inclusion and exclusion of identified studies.

Two study authors reviewed all of the obtained abstracts for inclusion or exclusion criteria. For each study meeting the inclusion criteria, data were extracted on first author, journal title, publication date, region, sample size, histological subtype, HPV specimen type (i.e., biopsy or exfoliated cells), PCR primers used to detect HPV and overall HPV prevalence. Type-specific HPV prevalence was also extracted when data were available from either the paper or from communication with the author. Data extraction was done by a single reviewer using a data extraction form that was validated and used in two previously published systematic reviews.13, 14 Data were later double-extracted by another individual, and disagreements were resolved by the two reviewers together. Authors were systematically contacted when data were not available on PCR primer type or if HPV prevalence was not available by country, HP type or histological subtype.

Data are presented for each study in the appendices. Hereafter, HSIL refers to cases reported as cytologically confirmed HSIL or histologically diagnosed anal intraepithelial neoplasia (AIN) II, AIN III, or carcinoma in situ. LSIL refers to cytologically confirmed LSIL or histologically confirmed AIN I. When studies amplified beta-globin to determine specimen quality, HPV prevalence calculations were limited to those cases that were beta-globin positive. If two or more studies were from the same population, the study with the largest sample size was included in data analyses. If an invasive cancer case was associated with an adjacent case of intraepithelial neoplasia, only data from the invasive cancer case were included in the analyses, and the adjacent neoplasm was not counted as a separate case in the HSIL or LSIL analyses.

HPV prevalence estimates were expressed as percentages of all cases tested for HPV DNA. Unless otherwise stated, male and female cases of invasive anal cancer, HSIL and LSIL are combined for prevalence estimates. Differences in HPV prevalence among invasive cancer and HSIL cases by gender were assessed with t-tests. Comparability of study results across strata of interest was assessed by stratifying HPV prevalence by factors related to HPV detection in previous studies (histological type, region, HPV DNA specimen type, primers used to amplify HPV DNA, year of study publication, gender and source of cases). Unadjusted and adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for the associations between HPV positivity and these covariates using STATA version 9. For adjusted analyses, gender was not included in the adjusted model due to the high number of missing values (N = 731). Multiple infections were separated into individual types, thus type-specific prevalence represents that in single or multiple infections.

Type-specific prevalence estimates were expressed as percent of all cases tested, and analyses were limited to those studies that typed for at least HPV 16 and 18 and obtained HPV DNA from biopsies, rather than only exfoliated cells. HPV types in biopsy specimens were included in the type-specific analyses to best represent the HPV types that were etiologically relevant. High-risk HPV was defined as types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66 and 68. Low-risk HPV was defined as types 6, 11, 32, 34, 40, 42, 43, 44, 53, 54, 55, 61, 70, 72, 73, 81, 83, 84, 89 and Pap155.

Results

A total of 437 abstracts were identified in Medline (Fig. 1). Thirty-seven studies met the inclusion criteria with a total of 1,824 cases (18% female, 41% male, 41% unspecified). Approximately half of these cases were invasive cancer (54.4%), a quarter were HSIL (25.9%) and 19.7% were LSIL (Table I). Overall HPV prevalence in invasive anal cancer, HSIL and LSIL was 71.2, 90.7 and 88.1%, respectively (Table I).

Table I. Distribution of Anal Cases by Histological Type and Country
Histological typeNo. of cases (No. of studies)Countries representedOverall HPV prevalence (%)
  • 1

    HSIL, high-grade squamous intraepithelial lesions; LSIL, low-grade squamous intraepithelial lesions.

Invasive cancer992 (22)Asia: China, Japan, Korea; Europe: Czech Republic, Denmark, France, Germany, Italy, Norway, Sweden, Switzerland, UK; North America: USA71.2%
HSIL1472 (21)Asia: Japan; Europe: Denmark, France, Germany, Sweden, UK; North America: USA; South America: Brazil90.7%
LSIL1360 (12)Europe: France, Germany, UK; North America: USA; South America: Brazil88.1%

Overall, most of the invasive cancer, HSIL and LSIL cases with HPV prevalence data were from Europe and North America (95.2%), with few from Asia (2.6%) and South America (2.1%) (data not shown). No data were available from Africa or Australia. The global distribution of invasive cancer and HSIL cases with available HPV typing data is presented in Figure 2. The countries with the most invasive cancer cases with HPV data were Sweden (N = 347), Denmark (N = 331), the United States (N = 268), Norway (N = 99) and the United Kingdom (N = 84). All countries with available data on invasive cancer also had data on HSIL, with the exception of Brazil, where only data on HSIL were available.

Figure 2.

Global distribution of anal cancer and HSIL (high-grade squamous intraepithelial lesions) cases with available human papillomavirus DNA typing data.

HPV DNA was extracted from biopsies in 72.7% of cases (100% of invasive cancer, 57.6% of HSIL, and 17.2% of LSIL) (data not shown). The remaining cases used exfoliated cells adjacent to the biopsies to isolate HPV DNA. HPV DNA was extracted from formalin-fixed or paraffin-embedded biopsies for most invasive cancer cases (92.9%). The remaining cases had DNA extracted from fresh or frozen biopsies. HPV prevalence in invasive anal cancer cases appeared to be similar for both types of biopsy specimens. When analyses for HSIL and LSIL were limited to biopsy samples after exclusion of studies where HPV was isolated from anal exfoliated cells, HPV prevalence was similar (HSIL: 88.2% for biopsies only vs. 90.7%, respectively; LSIL: 85.5% vs. 88.1%, respectively).

HPV positivity in invasive cancer was associated with histological subtype, region, primer type, study publication year and case source (population-based vs. hospital-based studies) (Table II). Odds ratios presented here are adjusted for these covariates as well as HPV DNA specimen type (fixed biopsies vs. fresh or frozen biopsies) and are designated “aOR.” Most invasive cancer cases (74.5%) were classified as squamous cell carcinoma (SCC). A smaller proportion of invasive cancers were classified as basaloid carcinoma (3.4%), adenocarcinoma (0.7%) or had an unspecified histological subtype (21.4%) (data not shown). Hereafter, SCC refers to invasive cancer cases identified as squamous cell carcinoma or those of unspecified pathology.

Table II. Prevalence of HPV in Biopsies of Invasive Anal Cancer and Prevalence Odds Ratios by Histological Type, Region, HPV DNA Specimen Type, Primer, Study Year, Case Source and Gender
 No. of casesHPV prevalence (%)OR (95% CI)1Adjusted2 OR (95% CI)
  • 1

    OR, odds ratio; CI, confidence interval; SCC, squamous cell carcinoma; ADC, adenocarcinoma; TS, type-specific.

  • 2

    Adjusted for histological type, region, HPV DNA source, primer and study year.

  • 3

    Referent.

  • 4

    PU-1M/31B (N = 33), SK38/39 (N = 6), TS L1 (N = 14).

Histological type
 SCC1 or unspecified95178.31.031.03
 Basaloid3482.31.3 (0.5–3.2)2.5 (0.8–7.6)
 ADC1742.90.2 (0.0–0.9)0.1 (0.0–0.8)
Region
 North America26876.91.031.03
 Asia4658.70.4 (0.2–0.8)0.2 (0.1–0.7)
 Europe67879.91.2 (0.9–1.7)0.5 (0.2–1.1)
HPV DNA Specimen Type
 Fixed biopsies92278.21.031.03
 Fresh or frozen biopsies7078.61.0 (0.5–1.8)1.0 (0.5–2.0)
Primer
 MY09/1121476.61.031.03
 GP5+/6+ or L1C1/C235886.92.0 (1.3–3.1)1.9 (0.8–4.3)
 TS1 (Early proteins)25678.51.1 (0.7–1.7)5.9 (2.6–13.2)
 Combo MY09/1111164.90.6 (0.3–0.9)1.0 (0.4–2.4)
 Other45352.80.3 (0.2–0.6)0.9 (0.3–2.6)
Study Year
 1989–199838064.71.031.03
 1999–200453887.03.6 (2.6–5.1)8.8 (3.4–22.6)
 2005–20077483.82.8 (1.5–5.4)7.4 (3.2–17.4)
Case Source
 Hospital-based studies48269.11.031.03
 Population-based studies51086.93.0 (2.1–4.1)0.7 (0.2–2.4)
Gender
 Males9076.71.03
 Females17186.51.9 (1.0–3.8)

Overall HPV prevalence in SCC cases was 78.3%. HPV prevalence was higher among basaloid cases than SCC (aOR = 2. 5, 95% CI: 0.8–7.6), although this association was not statistically significant. A lower HPV prevalence was found among a limited number of 7 ADC cases (aOR = 0.1, 95% CI: 0.0–0.8 vs. SCC). Crude HPV prevalence in invasive cancer by region was highest in Europe (79.9%), followed by North America (76.9%) and Asia (58.7%). Asia had significantly lower odds of HPV positivity than North America (aOR: 0.2, 95% CI: 0.1–0.7). HPV prevalence was similar among fresh or frozen biopsies (78.6%) compared to fixed biopsies (78.2%), and thus, the type of specimen used for HPV DNA testing did not appear to be associated with observed HPV DNA positivity (aOR: 1.0, 95% CI: 0.5–2.0).

The most common primer types used for HPV DNA detection were GP5+/6+ or L1C1/C2 (36.1%), early region type-specific (25.8%) and MY09/11 (21.6%). MY09/11 primers were used in combination with type-specific primers in 11.2% of cases. The remaining cases (5.3%) were typed using Hybrid Capture or PU-1M/31B, SK38/39 or type-specific L1 primers. HPV prevalence estimates were higher among GP5+/6+ or L1C1/C2 compared to studies that typed with MY09/11 primers only (aOR = 1.9, 95% CI: 0.8–4.3), although differences were not statistically significant. Early region type-specific primers appeared to have a significantly higher HPV DNA positivity than MY09/11 primers (aOR = 5.9, 95% CI: 2.6–13.2). Observed HPV prevalence for studies published between 1999 and 2004 and 2005 and 2007 was higher than in studies published between 1989 and 1998 (aOR = 8.8, 95% CI: 3.4–22.6 for 1999–2004 vs. 1989–1998; aOR = 7.4, 95% CI: 3.2–17.4 for 2005–2007 vs. 1989–1998). Population-based studies had a higher HPV prevalence compared to hospital-based studies (86.9% vs. 69.1%, respectively), although this difference was not statistically significant in the multivariate model (aOR = 0.7, 95% CI: 0.2–2.4).

HPV types found in invasive anal cancer, from most to least prevalent, were 16 (65.6%), 18 (5.1%), 6 (5.1%), 33 (3.3%), 31(1.6%), 11 (1.0%) and 45 (0.8%). The 6 most common types in HSIL cases match those of invasive cancer cases, with the following prevalences: 16 (64.0%), 18 (7.0%), 6 (5.8%), 33 (4.9%), 31 (4.2%) and 11 (3.5%). HSIL cases were also positive for types 58 (2.3%), 70 (1.7%), 45 (1.4%), 56 (1.2%), 61 (1.2%), 66 (1.2%) and 53 (0.7%). HPV types found in LSIL cases were 11 (40.8%), 6 (28.6%), 16 (16.1%), 70 (8.3%) and 31 (8.2%).

The combined HPV16 and/or 18 prevalence was 72.2% in invasive cancer, 68.7% in HSIL and 27.4% in LSIL among studies that reported prevalence of types 16 and 18, obtained biopsy specimens, and presented data on HPV types in multiple infections (data not shown). Figure 3 presents HPV prevalence by histological type for the most common HPV types in invasive cervical cancer (16, 18, 33, 31, 45, 35, 52 and 58).13 The prevalence of HPV 16 was highest in invasive anal cancer cases (68.5%), decreased slightly among HSIL cases (62.9%), and was notably lower in LSIL cases (6.1%). The prevalence of HPV 31 was lowest in invasive anal cancer cases (1.7%), increased among HSIL cases (4.5%), and was highest in LSIL cases (8.2%). The prevalences of HPV 18, 33 and 45 were also higher among HSIL cases than invasive anal cancer cases. LSIL cases were negative for types 18, 33, 45, 35, 52 and 58.

Figure 3.

Prevalence of the most common HPV type (type-specific prevalence estimates are restricted to studies that obtained HPV DNA from biopsies and typed for at least HPV 16 and 18) in biopsy specimens of invasive anal cancer (N = 810), HSIL (N = 178), and LSIL (low-grade squamous intraepithelial lesions, LSIL estimates include only 2 biopsied cases from women.) (N = 49).

When HPV prevalences were calculated as the percent of only HPV-positive invasive anal cancer cases, 85.1% of HPV positivity was attributable to HPV16, followed by HPV18 (7.2%) and HPV 6 (7.1%) (data not shown). The combined HPV16 and/or 18 prevalence was 89.7% when accounting for infections with 16/18 multiple infections.

Data on gender were available for a subset of 261 invasive cancer (26.3% of total) and 148 HSIL (31.4% of total) cases among studies that obtained HPV DNA from biopsies and typed for at least HPV 16 and 18 (Table III). The overall HPV prevalence estimates for invasive cancer were 86.5% for females and 76.7% for males (p < 0.001). HPV 16 prevalence was also higher among female invasive anal cancer cases compared to males (73.7% vs. 60.0%), although HPV 18 and 31 prevalences were comparable among females and males (HPV 18: 7.6% vs. 7.8%, respectively; HPV 31: 1.6% vs. 1.4%, respectively). HPV 33 prevalence among males was approximately double that among females (5.2% vs. 2.1%, respectively). The only low-risk HPV types present in invasive anal cancer cases were HPV 6 (3.4% in females and 0.0% in males) and 11 (3.4% in females and 8.3% in males). The overall HPV prevalence estimates for HSIL were higher than those for invasive cancer and closer in magnitude between the genders, at 91.2% HPV-positive for females and 93.4% HPV-positive for males (p < 0.001). HPV 16 prevalence for HSIL cases was higher among females than males (68.4% vs. 59.3%, respectively). HPV 18 and 33 prevalences for HSIL cases were comparable among the two genders, but HPV 31 prevalence was much higher among males than females (6.7% vs. 2.0%, respectively). Only two cases of LSIL were from women, and thus only results for male LSIL cases are presented in Table III. Among the 41 LSIL cases, HPV prevalence was 97.6%.

Table III. Comparison of Overall and Type-Specific HPV Prevalence in Biopsy Specimens of Invasive Anal Cancer and HSIL by Gender
HPV TypeInvasive Anal CancerHSILLSIL1
FemalesMalesFemalesMalesMales
nHPV (%)2nHPV (%)2nHPV (%)2nHPV (%)2nHPV (%)2
  • 1

    Only 2 cases of LSIL were female; only results for males are presented.

  • 2

    Type-specific prevalence estimates are restricted to studies that obtained HPV DNA from biopsies and typed for at least HPV 16 and 18; denominator is all cases tested.

  • 3

    Low-risk types may be included as a single infection, or in combination with other high-risk types.

All types17186.59076.75791.29193.44197.6
High risk 
1617173.79060.05768.49159.3417.3
181717.6907.8577.0917.7410.0
331432.1775.2575.3826.1110.0
311261.6701.4512.0906.7419.8
451220.8680.0..452.2110.0
52210.060.0170.0200.040.0
58210.060.0170.0462.2110.0
35170.040.0160.0190.040.0
39170.040.0160.0190.040.0
51170.040.0160.0190.040.0
56170.040.0160.0452.2110.0
59170.040.0160.0190.040.0
66170.040.0160.0452.2110.0
68170.040.0160.0190.040.0
Low risk3 
6293.4120.0577.0914.44122.0
11293.4128.3570.0913.34143.9
42210.060.0170.010.0..
53170.040.0500.0631.6342.9
70170.040.0160.0195.3425.0
73170.040.0160.0190.040.0

HIV serostatus was reported in only two of the included invasive anal cancer studies. These studies reported HIV seroprevalences of 18.2% in 11 cases and 0% in 6 cases.15, 16 Eleven of the 20 HSIL studies reported HIV seroprevalence data, with an average HIV seroprevalence of 48.8%. Among LSIL studies, half of the studies reported HIV seroprevalence, and the average HIV seroprevalence was 30.1%. When results were compared between studies reporting less than 50% HIV seroprevalence to those reporting over 50% HIV seroprevalence, overall HPV positivity and type-distribution were similar (data not shown).

Discussion

Overall HPV was detected in 71% of invasive anal cancers. Approximately 72% of invasive anal cancer cases were associated with HPV 16 and/or 18 infection. This estimate of 16 and 18 prevalence is similar to that found in invasive cervical cancer cases.13 One difference from invasive cervical cancer cases, however, is that a relatively high proportion of invasive anal cancer cases were HPV-negative (28.8%). Other high-risk HPV types present in the invasive anal cancer cases were HPV 31 (1.7%), 33 (4.3%) and 45 (0.8%). Overall HPV prevalence among invasive anal cancer cases did not vary much between Europe and North America, although Asia had a lower HPV prevalence. HPV positivity among invasive anal cancer cases was also associated with histological subtype, primer type and study publication year. Studies with later publication years were more likely to detect HPV in their samples, which likely reflect advances in PCR technique over time.

HPV prevalence in this review was higher among anal cancer cases with basaloid histology compared with those diagnosed as unspecified SCC. This is consistent with 2 previous studies that have shown significantly higher prevalence of high-risk HPV infection in anal cancers with basaloid features.17, 18 However, a study of HPV infection among 287 men with anal cancers from Denmark and Sweden found that anal canal cancers were much more likely to harbor high-risk HPV than perianal skin cancers whether or not basaloid characteristics were present.18 Anatomical location, therefore, may be a more important predictor of HPV positivity than histology. The lower HPV prevalence seen in 7 ADC cases compared to SCC here is consistent with a recent study of 42 anal cancers from the Czech Republic examined by GP5+/6+ PCR.19 Although sample size is limited, these data suggest that over half of adenocarcinoma of the anal cancer (58%) are likely not HPV-related.

HPV prevalence estimates in invasive anal cancer cases in this review were closer in magnitude between males and females than in 2 previous reports from Denmark and Norway, where high-risk HPV was found in ∼90% of female cases and 60% of male cases.18, 20 A further study of 93 women and 36 men in Washington state of in situ or invasive anal cancer found that 70% of female cases and 67% of male cases were positive for HPV types 16, 18, 6 and/or 11.21 However, the study from Washington state did not distinguish between cancers arising from different anatomical locations (anal canal vs. perianal skin cancers), whereas the studies from Denmark and Norway found a greater difference by gender when anal canal cancers were examined separately. HPV prevalence was higher in anal canal cancers compared to perianal skin cancers.17 It is therefore possible that some of the cancers included in this review are perianal skin cancers with relatively lower HPV prevalence, and therefore may have potentially lowered the overall HPV prevalence estimate found in anal cancers. The relatively higher HPV prevalence seen in biopsy samples of anal cancers from females may potentially be explained by the spread of HPV to the anal canal from existing HPV infections in the cervix or vulva.22, 23

HPV prevalence was higher among anal HSIL cases than among invasive anal cancer cases. One possible explanation for this could be a difference in the etiologies of the anal cancer. (Stoler MH, Personal Communication).24 However, the prevalence of HPV in anal HSIL is similar to that of cervical HSIL.13 It is the prevalence of HPV in LSIL anal cancer cases that is surprisingly low. This may be due to the use of PCR-based methods that amplify the viral capsid protein L1.23 In invasive cancer, the viral genome may be integrated into the host's DNA, leading to a potential underdetection of certain viral DNA sequences. However, this is not likely given that a higher prevalence of HPV in high-grade versus invasive cervical cancer has not been seen among studies using PCR-based methods. Thus the lower prevalence in anal cancer relative to high-grade lesions is difficult to explain.13 Invasive anal cancer or AIN2/3 studies with smaller sample sizes had similar HPV prevalence than larger studies, and thus observed results were likely not the result of publication bias.

Limitations of this systematic review with respect to ascertainment of anal cancer clinical outcomes and of HPV prevalence include variation in cytological and histological diagnoses across studies and in HPV type-specific sensitivity of different PCR-based methods.14 A limitation with respect to ascertainment of HPV prevalence is that many included studies did not type for HPV 16 and 18 types, resulting in a small sample size for HPV type-specific analyses. It was also difficult to determine the relative importance of low-risk HPV types (e.g., HPV 6) in disease carcinogenesis as few studies reported HPV types within multiple infections and broad-spectrum HPV primers were used in only a subset of studies on invasive anal cancer (68% of cases) or HSIL (81% of cases). In addition, many studies presented data on single rather than multiple infections, further limiting analyses of type-specific positivity. HPV prevalence could not be reliably stratified by age or HIV status as data were only available in 36.4 and 1.7%, respectively, of invasive cancer cases overall.

Another limitation is that these results are based on cross-sectional data and ascertained cases may not be representative of the natural history of anal disease. HPV detection alone is not sufficient to prove causation. However, given the similar histopathologic features of anal carcinoma and other anogenital carcinomas in women in which HPV infection has been accepted as a causative agent, it is likely that HPV plays the same etiologic role in these anal tumors.

When HPV prevalence was calculated as the percent of only HPV-positive invasive anal cancer cases, most HPV positivity (85.1%) was attributable to HPV16, followed by HPV18 (7.2%) and HPV 6 (7.1%). More than half (56%) of the HPV 6-positive cases were also infected with a high-risk HPV type. The combined HPV16 and/or 18 prevalence among HPV positive cases was 87.4% when accounting for multiple infections with 16/18. However, because many of the PCR systems used by the included studies amplify only a subset of HPV types, cases designated as HPV “negative” may actually be infected with other, unascertained types.14 We therefore did not restrict type-specific prevalence estimates to HPV-positive cases when comparing data across studies.

Primary anal cancer and HSIL prevention through HPV prophylactic vaccines could reduce the future incidence of anal carcinoma and anal HSIL, thus reducing morbidity and mortality among people that would otherwise require treatment. Chemoradiation and surgical excision or ablation are the primary forms of treatment of invasive anal cancer and HSIL, respectively.25, 26 Although treatment has been shown to decrease mortality, extensive anal surgery may lead to considerable morbidity, including anal stenosis or incontinence. In addition, most ablative therapies have considerable recurrence rates and require repeated surgical intervention in many cases.25, 26

A prophylactic HPV vaccine against HPV 16/18 has the potential to prevent more than two-thirds of invasive anal cancer and anal HSIL cases in women and in men. If ongoing clinical trials show efficacy in preventing anal HPV infection and associated anal lesions, prophylactic HPV vaccines may play an important role for the primary prevention of these cancers in both genders.

Acknowledgements

The authors thank Sudha Raman for her work in double-checking the abstracted data. They also thank those authors who made additional data available from their published studies.

APPENDIX

A1, AII, AIII

Table A1. Type-Specific Human Papillomavirus (HPV) Prevalence Among Men and Women with Invasive Anal Cancer, by Study
First authorReferenceCountryHPV DNA sourcePCR primers usedN casesHistology/ cytology% FemaleHPV prevalence (% of all cases tested)1
any1618334531585235595639516687366117082
  • 1

    The estimate for “any” HPV type was calculated using estimates from all studies. Type-specific prevalence estimates were calculated from studies that typed for at least HPV types 16, 18, 6, and 11 and obtained HPV DNA from biopsies.

  • 2

    Studies used to calculate type-specific prevalence estimates.

Crook TOncogene, 1991UKfrozen biopsiesConsensus NOS50SCC 80.076.08.0                 
Daling JR2Cancer, 2004 (Carter JJ et al., Cancer Research 2001)USAfixed biopsiesMY09/11166carcinoma62.786.772.36.63.00.61.8       9.0   0.6  
Frisch MCancer Research, 1999Denmark, Swedenfixed biopsiesGP5+/6+331SCC76.487.072.85.75.1 0.9              
Heselmeyer KBr J Cancer, 1997Swedenfixed biopsiesMY09/1116carcinoma68.856.350.0                  
Holm RMod Pathol, 1994Norwayfixed biopsiesTS (Early proteins)99SCC74.775.875.8                  
Indinnimeo M2J Exp Clin Cancer Res, 1999Italyfresh biopsiesTS (L1)14carcinoma 64.342.90.0          0.0   0.0  
Kagawa R2Surg Today, 2006Japanfixed biopsiesL1C1/C26SCC66.783.383.30.00.0 0.00.00.0     0.0   0.0  
Kiyabu MTAm J Surg Path, 1989USAfixed biopsiesTS (Early proteins)4SCC 100.0100.025.0                 
Koulos JMod Pathol, 1991USAfixed biopsiesTS (Early proteins)8carcinoma50.050.025.037.5  0.0  0.0           
Lai MD2WJG, 1998Chinafixed biopsiesMY09/1119SCC57.95.35.30.00.0         0.0   0.0  
Noffsinger AEMod Pathol, 1995USA, Canada, Chinafixed biopsiesMY09/11 and TS E650SCC 44.038.04.0          6.0      
Ogunbiyi OAJ Clin Pathol, 1993UKfixed biopsiesTS (Early proteins)34SCC 76.576.5                  
Palefsky JCancer Research, 1991USAfixed biopsiesTS (Early proteins)13carcinoma 84.676.90.07.7 23.1              
Poletti PAInt J Colorect Dis, 1998Switzerlandfixed biopsiesPU-1M/PU31-B33carcinoma100.039.4                   
Shroyer KRAm J Clin Pathol, 1995USAfixed biopsiesMY09/1413carcinoma76.976.961.515.40.0                
Shroyer KR2Arch Surg, 1992USAfixed biopsiesMY09/11 and E6 Consensus3carcinoma33.3100.033.333.366.70.00.0       0.0   33.3  
Tachezy R2APMIS, 2007Czech Republicfresh and fixed biopsiesGP5+/6+21carcinoma81.085.785.70.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.0
Varnai AD2Int J Colorectal Dis, 2006Germanyfixed biopsiesMY09/11 and GP5+/6+47SCC72.383.074.52.16.42.12.10.00.00.00.00.00.00.00.00.00.00.00.00.00.0
Vincent-Salomon A2Mod Pathol, 1996Francefixed biopsiesTS (Early proteins)27carcinoma70.470.463.07.40.0         0.0   0.0  
Youk EDis Colon Rectum, 2001Koreafixed biopsiesTS (Early proteins)21carcinoma33.3100.0100.00.0                 
Zaki SR2Am J Pathol, 1992USAfixed biopsiesMY09/11 and TS L111SCC63.672.727.39.10.0         9.1   9.1  
Zucchini CInt J Oncology, 2000Italyfresh biopsiesSK 38/396SCC83.3100.0100.00.0                 
TOTAL2    942  71.265.65.13.30.81.60.00.00.00.00.00.00.05.10.00.00.01.00.00.0
Table AII. Type-Specific Human Papillomavirus (HPV) Prevalence Among Men and Women with High-Grade Anal Intraepithelial Lesions, by Study
First authorReferenceCountryHPV DNA sourcePCR primers usedN casesHistology/ cytology% FemaleHPV prevalence (% of all cases tested)1
any1618334531585235595639516687366117082
  • 1

    The estimate for “any” HPV type was calculated using estimates from all studies. Type-specific prevalence estimates were calculated from studies that typed for at least HPV types 16, 18, 6, and 11 and obtained HPV DNA from biopsies.

  • 2

    Studies used to calculate type-specific prevalence estimates.

Chin-Hong PVJNCI, 2005USAexfoliated cellsMY09/1153HSIL0.083.0                   
Daling JR2Cancer, 2004 (Carter JJ et al., Cancer Research 2001)USAfixed biopsiesMY09/1169CIS49.392.873.97.27.2 4.3       5.8   1.4  
Fox PASex Transm Inf, 2005UKexfoliated cellsMY09/11 and GP5+/6+74AIN2/30.097.364.925.724.316.218.921.60.010.85.44.112.24.1       
Friedman HBJID, 1998USAexfoliated cellsMY09/11, HMB01, TS2HSIL0.0100.050.00.00.00.00.050.00.050.050.00.00.050.00.00.00.00.00.0  
Frisch MCancer Research, 1999Denmark, Swedenfixed biopsiesGP5+/6+55CIS89.192.7                   
Gagne SE2J Acquir Immune Defic Syndr, 2005USAfrozen biopsiesMY09/1126AIN2/30.0100.053.815.43.83.811.53.8   3.8  3.8  3.80.0  
Haga T2JAIDS, 2001USAfixed biopsiesMY09/1119AIN2/30.084.257.95.35.30.00.00.00.00.00.00.00.00.00.00.00.00.00.05.3 
Hampl M.2Obstetrics & Gynecology, 2006Germanyfixed biopsiesMY09/11 and GP5+/6+16AIN2/3100.087.581.30.00.0 0.00.00.00.00.00.00.00.00.00.00.00.00.00.00.0
Kagawa R2Surg TodayJapanfixed biopsiesL1C1/C22CIS50.0100.0100.00.00.0 0.00.00.0     0.0   0.0  
Kreuter AJ Am Acad Dermatol, 2005Germanyexfoliated cellsDegenerate13AIN2/30.0100.084.646.215.47.77.723.138.57.715.47.7  15.47.730.87.723.17.77.7
Manzione CR2Rev Assoc Med Bras, 2004Brazilfixed biopsiesTS L19HSIL0.0100.00.011.1  11.1       33.3   22.2  
Ogunbiyi OAJ Clin Pathol, 1993UKfixed biopsiesTS (Early proteins)39AIN2/3 66.766.7                  
Palefsky JCancer Research, 1991USAfixed biopsiesTS (Early proteins)6AIN2/3 50.033.316.70.0 16.7              
Palefsky JAIDS, 2006USAexfoliated cellsMY09/1114AIN2/30.0100.035.728.614.3 21.450.014.314.37.17.114.3 7.1   28.614.3 
Piketty CSex Trans Dis, 2004 (Piketty C et al., Ann Intern Med 2003)Franceexfoliated cellsMY09/1110HSIL0.090.030.030.030.010.010.00.00.00.010.00.00.00.030.00.010.010.010.020.0 
Surawicz CMAm J Gastroenterology, 1995USAexfoliated cellsMY09/11 and TS L16AIN2/30.0100.0100.050.0                 
Surawicz CMAm J Gastroenterology, 1995USAexfoliated cellsMY09/11 and TS L15CIS0.0100.0100.040.0                 
Varnai AD2Int J Colorectal Dis, 2006Germanyfixed biopsiesMY09/11 and GP5+/6+24AIN2/370.895.870.80.00.00.00.04.20.00.00.00.00.00.08.30.00.00.012.50.00.0
Walts AEDiagnostic Cytopathology, 2005USAexfoliated cellsN/A (HC)5HSIL0.0100.0                   
Wieland UArch Dermatol, 2006Germanyexfoliated cellsTS (L1)18AIN2/30.0100.088.950.022.211.116.727.833.311.116.711.10.00.05.65.622.211.116.711.15.6
Zaki SRAm J Pathol, 1992USAfixed biopsiesMY09/11 and TS L17CIS85.785.728.614.314.3         0.0   0.0  
TOTAL2    472  90.764.07.04.91.44.22.30.00.00.01.20.00.05.80.00.01.23.51.70.0
Table AIII. Type-Specific Human Papillomavirus (HPV) Prevalence Among Men and Women with Low-Grade Anal Intraepithelial Lesions, by Study
First authorReferenceCountryHPV DNA sourcePCR primers usedN casesHistology/ cytology% FemaleHPV prevalence (% of all cases tested)1
any1618334531585235595639516687366117082
  • 1

    The estimate for “any” HPV type was calculated using estimates from all studies. Type-specific prevalence estimates were calculated from studies that typed for at least HPV types 16, 18, 6, and 11 and obtained HPV DNA from biopsies.

  • 2

    Studies used to calculate type-specific prevalence estimates.

Chin-Hong PVJNCI, 2005USAexfoliated cellsMY09/11157LSIL0.084.1                   
Fox PASex Transm Inf, 2005UKexfoliated cellsMY09/11, GP5+/6+19AIN10.0100.063.215.826.321.10.010.50.00.00.00.00.05.3       
Friedman HBJID, 1998USAexfoliated cellsMY09/11, HMB01, TS41LSIL0.082.9                   
Gagne SE2J Acquir Immune Defic Syndr, 2005USAfrozen biopsiesMY09/117AIN10.085.714.30.00.00.00.00.0   0.0  28.6  0.042.9  
Haga T2JAIDS, 2001USAfixed biopsiesMY09/114AIN10.0100.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.00.050.025.0 
Kreuter AJ Am Acad Dermatol, 2005Germanyexfoliated cellsDegenerate (A5/A10 and A6/A8)7AIN10.0100.057.142.957.10.014.357.142.914.328.60.0  14.314.30.00.00.00.00.0
Manzione CR2Rev Assoc Med Bras, 2004Brazilfixed biopsiesTS (Late proteins)30LSIL0.0100.06.70.0  13.3       23.3   43.3  
Ogunbiyi OABr J Surgery, 1994UKfresh biopsiesTS (Early proteins)13AIN1 53.853.8                  
Piketty CSex Trans Dis, 2004 (Piketty C et al., Ann Intern Med 2003)Franceexfoliated cellsMY09/1118LSIL0.088.922.227.85.611.111.116.75.60.05.60.00.00.044.40.00.05.633.311.1 
Varnai AD2Int J Colorectal Dis, 2006Germanyfixed biopsiesMY09/11, GP5+/6+8AIN162.575.00.00.00.00.00.00.00.00.00.00.00.00.062.50.00.00.025.00.00.0
Walts AEDiagnostic Cytopathology, 2005USAexfoliated cellsHybrid Capture46LSIL0.0100.0                   
Wieland UArch Dermatol, 2006Germanyexfoliated cellsTS (Early proteins)10AIN10.0100.060.020.040.00.010.040.040.010.040.00.00.00.010.00.010.020.010.010.00.0
TOTAL2    360  88.116.10.00.00.08.20.00.00.00.00.00.00.028.60.00.00.040.88.30.0

Ancillary