| |
Prevalence of and Risk Factors for Anal Human Papillomavirus Infection in Heterosexual Men
|
| |
| |
The Journal of Infectious Diseases June 15, 2008;197:1676-1684
Alan Nyitray,1Carrie M. Nielson,2Robin B. Harris,1Roberto Flores,3Martha Abrahamsen,3Eileen F. Dunne,4 andAnna R. Giuliano3
1Arizona Cancer Center, Tucson; 2Oregon Health and Science University, Portland; 3H. Lee Moffitt Cancer Center and Research Institute, Tampa, Florida; 4Division of STD Prevention, Centers for Disease Control and Prevention, Atlanta, Georgia
In US men, the incidence of anal cancer, the primary cause of which is human papillomavirus (HPV) infection, has increased almost 3-fold in 3 decades; however, little is known about the epidemiology of anal HPV infection, especially in heterosexual men. In 2 US cities, behavioral data and anal biological specimens were collected from 253 men who acknowledged having engaged in sexual intercourse with a woman during the preceding year. On the basis of DNA analysis, overall prevalence of anal HPV infection was found to be 24.8% in 222 men who acknowledged having had no prior sexual intercourse with men. Of the men with anal HPV infection, 33.3% had an oncogenic HPV type. Risk factors independently associated with anal HPV were lifetime number of female sex partners and frequency of sex with females during the preceding month. These results suggest that anal HPV infection may be common in heterosexual men.
Between 1973 and 2004, the incidence of anal cancer in US men increased from 0.5 to 1.3 per 100,000 [1]. Although human papillomavirus (HPV) is known to be the primary cause of anal cancer [2, 3], the prevalence of and the risk factors for HPV infection in heterosexual men, particularly at anal sites, have received little attention.
Two studies have reported 8% and 1.2% prevalences of anal HPV infection in asymptomatic men who did not acknowledge sex with other men [4, 5]; however, the prevalence of anal HPV can be >50% in other male populations, including heterosexual men with HIV infection [6, 7] and men who have sex with men [8]. When the penis and/or scrotum are the focus of HPV DNA testing in heterosexual men, prevalence estimates range widely, from 9% to 70%, depending on the testing method, study population, and anogenital locations of sampling [4, 5, 9-12].
Although no studies thus far have assessed factors associated with anal HPV infection in healthy heterosexual men, 2 studies of heterosexual men with anal warts and of heterosexual men with HIV infection found no risk factors associated with anal HPV infection [6, 13].
Investigations of the risk factors for genital HPV infection in men may also be important for understanding the risk factors for anal HPV infection. Such investigations have reported inconsistent results for genital HPV risk factors but include younger age, nonwhite race, Hispanic ethnicity, younger age at first sexual activity, lack of a steady sex partner, single marital status, increased frequency of sex, higher lifetime number of female sex partners and/or higher number of recent female sex partners, lack of circumcision, lack of condom use, smoking, and presence of genital warts [10, 12, 14-19]. The purpose of the present study was to assess the prevalence of and risk factors for anal HPV infection in asymptomatic men from 2 US cities who acknowledge having sex only with women.
Results
Prevalence.
In table 1, sociodemographic data on the 222 men included in the study are compared with those on the 241 men who were part of the parent study but who were excluded from the study. Men were excluded if they acknowledged having had sex with men or had not been asked or had refused to answer questions about it. There were 7 factors in which the men included in the study were statistically different from the men excluded from it: city of residence, race/ethnicity, marital status, lifetime number of female sex partners, circumcision status, smoking status, and level of smoking.
Of the men included in the study, 71.6% were 18-29 years of age, 59.9% were non-Hispanic white and 18.0% were Hispanic, and most men were single, divorced, or separated (80.0%). None of the men in this group acknowledged having HIV infection. Although 9 men had visible warts or lesions at genital sites, none of the men had visible warts or lesions at anal sites. Participants reported having had sexual intercourse with a median of 1 woman during the preceding 3 months and a lifetime median of 9 women.
Table 2 presents HPV-type distribution, by anatomical site, in the 222 men included in the study. In these men, the prevalence of HPV infection at the anal canal and the perianal region was 16.6% (n=36) and 21.3% (n=45), respectively, whereas the overall prevalence of anal HPV was 24.8% (n=55). Of the 55 men with anal HPV infection, 16 (29.1%) had it only at an anal site, and the remaining 39 (70.9%) had it both at an anal site and either at another anatomical site or in semen (data not shown). The prevalence of any oncogenic HPV type at an anal site was 5.9% (n=13), and the prevalence of any nononcogenic type was 13.1% (n=29). With regard to the anal canal specifically, 12 (33.3%) of the 36 men with HPV infection had an oncogenic HPV type; with regard to the perianal site, 9 (20.0%) of 45 men with HPV infection had an oncogenic HPV type. The most commonly detected HPV types at anal sites were types 68 and CP6108. In contrast to the prevalence of HPV infection at only an anal site, the prevalence of HPV infection at any of the 7 sites sampled was 71.2%, and the prevalence of oncogenic types was 36.0%. Of the 55 men with anal HPV infection, 65.5% (n=36) had a single HPV type, whereas 34.5% (n=19) had multiple HPV types, at one or both anal sites (data not shown).
Figure 1 presents the distribution of oncogenic, nononcogenic, or unclassified HPV infection in men with HPV DNA detected only at anal sites (n=16) and in men with HPV DNA detected only at genital sites (n=95). Of the 16 men with HPV DNA detected only at anal sites, 5 (31.3%) had only oncogenic types, whereas 6 (37.5%) had only unclassified HPV; of the 95 men with HPV only at genital sites, 20 (21.1%) had only oncogenic types, whereas 15 (15.8%) only had unclassified HPV.
A total of 25 men had an HPV genotype detected at both anal and external genital sites. These men afforded 61 opportunities to document concordance of identical HPV types at both anal and genital sites; a match of HPV types occurred only 21 times, for a 34.4% concordance between HPV types at anal and those at genital sites. Type 68 was the HPV type most commonly found at both anal and genital sites, with 6 men having it at both sites, and 1 man had 5 HPV types detected at both anal and genital sites.
Risk Factors.
In univariate analyses (table 3), the lifetime number of female sex partners (OR, 3.66 [95% CI, 1.06-12.62] for 11-20 female sex partners [compared with 1-5 female sex partners) and the frequency of sex with females during the preceding month (OR, 3.89 [95% CI, 1.03-14.63] for 2-4 times per month [compared with 0-1 times per month]) were independently associated with anal HPV infection. Circumcision had a marginally statistically significant and inverse association with anal HPV infection (OR, 0.34 [95% CI, 0.11-1.01]).
Discussion
We believe that the present study is the largest to examine prevalence of anal HPV infection in men who do not acknowledge oral or anal sex with other men and is the first to investigate factors associated with anal HPV infection in asymptomatic heterosexual men. The prevalence of anal HPV infection in the present study was 24.8%, which suggests that HPV may be a common anal infection in sexually active heterosexual men who have no visible anal warts or lesions. This finding parallels research indicating that anal HPV infection is also common among sexually active women [23]. Of the men with anal HPV infection, 33.3% had an oncogenic HPV type.
The prevalence of anal HPV reported in the present study may be higher than that in 2 previous studies [4, 5] because of (1) differences in sampling technique, (2) genotyping of samples regardless of PCR-based HPV results, and (3) genotyping of 37 HPV types detectable in the assay [21]. Van Doornum et al. tested 85 men from a sexually transmitted disease clinic who acknowledged sex only with women and found that 1.2% of them were positive for anal HPV infection [5]. Those investigators detected DNA by using a 4-primer PCR system that detected 5 HPV types. Nicolau et al., using hybrid capture technology that detected 18 HPV types, reported an 8% prevalence of anal-canal HPV infection in 50 Brazilian men who were heterosexual partners of women with confirmed HPV infection [4].
Given the composition of the samples in these 2 previously published studies, the present studyfs use of a general community sample may not help to explain the higher prevalence that it found. Also, although the prevalence of anal HPV infection can be >50% in some male populations, including heterosexual men with HIV infection [6, 7] and men who have sex with men [8], none of the men in the present study acknowledged either having HIV infection or having ever had sex with other men. It is possible that we may have misclassified some of the menfs sexual behavior. Same-sex sexual behavior remains a highly socially stigmatized behavior, and some men avoid acknowledging it [24]. If we incorrectly classified bisexual men as men who only have sex with women, then the prevalence of anal HPV infection in heterosexual men may be lower. However, in the men who completed the expanded questionnaire and acknowledged ever having had oral or anal sex with another man (n=29), the prevalence of anal HPV infection was comparable to that in men who did not acknowledge ever having had sex with another man (20.7% vs. 24.8%;p=.63). In fact, there were no statistically significant differences between these 2 groups, with the exception that men who acknowledged ever having had sex with another man were more likely to be smokers (p=.005). However, a limitation of the present study is that its sample size was insufficient to allow adequate evaluation of the differences between men who acknowledged ever having had sex with another man and those who did not. Future studies are necessary to fill this information gap.
The most common HPV types detected at anal sites in the 222 men included in the study were types 68 (4.5%) and CP6108 (2.3%). Neither of these types is known to have oncogenic potential [22].
The estimate of the prevalence of HPV infection includes 37 types of HPV in addition to unclassified HPV types. Compared with men with only genital HPV infection, men with only anal HPV infection had a higher proportion of unclassified HPV types (37.5% vs. 15.8%; ). These results may be due to either different mechanisms being responsible for virus transmission or a different rate of transmission to different anatomical regions. Also, Hernandez et al. reported a higher proportion of unclassified HPV types in women with only anal HPV infection, compared with women with concurrent anal and cervical infection [23].
It is unlikely that these results are due to study-related contamination caused by inadvertent transfer of HPV DNA either from genital to anal sites or from the perianal region to the anal canal, because rigorous sampling methods were used by the study's clinicians. However, the unclassified HPV types observed in the present study may include types not thought to be sexually transmitted or may be false-positive results of PCR testing. For this reason, when we analyzed risk factors associated with anal HPV infection, we excluded men with only unclassified HPV types, thereby reducing, from 55 to 47, the number of men with anal HPV infection who were analyzed.
The risk factors that the present study found to be associated with anal HPV infection are similar to those which previous studies have found to be associated with genital HPV infection [10, 12, 16, 19, 21]. In the present study, the lifetime number of female sex partners and the frequency of sex during the preceding month were independently associated with anal HPV infection, whereas lack of circumcision was marginally associated with it. The result regarding the frequency of sex during the preceding month is noteworthy because it was limited to men who acknowledged a sex frequency of 2-4 times per month. Conversely, men who acknowledged the higher frequencies of sex-that is, 5 or more episodes-during the preceding month were not at increased risk for anal HPV infection. This finding may be spurious. Another explanation is that heterosexual men who have the most-frequent sex are more likely to be in a monogamous relationship and therefore are less likely to have exposure to HPV infection.
Finally, the data of the present study suggest the possibility that the prevalence of anal HPV infection is lower in heterosexual men in their 30s than in younger heterosexual men. Conversely, in their study of men who have sex with men, Chin-Hong et al. reported a stable age-specific prevalence of anal HPV infection [8]. If such a trend of decreasing age-specific prevalence in heterosexual men is confirmed by future studies, it could be considered to be due to a number of reasons, including a lack of persistence of anal HPV infection in older heterosexual men or a difference between the sexual behavior of younger heterosexual men and that of older heterosexual men.
There may be other, unmeasured factors that are associated with anal HPV infection, because none of the factors identified in the present study's univariate or multivariate analyses explain how HPV was transmitted to the anal region. Our questionnaire gathered data about penetrative sexual practices-specifically, vaginal, anal, and oral intercourse-implicated in HPV transmission between men and women; however, none of these behaviors seems able to transmit HPV to the anal region of a man who acknowledges having had sex only with women. Our questionnaire did not ask about other sexual behaviors, such as self-initiated or partner-initiated anal massage or anodigital insertion. It is noteworthy that one study has reported an association between nonpenetrative sex (finger-vulvar, penile-vulvar, and oral-penile) and genital HPV infection in female virgins [25]. In addition to sexual behaviors involving the perianal region or anal canal, nonsexual behaviors may also help explain the prevalence of anal HPV infection in heterosexual men. Three studies hint at nonsexual HPV transmission via hand carriage [26, 27] and objects [28].
If HPV DNA can be transmitted by fingers or objects, then self-transference from the penis or scrotum to the perianal region or anal canal may occur as a result of sexual or nonsexual behaviors. This explanation requires that male genital infection be included in the causal pathway between the risk factors that we measured and anal HPV infection. Accordingly, it is possible that the present study identified risk factors associated with genital, in addition to anal, HPV infection. Indeed, in univariate analysis, the presence of genital HPV DNA and the presence of anal HPV DNA were associated in these men; however, type-specific concordance between genital and anal HPV infection was limited.
Additional limitations to the present study include its sample size. It did not have sufficient power to detect risk factors for HPV infection at specific anatomical sites, such as the perianal region or the anal canal. Another limitation is that it is possible that detection of HPV DNA on the skin surface might occur in the absence of viral infection of the basal layer of the epithelium.
The subsample of men included in the study differed from the men excluded, with regard to the following 7 factors: city of residence, race/ethnicity, marital status, lifetime number of female sex partners, circumcision, smoking status, and level of smoking. Because recruitment occurred in Tampa and Tucson, variation in geography and/or recruitment practices may explain some differences in these variables [29, 30]. However, the prevalence (22.8%) of anal HPV infection in the excluded men was comparable to that in the included men, a finding that makes it less likely that exclusionary characteristics affected the results.
To our knowledge, the present study is an important first examination of prevalence anal HPV infection in a community sample of healthy heterosexual men. We do not believe that its data should be used, at this time, the basis for a change in clinical practice; however, further research that characterizes anal HPV infection in other populations of men who acknowledge sex only with women is necessary. We believe that the present study is also the first examination to assess risk factors for anal HPV infection in healthy heterosexual men. However, a clear understanding of the full complement of risk factors for anal HPV infection in heterosexual men will require more study. Future studies should collect sexual- and nonsexual-exposure information that may clarify routes of transmission to the perianal region and anal canal. Finally, it is important to assess factors associated with persistent oncogenic anal infection in prospective cohort studies of both younger and older men.
Subjects, Materials, and Methods
Study design and questionnaire.
Study design and participant recruitment have been reported elsewhere [20, 21]. In brief, using a cross-sectional study design, we recruited men in Tucson, Arizona, and Tampa, Florida, who acknowledged having had sexual intercourse with a woman within the preceding year. Participants were 18-40 years old; had no current diagnosis of sexually transmitted disease (STD), no pain during urination, and no penile discharge; and acknowledged no history of genital warts, anal cancer, or penile cancer. Recruitment was promoted via flyers posted on college and university campuses, newspaper advertisements, radio, direct mail, and face-to-face enrollment at an STD clinic. All participants consented to the study protocol, which was approved by the human subjects-protection committees of the University of Arizona, University of South Florida, Centers for Disease Control and Prevention, and US Department of Defense.
Participants completed a self-administered 51-item written questionnaire that included questions regarding demographic characteristics, alcohol and tobacco consumption, and sexual behavior. Approximately halfway through recruitment, 11 questions dealing with same-sex sexual behavior were added to the questionnaire.
Collection of biological samples.
Using a different saline-wetted Dacron swab for each anatomical site, clinicians sampled the following 6 anogenital sites from each of the men: urethra, coronal sulcus/glans penis, penile shaft (including prepuce if present), scrotum, perianal region, and lower anal canal between the anal verge and the dentate line. In addition, participants self-collected a semen sample for analysis. The urethral and semen sampling were discontinued in the third year of the study because the proportions of HPV-positive samples from these sites were much lower than those for other sites. For the 253 men who completed the expanded survey instrument, 2 clinicians completed all sampling of the men. Each clinician was trained in sampling techniques, which included instruction on how to avoid contamination of the sampling sites. The clinicians also drew blood and examined the anal region and genitals for lesions, abrasions, discharges, or other abnormalities. Lesions and warts were sampled via a wet Dacron swab, and their locations were recorded.
To preserve DNA for analyses of HPV, the clinician placed each wet Dacron swab into its own vial, along with Digene Specimen Transport Medium, and then immediately refrigerated the samples at 4°C. Samples were then transferred to a freezer at _70°C and were stored there until HPV testing was conducted.
HPV testing.
All samples were analyzed for HPV DNA and β-globin, as described elsewhere [20, 21]. In brief, to identify HPV DNA, laboratory staff used the polymerase chain reaction (PCR) consensus primer system (PGMY 09/11) to amplify a fragment of the HPV L1 gene. HPV genotyping was then conducted on all samples, by use of DNA probes labeled with biotin, to detect 37 HPV types: 6, 11, 16, 18, 26, 31, 33, 35, 39, 40, 42, 45, 51-56, 58, 59, 61, 62, 64, 66-73, 81-84, IS39, and CP6108. A sample was considered to be positive if either the PCR or genotyping tests detected HPV DNA. Oncogenic HPV types were 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 66 [22]. For any of the 37 types, a sample that was found to be positive by PCR but was not found to be positive by genotyping was labeled as having "unclassified" HPV. Accuracy and possible contamination were assessed by use of nontemplate negative control samples and CaSki DNA positive control samples. β-globin positivity by genotyping was 95.5% and 90.5% at the anal canal and perianal region, respectively.
Statistical analyses. Of the total study population of 463 men, 253 participants completed the expanded questionnaire. Of these 253 men, 29 men (11.6%) acknowledged having had anal and/or oral sex with another man, and another 2 men (0.8%) either refused to answer questions about sex with men or left the questions blank. The remaining 222 men, who denied having ever had anal and/or oral sex with men, were included in the study. Tests of proportion were used to assess sociodemographic differences between the study population and the men excluded from the study.
We combined test results from the anal canal and perianal region and then created a binary-outcome variable that indicated either the presence or absence of anal HPV DNA, regardless of the presence of genital, urethral, and seminal HPV infection. The case group consisted of men with anal HPV infection (at the anal canal and/or perianal region); the comparison group consisted of men who did not have anal HPV infection but may have had such infection at a genital site or in semen. Therefore, the comparison group consisted of both men with semen or a genital site positive for HPV and men with semen or a genital site negative for HPV.
A high proportion (7.7%) of men in the study had "unclassified" infection at anal sites. To lessen the potential for misclassification, the analysis of factors associated with anal HPV infection was restricted to men with HPV infection detected by genotyping, thus reducing the study population from 222 to 198 men.
Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated by univariate and multivariate logistic regression. Variables having statistically significant univariate associations with anal HPV infection (p<.05 ) and variables with p<.20 by a likelihood-ratio test were initially included in the model. Independent risk factors for anal HPV infection were identified by a backwards-elimination logistic regression with robust variances. Variables p>.10 with by a Wald test were individually removed until a final set of risk factors remained. Variables that previously had been rejected by the Wald test were again assessed both for significance (considered to be p<.10) in the final model and for their potential role as confounders. Confirmed confounders were then included in the final model.
Multivariate analyses were adjusted by time of laboratory analysis because a trend of increasing prevalence was observed during the first half of the laboratory-analysis period. This trend was attributed to improvements in laboratory methods. Data were analyzed by use of Intercooled Stata (version 9.2) for Windows and SAS (version 9.1; SAS Institute, Inc.) programs.
|
|
| |
| |
|
|
|
