Human papillomavirus (HPV) is transmitted by skin to skin and sexual contact and certain genotypes are now understood to be the cause of cervical cancer.1 Significant work has been done on understanding the prevalence and natural history of HPV in women2 and increasingly with men who have sex with men (MSM) populations.3 However, relatively little is known about HPV infections in heterosexual men.4 5 One of main issues for studying HPV in heterosexual men remains the challenge of specimen collection.6^–8 Data are evolving regarding which genital sites in men should be tested for HPV and how specimens should be collected in order to determine the epidemiology of HPV genital infections for the men themselves and for transmission to sexual partners. Because HPV is a virus that infects the epithelium, studies that have been conducted on HPV in heterosexual men rely on epidermal debriding followed by specimen collection by a saline-moistened Dacron swab.9 However, because men are often reluctant to participate in testing for sexually transmitted infections (STIs), and may find epidermal debriding unappealing, self-sampling presents an important option to increase uptake of testing both in clinical and research settings. In this study, we assessed the accuracy of self-collected HPV specimens in men compared with clinician-collected specimens and determined the prevalence of HPV subtypes at different male genital sites from heterosexual men in British Columbia.

METHODS

Men who reported never having sex with another man and who presented for STI screening at the Provincial STI Clinic at the British Columbia Centre for Disease Control in Vancouver, Canada, were invited to participate in this study. After providing informed consent, participants were administered a structured questionnaire by an experienced, trained healthcare provider to assess sexual history, risk factors for STIs, sociodemographic variables and self-reported history of STIs, including human papillomavirus, Chlamydia trachomatis, Neisseria gonorrhoea and herpes simplex virus. Study participants received honoraria of $CAN15.

Specimen collection

All participants self-collected their specimens and underwent clinician-collected specimen testing. Participants were randomly assigned to conduct self-collection or clinician-collected specimens first and to obtain their self-collected specimen from one of three sites. Clinical staff obtained specimen staff obtained specimens using emery paper followed by a saline-moistened Dacron swab from three genitourinary sites: glans penis/foreskin, penile shaft (ventral and dorsal surfaces) and scrotum.9 Participants received written instructions and took specimens from one of the three sites using the same technique as clinicians.

Laboratory analysis

Both the emery paper and the swab from each sampling site were placed together into a sample transport tube containing 1 ml of specimen transport medium (Digene Corporation, Gaithersburg, Maryland, USA), which was stored at room temperature (18–30°C) for no more than 48 hours until delivered to the laboratory. HPV testing was performed with the Roche Amplicor HPV test (Roche, Canada) following the manufacturer’s instructions. All samples found to be reactive with the Amplicor micro-well plate assay were subsequently tested with the Roche Linear Array HPV typing assay to establish the HPV genotype(s) in the sample. Both the Amplicor and Linear Array tests include a β-globin control to assess sample adequacy. Specimens testing negative for β-globin were considered to be inadequate and were not included in the analysis.

Statistical analysis

Demographic characteristics, sexual history and sexual practices from the study participants were determined. Rates of sample adequacy were compared between clinician-collected and self-collected specimens. Prevalence of HPV overall and for different genotypes and at different genital sites was determined. Overall prevalence rates were compared between study participants where clinicians collected specimens first and participants collected specimens first using χ² statistics. Characteristics of men who were positive for any HPV and for HPV-16 or HPV-18 were compared with men who were negative for HPV-16 or HPV-18 using χ² statistics for categorical variables and independent t test for continuous variables. Variables that achieve p<0.05 were offered for inclusion in a multivariable model to achieve a best-fit model. Backwards logistic regression was conducted to calculate adjusted odds ratios (AORs) to identify the factors that were predictive of overall HPV status and HPV-16 or HPV-18 status. Concordance in genotype-specific HPV detection between self-collected and clinician-collected specimens at different sites was determined using Cohen’s κ statistic. Analysis was performed using SPSS 14.0 for Windows.

RESULTS

Over a 10-month period, 297 men were recruited and, overall, 88% (n = 262) of the specimens were deemed adequate as confirmed by detection of β-globin in both the clinician-collected and the self-collected specimens and were included in the analysis. There was no difference in rates of specimen adequacy between clinician-collected and self-collected specimens (86.9% vs 89.5%; p>0.05). Study participants ranged in age from 16–69 years (median 29 years). Median age of first intercourse was 17 years, with a median of 15 lifetime sexual partners and 2 in the past 6 months. Just over a quarter of participants reported condom use all the time, but over 50% reported condom use during their last sexual encounter. Over 50% reported a history of some type of STI at some point in their life (table 1).

Overall prevalence of any HPV genotype from any site was 69.8% in clinician-collected specimens and 55.3% in self-collected specimens (table 2). Order of collection (clinician vs self-collected) did not impact on the prevalence of HPV in the specimens (65.9% vs 73.5%; p = 0.18). Prevalence of vaccine preventable genotypes (HPV 6, 11, 16 or 18) and high risk types (HPV 16, 18) in clinician-collected specimens was 35.1% and 24.0%, respectively, and 43.1% had evidence of infection with more than one HPV subtype at any site. Overall prevalence of HPV-16 was 18.7% and 8.4% for HPV-18. There was no consistent pattern in overall detection as prevalence at some sites was higher with clinician-collected specimens (penile shaft) and others higher with the self-collected specimens (glans/foreskin and scrotum). In bivariate analyses, individuals who were HPV positive were significantly more likely to have had their first sexual encounter at a younger age, have a higher number of lifetime sexual partners, have a higher number of sexual partners in the past 6 months, report never using condoms and self-report having had a STI. In multivariable modelling, individuals who were HPV positive were significantly more likely to have had their first sexual encounter at a younger age, have a higher number of lifetime sexual partners and were more likely to self-report having had a STI. Individuals who were positive for either HPV-16 or HPV-18 were significantly less likely to report that they always used a condom with sexual partners (table 3). In multivariable modelling, never having used a condom with partner in the last 6 months was significantly associated with HPV-16 or HPV-18 infection compared with those who always used a condom.

Correlation of type-specific HPV detection at the three different genital sites was evaluated between the clinician and self-collected specimens. The κ scores for agreement between clinician-collected and self-collected specimens ranged from fair to excellent (table 4). The best agreement was found for HPV-18 in the shaft (κ = 0.92) and the poorest agreement was found for HPV-16 for the glans/foreskin (κ = 0.36). Overall, there was better agreement between clinician-collected and self-collected specimens for HPV-18 (κ = 0.88 t 0.92) than for HPV-16 (κ = 036 to 0.62).

DISCUSSION

In this study of heterosexual men recruited at a provincial STI clinic, over 60% were HPV positive and almost a third were positive for HPV types contained in the quadrivalent HPV vaccine. In the self-collected specimens, where participants sampled only one site, overall prevalence was 55.3%. Using the emery paper and swab system, men were able to self-collect specimens from the glans/foreskin, penile shaft and scrotum for HPV and were able to provide adequate specimens based on β-globin detection. Order of specimen collection (clinician or self) did not appear to impact on the detection of HPV, indicating that self-collected specimen methods are not influenced by either viewing the nurse collect the specimen or by nurses debriding the epithelium prior to patient collection. However, site-specific agreement for specific HPV subtypes between clinicians and patients varied broadly and neither clinicians nor patients routinely obtained samples with consistently higher or lower prevalence at specific genital sites, indicating there are continued opportunities to improve techniques for male self-collected specimens for HPV, including overall acceptability of the method and development of improved sampling tools.

The overall prevalence for any HPV found in this study (69.8%) was higher than rates found in other studies where both younger men were included10 and in studies where men aged between 18 and 44 years were included.11 In studies that included glans/foreskin, penile shaft and scrotum, using similar specimen collection techniques, prevalence for any HPV infection ranged from 35%,9 to just over 40%12 to almost 60%.6 In this study, only 20% of participants were under the age of 25 years compared with 49.9% in Nielson’s study.7 In another study on HPV prevalence in men, where overall prevalence was 33.8%, the mean age of participants was 20.3 years compared with 32.0 years in this current study.13 Thus, given recent findings by Partridge, which showed the cumulative incidence of new HPV infections in men at 62.4%,14 our prevalence findings of 69.8% are not unexpected. We recruited men at the provincial STI clinic who were older than those recruited in other studies and who reported a high number of lifetime partners; thus, increasing their exposure to HPV than other studies of university students, the military or community based samples.

The reason for the difference in the level of agreement between self-collected and clinician-collected HPV genotypes is unclear and warrants further examination. Certainly, it has been established that there is variation in the duration of infection with different HPV genotypes in women as well as differing characteristics in terms of clearance rates for prevalent versus incident infections.15 Perhaps the level of integration into the epithelium or viral load in prevalent compared with incident infections impacted on the level of agreement for HPV-16. Further research on the characteristics and natural history of type-specific HPV infection in men will likely offer important insight into the potential reason for the variation in agreement between HPV genotypes.

Genital HPV infection is common and infects the genital region broadly. Similar to the studies by Nielson7 and Partridge,14 HPV prevalence for any type was highest on the penile shaft, then for the glans/foreskin and lowest in the scrotum, although this difference did not achieve statistical significance. Site-specific prevalence for HPV-16 overall, at the glans/foreskin, shaft and scrotum were 18.7%, 12.6%, 13.7% and 10.7%, respectively, in this study compared with 11.4%, 9.1%, 4.1% and 5.0% in Nielson’s study.7 This study indicates the high prevalence of vaccine-preventable HPV carriage in heterosexual men and, therefore, the possible benefit of the HPV vaccine for men to reduce the risk for HPV-associated penile carcinomas5 and to reduce HPV transmission to women. There is a clear need for continued study to establish the clinical efficacy of the HPV vaccine in heterosexual men and to consider the role of HPV vaccination in young men.16