Age-Specific Prevalence, Incidence, and Duration of Human Papillomavirus Infections in a Cohort of 290 US Men

Abstract

Background. Human papillomavirus (HPV) infections cause disease in men and women, and male-to-female HPV transmission influences the risk of cancer in females. The purpose of the present study was to describe the overall and age-specific incidence and clearance of HPV infections in men.

Methods. In a prospective cohort study of 290 men aged 18–44 years, participants were examined at baseline and every 6 months, with a mean duration of follow-up of 15.5 months.

Results. The period prevalence was 52.8% for any, 31.7% for oncogenic, and 30.0% for nononcogenic HPV infection. The 12-month cumulative risk of acquiring a new HPV infection was 29.2%. Incidences of HPV types 6, 11, 16, and 18 were 2.8, 0.5, 4.8, and 0.8 per 1000 person-months, respectively. The median time to clearance of any HPV infection was 5.9 months (95% confidence interval, 5.7–6.1 months), with comparable times to clearance for oncogenic and nononcogenic infections. Approximately 75% of men tested negative for any HPV 12 months after initial HPV detection. Age was not significantly associated with HPV incidence or duration of infection in men.

Conclusion. HPV infection in men was common, with relatively rapid rates of acquisition and clearance.

Infection with the sexually transmitted human papillomavirus (HPV) has been strongly linked to disease in men and women. In the United States in 2007, ∼25,000 cases of cancer in men and women were estimated to be attributable to infection with HPV [1, 2]. Such cancers include those of the cervix, vagina, vulva, anal canal, and penis and of the oral, head, and neck regions. Anogenital warts are a more common outcome of HPV, with 205 persons per 100,000 infected each year in the United States [3]. In addition, male-to-female HPV transmission significantly influences the risk of disease in females [4–10]. In case-control studies, the risk of cervical disease has been significantly higher among women whose male partners had a greater number of sex partners and HPV infection [4–10].

Because male sexual behavior clearly affects rates of HPV infection, cervical dysplasia, and invasive cervical cancer in female partners, a greater understanding of HPV infection in men is an essential component of cervical cancer prevention. This is especially true as we approach completion of trials testing the efficacy of HPV vaccines in men and as sex-neutral vaccination policies increasingly become possible [11]. Unfortunately, little is known of the natural history of HPV infections in men. To date, 3 small prospective studies of HPV infection in men have been conducted in Europe [12–14], 1 has been conducted in Latin America [15], and 1 has been conducted in the United States [16]. The purpose of the present study was to define the overall and agespecific incidence and duration of type-specific HPV infection among 18–44-year-old men residing in Tucson, Arizona.

Methods

Study design. A prospective study of the natural history of HPV infections among men aged 18–44 years was conducted in Tucson. Cohort participants were recruited via targeted community mailings and posters placed on university and college campuses, face-to-face recruitment in classrooms and county health clinics, publicity at mass events, and radio and newspaper advertisements from September 2003 through May 2005. Follow-up ended in December 2005. Participants underwent clinical examination at a preenrollment visit 2 weeks before the baseline visit and then at baseline and every 6 months thereafter, with follow-up visits at 6, 12, and 18 months. Because of variations in the actual date when men participated in these clinics, the maximum duration of follow-up was 24.7 months. At each visit, men completed a self-administered risk factor questionnaire, and penile and scrotal cell samples were collected. Clinical and laboratory measurements included HPV DNA detection by polymerase chain reaction (PCR) and genotyping and visual inspection of the skin and external genitalia for the presence of warts and lesions. To encourage compliance with follow-up, men were paid for their participation.

Study population. Men were eligible for participation if they (1) were 18–44 years old, (2) were residents of southern Arizona, (3) reported no prior diagnosis of penile or anal cancer, (4) reported no past diagnosis of genital warts, (5) reported no diagnosis or treatment of sexually transmitted infections, (6) were willing to comply with 4 scheduled visits at 6-month intervals, and (7) had no plans to relocate within the next 2 years. Men called a study-specific telephone number and were interviewed by phone for their interest in the study, to assess eligibility, and to schedule an appointment to complete the first clinical visit, the run-in (or preenrollment) visit. Recruitment generated 676 calls of inquiry. In addition, 464 men were approached for recruitment at the local county health department clinic. A total of 618 men met the eligibility criteria (54.2%). The run-in visit allowed potential study participants to experience all aspects of the study before formally enrolling at the baseline clinic visit scheduled 2 weeks later. Only men returning for the baseline visit were considered cohort members. A total of 377 men completed the run-in visit (61.0% of eligible men), 337 of whom returned ∼2 weeks later for the baseline study visit (89.4%). Sixteen of the 337 (4.7%) who returned had inadequate samples for HPV analysis, leaving 321 enrolled men with adequate samples. A total of 302 men returned (89.6% retention from baseline) for the third visit at 6 months, 12 (4.0%) of whom had inadequate specimens for HPV DNA detection. The present analysis includes the 290 men who completed at least 2 clinic visits scheduled 6 months apart and who had adequate samples for HPV DNA detection.

HPV penile and scrotal sampling. To avoid interspecimen contamination, warts or lesions found on the genitals were sampled by swabbing the wart or lesion with a saline-prewetted Dacron swab before sampling of normal genital skin. The swab was placed in 450 μL of fixative solution designed to preserve DNA (specimen transport medium; Digene). The next 3 swabs were placed in separate vials labeled by anatomic site, each containing 450 μL of specimen transport medium, as described and validated elsewhere [17]. The clinician used a new saline-prewetted swab and swept 360 degrees around the coronal sulcus and then another 360 degrees around the glans penis. A second swab was used to sample the entire skin surface of each of the quadrants of the shaft of the penis, and a third swab was used for scrotal sampling. All HPV samples were stored at −70°C until PCR analyses and genotyping were conducted. Before DNA extraction, the 3 samples of anogenital skin were combined to produce 1 DNA extract per participant clinic visit. We have demonstrated elsewhere that this method has high sampling reproducibility for the detection of HPV DNA [18].

HPV analyses. HPV testing of swabbed cellular material was conducted using PCR for amplification of a fragment of the L1 gene [19]. DNA extraction was performed using the QIAamp DNA Mini Kit (Qiagen) according to the manufacturer's instructions. DNAwas eluted with 50μL of 10 mmol/L Tris-EDTA buffer (pH 7.5) at 60°C and stored at −20°C until use.

Specimens were tested for the presence of HPV by amplifying 5 μL of the DNA extracts by means of the PGMY09/11 L1 consensus primer system [19]. HPV genotyping was conducted using the reverse line blot method [20] for all samples, regardless of the HPV PCR result (Roche Molecular Diagnostics). Only samples that tested β-globin positive (99% at baseline) were deemed adequate and included in this analysis. Before genotyping, 2% agarose gels were run to visualize a 450-bp band corresponding to HPV amplification, to identify samples that may harbor HPV other than the 37 types included in the genotyping assay. Samples that amplified HPV by PCR but did not hybridize with a specific HPV type at genotyping (e.g., unclassified infections) were categorized as HPV negative in this analysis. The following 13 HPV types were categorized as oncogenic: 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 66 [21]. The other (nononcogenic) types detected by the line blot methodology of Roche were 6, 11, 26, 40, 42, 53, 54, 55, 61, 62, 64, 67–73, 81–84, IS39, and CP6108.

Statistical analysis. Sociodemographic and sexual behavioral characteristics were compared between participants and nonparticipants by Pearson's χ² or Fisher's exact test for categorical variables and by Student's t test for continuous variables. Prevalence at enrollment, period prevalence, and incidence rates were estimated for any, oncogenic, and nononcogenic HPV type and by specific HPV type. HPV infections with single or multiple oncogenic HPV types or mixed infections with both oncogenic and nononcogenic HPV types were classified as oncogenic infections, and single or multiple infections with nononcogenic types only were classified as nononcogenic infections. Period prevalence was defined as the proportion of participants who tested positive at any study visit, including the baseline visit. For incidence rate estimates, only participants who were free of any HPV or a specific HPV type at baseline were included. Two types of incidence rate estimates were calculated for any HPV, oncogenic HPV, and nononcogenic HPV infections: (1) the incidence of total type-specific infections acquired during the follow-up period with the entire follow-up period used for person-time estimation, because an individual can be at risk for infection with a different genotype throughout the follow-up period, and (2) the incidence of first-acquired new HPV infections with the time elapsed from study entry to the date of the first HPV DNA detection used for person-time estimation, assuming that a new infection occurred at the date of detection. The exact 95% confidence interval (CI) calculated for an incidence estimate was based on the number of events modeled as a Poisson variable over the total number of person-months [22].

The cumulative risk of acquiring a new infection was estimated for any HPV, HPV-6 or -11 (HPV-6/11), HPV-16 or -18 (HPV-16/18), oncogenic HPV, and nononcogenic HPV infections, using the Kaplan-Meier method and the log-rank test [23]. The time to infection was measured in 2 ways: (1) from the date of study entry to the date of the first positive HPV test result and (2) as the total number of new infections observed from study entry throughout the follow-up period. Participants who were free of HPV infection throughout the follow-up period were censored at the date of the last study visit. The duration of infection was estimated for any HPV, HPV-6/11, HPV-16/18, oncogenic HPV, and nononcogenic HPV infections, again using the Kaplan-Meier method [23]. Prevalent HPV infections from baseline and incident infections with at least 1 subsequent follow-up visit were included in these analyses. The time to clearance or duration of an HPV infection was defined as the time elapsed from the date of the first positive HPV test result to the date of the first negative HPV test result after the last detection of HPV infection. Participants whose infection(s) did not clear were censored at the date of the last positive HPV test result. Age-specific incidence, duration of infection, and prevalence were estimated for any HPV, oncogenic HPV, and nononcogenic HPV infections. Age-specific incidence and duration of infection were also calculated for type-specific infections. Differences in HPV incidence by age were assessed on that basis of the CIs derived from the Poisson distribution. The log-rank test was used to test differences in the clearance of HPV infections by age. Differences in point prevalence and period prevalence across age groups were assessed by Pearson's χ² test. All analyses were performed with SAS software (version 9.1.3; SAS Institute).

Results

The median duration of follow-up for the 290 study participants was 15.5 months (range, 3.7–24.7 months), with a median duration of 5.3 months between follow-up visits. There were no significant differences between participants and nonparticipants (those who did not return for a baseline examination) with respect to demographic characteristics or sexual behavior factors. However, uncircumcised men were more likely than circumcised men to drop out after the run-in visit (70.0% circumcised among nonparticipants vs. 87.6% among participants; P = .006). Table 1 presents the distribution of sociodemographic, clinical, and sexual behavior risk factors at baseline among men included in this study. The mean age of participants was 29.7 years (SD, 8.0 years). Most study participants were white (86.4%), non-Hispanic (80.1%), circumcised (87.6%), and nonsmokers at enrollment (68.8%). At enrollment, 3.2% of men reported having 0 lifetime sex partners, 46.4% reported 1–8, and 50.4% reported ⩾9. The majority of men reported sexual activity with women only (87.8%), and 63.4% of men reported no new sex partners during the past 3 months. Within the past 3 months, 40.9% reported never using condoms.

Table 1

Sociodemographic, clinical, and behavioral characteristics of the study participants at baseline (n = 290).

Table 2 presents the period prevalence and incidence of HPV infection for the study period. The prevalence of any HPV type at enrollment was 30.0%, with a period prevalence of 52.8% (31.7% for oncogenic and 30.0% for nononcogenic infections). The incidence of any HPV infection over the study period was 42.3 per 1000 person-months and 29.4 per 1000 person-months for the first HPV type detected. The incidences of HPV-6, -11, -16, and -18 infections were 2.8, 0.5, 4.8, and 0.8 per 1000 person-months, respectively. The cumulative incidence of HPV infection among men is presented in figure 1. The 12-month incidence of a new HPV infection was 29.2%, and the cumulative incidence was 42.3%. Overall, the proportion of the population acquiring an oncogenic infection by 12 months was higher than the proportion acquiring a nononcogenic infection (19.0% vs. 16.0%). The proportion of the population acquiring a new HPV-16/18 infection by 12 months was 5.1%, compared with 3.4% for HPV-6/11 infections.

Table 2

Prevalence at enrollment, period prevalence, and incidence of human papillomavirus (HPV) infections among men.

Figure 1

Kaplan-Meier estimates of the cumulative incidence of any human papillomavirus (HPV) infection (A), HPV-6/11 vs. -16/18 infection (B), oncogenic HPV infection (C), and nononcogenic HPV infection (D). CI, confidence interval.

The time to clearance of infection among men who had a baseline HPV infection or who acquired an infection during the follow-up period is presented in figure 2. The median time to clearance of any HPV infection (defined as the time for 50% of infections to be cleared) was 5.9 months (95% CI, 5.7– 6.1 months), with ∼75% of infections cleared by 12 months after initial HPV DNA detection. The durations of infections were similar for nononcogenic and oncogenic HPV types, with a median time to clearance of 5.8 months (95% CI, 5.5–6.1 months) for oncogenic infections and 6.0 months (95% CI, 5.5–6.5 months) for nononcogenic infections. The median times to clearance of infection were similar for HPV-6/11 (5.4 months; 95% CI, 5.1–5.7 months) and HPV-16/18 (6.0 months; 95% CI, 5.2–6.8 months). At 6 months, 44.5%, 41.7%, and 46.5% of subjects remained positive for any HPV, oncogenic HPV, and nononcogenic HPV, respectively, compared with 25.2%, 19.0%, and 29.3% at 12 months and 10.7%, 0%, and 14.4% at 18 months.

Figure 2

Kaplan-Meier estimates of clearance and median time to clearance (t₅₀) for any human papillomavirus (HPV) infection (A), HPV-16/18 vs. -6/11 infection (B), oncogenic HPV infection (C), and nononcogenic HPV infection (D). CI, confidence interval.

Overall and oncogenic HPV prevalence at enrollment and period prevalence did not differ by age (figure 3). Significant differences in period prevalence (P = .009) were observed for nononcogenic infections by age. There were significant differences by age in the rate of acquisition of HPV infections and in the median durations of infections (figure 4). However, there was a nonsignificant difference in nononcogenic HPV infection acquisition by age, with rates increasing after age 30 years, similar to the period prevalence trend observed in age groups >30 years. In addition, a nonsignificant spike was observed in the median durations of any HPV and nononcogenic HPV infections among 31–35-year-old men.

Figure 3

Age-specific prevalence of human papillomavirus (HPV) at enrollment (A–C) and period prevalence (D–F), by oncogenic potential. The period prevalence for nononcogenic HPV infection (F) differed significantly by age group (P = .009).

Figure 4

Age-specific incidence of human papillomavirus (HPV) infection (A–C) and median duration of infection (D–F), by oncogenic potential.

Discussion

To our knowledge, this is the first report of the natural history of HPV infections in men over a broad age range (18–44 years) who had >2 repeated measures of HPV and a high proportion (99%) of adequate samples (β-globin positive) for HPV testing. Although there have been several reports of HPV prevalence among men from cross-sectional studies (reviewed by Dunne et al. [24]), only 3 published studies have estimated the incidence and persistence of HPV infections in men [14–16]. In a study conducted by Lajous et al. [15] among Mexican military men, the incidence of any type of infection was 17.9 per 1000 personmonths [15]; 1030 men were enrolled in that study, but only 336 were reevaluated for HPV infection at a second visit scheduled 12 months after baseline. Given our observed median time to clearance of ∼6 months, this study may have missed the acquisition and subsequent clearance of some infections in the study population, which may have led to a reduced observed estimate of HPV incidence and would explain the lower incidence observed compared with that in our study (29.4 per 1000 personmonths). Interestingly, Lajous et al. reported a period prevalence of 50.9%, similar to that observed in the present study (52.8%).

Kjaer et al. [14] evaluated a cohort of 374 Danish military male conscripts for the presence of HPV infection at 2 time points 6–8 months apart. Because of losses to follow-up and inadequate samples (∼10% were β-globin negative), the final sample size for analysis of acquisition and persistence was 250 men. Of these 250 men, 13.8% acquired a new infection during a 6–8-month period, and ∼25% had a persistently detected infection at the follow-up visit. Partridge et al. [16] examined rates of HPV acquisition among a cohort of 240 University of Washington students, of whom 62.4% acquired a new HPV infection during a 24-month period. Similarly, in our study, 29.2% of men acquired a new HPV infection during a 12-month period. Because of differences in populations, sampling protocols, follow-up intervals, and length of follow-up, it is difficult to compare results across studies.

Men have a lower incidence of HPV-associated cancer than do women, which has led to the speculation that the incidence and/or duration of HPV infections in men is lower than that in women [2]. Unfortunately, there are no published studies that allow for a direct comparison of the natural history of HPV infections in men and women. We have conducted a study of women aged 18–35 years residing in Tucson [25], using a study design and methods of HPV detection similar to those used in the present study of men. Among these women and the men in the present study, we observed similar period prevalences for any HPV infection (52.8% in men and 53.8% in women), and similar HPV incidence rates (29.4 per 1000 person-months in both men and women). The probability of acquiring HPV infections in the present study of men was relatively similar for the 2 categories of infection (12-month probability, 0.19 and 0.16 for oncogenic and nononcogenic infections, respectively). This is in contrast to what we [25] and others [26] have observed in women, in whom the probabilities for HPV acquisition were significantly higher for oncogenic than for nononcogenic infections. In the present study, we observed comparable rates of clearance across categories of HPV infection in men, again in contrast to what has been observed in women, for whom oncogenic infections tend to last longer than nononcogenic infections [25–27]. In summary, our results indicate that in men, unlike in women, the rates of acquisition and the durations of oncogenic and nononcogenic infections are comparable.

To our knowledge, this is the first study to examine the rate of HPV acquisition and clearance by age in men. Other studies among women have indicated that younger age is associated with a higher rate of HPV acquisition and clearance, with longer durations of infection in the older groups [28, 29]. No clear age pattern in HPV acquisition and clearance in men was observed when oncogenic infections were examined. However, we observed nonsignificant differences in the acquisition of any HPV type and nononcogenic HPV, with both increasing among men >30 years old. A similar distribution pattern was observed when we examined period prevalence, with age differences reaching statistical significance only for nononcogenic HPV infections. Unlike what has been observed in women [28, 29], we found no evidence for an association between age and duration of HPV infection in men. We did, however, observe a spike in the median duration of infections in men 31–35 years old. These data should be interpreted with caution, given the relatively small sample size used to estimate duration for each age group. Overall, observations from the present study of men are concordant with the findings of published HPV prevalence studies, which have indicated a relatively constant prevalence of HPV infections in men across age groups [24].

As with any epidemiological study, limitations must be considered in interpreting results. These include a relatively small sample size and short duration of follow-up, which resulted in lower statistical power for examining the acquisition and clearance of multiple infections and single-type HPV infections and influenced the estimates of the incidence and persistence of oncogenic and nononcogenic infections. In our calculation of duration of infection, we included prevalent infections at baseline as well as incident infections detected at follow-up visits. The inclusion of incident infections may have biased estimates of clearance time, because incident infections could be followed up only for a shorter period of time, and clearance of infection may not have been observed before the end of the study. Inclusion of prevalent infections at baseline may have biased estimates of the duration of infection, because it is uncertain how long such infections were present before enrollment. Because of the interval sampling, there was an artificially fixed lower bound in the duration of infection we could observe. This may have biased our estimates of duration. In addition, because of the relatively short follow-up period, we were limited to defining HPV clearance as the time to the first negative test result rather than confirming clearance with a subsequent negative test result. Finally, although we attempted to include a representative sample of the broader underlying community, more educated and informed men may have tended to enroll in the study. Given the methods of recruitment and the rigors of study participation, the cohort presented here is likely a select population of men, limiting the generalizability of our findings.

In 2006, the Food and Drug Administration approved the first prevention vaccine against HPV-6, -11, -16, and -18 for use in females. Vaccine efficacy testing is currently under way internationally in men [11]. Depending on the results of this trial, there will be discussion and debate regarding the effectiveness of deploying the HPV vaccine in both males and females. To enable fully informed US policy decisions, more information is needed regarding the natural history of HPV infections, specifically the incidence and duration of type-specific HPV infections across a broad age range and in a more diverse population of men from a larger US study.

Acknowledgments

We recognize Emily Jolles, MPH, and Melody Schiaffino, MPH, for their contribution to the editing and formatting of the manuscript; Danelle Smith, MS, Steven McAnany, and Shannon McCarthy for their contributions to the laboratory analysis; and Roberta Kline, ARNP, for her assistance with sample collection and clinical examinations.

Potential conflicts of interest: A.R.G. receives support from Merck, manufacturer of HPV vaccine, for the Phase III Quadrivalent HPV Vaccine Trial among Men (Tampa, Florida, site). She is also a member of the Merck Young Women's Advisory Board and is on the speakers' bureau for Merck. The research support she receives from Merck is <5% of her total funding for HPV research. All other authors report no potential conflicts.

Presented in part: 24th International Papillomavirus Conference and Clinical Workshop, Beijing, 3–9 November 2007 (abstract 7B-01).

Financial support: Arizona Disease Control Research Commission (contract 7016; funding period, 2003–2006). Digene donated specimen transport medium, and Roche donated reagents for HPV detection.