High Rates of Incident and Prevalent Anal Human Papillomavirus Infection Among Young Men Who Have Sex With Men

Abstract

Background. There are few published estimates of anal human papillomavirus (HPV) infection rates among young men who have sex with men (YMSM).

Methods. We estimated incidence and prevalence of type-specific anal HPV infection using clinician-collected anal swabs for HPV DNA testing obtained during a 1-year prospective study of 94 YMSM (mean age, 21 years) in Seattle.

Results. Seventy percent of YMSM had any HPV infection detected during the study, and HPV-16 and/or -18 were detected in 37%. The incidence rate for any new HPV infection was 38.5 per 1000 person-months and 15.3 per 1000 person-months for HPV-16/18; 19% had persistent HPV-16/18 infection. No participant tested positive for all 4 HPV types in the quadrivalent vaccine. The number of lifetime male receptive anal sex partners was significantly associated with HPV infection. The prevalence of HPV-16/18 was 6% among YMSM with a history of 1 receptive anal sex partner and 31% among YMSM with ≥2 partners.

Conclusions. Although the high prevalence of HPV among YMSM highlights the desirability of vaccinating all boys as a strategy to avert the morbidity of HPV infection, most YMSM appear to remain naive to either HPV-16 or -18 well into their sexual lives and would benefit from HPV immunization.

Human papillomavirus (HPV) is a highly prevalent sexually transmitted infection (STI) and a cause of the vast majority of cervical and anal cancers [1–5]. Among men, men who have sex with men (MSM) have a higher incidence of anal HPV infection and anal cancer than men who have sex with women [6, 7]. The quadrivalent HPV vaccine (Gardasil, Merck Research Laboratories), which includes protection against the 2 most common oncogenic HPV strains (16 and 18), has received Food and Drug Administration approval for cancer prevention in both women and men [8, 9]. The Centers for Disease Control and Prevention's Advisory Committee on Immunization Practices (ACIP) recommends that all males aged 11–21 years receive the quadrivalent HPV vaccine, as well as MSM up to age 26 [10].

At present, there are few published estimates of the incidence and prevalence of type-specific anal HPV infections among young MSM (YMSM). The existing literature either does not provide age-stratified estimates for HPV-16 and -18 or includes only small numbers of MSM age <30 [6, 11–15]. Additional data on YMSM would be helpful in informing HPV vaccination recommendations and defining how early immunization might need to occur be effective. In this study, we estimated the prevalence and incidence of type-specific anal HPV infections in a contemporary cohort of YMSM (aged 16–30 years) in Seattle using comparable analytic methods to those used in a recent study by Nyitray et al [6]. We also examined demographic and behavioral risk factors for both prevalent and incident high-risk anal HPV infections. Finally, we assessed data from this population regarding knowledge about HPV and willingness to receive HPV vaccine.

METHODS

Study Population and Procedures

The Development and Sexual Health (DASH) study was a 1-year prospective cohort study designed to describe the early sexual behavior patterns and associated human immunodeficiency virus (HIV)/STI risk of YMSM enrolled relatively soon after sexual debut. Young men were eligible to participate if they were aged 16–30 years, had ever had sex with another male (defined as mutual masturbation, oral sex, or anal sex) and either reported ≤10 lifetime male partners or were within 5 years of their same-sex sexual debut. Participants were recruited through targeted Facebook advertisements (36% of eligible participants), paid peer referral (21%), local community- and college-based organizations (14%), the Public Health–Seattle & King County (PHSKC) STD (sexually transmitted disease) Clinic (10%), and other means (19%).

At their baseline visit, eligible participants provided written informed consent. Study staff granted each participant access to a password-protected online account administered using Illume software (DatStat, Inc). Through this website, participants completed retrospective surveys every 3 months, including questions about demographics, sexual behaviors, and cigarette smoking. (Additional study methods and results are reported elsewhere [16]). At baseline, 6 months, and 12 months, participants received a standard HIV/STI screening exam at the PHSKC STD Clinic. At this exam, the clinician collected an anal swab for HPV DNA testing using the APTIMA vaginal swab collection swab and transport medium (Gen-Probe, Inc).

Ninety-five eligible YMSM enrolled, and >85% attended each 6-month HIV/STI testing exam. One participant was later found to be ineligible and was excluded from further analysis. Among the remaining 94 YMSM, 67 (71%) provided 3 swabs, 17 (18%) provided 2 swabs, 8 (9%) provided 1 swab, and 2 (2%) provided no swabs over 1 year of follow-up. The reasons for the missing swabs included loss to follow-up (21 swabs), participant refusal (10 swabs), insufficient sample for assay (3 swabs), and a clinical condition that precluded specimen collection (1 swab). In addition, at the baseline visit, study staff did not collect swabs from 4 YMSM who reported no history of receptive anal sex. After changing this protocol, we asked all participants to provide swabs at subsequent visits. The Human Subjects Division at the University of Washington approved all study procedures.

Laboratory Procedures

HPV genotyping was performed on specimens using liquid bead microarray assay (LBMA) [17]. In brief, DNA was isolated using the QIAamp DNA blood mini column (Qiagen) according to the manufacturer's protocol. Five microliters of purified DNA was polymerase chain reaction amplified and subsequently genotyped using the Luminex-based LBMA assay for the following 37 HPV types: 6/11/16/18/26/31/33/35/39/40/42/45/51/52/53/54/55/56/58/59/61/62/64/66/67/68/69/70/71/72/73/81/82/83/84/IS39/cp6108.

Analysis

We calculated the baseline prevalence, period prevalence, persistent infection prevalence, and incidence rate for each type-specific HPV. In addition, we further grouped these estimates as follows: any HPV; 16 and/or 18; 6, 11, 16, and/or 18; any oncogenic (16/18/31/33/35/39/45/51/52/56/58/59/68); and any nononcogenic (6/11/26/40/42/53/54/55/61/62/64/66/67/69/70/71/72/73/81/82/83/84/IS39/cp6108) [18].

Baseline prevalence included the proportion of specimens collected at baseline (n = 82) that were positive for type-specific HPV. Period prevalence was defined as the proportion of participants who provided ≥1 swab throughout the year of follow-up (n = 92) and had any type-specific-positive HPV specimen. Persistent infection was the proportion of participants who provided ≥2 swabs during follow-up (n = 84) and had a type-specific-positive specimen in consecutive swabs. Finally, we estimated the incidence of new type-specific HPV infection. For each type-specific rate, we determined the number of at-risk participants as those who provided ≥2 swabs during follow-up and were negative for the type-specific HPV in the earliest collected swab. (For groupings that included multiple type-specific HPVs—for example, 16/18—the earliest swab had to be negative for all applicable HPV types.) The incidence rate was then calculated by the number of at-risk participants who had a new type-specific-positive swab during follow-up divided by the number of at-risk person-months. Similar to Nyitray et al's analysis [6], we calculated 95% confidence intervals (CIs) using a Poisson distribution.

We stratified baseline prevalence estimates by lifetime number of male receptive anal sex partners reported at baseline, and stratified period and persistent infection prevalence by lifetime number of male anal sex partners, which included new partners reported during the study period. We evaluated correlates of anal infection with any oncogenic HPV type using 3 outcomes: prevalent, incident, and persistent infection. Because each outcome, as well as some covariates, was evaluated at multiple time points, we used generalized estimating equation with an independent correlation structure and robust standard errors to calculate the odds ratios (ORs) and 95% CI for several potential risk factors. Some covariates were measured only at baseline, whereas others were measured at baseline, 6 months, and 12 months.

HIV infection may modify the prevalence and persistence of HPV infection [19, 20]. Therefore, we conducted a sensitivity analysis that excluded HPV results and censored follow-up time starting at the time of HIV diagnosis from participants whose HIV infection was known (n = 1) or newly diagnosed (n = 3). This did not result in a meaningful change in the HPV incidence, prevalence, or persistence rates; thus, we presented estimates from the full sample.

All analyses were conducted using Intercooled Stata 12.1 (StataCorp).

RESULTS

Study Population

The median age of the 94 YMSM was 21 years (interquartile range [IQR], 19–23 years), and 40% identified as a race/ethnicity other than non-Hispanic white. The median age at male sexual debut was 17 years (IQR, 15–19 years). Nearly all participants reported a history of at least 1 episode of insertive (88%) and receptive (92%) anal sex. The median age of first anal sex with another man was 18 years (IQR, 16–20 years). Among all participants, YMSM reported a median of 2 (IQR, 1–6) male lifetime insertive anal sex partners and 3 (IQR, 1–5) male lifetime receptive anal sex partners. In the 3 months prior to baseline, participants had a median of 1 (IQR, 0–1) male partner for both insertive and receptive anal sex. At baseline, 30% of YMSM reported ever having had a female partner (median, 0 [IQR, 0–1]), and 40% reported regular cigarette smoking. One participant had previously been diagnosed with HIV infection; during the study 15% were newly diagnosed with HIV (n = 3), chlamydia (n = 7), and/or gonorrhea (n = 6).

Anal HPV Prevalence, Incidence, Persistence, and Clearance

More than two-thirds (69.6%) of YMSM in this study had any HPV infection detected over the 1 year of follow-up (Table 1). Among men with no HPV detected in their first swab, the incidence rate for any new HPV infection was 38.5 per 1000 person-months and the incidence proportion was 40%. HPV-16 and/or -18 was detected in 37% of study participants over 1 year, with an incidence rate of 15.3 per 1000 person-months. Nearly 1 in 5 (19%) participants had persistent HPV-16 and/or -18, with the same strain detected in consecutive swabs. One-half of YMSM had ≥1 of the 4 vaccine-preventable HPV strains (6/11/16/18) detected at some point during the study, but only 3% of participants had evidence of acquiring both HPV-16 and -18 at any time during follow-up, and no participant tested positive for all 4 HPV types included in the quadrivalent vaccine. HPV-16 was the strain with the highest period prevalence and was detected in 28% of participants; the other most common type-specific strains were HPV-51 (16%), HPV-6 (16%), HPV-56 (14%), and HPV-66 (14%). HPV-51 had the highest incidence rate (11.4 per 1000 person-months), followed by HPV-16 (10.3 per 1000 person-months), HPV-56 (9.0 per 1000 person-months), HPV-66 (6.9 per 1000 person-months), and HPV-84 (6.9 per 1000 person-months).

Table 1.

Prevalence, Persistence, and Incidence of Type-Specific Anal Human Papillomavirus Infection Among Men Who Have Sex With Men, Aged 16–30 Years, Seattle, Washington, 2009–2010

HPV Type	Baseline Prevalencea (n = 82), No. (%)	Period Prevalenceb (n = 92), No. (%)	Persistent Infectionc (n = 84), No. (%)	At Riskd, No.	New Infections, No.	Person-months, No.	Incidence Rate, per 1000 Person-months (95% CI)
Any HPV	43 (52.4)	64 (69.6)	35 (41.7)	45	18	467	38.5 (38.0–39.1)
16/18 (either)	21 (25.6)	34 (37.0)	16 (19.1)	65	11	720	15.3 (15.0–15.6)
6/11/16/18 (any)	28 (34.2)	46 (50.0)	25 (29.8)	59	16	629	25.5 (25.1–25.9)
Oncogenic HPV	33 (40.2)	54 (58.7)	25 (29.8)	54	18	588	30.6 (30.2–31.1)
16	17 (20.7)	26 (28.3)	12 (14.3)	70	8	777	10.3 (10.1–10.5)
18	6 (7.3)	11 (12.0)	4 (4.8)	79	4	878	4.6 (4.4–4.7)
31	2 (2.4)	4 (4.4)	2 (2.4)	83	2	919	2.2 (2.1–2.3)
33	2 (2.4)	2 (2.2)	2 (2.4)	82	0	911	0.0 (.0–4.1)
35	6 (7.3)	7 (7.6)	1 (1.2)	79	1	892	1.1 (1.1–1.2)
39	3 (3.7)	8 (8.7)	5 (6.0)	81	5	882	5.7 (5.5–5.8)
45	3 (3.7)	6 (6.5)	2 (2.4)	81	3	899	3.3 (3.2–3.5)
51	4 (4.9)	15 (16.3)	4 (4.8)	79	10	880	11.4 (11.1–11.6)
52	3 (3.7)	5 (5.4)	3 (3.6)	80	1	893	1.1 (1.1–1.2)
56	3 (3.7)	13 (14.1)	5 (6.0)	80	8	886	9.0 (8.8–9.2)
58	2 (2.4)	2 (2.2)	2 (2.4)	82	0	911	0.0 (.0–4.1)
59	2 (2.4)	4 (4.4)	1 (1.2)	82	2	916	2.2 (2.1–2.3)
68	0 (0.0)	1 (1.1)	0 (0.0)	84	1	935	1.1 (1.0–1.1)
Nononcogenic HPVe	26 (31.7)	43 (46.7)	22 (26.2)	58	15	629	23.9 (23.5–24.2)
6	9 (11.0)	15 (16.3)	9 (10.7)	74	5	823	6.1 (5.9–6.2)
11	5 (6.1)	9 (9.8)	3 (3.6)	80	4	886	4.5 (4.4–4.7)
26	0 (0.0)	0 (0.0)	0 (0.0)	84	0	935	0.0 (.0–3.9)
40	0 (0.0)	2 (2.2)	0 (0.0)	84	2	935	2.1 (2.0–2.2)
53	3 (3.7)	7 (7.6)	1 (1.2)	80	3	905	3.3 (3.2–3.4)
54	3 (3.7)	5 (5.4)	2 (2.4)	81	2	898	2.2 (2.1–2.3)
55	0 (0.0)	2 (2.2)	0 (0.0)	84	2	923	2.2 (2.1–2.3)
61	3 (3.7)	6 (6.5)	4 (4.8)	81	3	892	3.4 (3.2–3.5)
62	2 (2.4)	5 (5.4)	2 (2.4)	82	3	906	3.3 (3.2–3.4)
64	0 (0.0)	1 (1.1)	0 (0.0)	84	1	935	1.1 (1.0–1.1)
66	7 (8.5)	13 (14.1)	5 (6.0)	78	6	867	6.9 (6.7–7.1)
70	1 (1.2)	2 (2.2)	2 (2.4)	83	1	916	1.1 (1.0–1.2)
72	3 (3.7)	7 (7.6)	2 (2.4)	81	4	907	4.4 (4.3–4.6)
73	0 (0.0)	1 (1.1)	0 (0.0)	84	1	928	1.1 (1.0–1.1)
81	1 (1.2)	2 (2.2)	0 (0.0)	83	1	929	1.1 (1.0–1.1)
82	1 (1.2)	4 (4.4)	1 (1.2)	83	3	916	3.3 (3.2–3.4)
83	1 (1.2)	5 (5.4)	0 (0.0)	83	4	929	4.3 (1.2–4.4)
84	5 (6.1)	11 (12.0)	3 (3.6)	79	6	875	6.9 (6.7–7.0)
89/CP6108	2 (2.4)	7 (7.6)	1 (1.2)	82	5	893	5.6 (5.4–5.8)

Abbreviations: CI, confidence interval; HPV, human papillomavirus.

^a Defined as having a type-specific infection at the baseline visit. Denominator includes participants who provided a swab at this visit.

^b Defined as having a type-specific infection in any swab. Denominator includes participants who provided ≥1 swabs over 1 year of follow-up.

^c Defined as having a type-specific infection in consecutive swabs. Denominator includes participants who provided ≥2 swabs over 1 year of follow-up.

^d Defined as providing ≥2 swabs and being negative for the type-specific infection in the earliest collected swab over 1 year of follow-up. For groupings that include multiple strain-specific infections (eg, 16/18, oncogenic), defined as being negative for all applicable type-specific infections.

^e Our sample did not have any detectable HPV-42, -67, -69, or -71.

YMSM with a prior female sex partner had a higher baseline prevalence of any HPV infection (69.2%) compared with those who only had male partners (44.6%, P = .038), although there was a greater difference between the prevalence of low-risk strains (46.2% vs 25.0%, respectively, P = .055) compared to high-risk strains (50.0% vs 35.7%, respectively, P = .220).

Among the 48 YMSM who provided ≥2 swabs and had any prevalent HPV infection, 81% cleared ≥1 of these strains during the study (Table 2). Nine of 18 (50%) YMSM cleared an HPV-16 infection, and 4 of 6 (67%) cleared HPV-18.

Table 2.

Type-Specific Clearance of Human Papillomavirus Among Men Who Have Sex With Men, Aged 16–30 Years, Seattle, Washington, 2009–2010

HPV Type	HPV Clearance During Study Follow-up
	Prevalent, No.a	Cleared, No. (%)b
Any HPV	48	39 (81.3)
6	13	4 (30.8)
11	5	2 (40.0)
16	18	9 (50.0)
18	6	4 (66.7)
Any oncogenicc	36	24 (66.7)
Any nononcogenicd	34	26 (76.5)

Abbreviation: HPV, human papillomavirus.

^a Among those who had ≥2 swabs over 1 year of follow-up (n = 84), defined as having a type-specific infection in an earlier swab.

^b Among those who had ≥2 swabs over 1 year of follow-up, defined as clearance of a type-specific infection in a later swab.

^c Includes HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68.

^d Includes HPV types 6, 11, 26, 40, 42, 53, 54, 55, 61, 62, 64, 66, 67, 69, 70, 71, 72, 73, 81, 82, 83, 84, and 89/CP6108.

The prevalence of any HPV and HPV-16/18 increased with higher numbers of male anal sex partners (Figure 1). Specifically, among YMSM reporting a history of just 1 male receptive anal sex partner at baseline, 35% had any HPV infection detected at baseline and 6% were infected with HPV-16 or -18, whereas nearly one-third of YMSM with ≥2 partners had prevalent HPV-16 or -18. We observed a similar pattern when stratifying by lifetime number of male anal sex partners (ie, not only receptive partners), with persistent HPV-16 or -18 detected during the study period in approximately 20% of YMSM with ≥5 partners.

Figure 1.

A, Baseline prevalence of type-specific anal human papillomavirus (HPV) infection by lifetime number of receptive anal sex partners and (B) period prevalence and persistence of type-specific anal HPV infection by lifetime number of male anal sex partners among men who have sex with men, aged 16–30 years, Seattle, Washington, 2009–2010.

Correlates of Anal Oncogenic HPV Infection

As shown in Table 3, lifetime number of male receptive anal sex partners reported at baseline was consistently and significantly associated with anal HPV infection (prevalent, incident, or persistent infection), and the number of these partners in the past 3 months was associated with both prevalent and incident HPV. Specifically, for each additional partner in the past 3 months, the odds of prevalent infection increased by 24% (OR, 1.24 [95% CI, 1.00–1.57]) and the odds of incident infection increased by 32% (OR, 1.32 [95% CI, 1.03–1.70]). Moreover, the odds of prevalent HPV were significantly higher among YMSM who reported any unprotected receptive anal sex in the past 3 months. The odds of persistent infection were lower for YMSM with higher numbers of recent insertive anal sex partners. Regular cigarette smoking was significantly associated with incident HPV. Newly diagnosed HIV/STI was positively, but not significantly, associated with each HPV outcome.

Table 3.

Demographic and Behavioral Correlates of Prevalent, Persistent, and Incident Oncogenic Anal Human Papillomavirus Infection Among Men Who Have Sex With Men, Aged 16–30 Years, Seattle, Washington, 2009–2010

Characteristic	Prevalent Oncogenic-Typea HPV Infection, OR (95% CI)	Incident Oncogenic-Typea HPV Infection, OR (95% CI)	Persistent Oncogenic-Typea HPV Infection, OR (95% CI)
Age ≥21 y (vs <21 y)b	1.79 (.88–3.66)	1.45 (.66–3.20)	2.06 (.80–5.33)
Nonwhite race/ethnicityb	0.91 (.44–1.88)	1.32 (.60–2.87)	0.93 (.35–2.46)
Age at same-sex sexual debut, yb	0.96 (.85–1.08)	1.06 (.95–1.19)	0.90 (.76–1.06)
Age at insertive anal sex debut, yb	1.08 (.94–1.25)	1.05 (.93–1.19)	1.09 (.91–1.30)
Age at receptive anal sex debut, yb	0.99 (.86–1.12)	1.03 (.91–1.16)	0.95 (.80–1.12)
No. male insertive anal sex partners, lifetimeb	1.08 (1.01–1.15)	1.03 (.97–1.10)	1.09 (1.00–1.19)
No. male receptive anal sex partners, lifetimeb	1.16 (1.05–1.28)	1.08 (1.01–1.15)	1.14 (1.03–1.27)
No. male insertive anal sex partners, past 3 moc	0.80 (.56–1.12)	1.42 (.91–2.23)	0.41 (.20–.83)
No. male receptive anal sex partners, past 3 moc	1.24 (1.00–1.57)	1.32 (1.03–1.70)	1.02 (.77–1.35)
No. new male anal sex partners, past 3 moc	1.02 (.86–1.22)	1.03 (.81–1.30)	0.99 (.79–1.24)
Any unprotected receptive anal sex partner, past 3 moc	1.88 (1.07–3.32)	1.24 (.60–2.60)	2.09 (.83–5.24)
Smoked cigarettes regularly, past 3 moc	1.84 (.94–3.63)	2.55 (1.12–5.82)	2.04 (.81–5.14)
Any HIV/STD diagnosed at study visitd	2.08 (.71–6.06)	1.21 (.39–3.76)	2.36 (.66–8.50)

Abbreviations: CI, confidence interval; HIV, human immunodeficiency virus; HPV, human papillomavirus; OR, odds ratio; STD, sexually transmitted disease.

^a Includes HPV types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68.

^b Measured at baseline only.

^c Measured at baseline and at 6 and 12 months.

^d Includes new diagnosis of HIV, chlamydia, gonorrhea, or primary syphilis.

Knowledge and Opinions About HPV and HPV Vaccine

At their baseline visit, 93 participants answered questions about their knowledge and opinions of HPV. Most (76%) of these YMSM had heard of HPV before; 38% correctly identified it as a cause of genital warts, 34% identified it as a cause of anal cancer, and 6% identified it as a cause of oropharyngeal cancer. On a scale of 1–10, 14% indicated that genital HPV infection is a serious problem for guys like them (score 8–10), whereas 51% said it was somewhat of a problem (score 4–7), 30% said it was not a problem (score 1–3), and 5% did not answer. Three-quarters (75%) of participants had “ever heard of the HPV vaccine, also known as Gardasil.” Following this question, we informed participants that HPV can cause anal cancer, particularly for people who have receptive anal sex, and that the HPV vaccine may prevent most cases of this cancer. We stated that the vaccine requires 3 shots over 6 months and is very safe. Asked to assess their interest, on a scale of 1–10, in being immunized if the vaccine were available and free, 78% of participants indicated that they would be extremely interested in getting the vaccine (score 8–10), 15% would be somewhat interested (score 4–7), and 7% would not be interested (score 1–3).

DISCUSSION

In this cohort of YMSM, we observed a high prevalence and incidence of anal HPV infection. HPV-16 and/or -18, the 2 strains that cause the majority of anal cancers, was detected in more than one-third of study participants at least once during the year of follow-up, and nearly 20% of all YMSM in our study had evidence of persistent infection with 1 of these 2 strains. The strongest and most consistent predictor of anal HPV was lifetime number of male receptive anal sex partners. In addition, there was a 32% increased odds of incident infection with any oncogenic strain per additional receptive anal sex partner in the 3 months preceding testing. By the time men in our cohort had had >1 male receptive anal sex partner, a level of sexual experience that 73% of these MSM had achieved, nearly one-third had acquired either HPV-16 or -18. On the other hand, among men in our study, all appeared to remain susceptible to at least 1 of the HPV types included in the vaccine, even among those reporting up to 25 male receptive anal sex partners. Given the high prevalence of HPV occurring very early in the sexual life course among MSM, our findings highlight the need to immunize YMSM prior to their sexual debut, something that will likely require universal male immunization. At the same time, given the fact that most YMSM appeared to remain susceptible to at least some HPV types included in the vaccine, catch-up immunization programs do offer YMSM some benefit. As in prior studies, approximately 75% of MSM indicated their willingness to receive the HPV vaccine [21, 22].

We observed higher incidence and period prevalence rates of anal HPV infection in this cohort of YMSM than those reported in other studies of MSM [6, 20, 23–25]. (The Supplementary Table provides a more detailed summary of the sample populations, HPV specimen collection procedures, and HPV assays in each study.) Although all of these studies used clinician-collected specimens and similar collection devices and transport media, HPV detection assays have become more sensitive over time with greater ability to differentiate between larger numbers of HPV types. These differences, as well as differences in study populations, may explain some or all of the variance in findings between studies. Nyitray et al reported that the incidence of any HPV type was 25.9 per 1000 person-months (vs 38.5 in our study), 21.2 per 1000 person-months (vs 30.6) for any oncogenic HPV, 4.8 per 1000 person-months (vs 10.3) for HPV-16, and 3.8 per 1000 person-months (vs 4.6) for HPV-18 [6]. Our cohort's higher rates were likely due to its lower age range and perhaps, in part, to geographic differences in the background prevalence of HPV and/or sexual behavior. Palefsky et al calculated an incidence rate of 4.1 per 1000 person-months for HPV-16 (vs 10.3 in our study) and 2.3 per 1000 person-months for HPV-18 (vs 4.6) among MSM in the placebo group [24]. Because this study had a similar age distribution as ours, the lower HPV rates are more likely to be attributed to the sample's lower number of male sex partners or its broader geographic distribution. In a study by Chin-Hong et al of MSM age <25 who were enrolled in the EXPLORE study, high-risk HPV strains (ie, nearly analogous to the oncogenic classification used in our study) were detected in 15%–20% of YMSM [25], which is lower than our baseline prevalence (40.2%). As these samples were collected in the early 2000s, this difference may be a function of a secular change in HPV prevalence over time. Finally, a study conducted in Seattle in the mid-1990s found that 65.9% of HIV-negative MSM (mean age 35) had any HPV DNA detected at enrollment [20]. Unlike the other studies, this was a higher prevalence than what we observed (52.4%). However, those participants were older and had a substantially higher lifetime number of male sexual partners (45% reported >50 partners), indicating far greater opportunity for exposure to HPV.

The type-specific clearance rates we observed were slightly higher than those observed in Nyitray et al's study of MSM (aged 18–70), which used analytic methods similar to ours [6]. For example, they observed 6-month clearance rates of 57.1% for any HPV, 55.8% for any oncogenic HPV, 27.2% for HPV-16, and 62.5% for HPV-18 [6]. Over 1 year, we observed respective clearance rates of 81.3%, 66.7%, 50.0%, and 66.7%. Our higher clearance rates are likely due, in part, to our additional 6 months of follow-up. Others have observed an association between cigarette smoking HPV persistence and/or prevalence [6, 26], whereas we only found a significant association with incidence.

In our analysis of correlates of high-risk HPV infection, the number of receptive anal sex partners was the most consistent predictor of HPV infection. This pattern—a strong association between HPV infection and number of partners, but not age—has been repeatedly observed in other studies as well [6, 23, 25, 27]. In our study, it is possible that this result is an artifact of our inclusion criteria regarding sexual experience. Specifically, in this sample of 16- to 30-year-olds, our older participants were not necessarily representative of most MSM their age as they either had to be within 5 years of their same-sex sexual debut or report ≤10 lifetime partners. Because of this, in our sample there is likely to be more variation in HPV risk attributable to sexual behavior than to age. That said, the consistency in this finding across studies as well as its intuitive biological plausibility suggests that the observed association is likely real. Given the very high prevalence of HPV infection among YMSM, many (if not most) have been exposed to infection from one of their first few partners, a finding that highlights the desirability of immunizing YMSM before they become sexually active.

To our knowledge, this is the first study to assess the epidemiology of type-specific HPV infection in YMSM. Specifically, the longitudinal nature of the study facilitated the calculation of incidence rates, and we used a state-of-the-art assay with the ability to detect several different types of genital HPV. This study also has limitations. First, the sample size was relatively small, which precluded our ability to detect more stable type-specific clearance rates as well as predictors of type-specific incidence and clearance. Related to this, our finding that YMSM who had ever had a female sex partner had a higher baseline prevalence of any HPV infection needs to be confirmed in a larger study that can control for potential confounding by demographic and behavioral factors. Second, the study's behavior-based inclusion criteria may have reduced the generalizability to all YMSM. Our goal in developing our inclusion criteria was to recruit a relatively sexually inexperienced population. However, several of the YMSM reported high levels of sexual activity, and the age at anal sex debut of our participants was consistent with both local [28] and national samples of YMSM [29]. This suggests that our prevalence estimates are likely fairly representative of the broader YMSM population. Third, some men in our study may have acquired and cleared HPV infections prior to study enrollment or between episodes of specimen collection. This would have led us to underestimate the population's true cumulative lifetime risk of HPV and perhaps the incidence of HPV during the study. Finally, we only have 1 year of follow-up data from participants. Additional follow-up time would provide more detailed short- and long-term estimates of HPV persistence and clearance rates.

The ACIP currently recommends that all boys aged 11–21 years receive the quadrivalent HPV vaccine, and that all MSM aged ≤26 be immunized. However, to date, vaccine uptake in the United States has been relatively low among girls [10], and is likely much lower among boys. Most YMSM appeared to remain uninfected with at least 1 HPV type included in the quadrivalent vaccine, suggesting that HPV immunization of sexually experienced YMSM might be worthwhile. However, our findings demonstrate that initial HPV infections, including infection with HPV-16, the most common cause of anal cancer, frequently occur with a man's first receptive anal sex partner. This highlights the need to immunize YMSM before they become sexually active. Given the stigma that can be associated with homosexuality and the challenges adolescent MSM face, we believe that specifically targeting YMSM for immunization prior to their first episode of receptive anal sex, although potentially worthwhile, is likely to be met with limited success. Instead, much greater effort should be placed on immunizing all boys. Australia has achieved high levels of HPV immunization among girls through school-based immunization [30], demonstrating that success is possible. (As of February 2013, boys were also included in Australia's National Immunisation Program.) The United States' failure to date to achieve high levels of HPV immunization among adolescents is perhaps the single greatest missed opportunity in the area of STD prevention. Increasing HPV immunization, including HPV immunization among YMSM, requires a much more aggressive public health and research effort.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.