Valleroy LA, MacKellar DA, Karon JM, Rosen DH, McFarland W, Shehan DA, Stoyanoff SR, LaLota M, Celentano DD, Koblin BA, Thiede H, Katz MH, Torian LV, Janssen RS, for the Young Men's Survey Study Group . HIV Prevalence and Associated Risks in Young Men Who Have Sex With Men. JAMA. 2000;284(2):198–204. doi:10.1001/jama.284.2.198
Author Affiliations: Division of HIV/AIDS Prevention–Surveillance and Epidemiology, National Center for HIV, STD, and TB Prevention, Centers for Disease Control and Prevention, Atlanta, Ga (Drs Valleroy, Karon, Rosen, and Janssen, and Mr MacKellar); San Francisco Department of Public Health, San Francisco (Drs McFarland and Katz), and Los Angeles County Department of Health Services, Los Angeles (Ms Stoyanoff), Calif; University of Texas Southwestern Medical Center at Dallas (Mr Shehan); Florida Department of Health, Tallahassee, (Ms LaLota); Johns Hopkins University School of Hygiene and Public Health, Baltimore, Md (Dr Celentano); The New York Blood Center (Dr Koblin), and the New York City Department of Health (Dr Torian), New York, NY; and Public Health—Seattle and King County, Seattle, Wash (Dr Thiede).
Context Studies conducted in the late 1980s on human immunodeficiency virus
(HIV) infection among older men who have sex with men (MSM) suggested the
epidemic had peaked; however, more recent studies in younger MSM have suggested
continued high HIV incidence.
Objective To investigate the current state of the HIV epidemic among adolescent
and young adult MSM in the United States by assessing the prevalence of HIV
infection and associated risks in this population in metropolitan areas.
Design The Young Men's Survey, a cross-sectional, multisite, venue-based survey
conducted from 1994 through 1998.
Setting One hundred ninety-four public venues frequented by young MSM in Baltimore,
Md; Dallas, Tex; Los Angeles, Calif; Miami, Fla; New York, NY; the San Francisco
(Calif) Bay Area; and Seattle, Wash.
Subjects A total of 3492 15- to 22-year-old MSM who consented to an interview
and HIV testing.
Main Outcome Measures Prevalence of HIV infection and associated characteristics and risk
Results Prevalence of HIV infection was high (overall, 7.2%; range for the 7
areas, 2.2%-12.1%) and increased with age, from 0% among 15-year-olds to 9.7%
among 22-year-olds. Multivariate-adjusted HIV infection prevalence was higher
among blacks (odds ratio [OR], 6.3; 95% confidence interval [CI], 4.1-9.8),
young men of mixed or other race (OR, 4.8; 95% CI, 3.0-7.6), and Hispanics
(OR, 2.3; 95% CI, 1.5-3.4), compared with whites (referent) and Asian Americans
and Pacific Islanders (OR, 1.1; 95% CI, 0.5-2.8). Factors most strongly associated
with HIV infection were being black, mixed, or other race; having ever had
anal sex with a man (OR, 5.0; 95% CI, 1.8-13.8); or having had sex with 20
or more men (OR, 3.0; 95% CI, 2.0-4.7). Only 46 (18%) of the 249 HIV-positive
men knew they were infected before this testing; 37 (15%) were receiving medical
care for HIV, and 19 (8%) were receiving medical drug therapy for HIV. Prevalence
of unprotected anal sex during the past 6 months was high (overall, 41%; range,
Conclusions Among these young MSM, HIV prevalence was high, underscoring the need
to evaluate and intensify prevention efforts for young MSM, particularly blacks,
men of mixed race or ethnicity, Hispanics, and adolescents.
By the late 1980s, the human immunodeficiency virus (HIV) epidemic among
men who have sex with men (MSM) in the United States appeared to have peaked.1- 6
However, studies in the 1980s mainly sampled men 30 years of age and older;
analyses of the small subsamples of men younger than 30 years suggested higher
HIV incidence3,4 and more sexual
than among older men. In the 1990s, 6 cross-sectional surveys that focused
on small samples of young MSM found high prevalence of HIV and sexual risk
These and other findings involving sexually transmitted diseases (STDs) and
unsafe sex,14,15 and HIV seroincidence16- 18 signal a significant
and continuing HIV epidemic among MSM.
The 1992 through 1993 Young Men's Survey (YMS), conducted in San Francisco
and Berkeley, Calif, was 1 of the 6 cross-sectional surveys focusing on young
MSM in the 1990s.10 Of the 425 participants
aged 17 to 22 years, 9.4% were HIV positive, and 32.7% reported having unprotected
anal sex in the past 6 months. To determine whether this public health problem
was widespread, the Centers for Disease Control and Prevention (CDC) and local
public health officials adapted and expanded the YMS conducted in the years
1994 through 1998 to include the San Francisco Bay Area (San Francisco, Oakland,
and San Jose, Calif), Baltimore, Md; Dallas, Tex; Los Angeles, Calif; Miami,
Fla; New York, NY; and Seattle, Wash. In this first multisite report of the
1994 though 1998 YMS, we confirm that the 1992 through 1993 YMS findings held
true not only in the San Francisco Bay Area19
but also in the 6 other areas.
The 1994 through 1998 YMS was a cross-sectional, multisite, venue-based
survey.20 Our goal was to estimate the prevalence
of HIV and associated risk behaviors in adolescent and young adult MSM. Young
MSM are hard to reach with traditional household-oriented sampling methods,
and findings from surveys using convenience samples are not generalizable
to broader populations of young MSM. Thus, we developed a venue-based survey
method in which we sampled public venues frequented by young MSM and young
men attending these venues.
Throughout the survey, we tried to identify all venues frequented by
young MSM and the days of the week and times of day when young MSM frequented
the venues (times). During the study's start-up, we conducted formative research
to identify all potential venues and times by reviewing local gay publications,
interviewing key public health officials and community informants, and meeting
with focus groups of young men. Staff then visited these potential venues
during potential times, carrying out standardized counts and brief interviews
with young men. Thus, we established a sampling frame of all venues, along
with 4-hour periods, of where and when we might enroll at least 7 eligible
young men during those 4 hours. The minimum of 7 was selected for logistical
and cost efficiency.
We identified an array of venues, ranging from urban shopping blocks,
to dance clubs, to young gay organizations, ie, not just places where men
find sex partners. These venues can be categorized as street locations, dance
clubs, bars, businesses, social organizations, bathhouses, health clubs, and
other public places (eg, parks, beaches). Street locations were stretches
of sidewalk with considerable foot traffic near businesses and socializing
places. We did not include in the sampling frame venues attended primarily
by men with high HIV-related risks, eg, needle exchange programs, commercial
The sampling plan had 3 stages. First, each month, in each area, we
randomly selected (without replacement) 12 to 16 venues from the sampling
frame. Second, for each venue, we randomly selected 1 of the 4-hour periods
associated with that venue. We then scheduled sampled venues and times on
a monthly calendar. Third, we sampled young men during sampling events, ie,
times when team members would enroll young men at venues. In a van outfitted
for interviewing and phlebotomy, 4 to 5 team members arrived at a venue and
sampled young men passing through the venue intercept area (ie, a defined
area at each venue). Young men were approached and asked to give their age
in years and their county of residence. Respondents were eligible if they
were aged 15 to 22 years and residents of specified counties in that area.
Young men were ineligible if they approached team members for enrollment or
had enrolled previously. Sexual experience and orientation were not eligibility
criteria. During the brief interview, team members explained survey procedures.
We collected limited demographic data on men who did not enroll and counted
all young men passing through the venue intercept area during the sampling
The men who enrolled were then interviewed in the survey van. The survey
was anonymous: names were never linked to questionnaires, specimens, or test
results. After obtaining informed consent, an interviewer administered a 45-minute
standardized questionnaire and then conducted an HIV/AIDS (acquired immunodeficiency
syndrome) counseling session and performed phlebotomy. After the interview,
counseling session, and blood draw, participants were paid $40 to $50 for
their time, given a survey identification number, and were scheduled to return
in 2 weeks for test results, posttest counseling, and service referrals.
To prevent duplicate enrollees, we used several methods, beginning at
the eligibility interview and continuing through final data management procedures.
At the eligibility interview, for example, men who enrolled were introduced
to all available team members. The final method was the Miragen Assay,21,22 an antibody-profile assay used to
distinguish persons. If 2 or more participants from an area had identical
birth dates and race or ethnicity, we used the Miragen Assay to test their
blood specimens. When antibody profiles matched, we assumed the participant
had enrolled more than once and analyzed data from his first enrollment.
The YMS protocol was approved by the CDC institutional review board
and by institutional review boards in each of the areas.
The questionnaire was used to obtain information on demographic characteristics,
frequency of attendance at the venues, HIV-related risk behaviors, and factors
potentially associated with these risk behaviors. For many behavioral questions,
a 6-month recall period was used. Human immunodeficiency virus–related
risk behavior questions were used to measure sexual behavior; condom use;
and the use of alcohol, drugs, and needles or syringes. Other questions concerned
factors potentially associated with HIV-related risk behaviors: personal history,
medical history, and psychosocial factors.
In our analysis, we focused on age, race and ethnicity, sexual identity,
and lifetime behaviors and experiences as potential factors associated with
HIV infection. Lifetime behaviors and experiences included having only male
or both male and female sex partners, the number of male sex partners, having
anal sex with a man, using illicit "party" drugs, injecting drugs, having
had an STD, having run away or removed from home, and having been forced to
have sexual contact. Participants were asked if they were black; Asian, Asian
American, or Pacific Islander; Hispanic; American Indian; white; or from multiple
racial backgrounds (mixed). Participants were asked if they considered their
sexual identity to be straight (heterosexual), bisexual, gay (homosexual),
or transgender. Men who identified with the feminine gender defined themselves
as transgender. For those who had male or female sex partners, sex was defined
as oral, vaginal, or anal. Illicit party drugs selected for analysis were
amphetamines, cocaine, lysergic acid diethylamide (LSD), methylenedioxymethamphetamine
(ecstasy), and nitrites. For run away or removed from home, participants were
asked whether they had ever run away from home or been removed from their
parents' or guardians' home. For forced sexual contact, they were asked if
anyone had ever forced them to have sexual contact as defined by themselves.
There were no commercial sex-related questions. However, there were questions
about exchange of sex for food, shelter, drugs, or money in the past 6 months
(L.A.V., unpublished data, 1999).
We also compared sexual risk behaviors during the past 6 months of men
who were HIV negative and men who were HIV positive. Sexual risk behaviors
selected for analysis were having male sex partners for oral or anal sex and
not always using condoms during anal sex with male partners. Serostatus for
HIV was assessed as described below. Knowledge of HIV serostatus was assessed
by asking participants if they had ever been tested for HIV, and, if so, what
their last test results were.
We used assays licensed by the US Food and Drug Administration to test
all specimens for HIV antibody, and for evidence of past or current hepatitis
B virus infection (antibody to hepatitis B core antigen and hepatitis B surface
antigen) and syphilis (VDRL test or rapid plasma reagin and microhemagglutination
assay for Treponema pallidum). We used an enzyme
immunoassay to screen blood specimens for the HIV antibody. Repeatedly reactive
specimens were confirmed by Western blot or indirect immunofluorescence.23 Past or current hepatitis B infection was defined
as the presence of antibody to hepatitis B core antigen or hepatitis B surface
We designed the survey as a venue-based survey, intending to weight
each participant's data according to his probability of being sampled.20 If frequency of venue attendance had been positively
associated with HIV prevalence, weights would have been needed to adjust for
the fact that men who went out frequently had a greater chance of being sampled;
their higher prevalence would have inflated our prevalence estimates. Because
we found no association between frequency of venue attendance and HIV prevalence,
we did not weight the data.
We used logistic regression and likelihood ratio tests to determine
whether there was significant variation in the prevalence of HIV and hepatitis
B virus, and in syphilis among the areas. We used the Breslow-Day test24 to test the homogeneity of the association between
HIV status and each of the variables involving age, race or ethnicity, sexual
identity, and lifetime behavior and experience across the 7 areas. Because
the test results showed homogeneity, we combined the data over the 7 areas.
We examined the association between HIV prevalence and individual factors
using the Mantel-Haenszel χ2 test.25
We assessed the association between HIV prevalence and demographic and risk
behavior variables via logistic regression analysis, entering into the model
all variables identified by other studies as significant factors. Variables
with more than 2 levels were coded as a group of dichotomous indicator variables.
We derived the final model by removing insignificant variables as determined
by the likelihood ratio test. All indicator variables in a group were removed
together. We calculated adjusted odds ratios (ORs) and confidence intervals
(CIs) for the demographic and sexual risk behavior variables that remained
in the final model.
We examined the association between sexual risk behaviors and HIV serostatus
and knowledge thereof using risk prevalence ratios. We analyzed all data using
SAS version 6.12 (SAS Institute, Cary, NC).
In the 7 areas, 38,622 men (who appeared to be young) entered our venue
intercept areas during 1592 sampling events at 194 venues. Of the 23,881 men
(62%) who were approached for eligibility interviews, 21,096 (88%) completed
the interviews and 6866 (33%) were eligible, of whom 4272 (62%) enrolled.
The eligibility percentage was low (33%) because many men who appeared to
be young were older than 22 years, and some men were nonresident visitors
to these metropolitan areas. Enrollment rates differed by area (Table 1) and by age, race or ethnicity, and venue. More of those
aged 15 to 19 years enrolled than did those aged 20 to 22 years (69% vs 57%; P=.001). Compared with men of all other races or ethnicities,
more men of mixed race enrolled (77% vs 61%; P=.001)
and fewer Asians enrolled (52% vs 63%; P=.001). Compared
with men recruited from all other venues, more men at social organizations
enrolled (79% vs 61%; P=.001), and fewer men at dance
clubs enrolled (53% vs 66%; P=.001).
After removing duplicate enrollments from the database, we found that
3% of the 4111 enrolled men had never had sex, and another 11% had never had
sex with a man. We limited our analyses to the 3492 men who reported ever
having had sex with a man. Of these men, 30% enrolled at street locations,
29% at dance clubs, 12% at bars, 10% at social organizations, 9% at businesses,
4% at parks, 4% at bathhouses or health clubs, and 4% at other locations (eg,
Overall, HIV infection prevalence was 7.2% and was lower in Seattle
(2.2%) and higher in New York City (12.1%) vs the other areas (P<.001) (Table 2). The
prevalence of markers for hepatitis B virus was 10.7% overall and was lower
in Seattle (5.8%) vs the other areas (P=.001). Prevalence
of markers for syphilis was 0.7% overall and was similar in all areas (P=.30).
Univariate analyses showed that HIV prevalence was higher among 20-
to 22-year-olds than among 15- to 19-year-olds (Table 3). None of the 15-year-olds were HIV infected (n=56); 9.7%
of the 22-year-olds were (n=619). Prevalence of HIV was higher among blacks,
Hispanics, and men of mixed or other race than among whites. Among men of
mixed race, HIV prevalence was higher among those who reported black backgrounds
(n=172) than among those who did not (n=212; 16.9% vs 9.1%; OR, 2.0; 95% CI,
1.1-3.7). Sexual orientation was measured by reported sexual identity and
by reported lifetime sexual activity with men only or with both men and women;
HIV prevalence was higher only among transgender persons. Prevalence of HIV
increased with greater number of lifetime male sex partners. When we compared
lifetime behaviors and experiences, HIV prevalence was higher among men who
reported anal sex with men, injecting drugs, having had an STD, or having
run away from home.
In multivariate analysis, the demographic and lifetime behavioral characteristics
and experiences associated with HIV infection were being aged 20 to 22 years
old; being black, Hispanic, mixed or other race; having had 20 or more male
sex partners; having had 5 to 19 male sex partners; having had anal sex with
men; injecting drugs; or having had an STD (Table 3). The results from multivariate analysis controlling for
area were not substantially different from these results.
To determine which factors were associated with HIV infection by race
or ethnicity group, we entered all the variables in Table 3 except race or ethnicity in 3 logistic regression models
for race or ethnicity groups with sample sizes of more than 500. For blacks,
the associated factors were having had 20 or more male sex partners (OR, 3.4;
CI, 1.8-6.4) or having had 5 to 19 male sex partners (OR, 2.9; CI, 1.3-6.2).
For Hispanics, the associated factors were having had 20 or more male sex
partners (OR, 4.6; CI, 2.1-10.2), having had an STD (OR, 2.4; CI, 1.3-4.4),
having been forced to have sex (OR, 0.5; CI, 0.3-0.9), or being aged 20 to
22 years (OR, 2.0; CI, 1.1-3.6). For whites, the associated factors were having
had 20 or more male sex partners (OR, 6.5; CI, 1.4-28.9), having had 5 to
19 male sex partners (OR, 6.4; CI, 1.4-28.3), having had an STD (OR, 3.7;
CI, 1.8-7.5), or having been forced to have sex (OR, 2.8; CI, 1.4-5.7).
All 3492 young men received HIV counseling and testing as part of YMS:
2268 (65%) reported having been tested previously, and 2186 (63%) knew the
results. Of the 249 young men who tested positive in YMS testing, 79% had
been tested previously, compared with 64% of the men who tested negative in
YMS testing (OR, 2.1; CI, 1.5-2.8). Of the 249 HIV positive men, 189 (76%)
knew their previous test results, but only 46 (18%) knew that they were currently
HIV positive. Of these 249 men, 37 (15%) were receiving medical care for HIV
infection, and 19 (8%) were taking medical drugs for HIV infection or to prevent
Among these 3492 young men, 90% reported having had sex with at least
1 man, and 23% had had sex with at least 5 men during the past 6 months. Overall,
the prevalence of unprotected anal sex (insertive or receptive) was 41% (range,
33%-49%); 30% had had unprotected insertive anal sex, and 31% had had unprotected
receptive anal sex (Table 4).
We compared the sexual risk behaviors of the HIV-negative and HIV-positive
men (Table 4). Compared with HIV-negative
men, HIV-positive men who did not know they were infected were more likely
to have had unprotected insertive or receptive anal sex during the past 6
months. Compared with HIV-negative men, HIV-positive men who had known for
6 months or more that they were infected were just as likely to have had unprotected
insertive or receptive anal sex during the past 6 months.
In all 7 US areas, the prevalence of HIV infection and associated sexual
risk behaviors were high among these young MSM. Considering their youth, their
recent initiation of sexual activity, and the high prevalence of recent unsafe
sex, many of the HIV-positive men probably were infected recently. Considering
their youth, the high prevalence of recent unsafe sex, and the high HIV prevalence
in US MSM, many of the HIV-negative men are likely to become HIV-infected
in the near future.
Results from this large sample of young men are consistent with the
high prevalence of HIV suggested from smaller samples of young men in single
the 1992 through 1993 YMS in San Francisco and Berkeley.10
The results are alarming in light of the men's youth and compared with the
HIV prevalence for samples of primarily heterosexual youth in the United States.
For example, according to 1996 data on 33,989 males aged 16 to 21 years entering
the US Job Corps, a training program for disadvantaged out-of-school youth,
HIV prevalence was 0.22% among black men, 0.10% among Hispanic men, and 0.04%
among white men.26 Data on 205,026 males aged
17 to 22 years applying for US military service during 1996 showed HIV prevalence
of 0.11% among black men, 0.01% among Hispanic men, and 0.01% among white
men (L.A.V., unpublished data, 1999).
Being black, mixed or other race, having ever had anal sex with a man,
or having had sex with 20 men or more in their lifetimes were the demographic
and lifetime behavioral characteristics and experiences most strongly associated
with HIV infection. The prevalence of HIV infection in this large sample of
young blacks was 14%. This high prevalence among blacks was suggested from
data from very small samples of blacks in earlier surveys.8,10- 12
Our data show very high HIV prevalence among men of mixed race who reported
black backgrounds (16.9%). Prevalence of HIV was also higher among men of
mixed or other race (12.6%) and Hispanics (6.9%), than among Asian Americans
and Pacific Islanders, and whites. These findings display the large racial
gap in the current HIV epidemic in the United States and point to the need
for HIV intervention research, prevention programs, and early care programs
for young men of color who have sex with men.
Considering their youth, the proportion of these men who had been tested
for HIV was high: almost two thirds of all the men and almost four fifths
of the HIV-positive men reported having been tested previously. High proportions
of all the men and of the HIV-positive men also reported knowing these previous
test results. However, only 18% of the men who tested HIV positive in the
YMS knew they were currently HIV-infected, only 15% were receiving medical
care for HIV, and only 8% were taking medical drugs for HIV. Many of these
young men may have become infected since their most recent test. Prevention
efforts should continue to make HIV counseling and testing services available
and attractive for young MSM, so that these men can learn their serostatus,
receive appropriate counseling, and be referred for care if necessary.
We also examined recent HIV-related sexual risk behaviors in light of
HIV serostatus and self-knowledge of serostatus. It is sobering that 41% of
all the men had had unprotected anal sex during the past 6 months. It is also
sobering that 37% of the HIV-infected men who did not know they were infected
(n=75), and 13% of the HIV-infected men who did know they were HIV infected
(n=4), reported unprotected insertive anal sex during the past 6 months. Although
we asked the men if they knew their own serostatus, we did not ask whether
they knew the serostatus of their sex partners, so we do not know the risk
of these sex acts. However, it is clear that the prevalence of unprotected
anal sex was high, especially among HIV-positive men who had not known they
There are 2 potential biases to consider concerning our estimates of
HIV prevalence and its associated risks. The first is potential enrollment
bias. Our enrollment rates were high, considering that we sampled through
outreach at venues. However, there was some variation in enrollment rates
by age and race. Younger men (among whom seroprevalence was lower) were more
likely to enroll; older men (among whom seroprevalence was higher) were less
likely. Men of mixed race (among whom seroprevalence was higher) were more
likely to enroll; Asian Americans (among whom seroprevalence was lower) were
less likely. However, since approximately equal numbers of younger and older
men enrolled, and since approximately equal numbers and relatively few Asian
American and men of mixed race enrolled, we do not think that enrollment rate
variations biased our prevalence estimates markedly. We have no data on whether
enrollment rate was associated with HIV testing history.
Our estimates may be affected by sampling bias. We sampled young MSM
who went to public venues in 7 metropolitan areas. We did not sample young
men who did not go to venues or young men who did not live in these metropolitan
areas; presumably, our prevalence estimates would have been lower if we had
also sampled these men. However, in conducting a venue-based survey, we assumed
that many young MSM who live in these areas do go to venues and do so frequently,
as shown by data from the San Francisco Young Men's Health Study, a population-based,
randomized household survey of young men. According to this study, 91% of
the 18- to 23-year-old young MSM had gone to a gay bar in San Francisco in
the past 6 months (Kimberly Page Shafer, PhD, MPH, oral communication, June
6, 2000).10 Similarly, the Urban Men's Health
Survey, a population-based telephone sample of MSM in San Francisco, Los Angeles,
New York, NY, and Chicago, Ill, found that 74% of the 18- to 22-year-olds
sampled had gone to a bar, night club, or dance club at least once a month
in the past 12 months (Joseph Catania, PhD, oral communication, July 19, 1999).
Even though we sampled only young men who went to venues, we tried throughout
the survey period to identify all the venues frequented by young MSM. We did
not just sample young men at bars and dance clubs; approximately half the
young men we sampled enrolled at street locations, social organizations, and
Despite the young age of those participating in our study and their
presumed exposure to HIV/AIDS prevention educational initiatives while they
were growing up, HIV prevalence was high, particularly among blacks, men of
mixed race, and Hispanics. The high prevalence of recent sexual risk-taking
portends additional HIV infections. Our findings signal a critical and widespread
public health problem and underscore a need to evaluate and intensify prevention
efforts for young MSM. Particular attention should be paid to reaching blacks,
men of mixed race, Hispanics, and young adolescents before they begin having
sex with men.