Oral presentation on May 1, 1998, in Quebec City, at the 7th annual conference of the Canadian Association for HIV/AIDS Research.

Unsafe Sex and Elevated HIV Incidence among Young Men having Sex with Men (MSM)

Presented by Mary Lou Miller, RN, on behalf of...

Steffanie A. Strathdee, PhD; Stephen L. Martindale; Robert S. Hogg, PhD; Peter G.A. Cornelisse, Msc; Mary Lou Miller, RN; Darrel Cook, PhD; Bonnie Devlin; Julio S.G. Montaner, MD; Michael V. O'Shaughnessy, PhD; Martin T. Schechter, OBC, MD, FRCP, PhD; for the Vanguard Project; UBC and the BC Centre for Excellence in HIV/AIDS.

Acknowledgements:

I'd like to take this opportunity to acknowledge the Principal Investigators of this study, Dr. Martin Schechter, Dr. Steffanie Strathdee and Dr. Robert Hogg; and the National Health Research Development Program of Health Canada for project funding.

I would also like to acknowledge the project staff, the other participating doctors and clinics, members of the Community Advisory Committee and the participants themselves.

Primary Objective:

The Vanguard Project was launched in May, 1995, as an expansion of the Vancouver Lymphadenopathy-AIDS Study (VLAS), which has been running since 1982. The primary objective of the Vanguard Project is to monitor HIV incidence rates and associated risk factors among young men having sex with men (MSM).

The objective of the present analysis is to provide an update on risk behaviour, HIV prevalence and incidence, and preliminary results describing trends in behaviour change.

To estimate HIV seroincidence among young gay men and to investigate trends in high-risk behaviour.

Eligibility Criteria:

Eligible participants are males between the ages of 18 and 30, living in the Greater Vancouver area, who have not previously tested HIV-positive.

The study is open to all men who have sex with men, regardless of whether they self-identify as gay, bisexual or straight.

• 18 to 30 years old at baseline
• Living in Greater Vancouver area
• Gay, bisexual and/or MSM
• No previous positive HIV test result

Study Design:

Eligible subjects who agree to participate complete an annual self-administered questionnaire and submit a blood sample for HIV antibody testing, for which they provide informed consent.

• Open cohort
• Annual HIV antibody test
• Self-administered questionnaire

Questionnaire Data:

The questionnaire includes information on demographics, sexual behaviour with both men and women, and different types of male partners. Definitions are included in the questionnaire to differentiate between different types of partners and kinds of sexual activity.

A distinction is made in the questionnaire between regular and casual male partners; the former being defined as "men you have sex with at least once a month," and the latter being "men you have sex with less than once a month."

We also differentiate between consensual and non-consensual sex. The questionnaire also includes questions on recreational drug use, social support and depression.

• demographics
• sexual behaviour
• with men / women
• regular / casual partners
• consensual / non-consensual / paid sex
• Recreational drug use
• Institutionalization
• Social Support (26 item IES scale)
• Depression (7 item CES-D)

Recruitment:

This graph depicts recruitment for the Vanguard project, since it began in May 1995. Shown in red are subjects who have provided contact information for participation into the study, but who have not yet completed a baseline questionnaire. As of April 22, 1998, 684 participants had completed the baseline questionnaire (as shown in blue), and 388 participants had returned for their first follow-up visit (as shown in yellow), and 149 have completed a third questionnaire (as shown in orange).

Unprotected Anal Sex:

The level of unprotected anal sex with both regular and casual male partners at baseline was higher than anticipated. Among men with regular partners, almost half reported having at least one episode of unprotected anal sex, either insertive or receptive.

Of note is that among men with regular or casual partners, higher proportions reported having unprotected sex without ejaculation. Since this practice does not necessarily protect against transmission of HIV or other STDs, additional prevention messages are needed which target specific sexual activities among young MSM to counter any myths about safer sex practices.

It is also of mention that while a smaller proportion of men engaged in unprotected anal sex with casual partners, fully one-quarter of respondents reported having done so at least once in the previous year.

REGULAR PARTNERS (n=503)	INSERTIVE	RECEPTIVE
Without ejaculation	151 (30%)	172 (34%)
With ejaculation	121 (24%)	134 (27%)
At all with a regular partner:  245 (49%)
CASUAL PARTNERS (n=537)	INSERTIVE	RECEPTIVE
Without ejaculation	85 (16%)	74 (14%)
With ejaculation	51 (9.5%)	31 (5.8%)
At all with a casual partner: 140 (26%)

* In the year prior to baseline.

HIV Prevalence and Incidence:

HIV prevalence at baseline was 2%. This figure should not be extrapolated to the wider community, however, as we excluded people who had previously tested positive.

More disturbing was the fact that HIV incidence was 1.49 per 100 person years (although confidence intervals remain wide due to small numbers). Although this incidence rate is somewhat lower now than was reported by our group one year ago, this rate of infection remains unacceptably high.

Over three-quarters of those due back have now returned for at least one follow-up visit.

*Participants >2 months late considered lost to follow-up.

Profile of Seroconverters:

Of particular interest is the profile of the 10 seroconverters observed to date. Note that the median age of these men (23 years at baseline) is actually lower than the median age of 25 for the entire cohort.

Two of the seroconverters (#8 & #9 on the slide) are now deceased from non-HIV-related causes, and the fact that they had seroconverted was drawn from the coroner's reports. Follow-up questionnaires are, however, available for the remaining eight seroconverters.

Five of the seroconverters reported having been paid for sex. Four of the seroconverters reported injection drug use, two of whom reported having shared needles.

Note that four of the seroconverters reported having engaged in unprotected anal sex with a male they knew at the time was HIV+, suggesting that serodiscordance is a reality in many gay male relationships, and special prevention efforts are needed to target both HIV+ MSM and those with HIV+ partners.

* Either in year prior to baseline or during follow-up interval.
** Reported sharing needles.

Follow-up questionnaires available for all seroconverters except: #8 & #9 (deceased).

Objective of Relapse Analysis:

I will now highlight one of the analyses which has been conducted on this cohort to date which examines sexual behaviours.

To investigate trends in condom use during anal intercourse among young men having sex with men (MSM) at baseline vs. follow-up.

Relapse Analysis:

To investigate current trends in sexual behaviours, we studied 366 participants who had returned for follow-up as of April 1998. Behaviours pertained to the year prior to baseline, and the period from baseline to follow-up. We compared baseline and follow-up responses with respect to unprotected anal insertive or receptive intercourse with regular and casual male partners, as defined earlier.

Relapse defined as < 100% condom use during anal sex at follow-up, compared to 100% at baseline.

Odds ratios for relapse tested using McNemar's test for matched pairs.

Regular partners:

Among men reporting always using condoms during anal intercourse prior to baseline, we defined "relapse" as any episode of unprotected anal intercourse reported at follow-up. Odds ratios (ORs) for relapse were tested using McNemar's test. Among men with regular partners (n=285), the odds of relapse at follow-up relative to baseline were approximately two-fold for both unprotected insertive and receptive anal intercourse. For example, among men with regular partners with whom they had insertive anal sex, the odds of relapse are reflected by a ratio of these two cells: 52 and 23, for a an odds ratio of 2.26.

*Based on McNemar's Test.

Casual partners:

Among men with casual partners (n=278), similar increased odds of relapse were observed for unprotected insertive anal intercourse, but there was no significant change for receptive anal intercourse. Taken together, these data suggest a disturbing trend towards increasing levels of unprotected anal sex among young gay and bisexual men in the Vancouver area.

Among other factors, complacency towards HIV infection could reflect undue optimism surrounding recent advances in therapy, feelings of fatalism and inevitability, or lack of direct experience of the AIDS epidemic among a newer generation of gay men. In any case, these data underscore the urgent need for HIV interventions among young gay men who remain at high risk for infection despite a decade of prevention efforts.

*Based on McNemar's Test.

Limitations:

It is important to acknowledge the limitations of this study, as well as the preceding analysis.

First, the definitions used for casual and regular partnerships can be considered somewhat restrictive.

Second, since behaviours are self-reported, the validity and reliability of measures, although shown to be adequate in previous studies, is nevertheless a concern.

Differential follow-up may have underestimated HIV incidence estimates and rates of relapse. Efforts to reduce this bias are underway.

Finally, this study is comprised of volunteers and represents a city with a different ethnic mix than other urban cities, and thus caution should be exercised when generalizing these results.

• Definitions (eg. risk taking, partnerships, non-consensual sex)
• Self-reported risk behaviours
• Differential follow-up
• Generalizability

Conclusions:

Despite these limitations, continued high levels of unprotected anal sex are consistent with a high HIV incidence.

Seroincidence data revealed the need to take into account HIV+ men and men in serodiscordant relationships, in prevention programmes aimed at this population.

Finally, there is evidence of a relapse to unsafe sex among men who were previously reporting consistent safe sex practices.

• High levels of risk behaviour consistent with high HIV incidence
• Importance of serodiscordant relationships
• Disturbing trends suggest "relapse" to high risk sex

Web site:

This presentation, along with all other Vanguard Project presentations and press coverage, can be found on the Vanguard web site, which can be accessed via the menu at <www.hivnet.ubc.ca>.