This talk was presented at the AIDS Impact conference in Brighton, England, in July 2001. A poster presentation of these data was also displayed at the conference.

Increasing rate of new HIV infections among young gay and bisexual men in Vancouver, 1995-99 vs. 2000

Steve Martindale, Kevin JP Craib, Keith Chan, Mary Lou Miller, Darrel Cook*, Robert S. Hogg.

The Vanguard Project, BC Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada.
* BC Centre for Disease Control, Vancouver, BC, Canada.

Good morning. My name is Steve Martindale and I'm the coordinator of the Vanguard Project, which is an ongoing study of HIV rates and risk factors among young gay and bisexual men in the Greater Vancouver region of British Columbia, Canada.

I'm here to speak to you today about the recent increase in HIV incidence that we've documented in our cohort in the year 2000.

I'd like to start by acknowledging my co-authors as listed on the slide, all of whom are my colleagues at the BC Centre for Excellence in HIV/AIDS at St. Paul's Hospital in Vancouver, with the exception of Darrel Cook who is with the BC Centre for Disease Control.

BACKGROUND:

• Young MSM still at high risk for HIV infection
• Increased risk behaviours
• Increased HIV rates in Canadian cities outside of Quebec

As the HIV epidemic enters its third decade, gay and bisexual men continue to have among the highest rates of HIV infection in Canada.

Previous studies have highlighted the decline in HIV incidence and risk behaviour among gay and bisexual men. However, several studies ñ including our own ñ have suggested that young gay and bisexual men continue to engage in unprotected sexual behaviours.

Recent reports have documented an increase in the incidence of HIV among gay and bisexual in several urban centres in Canada, with the exception of Montreal.

OBJECTIVES:

• To determine changes in HIV incidence among young gay and bisexual men in an open cohort
• To identify demographic characteristics and risk factors related to recent HIV seroconversions

The primary objective of the Vanguard Project is to monitor HIV incidence rates and associated risk factors among young men who have sex with men.

The objectives of this analysis were initially to determine recent trends in the incidence of HIV among young gay and bisexual men in our cohort in Vancouver, and subsequently to identify demographic characteristics and risk factors related to recent HIV seroconversions.

THE VANGUARD PROJECT:

• Prospective study of over 900 MSM
• Living in Vancouver area at baseline
• 15 to 30 years of age at baseline
• No previous HIV-positive test result
• Recruited through outreach, publicity, medical clinics and physicians
• Annual self-administered questionnaire
• HIV testing and stored blood specimens

The Vanguard Project is a prospective study of gay and bisexual men aged 15 to 30 who live in the Greater Vancouver region at baseline.

To date over 1000 participants have been recruited through outreach, publicity, medical clinics, and physicians' offices, and over 900 have completed baseline questionnaires.

The study is open to men who self-identify as gay or bisexual, as well as men who have sex with other men. To be eligible for this study, participants must have not previously tested positive for HIV.

Since the study began in 1995, Vanguard participants have completed annual self-administered questionnaires and provide regular blood samples for HIV antibody testing and storage.

STATISTICAL METHODS:

• HIV incidence calculated using person-time methods
• May 1995 to December 2000
• Wilcoxon rank-sum test (continuous variables)
• Pearson's Chi-squared test (categorical)
• 95% confidence intervals calculated based on Poisson distribution

Among men who had returned for follow-up, HIV incidence was calculated per annum as the number of new infections divided by the total person-time under observation for each calendar year, from 1995 to 2000.

Conventional statistical methods were used to compare continuous and categorical variables, and 95% confidence intervals for the incidence estimates were calculated.

RESULTS: HIV incidence by year and category

As of December 31, 2000, 668 participants had completed at least one follow-up visit and more than one HIV test.

This table provides a summary of HIV incidence in the cohort by year since 1995. A total of 26 new HIV infections were observed. The overall HIV incidence rate in the cohort was 1.4 per 100 person-years, as highlighted in the purple bar.

Participants who reported ever injecting drugs had a higher overall incidence rate than those who did not (3.9 vs. 1.0 per 100 person-years as shown in the red bar). Note, however, that our sample size of injection drug users is small and confidence intervals for the IDU comparison are wide.

A significant increase in new HIV infections was observed between the first four year of the study (1995 to ë99) and the fifth year of the study (2000) within the entire cohort, as shown in the green bar; this remained true when participants who had injected drugs were excluded from the analysis, as shown in the blue bar.

RESULTS: HIV incidence by year and category

These same data are presented in this graph.

As shown here in the red line, the rate of new infections within the entire cohort increased from an average of 0.9 per 100 person years in 1995-99 to 5.0 per 100 person years in the year 2000.

A similar increase is seen among those who have never injected drugs, as shown here in the blue line: the rate of new infections increased from an average of 0.6 per 100 person years in 1995-99 to 3.7 per 100 person years in the year 2000.

BCCDC RESULTS: Newly positive HIV tests among MSM in BC

Our findings are supported by data from the BC Centre for Disease Control.

As shown in this chart, between 1999 and 2000, the BC CDC found a 31% increase among MSM in British Columbia. Approximately three-quarters of these men live in Vancouver.

If this category is restricted to exclude men who inject drugs and/or "work the street,î we see a similar trend, with a 36% increase in newly positive HIV tests.

The BC CDC also reports dramatic increases in newly positive HIV tests among both younger men and older men. Although for the past five years the greatest number of MSM testing positive for HIV have been in their 30s, between 1999 and 2000 the BC CDC found a 66% increase among MSM over the age of 40 and a 71% increase in newly positive HIV tests among MSM under the age of 30.

NESTED CASE-CONTROL ANALYSIS: Methods

• To identify demographic characteristicsand risk factors for HIV seroconversion
• 23 HIV-positive cases
• Index visit: last HIV-negative Q
• 69 HIV-negative controls
• Index visit: matched Q within 3 months

To identify demographic characteristics and risk factors associated with new HIV seroconversions, we performed a nested case-control analysis.

23 cases were identified. These were participants who had seroconverted between May 1995 and December 2000. For each case, three HIV-negative control participants were randomly selected from all participants who remained seronegative as of December 2000.

Risk factor data for cases were taken from an index visit, which was defined as the questionnaire completed at the last visit at which the participant tested HIV-negative. This was done in order to assess risk behaviours that took place prior to seroconversion, as testing positive for HIV may affect a participant's behaviour.

For controls, risk factor data was obtained from a matched visit which occurred within three months of the index visit for the corresponding case.

NESTED CASE-CONTROL ANALYSIS: Socio-demographic characteristics

	HIV+ (n=23)	HIV­ (n=69)	Odds Ratio	p-value
Median age (IQR)	24 (22-27)	27 (24-29)	0.9 (0.8-1.0)	0.065
Non-Caucasian	4 (17%)	21 (31%)	2.1 (0.6-7.0)	0.210
Low education*	7 (30%)	3 (4%)	9.6 (2.2-41.4)	< 0.001
Unstable housing	6 (26%)	5 (7%)	4.4 (1.2-16.4)	0.017
Income < $10,000	6 (33%)	17 (28%)	1.3 (0.4-3.9)	0.683

* Less than high school

In univariate analyses, cases were significantly more likely than controls to report having less than a high school education and to lack stable housing.

NESTED CASE-CONTROL ANALYSIS: Sexual behaviours in the past year

	HIV+ (n=23)	HIV­ (n=69)	Odds ratio	p-value
Median # of regular partners (IQR)	2 (0-4)	1 (0-2)	1.7 (1.2-2.4)	0.023
Median # of casual partners (IQR)	5 (0-40)	2 (0-13)	1 (1.0-1.0)	0.453
Sold sex	7 (30%)	8 (12%)	3.3 (1.1-10.6)	0.034
INSERTIVE sex with REGULAR partner*	13 (57%)	16 (24%)	4.2 (1.6-11.4)	0.003
RECEPTIVE sex with REGULAR partner*	15 (65%)	14 (21%)	7.2 (2.6-20.5)	0.001
INSERTIVE sex with CASUAL partner*	10 (43%)	10 (15%)	4.5 (1.5-12.9)	0.004
RECEPTIVE sex with CASUAL partner*	8 (35%)	5 (7%)	6.7 (1.9-23.5)	< 0.001

* Unprotected anal sex

Cases were also more likely than controls to have sold sex in the previous year, to have engaged in unprotected anal insertive or receptive sex with a regular or casual partner and to have had a greater number of both regular and casual partners.

We defined regular partners as men you have sex with at least once a month in the past year, and casual partners as men you have sex with less than once a month.

Injection drug use in the previous year was not found to be statistically different between cases and controls.

NESTED CASE-CONTROL ANALYSIS: Independent predictors of HIV seroconversion

	Unadjusted Odds Ratio	95% C.I.	Adjusted Odds Ratio	95% C.I.
Unprotected anal receptive sex with REGULAR partner	7.2	(2.6-20.5)	6.6	(2.2-19.6)
Unprotected anal receptive sex with CASUAL partner	6.7	(1.9-23.5)	5.7	(1.4-22.7)

Multivariate analysis revealed unprotected anal receptive intercourse with both regular and casual partners to be independently associated with the risk of seroconversion.

This of course supports what we already know about the risk of unprotected anal sex, but it's worth noting that we found unprotected sex with regular partners carries a similar odds ratio to unprotected sex with casual partners, which may have implications for gay and bisexual men practising "negotiated safety" within their primary relationships.

CONCLUSIONS:

• Recent increase in HIV rate among young MSM in our cohort
• In 2000, a six-fold increase in HIV infection rate among non-IDU MSM
• MSM at risk for HIV from unprotected receptive anal sex with both regular and casual partners

Our data indicate a recent significant increase in the rate of new HIV infections in young gay and bisexual men in our cohort. Among non-injection drug using participants we observed a six-fold increase in infection rates in the year 2000.

These results corroborate previous reports from other large North American cities, suggesting the need for continued surveillance.

In our investigation of the determinants of seroconversion, we found evidence that prevention efforts should incorporate issues related to unprotected sex with regular partners in order to reduce the risk of seroconversion among young gay and bisexual men.