METHODS

Study population

The Vancouver injection drug users study (VIDUS) is a prospective study of over 1400 injection drug users that began in May 1996 during an explosive outbreak of HIV in Vancouver's DTES. The study established a storefront office, with participants coming to the study through local advertisements and word of mouth. For study enrolment participants had to give a history of injection drug use in the proceeding month and provide written informed consent. Blood samples for HIV and hepatitis C antibody testing are collected, and each participant undergoes an interview administered questionnaire at the baseline visit and semi-annually thereafter. Participants are reimbursed Cdn$20 for each study visit, at which time referrals are provided for medical care, HIV/AIDS care, drug and alcohol treatment, and counselling. Detailed information is collected on demographics, sexual behaviours, types of drugs injected, frequency of drug use, incarceration, housing, and the utilisation of health services such as emergency departments, clinics, and hospitals.

Statistical analysis

Univariate statistical techniques were applied to determine factors associated with female sex work. Categorical explanatory variables were analysed using Pearson's χ² test, and the Wilcoxon rank-sum test was used for continuous variables. For this comparison, we identified 234 female sex workers, defined as those who were sexually active and reported selling sex in the six months before study enrolment. The condom use and sex partner information in table 3 was taken from the questionnaire administered at the most recent follow up visit (June to December 2000) and identified 163 women actively engaged in commercial sex work. Drug use and other behavioural variables referred to activities in the past six months.

RESULTS

Demographics

The demographic characteristics of the VIDUS cohort are shown in table 1. Male participants are older and have longer injection careers than female participants. The ethnic make-up of the cohort generally reflects the community, with a disproportionate number of female subjects among the Canadian aboriginal participants. The majority of the Hispanic population is male. The most common housing arrangements are single occupancy hotel rooms that are rented by the month. Seventy per cent of the participants receive government welfare payments for housing and other support. The prevalence of HIV-1 is higher among women (35% v 26%) and the prevalence of hepatitis C is similar among women and men.

Sexual activity and reported condom use

Table 2 summarises the reported sexual activity collected from baseline questionnaires. During the six months before enrolment, 72% of men and 92% of women reported some sexual activity. The majority reported heterosexual partners, although same sex partners were not uncommon. Among male subjects, 39% reported fewer than 20 lifetime sex partners, while 78% reported fewer than 100. In contrast, 26% of female subjects reported fewer than 20 lifetime partners and 43% reported less than 100. The high proportion of women involved in the sex trade was reflected in the observation that 57% of women reported at least 100 lifetime partners.

Table 3 summarises the reported frequency of sexual activity, the type of sexual activity, and condom use, stratified by regular partners, casual partners, and paid partners. Of participants who reported any sexual activity in the past six months, 593 (56%) reported having a regular sex partner of the opposite sex, defined as “someone you have had a sexual relationship with for more than three months.” Women were more likely to report having a regular sex partner then men (63% v 51%, p < 0.001). Vaginal sex was reported by 99% of male and female participants. The reported frequency of vaginal intercourse was at least one to three times a week by 79% of men and 74% of women. Overall, 65% of participants never used condoms during vaginal intercourse with their regular partner. Seventy six per cent of participants had engaged in oral sex in the past six months. Condom use during oral sex with a regular partner was very uncommon. Anal sex with a regular partner was reported by 20% of participants but was practiced infrequently.

One or more casual sex partners of the opposite sex were reported by 471 (45%) of participants. Although vaginal intercourse was reported by 95% of respondents, 63% reported this activity less than once a week. Condoms were “always” used during vaginal sex with casual partners by about half of the participants. Oral sex was reported by 53% of this group but was also relatively infrequent. Condoms were generally not used during oral sex with casual partners. Ten per cent of male participants and only 5% of female participants reported anal sex with casual partners.

Nine per cent of men reported sex with other men in past six months. Oral sex was reported by 95% of this group, with receptive and insertive anal intercourse reported by 32%. For oral sex, condoms were “always” used with paying partners by 60%. Of the men who reported anal sex with paid contacts, 71% reported “always” using condoms. Of the 182 female participants who reported commercial sex work in the past six months, 163 (90%) had paying heterosexual partners or “tricks”. Vaginal sex was practiced by 90%, with 92% reporting a frequency of at least one to three times a week. Condom use was relatively high, with 82% reporting that they “always” used a condom. Oral sex was reported by 99%, with 94% having at least one to three contacts a day. Seventy five per cent reported that they “always” used a condom during oral sex. Anal sex was relatively uncommon and infrequent among female sex workers.

Female sex workers

Table 4 compares the baseline demographic and drug use characteristics of female sex workers with other women in the cohort. Female sex workers were less likely to have a regular sex partner. There was no statistically significant ethnic difference among women involved in sex work, although aboriginal women tended to be less involved. Female sex workers were more likely to live in the DTES, to have unstable housing arrangements, and to report incarceration in the previous six months. Regular heroin use was marginally higher among female sex workers, although frequent injection cocaine and crack cocaine use was much more common in this group. The HIV prevalence was higher among women who sold sex. At baseline, HIV prevalence was 32% among female sex workers and 21% among non-sex-workers. HIV seroconversion during follow up was documented in 10% of active female sex workers and 7% of the other female participants. The reported occurrence of an STD was relatively uncommon overall, although marginally higher among female sex workers.

STD history

STD history among male and female participants is shown in table 5. Compared with the men, the women in the cohort are more likely to have had an STD in the past. This association is consistent for both non-ulcerative (gonorrhoea and chlamydia) and ulcerative infections (syphilis and herpes). For the 685 participants reporting a previous STD, the rate of HIV was marginally higher (53% v 46%; p = 0.05). This association was similar when stratified by sex.

Sexual networks

Figure 1 is a compartment model from the DTES with female sex workers as the core group. Although many of these women also use injection drugs, there is a sizeable minority who do not. Endemic rates of STDs are sustained among the female sex workers by sexual activity with males who reside in the DTES. These men can be classified into four general categories. Group 1 consists of regular male sex partners who use injection drugs and also reside in the DTES. The money generated by their female partner's sex work also helps to support their drug use. Group 2 consists of regular male sex partners who do not use injection drugs themselves. Some of these arrangements are based on the organised recruitment of women through drugs and intimidation. Group 3 consists of resident males who purchase sex and do not inject drugs. These are mainly older men who are long term residents of the community and are marginalised because of various physical and mental ailments. Group 4 consists of resident males who are injection drug users and purchase sex. As most men in the area do not have regular sex partners, much of their sexual activity occurs with paid partners. This may also include sex-for-drugs arrangements.

The bridging to populations outside the DTES occurs primarily through men who visit the DTES specifically for sex. The sex may occur within the hotel system in the DTES or by driving to locations outside the area. These men are usually non-injection drug users and have both casual and regular sex partners residing outside of the DTES. A second important bridging population is foreign visitors to the area, either tourists or ship workers stopping at the Vancouver port. This type of sexual contact has far reaching implications for the transmission of STDs/HIV to other cities. It is also a potential route for STD pathogens to be introduced into the community.

DISCUSSION

The complex interaction between illicit drug use and commercial sex work presents major challenges for STD prevention and policy.^{3, 4} The DTES of Vancouver provides a graphic example of how an explosive outbreak of HIV infection occurred among a population of injection drug users despite a range of public health interventions including a legal needle exchange, methadone maintenance treatment, and free health care.¹⁵ The transmission of other STDs and the contribution of sexual transmission to the HIV outbreak have received less attention. A recent outbreak of syphilis in the province of British Columbia, however, traced the epicentre of transmission to the DTES, leading to a syphilis mass treatment campaign in the neighbourhood.

The transmissibility of HIV through blood to blood contact, as occurs among injection drug users, is thought to be severalfold higher than through sexual transmission.¹⁸ Therefore the actual proportion of HIV resulting from sexual transmission in this population is difficult to determine. However, the concentration of female sex workers, the large numbers of sex partners, the complex networks that have evolved, and the relatively low rates of condom use create an environment where sexual transmission of HIV does occur.^{19, 20} In addition, sexual relationships and intimacy are closely linked to drug using patterns and needle sharing.^{21, 22} Rothenberg et al propose that the formation of social networks may act as a barrier or a facilitator for disease transmission and that the context for risk behaviours influences transmission dynamics.¹⁴ The geographical concentration of injection drug users in the DTES lends itself to these influences, and understanding these dynamics has implications for prevention strategies.

The relatively low uptake of condoms and the large number of sex partners documented in this cohort study are consistent with other injection drug user populations.^{4, 23} Female injection drug users who also sell sex continue to put themselves at high risk, knowing they are in an environment with extremely high rates of HIV and other STDs. Drug addiction propels many women into sex work in order to provide the money needed to maintain their drug habits. As demonstrated in this cohort, women injection drug users who sold sex were more likely to be resident in the DTES and to have unstable housing situations, and were less likely to have regular sex partners. The intensity of drug use was also significantly higher in this group, compared with women who did not sell sex. Drugs impair judgement, making attempts to use condoms consistently or avoid high risk sexual behaviours difficult, even for the most motivated and well informed. As with other studies in drug using populations, we found that injection cocaine and crack cocaine were commonest among female sex workers.²⁴

Several interventions designed to reduce sexual risk behaviours among injection drug users have been developed. Outreach and community based programmes that promote safer sexual practices can be successful in some situations.²⁵ Further, it has been shown that those in drug treatment programmes are less likely to engage in high risk sexual activities than those out of treatment.²⁶ Methadone maintenance treatment participants have also been shown to have fewer sexual partners and in some cases are able to disengage from commercial sex work.²⁷ Certainly these initiatives need to be expanded and applied to other communities.

In the dynamic topology of STD epidemics described by Wasserheit and Aral,²⁸ the population in the DTES has the characteristics consistent with phase IV epidemics. This includes a geographically concentrated subpopulation with high rates of partner change and little contact with the health care system. Further, external factors such as poverty, low levels of education, and illicit drug use serve to sustain the epidemic in this community. The VIDUS cohort illustrates how these factors can converge to push individuals to the margins of society where interventions are more difficult to initiate.

Strategies that primarily target “spread networks” have been a priority in this community. In a phase IV epidemic environment this would include targeted health promotion activities. In the DTES there has been a wide range of targeted health promotion initiatives including safe sex posters and pamphlets distributed by health clinics and community groups, peer risk reduction counselling delivered through support groups for sex workers and injection drug users, and an active street nurse programme that delivers information directly to individuals who engage in high risk behaviours. The introduction of partner notification is particularly difficult in this community. There is a general reluctance to cooperate with health officials in divulging names and addresses, as such people are perceived to be associated with the criminal justice system. Many of the highest risk individuals have previous drug related convictions and fear arrest and conviction on illicit drug charges.

As the epidemic progresses through phases, the subpopulations become characterised by progressively higher rates of partner change, less effective contact with the health care system, and lower educational and social status.²⁸ This is very characteristic of injection drug user populations, when drug users spend increasing resources on illegal drugs and are driven further away from social support systems. The large proportion of individuals who receive a criminal conviction for drug related offences (more than 70% in our cohort) may also limit the future potential for housing and employment.

Intervention research is imperative to help better define optimal approaches to STD control in this population. This is important not only to reduce the sexual transmission of HIV within the community but also to prevent transmission of infections to the bridging populations. As illustrated in fig 1, the central location of this community allows contacts with a range of non-residents. The open illicit drug markets, as well as the commercial sex trade, draw people to this neighbourhood. As described by Wasserheit and Aral, research on bridge populations is crucial as this group serves to maintain transmission to other populations.

This study has some limitations inherent to prospective cohort studies. Although follow up exceeds 80%, there is a loss to follow up that may introduce bias. It is probable that the higher risk participants are less likely to return, and therefore high risk behaviours are underestimated. Further, we rely on participants identifying themselves to the study and this may attract higher risk individuals and limit the extent to which the results can be generalised to the injection community as a whole. Self reporting drug use and sexual behaviours may be prone to socially desirable responses and poor recall. However, there is some evidence that the validity of self reported drug use behaviours is not seriously compromised in this population.²⁹ In any case this would be likely to lead to an overreporting of condom use and an underreporting of high risk sexual activities, making the study observations even more compelling. Finally, we did not routinely perform diagnostic STD testing and could not determine with certainty the pathogen specific phases of the epidemic.

The transmission of STDs in drug using communities requires urgent study. It is clear that female drug users who also sell sex are a population in need of intensive STD prevention interventions. While providing condoms, free diagnostic testing, and antimicrobial treatment are important components of prevention, they have not been sufficient. Existing programmes should be strengthened, but innovative approaches to STD surveillance, diagnosis, and prevention are needed. One approach would be to form closer links between STD prevention and existing harm reduction initiatives. Reducing STD transmission among injection drug users has far reaching implications for STD control in general and remains a major public health challenge.