In developed countries, sexually transmitted infections (STI) and bloodborne pathogens (BBP) disproportionately affect marginalized populations. In the United States, Australia, and Canada the combined impact of poverty, lack of access, and historical and systemic oppression have resulted in overrepresentation of indigenous populations in national HIV/AIDS and STI statistics, especially amongst females and youth123456. Within Canada, injection drug users (IDU) account for a significant proportion of prevalent HIV and other BBP (such as hepatitis C [HCV]) infections, and are an especially important risk group sustaining endemicity of these pathogens within Aboriginal populations4789. However, despite progress in, and substantial efforts towards both understanding, and addressing BBP epidemics in Canadian Aboriginal populations7, the transmission of some BBP, such as HIV and HCV, appear to be growing unabated101112. This paradox has motivated researchers to examine heterogeneity in marginalized subpopulations, with the intention of finding and describing subpopulations that may be at particularly high risk of BBP transmission, as well as the environmental contexts within which they are embedded13141516.

To this end, solvent abuse has been shown to be a particularly serious and destructive issue affecting Aboriginal populations in Canada, and elsewhere171819202122232425. In North America, the lifetime use of solvents has been reported to be as high as 44% in some high-risk groups26, with some studies finding the prevalence of lifetime use at 17% by the eighth grade27. Solvent use is a term broadly applied to the self-administered inhalation of a variety of volatile, psychoactive substances that are found in many common products, including gasoline and adhesive glue2428. Solvent users have elevated rates of negative health outcomes including mental illness2930, damage to the central nervous system, heart and lungs 283132, as well as mortality3233. Contributing to its perniciousness, solvents are primarily legal and easily obtainable1834. As well, multiple factors have been identified as being associated with solvent use, including age, sex, ethnicity, education level, co-existing alcohol or other substance use disorders, and child and physical abuse3536373839. In youth, solvent use has been linked to broader societal issues such as higher school drop-out rates40, delinquency (including criminal activity)3639 and family conflict3941. Salient to this study, an association between chronic solvent use in adolescence and injection drug use among the most marginalized of populations has been demonstrated20424344.

On the treatment side, a particular defining feature of chronic solvent use is that it is typically associated with the most marginalized populations, with, for example higher levels of anti-social behaviour, trauma-exposure and psychiatric morbidities192445. In response to the burgeoning need for Aboriginal-specific programs, Canada has over a dozen solvent abuse treatment centres spread across the country46. These centres operate under a continuum of interventions, including prevention, early intervention, residential treatment and environmental deterrence. Furthermore, evidence suggests IDU with a solvent use background have a "specific course of addiction"39, often with much more detrimental outcomes, and a particular intransigency to treatment3947. This "deviant group within a deviant group" has been recognized since the late 1970s47, but is still poorly understood, relative to other IDU groups.

Despite the link observed between solvent use and IDU, and the disproportionate burden of both solvent abuse and STI/BBP infection in Aboriginal populations, there is little published research on solvent use among Aboriginal IDU. We therefore undertook this study to examine the association between solvent use and socio-demographic and drug-related risk factors in Aboriginal IDUs in Manitoba, Canada. We were also interested in examining the relationship between solvent use and injection of other types of illicit substances, as well as being infected with a BBP (i.e., HIV and HCV).

A total of 272 IDU identified as Aboriginal. An additional 6 that identified as transgendered were excluded from the analyses due to small numbers, leaving a total sample size of 266. Overall, 44 (16.5%) of the study sample reported solvent use in the last 6 months. Table 1 displays a comparison of characteristics of solvent and non solvent-using IDU. Broadly speaking, the two groups differed significantly, at least at the p < .05 level, by age, injection locations, injection risk behaviours, type of drugs injected and BBP status (Table 1).

This study examined the association between solvent use in Aboriginal IDU and socio-demographic factors, drug-related risk factors, use of other illicit substances and BBP infection. We found that after adjusting for other variables including sex, solvent use was significantly associated with Talwin & Ritalin injection, HCV status and age in this population.

Some important limitations of the study should be stated at the outset. First and foremost, ours was a cross-sectional study, and a causal linkage between solvent use and injection drug use cannot be inferred from the data. Although both likely share determinants, our data are insufficient to establish causality. Aboriginal individuals in Canada face a combination of socially and structurally determined vulnerabilities, including high rates of entrenched poverty, unemployment, homelessness and sexual and physical abuse25253. Many of these factors stem from a history of colonization, oppression, systemic racism and discrimination in Canadian society and have resulted in Aboriginal Canadians having unequal access to a variety of resources 254. Thus, the perniciousness of both solvent and injection drug use within Aboriginal populations is more likely a result of these determinants. Second, solvent use was measured broadly. The measure used was not precise enough to discriminate between chronic and casual use. Similarly, different types of solvents were not captured in this study. Third, since a sampling frame was not possible to construct for this marginalized and hidden population, the sample was not randomly generated and may not be representative of Aboriginal IDUs in other settings, or in Winnipeg. Fourth, social desirability bias, or high non-response rate is always an issue with self-reported data; however, it is likely that this would have served to underestimate associations toward the null. Finally, the sample size was relatively small and thus may have not had power to detect significant findings.

Previous studies in Winnipeg have reported Talwin & Ritalin injection as being strongly associated with both Aboriginal ethnicity5055 and high HCV prevalence49. That HCV infection is three times more likely in the population of solvent-using Aboriginal IDU, after controlling for Talwin & Ritalin injection and risky injection practices, strongly suggests the existence of pockets of higher risk even amongst an already high-risk subpopulation3947. It was also demonstrated that these qualitatively distinct 'higher-risk' groups can be distinguished when both injectable and non-injectable drug use is considered.

The relatively low prevalence of both HIV and HCV among IDU in our geographic setting has motivated researchers to ask what role, if any, public health responses in Winnipeg may have contributed to lower prevalence56. Both HIV and HCV prevalence in the subset of solvent-using IDU are relatively higher than other IDU in our sample; and at 18% and 81% respectively, are in closer alignment with the prevalence observed in other jurisdictions5758. This dichotomy in prevalence reinforces the exceptionally high risk faced by solvent-using IDU, and their real or potential ability to be missed by what otherwise may be an effective public health response. This higher-risk group is particularly relevant given the recent attention paid to especially high rates of HIV in Aboriginal populations in central Canada1112, and serve to illustrate that BBP epidemics in Canada are not homogeneous.

Solvent use is an issue where there are no easily-identifiable solutions2324. Solvent users are at the bottom of a drug-using hierarchy, in terms of perception by other substance users and practitioners, and by the sheer volume of their social and personal challenges29394247. Thus, given the already difficult lifestyle and behavioural issues related to injection drug use5859, a combination of solvent use and injection drug use within Aboriginal populations may present considerable, and specific challenges for treatment3947. For example, although there is well-established literature on the effectiveness of harm-reduction efforts such as needle-exchange programs in curtailing the spread of BBPs6061, the constituents of an equivalent and appropriate harm reduction strategy for solvent users have not been well articulated in the literature 24, although practical advice may include using solvents in groups, and using clean rags or sponges. As well, outreach efforts to these populations may be unduly hampered by the considerable stigma attached to chronic solvent use. Similar to recent Canadian research demonstrating that IDU who also smoked crack cocaine were at higher risk of HIV seroconversion10, perhaps an especially chaotic lifestyle is contributing to the higher HCV prevalence in our solvent-using subpopulation.

In conclusion, although addressing social or peer group norms has long been advocated as part of an effective prevention and treatment strategy for IDU, perhaps structural-level interventions are especially indicated for solvent-using Aboriginal IDU. At a time when rates of HIV and other BBPs are escalating in Canadian Aboriginal populations, studies like this one can help inform targeted strategies, as well as motivate harm reduction research in very marginalized populations. The strong socially-constructed vulnerabilities of Aboriginal populations, the illegality of injection drug use, the obduracy of solvent use to traditional regulation and control, and the extreme marginalization of solvent users may be interacting to create a 'perfect storm' for those IDU already infected, and those at high risk for infection to slip through the cracks in public health systems.

The authors declare that they have no competing interests.

SS was responsible for conceptualization of the study, analysis, interpretation of data and writing of the manuscript. KD made substantial contributions to data analysis and interpretation and revised the manuscript critically, and made important intellectual contributions to the manuscript. AJ and JW conceived the study, acquired the data and helped draft the article, as well as revised it critically. All authors have given final approval for this version of the manuscript.