CMAJ Research Crystal methamphetamine and initiation of injection drug use among street-involved youth in a Canadian setting Dan Werb MSc, Thomas Kerr PhD, Jane Buxton MBBS MHSc, Jeannie Shoveller PhD, Chris Richardson PhD, Julio Montaner MD, Evan Wood MD PhD Abstract Competing interests: Julio Background: Although injection drug use is involved youth provided 1434 observations, with Montaner has received grants from, served as an known to result in a range of health-related 64 (16.2%) participants initiating injection drug ad hoc adviser to or spoken harms, including transmission of HIV and fatal use during the follow-up period, for a cumula - at events sponsored by overdose, little is known about the possible tive incidence of 21.7 (95% confidence interval Abbott, Argos Therapeutics, role of synthetic drugs in injection initiation. [CI] 1.7–41.7) per 100 person-years. In multivari - Bioject Inc., Boehringer We sought to determine the effect of crystal able analysis, recent noninjection use of crystal Ingelheim, BMS, Gilead methamphetamine use on risk of injection in - methamphetamine was positively associated Sciences, GlaxoSmithKline, Hoffmann-La Roche, itiation among street-involved youth in a with subsequent injection initiation (adjusted Janssen-Ortho, Merck Canadian setting. hazard ratio 1.93, 95% CI 1.31–2.85). The drug of Frosst, Panacos, Pfizer Ltd., first injection was most commonly reported as Schering, Serono Inc., Methods: We used Cox regression analyses to crystal methamphetamine (14/31 [45%]). TheraTechnologies, Tibotec identify predictors of injection initiation (J&J) and Trimeris. He has among injection-naive street-involved youth Interpretation: Noninjection use of crystal also received grant funding enrolled in the At-Risk Youth Study, a prospec - methamphetamine predicted subsequent from the BC Ministry of tive cohort study of street-involved youth in injection initiation, and crystal methampheta - Health, the US National mine was the most commonly used drug at Institute on Drug Abuse, Vancouver, British Columbia. Data on circum - Merck, Janssen and ViiV stances of first injection were also obtained. the time of first injection. Evidence-based Healthcare for the STOP strategies to prevent transition to injection (Seek and Treat for Optimal Results: Between October 2005 and November drug use among crystal methamphetamine outcomes and Prevention) 2010, a total of 395 drug injection–naive, street- users are urgently needed. in HIV & AIDS in IDU initiative. None declared for Dan Werb, Thomas Kerr, Jane Buxton, Jeannie 12 treet-involved youth are at high risk of increased risk of injection initiation. Globally, Shoveller, Chris Richardson initiating injection drug use. 1 This situa - amphetamine-type stimulants have emerged as one or Evan Wood. tion is of concern, given that injection of the most commonly used groups of illicit drugs, This article has been peer S 13 drug use has been associated with increased second only to cannabis. This increase in amphet - reviewed. risk of transmission of HIV and hepatitis C amine use is reflected in the epidemiology of drug Correspondence to: virus 2,3 and fatal overdose, 4 as well as a range of use in some Canadian settings. For instance, in Evan Wood, other serious negative health and social out - Vancouver, British Columbia, rates of injection use uhri-ew @cfenet.ubc.ca comes. Newly initiated injection drug users of crystal methamphetamine have increased sig - CMAJ 2013. DOI:10.1503 have also been identified as a subpopulation at nificantly among adult injection drug users. 14 This /cmaj.130295 particularly high risk of injection-related pattern is of substantial public health concern, harm. 5–8 Unfortunately, despite recent calls to given that crystal methamphetamine has been prioritize interventions to prevent the initiation associated with a range of health and social harms, of injection drug use, 9 there are few evidence- including the potential to drive high-risk drug-use based strategies to prevent injection initiation patterns, including injection. 15 Given these con - among street-involved youth. cerns and the well-established health-related harms This situation relates, in part, to the fact that lit - of injection drug use, we investigated the possible tle is known about the risk factors for injection in - role of crystal methamphetamine use in the inci - itiation within this population. Prospective dence of first injection drug use within a cohort of research from Montréal, Quebec, has alluded to street-involved youth in a Canadian setting. the role that crack and powder cocaine may play in promoting subsequent injection initiation, 10 as has Methods retrospective research conducted among drug users in Baltimore, Maryland. 11 Much less is known The At-Risk Youth Study, established in 2005, is about the possible role that synthetic drugs, such as an open and ongoing prospective cohort study of methamphetamine, may play in contributing to an street-involved youth aged 14 to 26 years based © 2013 Canadian Medical Association or its licensors CMAJ, December 10, 2013, 185(18) 1569 Research in Vancouver. 12 All study participants were itiation, stratified by crystal methamphetamine use. recruited through street outreach and self-refer - We then employed an a priori approach whereby ral. Consistent with previous studies of street- basic sociodemographic variables (i.e., age, sex involved youth populations, 10,16,17 potential partici - and ethnicity) were included in the final multivari - pants were eligible if they had used illicit drugs able model, along with potential confounders sig - other than marijuana in the past 30 days, if they nificant at the p < 0.05 level in univariable testing. provided written informed consent and if they This approach reflects recent guidance from Roth - were “street-involved.” Although there was no man and Greenland 21 encouraging specification of specific requirement that youth spend a mini - minimal models compatible with available infor - mum amount of time on the street to qualify for mation. Finally, we undertook a subanalysis inves - the study, in practice, the street-based recruit - tigating the circumstances of first injection in an ment produced a sample of youth who spent effort to identify specific microsetting phenomena extensive time on the street, a large proportion of accompanying injection initiation events. All sta - whom were homeless. Nevertheless, because our tistical analyses were performed using SPSS soft - study lacked an explicit requirement that partici - ware version 17.0 (SPSS, Chicago, IL). pants be living on the street, we use the term “street-involved youth,” rather than “street Results youth,” since the latter term is generally applied to youth known to live full-time or part-time on A total of 991 participants completed the At-Risk the street. 18 Once enrolled, participants com - Youth Study baseline questionnaire between pleted an interviewer-administered questionnaire October 2005 and November 2010, of whom 395 (available by request to the corresponding (39.9%) reported using crystal methamphetamine author ) at baseline, as well as physical and men - at baseline and 390 (39.4%) reported injecting tal health assessments, including blood sampling drugs at baseline. Those with a history of injection for diagnostic testing and assessment for phys - drug use at baseline were more likely to use crys - ical stigmata of injection drug use (i.e., track tal methamphetamine than those without a history lines). Participants subsequently underwent of injection drug use (183/390 [46.9%] v. 204/601 physical examinations and completed inter - [33.9%], p < 0.001). Among the 601 individuals viewer-administered questionnaires semiannual - who were injection-naive at baseline, 395 (65.7%) ly; the questionnaires solicited basic sociodemo - had a follow-up visit to assess for injection initia - graphic and drug-use data, along with data on tion and were therefore eligible for inclusion in other relevant behaviours, including injection of the present study. The 206 injection-naive partici - drugs. Participants were given a $20 honorarium pants who were ineligible for the present study at each visit. The study has been approved by the because they had not yet returned for a follow-up University of British Columbia/Providence Health Care Ethics Review Board. The study presented here includes data collected Pa rti cipants rec ruite d to ARYS and from street-involved youth who were drug injec - complet ed baseli ne interview betwe en Oct ober 2005 and Nove mber 20 10 tion–naive, who were enrolled in the At-Risk Youth n = 991 Study cohort between Oct. 1, 2005, and Nov. 30, 2010, and who completed at least one follow-up Exc luded n = 390 interview. The outcome of interest was a first • Reported injection episode of any injection drug use. The primary drug use at bas eline independent variable of interest was any noninjec - tion use of crystal methamphetamine (i.e., smoking Pa rticipants drug inje cti on –naive at baseline and/or intranasal use) in the previous 6 months. n = 601 Potential confounders identified in the literature 10 and included in the study were age, sex (i.e., male Excluded n = 20 6 v. female or male-to-female transgender), ethnicity • Did not return for (white v. other), 19,20 recent (i.e., in the previous 6 foll ow- up intervi ew months) noninjection heroin use, 10 recent powder cocaine use (i.e., smoking and/or intranasal use), 10 Pa rti cipants drug injecti on –naive at recent crack cocaine smoking, 10,19,20 recent cannabis baseli ne wi th ≥ 1 fo llow-up intervi ew n = 395 use 19,20 and recent alcohol use. 20 First, we generated correlations between vari - ables and tested for significant correlations using Figure 1: Flow diagram of recruitment and retention 2 Pearson χ tests . We used Kaplan–Meier methods of participants in the At-Risk Youth Study (ARYS) to calculate the cumulative hazard of injection in- and inclusion of participants in the current study. 1570 CMAJ, December 10, 2013, 185(18) Research interview did not differ significantly from the than among those who did not report use of this study sample in terms of sex (142/206 [68.9%] v.