The NHBS system was established in 2003 by the Centers for Disease Control and Prevention (CDC) to monitor risk behaviors among three populations at highest risk for HIV infection in the United States: gay, bisexual, and other men who have sex with men, persons who inject drugs (PWID), and heterosexuals at increased risk for HIV. NHBS involves rotating 12-month cycles of surveillance activities in these three populations. Surveillance activities include ethnographic formative research, an in-depth behavioral survey, and HIV testing during each cycle. The analyses presented in the current short communication are from the 2015 NHBS cycle among PWID in Denver, Colorado.

Between June and December 2015, participants were recruited using respondent-driven sampling (RDS), a peer-referral sampling methodology.¹¹ Each eligible participant was allowed to refer up to five persons from their network. Participants were instructed to recruit someone they knew who injects drugs and who they had seen in the past 30 days. Persons were eligible to participate if they were at least 18 years or older, had injected drugs during the preceding 12 months, resided in the Denver metropolitan statistical area, had not previously participated in the 2015 NHBS cycle, and were able to provide informed consent. Additional eligibility criteria included having physical evidence of recent injection (fresh track marks) or having current knowledge of drug preparation and injection technique and syringe description.

Verbal informed consent was obtained from eligible participants. Participants completed a standardized interviewer-administered behavioral risk survey using handheld computers.

The behavioral risk survey includes questions about sexual behaviors, injection behaviors, substance use, and HIV testing. In addition to the behavioral risk survey, which is the core NHBS survey and administered nationally across all sites, participants completed a shorter local questionnaire. Local questionnaires are developed during the formative research phase of NHBS, which includes interviews with key informants and focus groups to identify salient issue of concern or interest for the target population. The following question was included in the local questionnaire to capture information on nonmedical use of prescription opioids prior to injection: “Think back to the very first time that you injected any kind of drug. Were you hooked on painkillers, such as Oxycontin®, Vicodin®, morphine, or Percocet®, before you injected drugs for the very first time?” Similar to a group in Seattle, we used the term “hooked on” to avoid the word “dependence” which can be interpreted in different ways by non-clinicians. ¹²

Participants were also offered a rapid HIV test in addition to a rapid test for Hepatitis C Virus (HCV). Participants were given $35 for completing the survey, $35 for HIV and HCV testing, and $20 for each eligible person they recruited. All NHBS activities were voluntary and no names were collected. The study protocol was reviewed and approved by the Colorado Multiple Institutional Review Board.

Chi-square (Χ²) statistics were calculated to assess differences based on being hooked on prescription opioids prior to injection drug use. Generalized linear models with log link and binomial error distribution were used to estimate prevalence ratios (PR) and associated 95% confidence intervals (CI) for the association between prior nonmedical use of prescription opioids and overdose. All analyses were conducted using StataSE Version 12 (StataCorp, College Station, TX).

A total of 712 persons were screened for the 2015 cycle of NHBS in Denver, of whom 603 (84.7%) were eligible to participate and 599 (84.1%) completed the local questionnaire. The final sample included 19 (3.2%) seed participants, all of whom were 29 years or younger. Younger seeds were purposefully recruited in an attempt to generate a sample which more accurately represents the PWID population. Approximately three quarters (70.4%) of participants were male and more than half (58.2%) identified as non-Hispanic white (Table 1). Nearly one-quarter (22.0%) were 29 years old or younger. A large proportion of participants reported experiencing homelessness with 66.4% reporting being currently homeless and 15.1% reporting being homeless, but not in the past 12 months.

Overall, 192 (32.1%) of participants reported being hooked on painkillers, such as Oxycontin, Vicodin, morphine, or Percocet, before they injected drugs for the very first time (Table 1). Compared to participants who did not report being hooked on prescription opioids before they injected, those who were hooked on prescription opioids were significantly more likely to be younger (under age 29), to be new injectors (i.e. less than 1 years since their first injection), to have injected heroin as their first drug, and to report heroin as the drug most frequently injected. Though those who reported being hooked on prescription opioids were more likely to be under the age of 29, they were also more likely to report a slightly older age of first injection with the mean age at first injection being 24.6 (95% CI: 23.5, 25.7) compared to a mean age of 22.4 (95% CI: 21.5, 23.2) for those who were not hooked on prescription opioids prior to their first injection.

Participants were asked about the number of times they had overdosed in the past 12 months. A definition of overdose was not included in the question. Overall, 71 (12.0%) reported overdosing once in the past 12 months, 26 (4.3%) reported overdosing twice, and 31 (5.2%) reported overdosing 3 or more times in the past 12 months (Table 2). Reporting being hooked on prescription opioids prior to injection drug use was associated with a 1.55 (95% CI: 1.14, 2.10) fold increase in the risk of at least one overdose in the past 12 months. To control for factors that could potentially confound this association, we included variables that were associated with being hooked on prescription opioids prior to injection that might also be associated with the risk of overdose in a multivariate model. The association between being hooked and overdose was attenuated when we adjusted for age, years since first injection, non-injection use of heroin or prescription opioids, use of benzodiazepines, and participation in a treatment program in the past 12 months (prevalence ratio 1.36, 95% CI: 0.97–1.91). On the other hand, participants who reported being hooked on prescription opioids prior to injection drug use were also more likely to report obtaining a take-home naloxone kit in the past 12 months (44.3% vs 26.5%, p<0.001).

Participants who reported being hooked on prescription opioids prior to injection drug use were more likely to report more frequent (i.e. more than once a day) use of benzodiazepines, non-injection prescription opioids, and non-injection heroin, 13.6%, 13.1%, and 15.7% compared to those who were not hooked, 4.7%, 5.4%, and 7.7%, respectively (Table 2). Participants who reported being hooked on prescription opioids before injecting were also slightly more likely to report injecting one or more times a day compared to those who did not report being hooked (84.3% vs. 75%, p=0.024).