Answer all questions in complete sentences and use proper APA citation when appropriate

Contents lists available at ScienceDirect

Addictive Behaviors

journal homepage: www.elsevier.com/locate/addictbeh

Short Communication

Increased use of heroin as an initiating opioid of abuse

Theodore J. Cicero⁎, Matthew S. Ellis, Zachary A. Kasper Washington University in St. Louis, Department of Psychiatry, Campus Box 8134, 660 S. Euclid Avenue, St. Louis, MO 63110, United States

H I G H L I G H T S

• Heroin as the first opioid of abuse has grown significantly in the past decade. • Heroin as an initiating opioid now exceeds hydrocodone and oxycodone. • Such increases among inexperienced opioid users could lead to increased risk of overdose.

A R T I C L E I N F O

Keywords: Heroin Heroin overdose Prescription opioid abuse Opioid abuse Opioid initiation

A B S T R A C T

Introduction: Given the relatively recent growth in access to heroin and a more permissive atmosphere sur- rounding its use, we hypothesized that an increasing number of persons with limited experience and tolerance to opioids would experiment with heroin as their first opioid rather than more common prescription opioid an- algesics. Methods: Individuals entering substance abuse treatment for an opioid use disorder in the period 2010–2016 (N = 5885) were asked about the specific opioid they first regularly used to get high. To limit long-term recall and survival bias, analyses was restricted to opioid initiation that occurred in the past ten years (2005–2015). Results: In 2005, only 8.7% of opioid initiators started with heroin, but this sharply increased to 33.3% (p < 0.001) in 2015, with no evidence of stabilization. The use of commonly prescribed opioids, oxycodone and hydrocodone, dropped from 42.4% and 42.3% of opioid initiators, respectively, to 24.1% and 27.8% in 2015, such that heroin as an initiating opioid was now more frequently endorsed than prescription opioid analgesics. Conclusions: Our data document that, as the most commonly prescribed opioids – hydrocodone and oxycodone – became less accessible due to supply-side interventions, the use of heroin as an initiating opioid has grown at an alarming rate. Given that opioid novices have limited tolerance to opioids, a slight imprecision in dosing in- herent in heroin use is likely to be an important factor contributing to the growth in heroin-related over dose fatalities in recent years.

1. Introduction

The United States has been burdened by an epidemic of opioid abuse and overdose deaths over the past two decades (Manchikanti, Fellows, Ailinani, & Pampati, 2010). It began in earnest in the 1990s with sudden increases in the number of prescribed opioids, in particular novel extended release opioids not adulterated with acetaminophen or NSAIDs (e.g., OxyContin®) (U.S. OxyContin abuse and diversion and efforts to address the problem [Report to Congressional Requesters, #GAO-04-110], 2003). As the diversion of prescription opioids in- creased, they were perceived by abusers to be safer due to their legality and readily apparent brand and dose specificity, which helped avoid accidental overdose (Daniulaityte, Falck, & Carlson, 2012). In an effort to address this growing problem, a number of “supply” reduction efforts

were launched by federal agencies and the pharmaceutical companies benefitting from the sale of these products. These include, but are not limited to, statewide prescription monitoring programs intended to discourage doctor shopping and “script doctors” (Brady et al., 2014), increased physician education on the appropriate use of opioids (Alford, 2016), and the development of abuse deterrent formulations of opioids which make it difficult to crush or solubilize tablets for insuf- flation or IV injection (Cicero & Ellis, 2015a).

Recent studies have shown that these supply-reduction efforts have been modestly successful in reducing the supply of prescribed opioids and subsequent diversion for non-therapeutic purposes (Dart et al., 2015a). Thus, those already dependent on prescription opioids were faced with a dilemma: find more money to buy harder to find and more expensive prescription opioids, or find a cheaper alternative. For many,

http://dx.doi.org/10.1016/j.addbeh.2017.05.030 Received 20 March 2017; Received in revised form 18 May 2017; Accepted 22 May 2017

⁎ Corresponding author. E-mail address: [email protected] (T.J. Cicero).

Addictive Behaviors 74 (2017) 63–66

Available online 23 May 2017 0306-4603/ © 2017 Published by Elsevier Ltd.

MARK

the solution was a transition to heroin, a popular alternative given its steadily lower price, making it more widely accessible and with a high comparable, if not stronger, than prescription opioids (Cicero, Ellis, Surratt, & Kurtz, 2014; Compton, Jones, & Baldwin, 2016). Thus, to meet increased demand, there has been a significant dealer-driven in- crease in supply. As a result of this increased supply of cheap, accessible heroin, we hypothesized a spillover effect, where an increasing number of persons inexperienced with opioids might begin to experiment with readily available heroin as their first opioid of abuse, rather than a less risky, but also less accessible prescription opioid. Given the imprecision in titrating doses and the potential for potent adulterants (e.g., fentanyl analogues), we anticipated that, should our hypothesis be supported, this emerging trend could very well be associated with an increase in heroin related overdose fatalities, particularly in novice opioid users who lack the degree of tolerance found in more experienced ones.

To determine whether there has been, in fact, an increase in the use of heroin as a first opioid, we analyzed data on opioid use initiation patterns (i.e., first opioid regularly used) using self-administered sur- veys in opioid-dependent patients (N = 5885) entering one of over 150 substance abuse treatment programs around the country from 2011 to 2016.

2. Methods

This report utilized data from the ongoing nationwide Survey of Key Informants' Patients (SKIP) Program, a key element of the Researched Abuse, Diversion and Addiction-Related Surveillance (RADARS®) System, a comprehensive series of programs that collect and analyze post-marketing data on the abuse and diversion of prescription opioid analgesics and heroin (Cicero et al., 2007; Dart et al., 2015b). The SKIP Program consists of a Key Informant network with annual participation of over 150 public and privately funded treatment centers, with a reasonable representativeness of the four census areas (Region [SKIP %, 2014 Census%]; Midwest [27.9%,21.2%], Northeast [15.5%,17.6%], South [33.6%,37.6%] and West [23.0%,23.6%]). Key Informants were asked to recruit clients (eighteen years and older) who were entering their substance abuse treatment program with a primary diagnosis of opioid use disorder, as defined by DSM-IV or V criteria, depending on the time of completion. Clients were asked to complete an anonymous paper survey centered on opioid abuse patterns and related behaviors, with an 85% response rate attained. The survey packet included a $20 Wal-Mart gift card and a self-addressed stamped envelope which, after completion, was used by the respondent to mail the survey (identified by a unique case number) directly to Washington University in St. Louis (WUSTL). All protocols were approved by the WUSTL Institutional Review Board.

In addition to demographics, SKIP respondents, analyzed from 2011 to 2016, were asked the specific opioid they first regularly used (i.e. 2+ times a month), categorized as ‘prescription opioid’ or ‘heroin’, and the age they began to regularly use opioids. The year regular use began was calculated (Year of survey completion − Age at survey completion + Age of first regular opioid use = Year of beginning regular opioid use), with the analyses restricted to those initiating use within the past ten years (2005–2015; N = 5885) to limit long-term recall and survival bias (no respondent who completed a survey in 2016 initiated opioid use in that same year), a time-period shown to have stable recall for opioid abuse (Shillington, Cottler, Mager, & Compton, 1995). Cochran- Armitage test for trend was used to determine significant changes in first opioids over time and differences in demographic characteristics between heroin and prescription opioid initiates were assessed using Pearson Chi-Square tests of comparison using IBM SPSS Statistics v24.

3. Results

3.1. First opioid

Fig. 1 shows the annual unadjusted proportion of total SKIP re- spondents who indicated initiating regular opioid use with prescription opioids (hydrocodone, oxycodone or other prescription opioids) or heroin from 2005 to 2015. Only 8.7% of opioid initiates who began regular use in 2005 started with heroin, but its use sharply increased thereafter to the point where in 2015, heroin as an initiating opioid was at its highest point, 33.3% (p < 0.001), with no evidence of stabili- zation. Hydrocodone and oxycodone were the most widely identified initiating opioids over the ten year period. While they were equally attractive for first time users in 2005 (42.4% and 42.3%, respectively), at levels in excess of heroin for several years (~10%), their rate of endorsements began to gradually decrease as initiation with heroin increased. By 2015, hydrocodone and oxycodone were at 24.1% (p < 0.001) and 27.8% (p = 0.13), respectively, below the 33.3% proportion of heroin – now the leading drug for new opioid initiates.

3.2. Demographic comparisons of recent heroin and prescription opioid initiates

As shown in Table 1, heroin initiates were compared to prescription opioid initiates across a number of demographic variables. While pre- scription opioid initiates were slightly older, had higher rates of college education, were slightly more likely to be white and tended to reside in more non-urban areas, all of these differences were very small (yet significant, given the sample size). Interestingly, there were no sig- nificant sex differences between the two groups.

4. Discussion

The rapid four-fold increase in the use of heroin by new initiates to opioid use from 2005 to 2015 is a striking finding with significant public health implications. Given the imprecision in estimating a cor- rect dose of heroin, largely due to uncertainty about its purity and potential adulterants (i.e., fentanyl analogues), the possibility of over- dose in opioid novices is considerable and seemingly inevitable given that they lack the ability to tolerate even small errors in calculating their initial dose of heroin. Obviously, we could not ascertain directly whether this was true in the users surveyed (i.e., with overdose death as the endpoint), but the possibility of overdose and death is a clear and present danger which must be addressed, particularly by agencies

Fig. 1. First opioid of regular use among opioid initiates from 2005 to 2015 (N = 5885). Cochran-Armitage trend tests showed significant changes for heroin (< .0.001), hydro- codone (< 0.001), other prescription opioids (< 0.001), but not oxycodone (p = 0.13).

T.J. Cicero et al. Addictive Behaviors 74 (2017) 63–66

64

charged with protecting the public health so that, as a country, we are much better prepared to deal with this aspect of the opioid problem than we were in recognizing the initial emergence of prescription opioid abuse in the 1990s and the recent transition to heroin.

While we lack the ability to determine the plausibility of our pos- tulate that those who initiate opioid use with heroin are more likely to overdose or die, it is noteworthy that the Centers for Disease Control has reported that the national overdose deaths attributable to heroin rose from 2009 deaths in 2005 to 12,989 in 2015 (National Institute on Drug Abuse, 2017; Rudd et al., 2016), the same time period analyzed in this study. While there are no data in the CDC databases to link those deaths with prior drug histories, the two variables – heroin as the first opioid of use and fatal overdose – seem to be highly correlated (See Supplemental Fig. 1, Spearman correlation rs = 0.81). Obviously, ad- ditional, more directed studies will be needed to establish causality, but these data are certainly suggestive.

The fact that we did not find major differences in the demographics of those initiating opioid use with prescription opioids or heroin in- dicates that the two groups are not clearly distinguishable, which makes interventions and prevention more difficult. What is probably equally, or more, important is that heroin and/or prescription opioids are being interchangeably used in roughly the same population (Cicero & Ellis, 2015b). In this connection, it should be noted, that, in the long-term, there has been a significant shift in the demographic of heroin users over the last fifty years. In the 1960s–1970s, heroin use was primarily an inner-city, minority issue, whereas today, the “epi- demic” has shifted to include a larger proportion of white, suburban- rural residents (Cicero et al., 2014).

It is clear from a number of studies that supply reduction efforts – e.g., prescription monitoring programs and abuse deterrent drug for- mulations – have been effective in reducing the supply of prescription opioids on the black market (Dart et al., 2015a). However, it also seems clear that the reduction in the supply of prescription opioids has had both intended and unintended consequences. The number of opioids prescribed by physicians has declined in recent years (Pezalla et al., 2017) and, if one assumes that this reflects better awareness by phy- sicians of the appropriate use of opioids, and a reduction in the “pill mills” and doctor shopping, these programs have met positive ex- pectations. There is, however, one potential unanticipated effect - physicians may be more reluctant to prescribe opioids to treat pain patients due to widely publicized negative aspects of these drugs.

The clear message from the foregoing discussion is that supply re- duction efforts alone will not completely solve our current epidemic of heroin use and overdose deaths. History has taught us that as long as there is a demand, dealers and clever, motivated users will find a way to acquire their preferred types of drugs or introduce newly manufactured ones (e.g., illicit fentanyl analogues) that are not subject to the supply side limitations currently in place (Boettke, Coyne, & Hall, 2013). Ob- viously, we are not arguing against targeted supply reduction efforts – they clearly can be helpful – but we need to stress how incomplete they are unless we begin to address demand reduction efforts as well.

There are important limitations to our studies. Most significantly, ours is a retrospective study. While there are potential issues of recall in such studies, it would be nearly impossible to practically carry out a prospective study linking first opioid use with heroin overdose deaths as a dependent variable. In addition, the time lag between onset of use and treatment entry resulted in a decrease in Ns available for recent years, which could impact the generalizability of our findings. Beyond these practical limitations, it needs to be stressed that our results po- tentially underestimate the extent of harm to those who initiate regular opioid use with heroin – so called survival bias (i.e., those who over- dosed early into their heroin use would obviously not be included in these studies). Neither we nor the CDC database can compensate for the inherent study limitations and suggest, albeit indirectly, that the ad- verse consequences of early heroin adoption are worse than our data suggest.

Additional limitations include the fact that, since ours is a treat- ment-based sample, one could also argue that the results are not re- presentative of those who use opioids “recreationally.” Furthermore, differences in the factors influencing the decision to enter treatment, such as family/court pressures and financial ability, could limit the heterogeneity of our sample. These differences could also vary over the analyzed period, as well as between those initiating heroin versus prescription opioids. Future studies should examine, at a deeper level, the potential impact of these confounders on treatment-seeking beha- vior.

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.addbeh.2017.05.030.

Role of funding source

The national data were collected from a subset of participants from the Survey of Key Informants' Patients (SKIP) Program, a component of the RADARS® (Researched Misuse, Diversion and Addiction-Related Surveillance) System. The RADARS System is supported by subscrip- tions from pharmaceutical manufacturers for surveillance, research and reporting services. RADARS System is the property of Denver Health and Hospital Authority, a political subdivision of the State of Colorado. Denver Health retains exclusive ownership of all data, databases and systems. Subscribers do not participate in data collection or analysis, nor do they have access to the raw data. Dr. Cicero serves as a paid consultant on the Scientific Advisory Board of the RADARS® System. None of the authors have a direct financial, commercial or other re- lationship with any of the subscribers of the RADARS® System.

Contributors

The corresponding author oversaw the development, implementa- tion and management of the studies involved and takes responsibility for the integrity of the data and the accuracy of the data analysis, which was conducted by Ellis and Kasper, in conjunction with the corre- sponding author. Authors Cicero and Ellis developed and wrote the manuscript. All authors have reviewed and approved the manuscript.

Conflict of interest

Author Cicero serves as a consultant on the Scientific Advisory

Table 1 Characteristics of heroin vs. prescription opioid initiates, 2005–2015.

Initiate Cohort, No. (%)

Heroin (n = 631)

Prescription opioid (n = 5254)

Sig.a

Age at survey completion (SE)

27.0 (0.28) 28.9 (0.11) < 0.001b

Gender 0.82 Male 299 (47.8%) 2519 (48.3%) Female 327 (52.2%) 2701 (51.7%)

Ethnicity 0.01 White 479 (78.0%) 4262 (82.2%) Non-white 135 (22.0%) 922 (17.8%)

Urbanicity of residence 0.01 Urban 280 (51.6%) 2095 (46.1%) Suburban/rural 263 (48.4%) 2454 (53.9%)

Highest completed education

< 0.001b

Some college or more 204 (32.7%) 2141 (41.0%) Education lower than college

409 (65.5%) 2994 (57.3%)

None 11 (1.8%) 90 (1.7%)

a Independent samples t-test used for age, Pearson's chi-squared was used for all other variables.

b Significant at a level of p ≤ 0.01.

T.J. Cicero et al. Addictive Behaviors 74 (2017) 63–66

65

Board of the non-profit post-marketing surveillance system, RADARS®. Authors Ellis and Kasper have no conflicts of interest to report.

Acknowledgements

The national data were collected from a subset of participants from the Survey of Key Informants' Patients (SKIP) Program, a component of the RADARS® (Researched Misuse, Diversion and Addiction-Related Surveillance) System. The RADARS System is supported by subscrip- tions from pharmaceutical manufacturers for surveillance, research and reporting services. RADARS System is the property of Denver Health and Hospital Authority, a political subdivision of the State of Colorado. Denver Health retains exclusive ownership of all data, databases and systems. Subscribers do not participate in data collection or analysis, nor do they have access to the raw data.

References

Alford, D. P. (2016). Opioid prescribing for chronic pain—Achieving the right balance through education. The New England Journal of Medicine, 374(4), 301–303.

Boettke, P. J., Coyne, C. J., & Hall, A. R. (2013). Keep off the grass: The economics of prohibition and US drug policy. Oregon Law Review, 91, 1069–1096.

Brady, J. E., Wunsch, H., DiMaggio, C., Lang, B. H., Giglio, J., & Li, G. (2014). Prescription drug monitoring and dispensing of prescription opioids. Public Health Reports, 129(2), 139–147.

Cicero, T. J., & Ellis, M. S. (2015a). Abuse-deterrent formulations and the prescription opioid epidemic in the United States: Lessons learned from OxyContin. JAMA Psychiatry, 72(5), 424–430.

Cicero, T. J., & Ellis, M. S. (2015b). Shifting patterns of prescription opioid and heroin abuse in the United States. The New England Journal of Medicine, 373(18), 1789–1790.

Cicero, T. J., Dart, R. C., Indiardi, J. A., Woody, G. E., Schnoll, S., & Munoz, A. (2007). The development of a comprehensive risk-management program for prescription

opioid analgesics: Researched Abuse, Diversion and Addiction-Related Surveillance (RADARS®). Pain Medicine, 8, 157–170.

Cicero, T. J., Ellis, M. S., Surratt, H. L., & Kurtz, S. P. (2014, Jul 1). The changing face of heroin use in the United States: A retrospective analysis of the past 50 years. JAMA Psychiatry, 71(7), 821–826.

Compton, W. M., Jones, C. M., & Baldwin, G. T. (2016). Relationship between nonmedical prescription-opioid use and heroin use. The New England Journal of Medicine, 374, 154–163.

Daniulaityte, R. R., Falck, R., & Carlson, R. G. (2012). “I'm not afraid of those ones just 'cause they've been prescribed”: Perceptions of risk among illicit users of pharma- ceutical opioids. The International Journal on Drug Policy, 23(5), 374–384.

Dart, R. C., Surratt, H. L., Cicero, T. J., Parrino, M. W., Severtson, S. G., Bucher-Bartelson, B., & Green, J. L. (2015a). Trends in opioid analgesic abuse and mortality in the United States. The New England Journal of Medicine, 372, 241–248.

Dart, R. C., Surratt, H. L., Cicero, T. J., Parrino, M. W., Severtson, S. G., Bucher-Bartelson, B., & Green, J. L. (2015b). Supplementary appendix: Trends in opioid analgesic abuse and mortality in the United States. The New England Journal of Medicine, 372, 241–248.

Manchikanti, L., Fellows, B., Ailinani, H., & Pampati, V. (2010). Therapuetic use, abuse and nonmedical use of opioids: A ten-year perspective. Pain Physician, 13(5), 401–435.

National Institute on Drug Abuse (2017). National overdose deaths from select pre- scription and illicit drugs, Data sourced from the Centers for Disease Control accessed on January 9, 2017. https://www.drugabuse.gov/related-topics/trends-statistics/ overdose-death-rates.

Pezalla, E. J., Rosen, D., Erensen, J. G., Haddox, J. D., & Mayne, T. J. (2017). Secular trends in opioid prescribing in the USA. Journal of Pain Research, 10, 383–387.

Rudd, R. A., Seth, P., David, F., & Scholl, L. (2016, Dec 30). Increases in drug and opioid- involved overdose deaths – United States, 2010-2015. MMWR. Morbidity and Mortality Weekly Report, 65(5051), 1445–1452.

Shillington, A. M., Cottler, L. B., Mager, D. E., & Compton, W. M. (1995). Self-report stability for substance use over 10 years: Data from the St. Louis Epidemiologic Catchment Study. Drug and Alcohol Dependence, 40, 103–109.

U.S. OxyContin abuse and diversion and efforts to address the problem [Report to Congressional Requesters, #GAO-04-110]. Washington, DC: General Accounting Office.

T.J. Cicero et al. Addictive Behaviors 74 (2017) 63–66

66