Benzodiazepines and opioid co-use among rural people who use drugs: findings from the Rural Opioid Initiative

Abstract

Background:

Benzodiazepines and opioids are among the most frequently misused psychoactive substances, but their patterns of co-use (polysubstance use) in rural areas are unclear. As resources to address substance use are disproportionally scarce in rural areas, a better understanding of this polysubstance use is critical to allocate and direct interventions.

Methods:

The Rural Opioid Initiative comprises eight research cohorts spanning 10 states and 65 rural counties. Participants were recruited from 1/2018–3/2020 and eligibility included past 30-day opioid use by any route or past 30-day injection of any substance. Analyses were restricted to participants reporting past 30-day opioid use and either benzodiazepine or stimulant use. We described bivariate cross-sectional associations between benzodiazepines+opioid use, compared to stimulant+opioid use, and substance use behaviors, health outcomes, injection drug use, addiction treatment, and criminal legal system involvement.

Results:

Of the 1107 ROI participants that met inclusion criteria, 10% (n=107) reported benzodiazepines+opioid use, and 90% (n=1000) reported stimulant+opioid use. The benzodiazepines+opioid group, compared to the stimulant+opioid group, had higher use of opioid pain medication (73% vs 55%), gabapentin (43% vs 23%), and clonidine (12% vs 4%) to get high and used these substances more frequently; they also reported more frequent heavy episodic drinking (6.1 days per 30 days, SD=9.4 vs 4.1 days, SD 7.5). The benzodiazepine+opioid group reported a lower prevalence in the past 6 months of law enforcement stop-and-search incidents (29% vs 48%), arrests (11% vs 28%), probation (22% vs 34%), jail/prison (18% vs 41%), and fewer days in jail/prison (4.7, SD=19.1 days vs 15.9, SD=35.7 days).

Conclusion:

We found that benzodiazepines+opioids use was associated with more heavy episodic drinking and gabapentin use, and lower prevalence of criminal legal system involvement. These data suggest that individuals reporting benzodiazepines+opioids use have distinct behavioral patterns and outcomes that require targeted interventions for rural populations.

INTRODUCTION

Benzodiazepines are among the most commonly prescribed drug classes in the United States¹ (US), with ambulatory prescriptions doubling in the past decade². Commonly prescribed benzodiazepines include lorazepam, clonazepam, diazepam, and alprazolam, which are most commonly prescribed to treat anxiety and sleep disorders³. Benzodiazepine misuse has increased steadily over the past two decades⁴. In the US, 17% of people prescribed benzodiazepines self-reported misusing the drug over the last year⁵ and benzodiazepine-related overdose mortality increased nearly five-fold from 1996 to 2013⁶. The rise in benzodiazepine related overdose deaths suggests that the increases in prescribed benzodiazepines and the misuse are compliment to each other⁵. Benzodiazepines have gained more attention amid the opioid overdose epidemic⁷, especially since benzodiazepine misuse is strongly associated with opioid misuse or dependence⁵.

Benzodiazepines and opioids are among the most frequently misused psychoactive substances⁸. The prevalence of benzodiazepine misuse among people who use opioids is high, ranging from 51–70%⁸. There are many reasons for benzodiazepine and opioid co-use including: to enhance the high of opioids^9,10, ease the symptoms of withdrawal^10–13, supplement use when opioids are unavailable¹⁰, and treat comorbid anxiety disorders^14,15. People who are co-prescribed opioids and benzodiazepine are at increased risk of overdose and overdose deaths^16–18 and 93% of benzodiazepine-related deaths involve opioids¹⁹. Among people who use opioids, benzodiazepine use is associated with poorer treatment retention²⁰, more emergency department visits²¹, and increased risk of HIV²² and hepatitis C virus (HCV) infections²³. These adverse outcomes are further exacerbated by lack of resources and services found in rural areas of the US²⁴.

Rural areas of the US continue to be disproportionately impacted by the opioid crisis^22,23, and polysubstance use within rural areas is common^23,24. Of polysubstance use typologies among people who also use opioids, combined opioid and prescription drug misuse is the most common among rural individuals, and polysubstance use involving prescription drugs was twice as common in rural areas than in urban areas²². Polysubstance use, including the combination of benzodiazepine and opioids, is frequently associated with higher rates of alcohol dependency^25,26. This co-use of multiple substances often leads to a higher likelihood of criminal legal system involvement²⁷. Older adults in rural areas are also ~1.5 times as likely to receive a prescription for a benzodiazepine, compared to their counterparts in urban areas²⁸, which has been attributed, at least in part, to the lack of medical specialists (e.g., psychiatrists, geriatricians, etc.) in rural areas and the reliance on family physicians and general internists to treat mental health concerns highly urbanized areas²⁸. Additionally, rural residence and number of existing prescription medications taken were predictive of benzodiazepine use²⁶. Some studies illustrated that people in rural areas had more frequent overlapping benzodiazepine and opioid use compared to people in non-rural areas^27,28. These findings highlight the dangerous role benzodiazepine use has within rural US communities, especially in the context of the opioid overdose epidemic. However, the impacts of polysubstance use in the form of benzodiazepine and opioid co-use remain relatively unexplored²².

This study aims to address critical gaps in our understanding of benzodiazepine and opioid co-use among rural populations, a group that is underrepresented in substance use literature. While the co-use of opioid and benzodiazepine has been extensively studied in urban settings, much less is known regarding rural communities, where context and circumstance of substance use differ notably, as do barriers to treatment. In this study, we described the characteristics of people from multiple rural areas across the US who used both opioids and benzodiazepines and compared this to those using both opioids and stimulants. This study, to our knowledge, is the only one performed to date that focuses on benzodiazepines+opioid misuse in a geographically diverse cohort of rural individuals.

Methods

Study Design

This analysis utilized cross-sectional, multi-study data from the Rural Opioid Initiative (ROI, described previously²³). In brief, ROI is a two-phase, multi-cohort project designed to better understand the epidemiology of opioid and other drug use in rural settings within the US in order to tailor interventions to address opioid use in these communities²⁹. The first phase, from which the data used in this analysis is derived, focused on rapid epidemiological assessments and fostering local collaborations, while the second phase trialed locally tailored interventions informed by phase one collaborations and asssessments²⁹. ROI includes 8 research cohorts across 10 states (Illinois, Kentucky, New England [Massachusetts, New Hampshire, Vermont], North Carolina, Ohio, Oregon, Wisconsin, West Virginia) and 65 rural counties. A county’s rurality status was confirmed by the Health Resources and Services Administration (HRSA) “Am I Rural” website²⁹. Participants were recruited from January 2018 to March 2020.

Study Participants

Eligibility criteria into the ROI included a minimum age (≥18 for 6 studies and ≥15 for 2 studies), residence in one of the 65 participating rural counties, and reporting past 30-day use of any opioid (heroin, fentanyl, prescription pain medication) ‘to get high’ by any method and/or past 30-day injection of any drug. All studies obtained local institutional review board approval for research activities and data sharing between research cohorts within ROI. Participants provided either written or verbal consent prior to the survey administration.

We restricted our analyses to people who reported polysubstance use to account for potential confounding by polysubstance use, since people who use benzodiazepines have a higher prevalence of polysubstance use²⁹. Our exposure of interest was polysubstance use with benzodiazepines and opioids compared to polysubstance use with stimulants and opioids. Stimulant and opioid use was selected as the comparison group because stimulants are the most common substance type used other than opioids in the ROI cohort²³. Additionally, stimulants and benzodiazepines have opposing effects on the central nervous systems^30,31 and differing demographics and risk profiles^32–34, thus allowing us to better measure associations between these groups and our outcomes of interest. Participants could have reported the use of additional substances (e.g., benzodiazepine+opioid+clonidine), but were categorized based on either their benzodiazepine and opioid use or their stimulant and opioid use.

Benzodiazepines and opioid polysubstance use was defined by self-reported use of benzodiazepines, and not stimulants, to get high in the past 30 days plus self-reported use of opioids (heroin, fentanyl, opioid pain medication, synthetics, buprenorphine, or methadone) to get high in the past 30 days. Stimulant and opioid polysubstance use was defined by self-reported use of stimulants (either methamphetamine/crystal meth or crack/cocaine), and not benzodiazepines, to get high in the past 30 days plus self-reported use of opioids in the past 30 days.

Study Procedures

Respondent-driven sampling was used to recruit participants^30,31. Each study recruited seeds to initiate peer recruitment from syringe service programs, local health departments, community health centers, and other agencies or community outreach programs. Seeds were provided up to six coupons to recruit peers and received compensation for peer recruitment and study participation. Standardized surveys were administered to participants through audio computer-assisted self-interview³⁵ (A-CASI), computer-assisted personal interviews³⁶ (CAPI), or computer-assisted self-interview³⁷ (CASI). Surveys included questions about drug use and use networks; socioeconomic factors; previous use/injection of drugs to get high; severity of dependence; access to and utilization of health care and harm reduction programs; and awareness of sexually transmitted infections and blood-borne diseases, such as HIV and HCV infection, among other domains.

Measures

The primary exposure of interest was self-reported past 30-day use of benzodiazepines and opioids for getting high. The comparison group was participants that self-reported past 30-day use of stimulants and opioids for getting high.

To compare past 30-day combined benzodiazepine and opioid use to stimulant and opioid use among participants, we evaluated outcomes in five domains:

1. self-reported past 30-day substance use behaviors (drug of choice, substances used [including heroin, fentanyl, opioid pain medication, buprenorphine, methadone, gabapentin, and clonidine], cigarettes, and heavy episodic drinking [more than 4 drinks per day for females, or 5 drinks per day for male]), frequency of use;

2. self-reported substance use-related health outcomes (current possession of naloxone, previous testing for HCV, ever told they tested positive for HCV, previous HCV clearance, witnessed an overdose, number of people known who died of an overdose, personal history of an overdose, number of times overdosed);

3. self-reported past 30-day injection drug use (IDU) and syringe/paraphernalia behaviors (current IDU, frequency of IDU, receptive syringe sharing [use of a syringe or needle previously used by someone else], use of supplies [e.g., cottons, cookers, spoons, or water] previously used by someone else, syringe mediated drug sharing [i.e., injecting substances that somebody else prepared, mixed, or divided with a used syringe], and use of multiple injections per injection episode;

4. self-reported past-30 day addiction treatment attainment and past 6 month number of days in addiction treatment (buprenorphine maintenance medication, methadone maintenance, naltrexone injection, and buprenorphine maintenance injection); and

5. self-reported past 6 months criminal legal system involvement (law enforcement stop-and-search, arrested and booked, probation, parole, supervised release, or community supervision, jail or prison).

Statistical Analyses

Descriptive statistics were utilized to describe demographic characteristics of participants that used either benzodiazepines+opioids or stimulants+opioids in the past 30 days. We then performed bivariate analyses for benzodiazepines+opioids or stimulants+opioids and the outcomes of interest. We used Chi-square tests for binary variables and t-tests for continuous variables to estimate statistically significant differences in outcomes between groups (p-values ≥ 0.05 were considered statistically significant). Additionally, we conducted a sensitivity analysis including participants who reported past 30-day use of benzodiazepine, gabapentin, and/or clonidine+opioids or stimulants+opioids that replicated the methods utilized in the main analysis to estimate significant differences in outcomes between groups. This sensitivity analysis was conducted because gabapentin and clonidine are often co-used with opioids and benzodiazepine^38,39 and can be prescribed off-label to treat opioid withdrawal⁴⁰. Analyses were conducted using Stata version 17.0 (StataCorp, College Station, TX).

RESULTS

Among the 3048 ROI participants, 1107 were included in the study population; 107 (10%) reported having used benzodiazepines and opioids without stimulants, and 1000 (90%) reported having used stimulants and opioids without benzodiazepines (Table 1).

Table 1.

Demographic characteristics of Rural Opioid Initiative Participants with self-reported past 30-day benzodiazepine/opioid and past 30-day opioid/stimulant use (N = 1107)

	Past 30-day substance use
	Benzodiazepine1 and Opioid2,3	Stimulants4 and Opioid2,5
	N (%)	N (%)
	107 (10)	1000 (90)	p-value
Age, mean (SD)	37.0 (11.4)	36.4 (10.3)	0.5
Female	55 (51)	401 (40)	0.02
Race/Ethnicity			0.1
Non-Hispanic White	90 (84)	833 (83)
Non-Hispanic Black	6 (6)	23 (2)
Non-Hispanic American Indian	3 (3)	76 (8)
Non-Hispanic Other/Unknown	5 (5)	33 (3)
Hispanic6	3 (3)	35 (4)
Education			0.6
Less than high school	27 (25)	210 (21)
High school or GED	47 (44)	499 (50)
Some college or technical school	31 (29)	268 (27)
Bachelor’s degree or above	2 (2)	21 (2)
Source(s) of income
Traditional7	73 (68)	654 (65)	0.6
Assistance8	35 (33)	217 (22)	0.01
Illegal9	20 (19)	280 (28)	0.03
Homelessness, past 6 months	37 (35)	520 (52)	<0.001
Geographic region			<0.001
Illinois	9 (8)	37 (4)
Kentucky	22 (21)	125 (3)
North Carolina	8 (7)	114 (11)
New England	30 (28)	199 (20)
Ohio	14 (13)	97 (10)
Oregon	1 (1)	85 (9)
Wisconsin	9 (8)	309 (31)
West Virginia	14 (13)	34 (4)

Abbreviations: SD, standard deviation.

¹Self-reported past 30 day prescription anxiety drugs – like Xanax, Valium, Klonopin, etc. – to get high

²Includes self-reported past 30-day heroin, fentanyl, opiate pain medication, synthetics (e.g., U47700, U4, or “Pink”), buprenorphine, and methadone used “to get high”

³Excludes self-reported past 30-day methamphetamine/crystal meth or crack/cocaine to get high

⁴Includes self-reported past 30-day methamphetamine/crystal meth or crack/cocaine to get high

⁵Excludes self-reported past 30 day prescription anxiety drugs – like Xanax, Valium, Klonopin, etc. – to get high

⁶Race/ethnicity are mutually exclusive categories.

⁷“Full-time work,” “Part-time work,” “Retirement check,” and/or “Someone supports me.”

⁸“Public assistance check” and/or “Disability check.”

⁹“Selling drugs,” “Selling sex,” and/or “Theft, shoplifting, or stealing.”

Participants in the benzodiazepine+opioid group had a mean age of 37.0 (standard deviation [SD] 11.4) and were 84% Non-Hispanic White, participants in the stimulant+opioid group had a mean age of 36.4 (10.3) and 83% were Non-Hispanic White, indicating no meaningful difference in age and racial/ethnic makeup between the two groups (Table 1). However, more participants in the benzodiazepine+opioid group were female, 51% compared to 40% of participants in the stimulant+opioid group.

There was a higher prevalence of government assistance as the main form of income for the benzodiazepine+opioid group, compared to the stimulant+opioids group, (33% vs 22%) and a lower prevalence of income by illegal means (19% vs 28%). The benzodiazepine+opioid group had a lower prevalence of experiencing homelessness in the last six months (35% vs 52%).

The benzodiazepine+opioid group more often reported opioid pain medications as their preferred substance, compared to the simulant+opioid group (37% and 7%, respectively) (Table 2). Participants in the benzodiazepine+opioid group had a lower prevalence of heroin use (70% vs 81%) and used heroin less frequently compared to the stimulant+opioid group, whereas participants in the benzodiazepine+opioid group more often reported use of opioid pain medication (73% vs 55%), gabapentin (43% vs 23%), and clonidine (12% vs 4%) to get high and used these substances more frequently. We found no difference in the prevalence of heavy episodic drinking between the benzodiazepine+opioid and stimulant+opioid groups (46% vs 49%, respectively), however we found participants in the benzodiazepine+opioid group reported more frequent heavy episodic drinking in the past 30-day (6.1 days, SD=9.4 vs 4.1 days, SD 7.5)

Table 2.

Bivariate associations of Rural Opioid Initiative Participants with self-reported past 30-day benzodiazepine/opioid and past 30-day opioid/stimulant use by outcomes of interest.

	Past 30-day substance use
	Benzodiazepine1 and Opioid2,3	Stimulants4 and Opioid2,5
	N (%)	N (%)	p-value
Substance Use Behaviors, past 30-days	107 (10)	1000 (42)
Drug of choice for getting high			<0.001
Heroin	39 (36)	431 (43)
Fentanyl/cartfentanil	4 (4)	14 (1)
Opioid pain medication	40 (37)	73 (7)
Synthetic opioids	0 (0)	7 (1)
Buprenorphine	6 (6)	31 (3)
Methadone	3 (3)	8 (1)
Prescription anxiety medication	6 (6)	0 (0)
Cocaine/crack	2 (2)	76 (8)
Methamphetamine/crystal meth	4 (4)	334 (33)
Gabapentin or Neurontin	2 (2)	2 (<1)
Clonidine	1 (1)	1 (<1)
Other	0 (0)	21 (2)
Number of substances used	2.6 (0.6)	2.6 (0.7)	0.9
Heroin	75 (70)	813 (81)	0.005
Number of days used, mean (SD)	11.3 (11.7)	12.5 (12.0)	0.3
Fentanyl	37 (35)	374 (37)	0.9
Number of days used, mean (SD)	4.3 (8.1)	4.7 (8.8)	0.7
Opioid Pain Medication	78 (73)	551 (55)	<0.001
Number of days used, mean (SD)	8.6 (8.7)	4.4 (7.2)	<0.001
Buprenorphine6	47 (44)	384 (38)	0.4
Number of days used, mean (SD)	5.6 (9.3)	4.6 (8.9)	0.2
Methadone6	23 (22)	189 (19)	0.6
Number of days used, mean (SD)	1.7 (5.5)	2.1 (6.6)	0.6
Gabapentin	46 (43)	234 (23)	<0.001
Number of days used, mean (SD)	5.2 (9.3)	1.9 (5.3)	<0.001
Clonidine	13 (12)	39 (4)	<0.001
Number of days used, mean (SD)	1.4 (5.4)	0.3 (2.5)	0.0006
Cigarettes	87 (81)	909 (91)	0.003
Number smoked in a day, mean (SD)	14.4 (1.1)	15.5 (11.3)	0.4
Heavy Episodic Drinking7	52 (49)	455 (46)	0.6
Number of days used, mean (SD)	6.1 (9.4)	4.1 (7.5)	<0.01
Health Outcomes
Current possession of naloxone	37 (35)	383 (38)	0.4
Previous testing for HCV	75 (71)	745 (75)	0.4
Ever tested positive for HCV	34 (32)	351 (35)	0.3
Ever cleared HCV8	11 (32)	99 (28)	0.7
Witness overdose	80 (75)	753 (75)	0.7
Number of people know died of overdose, mean (SD)	3.5 (4.4)	3.8 (6.7)	0.6
Personal history of overdose	45 (42)	476 (48)	0.4
Number of times overdosed, mean (SD)	4.5 (24.9)	2.9 (18.7)	0.4
Injection Drug Use and Syringe/Paraphernalia Behaviors, past 30 days
Current injection drug use	61 (57)	861 (86)	<0.001
Daily or more injection drug use frequency9	30 (49)	599 (70)	0.001
Receptive syringe sharing9	21 (34)	345 (40)	0.2
Number of days, mean (SD)9	3.1 (9.1)	6.3 (21.7)	0.3
Used supplies that was used by someone else9	25 (41)	410 (48)	0.2
Number of days, mean (SD)9	5.5 (10.6)	8.9 (22.6)	0.2
Practiced syringe mediated drug sharing9	26 (43)	403 (47)	0.4
Number of days, mean (SD)9	3.7 (8.2)	7.3 (16.9)	0.1
Practiced multiple injection per injection episode9	47 (77)	599 (70)	0.5
Number of days, mean (SD)9	8.4 (11.1)	13.8 (26.3)	0.1
Criminal Justice History, past 6 months
Law enforcement stop and search	31 (29)	480 (48)	<0.001
Number of times, mean (SD)	1.2 (3.9)	4.2 (14.8)	0.04
Arrested and booked	12 (11)	281 (28)	<0.001
Number of times, mean (SD)	0.7 (4.3)	1.9 (9.1)	0.2
Probation, parole, supervised release, or community supervision	23 (22)	336 (34)	0.008
Jail or prison	19 (18)	410 (41)	<0.001
Number of days, mean (SD)	4.7 (19)	15.9 (35.7)	0.002

Abbreviations: SD, standard deviation; HCV, Hepatitis-C virus.

¹Self-reported past 30 day prescription anxiety drugs – like Xanax, Valium, Klonopin, etc. – to get high

²Includes self-reported past 30-day heroin, fentanyl, opiate pain medication, synthetics (e.g., U47700, U4, or “Pink”), buprenorphine, and methadone used “to get high”

³Excludes self-reported past 30-day methamphetamine/crystal meth or crack/cocaine to get high

⁴Includes self-reported past 30-day methamphetamine/crystal meth or crack/cocaine to get high

⁵Excludes self-reported past 30 day prescription anxiety drugs – like Xanax, Valium, Klonopin, etc. – to get high

⁶Buprenorphine and/or methadone used “to get high.”

⁷Defined as having 4 (females)/5 (males) or more drinks in a day.

⁸Among participants reporting having ever being told they tested positive for HCV.

⁹Among participants reporting current injection drug use.

Additionally, participants in the benzodiazepine+opioid group, compared to the stimulant+opioid group, reported current injection drug use less (57% vs 86%) and reported less use of injecting substances daily or multiple times per day (49% vs 70%). However, we did not observe significant differences in receptive syringe sharing, sharing of drug supplies, syringe-mediated drug sharing, or multiple injections per injection episode (Table 2).

Participants in the benzodiazepine+opioid group, compared to the stimulant+opioid group, were less likely to have experienced stop-and-search by law enforcement in the past 6 months (29% vs 48%) and experienced it less often, 1.2 days (SD=3.9) compared to 4.2 days (SD=14.8). In addition, fewer benzodiazepine+opioid participants experienced an arrest in the past 6 months compared to the stimulant+opioid participants (11% vs 28%). Fewer participants in the benzodiazepine+opioid group reported in the past 6 months they were under probation, parole, supervised release, or community supervision (22% vs 34%); spent any time in jail or prison (18% vs 41%) and spent fewer days in jail or prison (4.7, SD=19.1 days vs 15.9, SD=35.7 days) (Table 2).

We found no significant between-groups differences pertaining to substance use-related health outcomes or addiction treatment outcomes.

The sensitivity analysis that included participants who reported past 30-day use of benzodiazepine, gabapentin and/or clonidine+opioids, compared to participants who reported past 30-day use of stimulants+opioids, had no meaningful differences when compared to the main analysis (Supplemental Table 1 & Supplement Table 2).

Discussion

Our findings suggest that benzodiazepine+opioid use differs in substance use behaviors, including higher use of opioid pain medication, gabapentin, and clonidine, as well as more frequent heavy episodic drinking occurrences, compared to stimulant+opioid use. The benzodiazepines+opioid group had a lower prevalence of criminal legal system involvement, such as stop-and-search incidents, arrests, probation, and jail/prison. These findings underscore the unique challenges faced by rural people who use drugs (PWUD) amidst the opioid overdose epidemic and emphasize the need for targeted intervention to address the specific risk associated with benzodiazepine and opioid co-use.

We observed elevated use of gabapentin to get high among the benzodiazepine+opioid group. Since its release, prescribers have generally presumed that gabapentin has ‘no abuse potential’⁴⁵, and gabapentinoids are not classified as controlled substances by the federal government⁴⁶. Only seven states consider gabapentin a schedule V controlled substance, and only 12 have mandated gabapentin reporting⁴⁶. While gabapentin misuse has been better documented in the last decade, prescription rates have continued to rise, including for off-label use ^45,47. Among people with opioid use disorder, 15–22% also misused gabapentin⁴⁵, which is associated with a 60% higher odds of opioid-related death⁴⁸. Co-prescription of opioids, benzodiazepines, and gabapentin is common, despite the high overdose death rates associated with their combined use^49,50. Particularly concerning is recent research showing that gabapentin use can attenuate naloxone reversal of heroin-induced respiratory depression, which has the potential to greatly increase the risk of overdose-related mortality among people who use both opioids and gabapentin⁵¹. Regardless, combined use of opioids, benzodiazepines, and gabapentin may have grave consequences among rural populations, particularly if myths about its ‘non-abuse potential’ continue to be propagated.

While nearly half of both groups affirmed past 30-day heavy episodic drinking (46% and 49%, respectively), we found that the frequency of past-30 day heavy episodic drinking was higher in the benzodiazepine+opioid group than the stimulant+opioid group (6.1 days vs 4.1 days, respectively) (Table 2). The simultaneous use of benzodiazepines and alcohol is common, with one study reporting that 61% of individuals who reported past-year benzodiazepine misuse also used alcohol at that time¹⁴. Heavy episodic drinking poses elevated dangers for people who use benzodiazepines and opioids due to the synergistic effects between all three drugs^8,41,42: combined use of benzodiazepines and alcohol has previously been associated with increased rates of overdose and death^41,43, and in 2017, alcohol was involved in more than 25% of benzodiazepine-related U.S. emergency department visits⁴⁴. Fatal poisonings involving opioids are also frequently associated with alcohol use due to the significant respiratory depression they can induce alone and in combination⁴⁵; and heavy alcohol consumption can interfere with treatment for opioid use⁴⁶.

Given that all three drug classes have depressant effects, motivations for simultaneous use may be similar: misuse of benzodiazepines with alcohol and/or opioids is common due to enhanced feelings of relaxation and euphoria, as is combining benzodiazepines/alcohol to cope with anxiety, depression, and insomnia¹⁴. The relationship between benzodiazepines and alcohol use is further complicated by their clinical use: benzodiazepines are commonly prescribed for alcohol withdrawal⁴⁷. Following, patients with unhealthy alcohol use are more likely to be prescribed benzodiazepines⁴⁸. Taken together, these findings indicate a possibility for implementing better screening practices for alcohol use and heavy episodic drinking among people receiving benzodiazepines prescriptions^43,49.

In our study, we observed benzodiazepine+opioid use was associated with a lower prevalence of criminal legal system involvement. This contrasts with previous studies suggesting that benzodiazepine misuse is associated with higher criminal legal system involvement⁵⁰. This discrepancy could be because people who use stimulants—our comparison group—are more likely to have criminal legal system involvement^51–53. A recent study among the ROI cohort showed that 45% of individuals who used methamphetamine in the last 30 days had been in jail or prison in the last 6 months, compared to 28% of individuals who had not used methamphetamine⁵⁴. Methamphetamine use has also been shown to increase the risk of recidivism after first interaction with the criminal legal system⁵¹, reflecting the complex relationship between substance use behaviors and legal consequences. Moreover, the social stigma attached to IDU, which is a common route of admission amongst people who use methamphetamine, may contribute to higher scrutiny and harsher legal outcomes within the criminal legal system^7,55. The observed lower prevalence of criminal legal system involvement in the past 6 months among people who use benzodiazepine+opioids in our study prompts further investigation into the social and contextual factors influencing legal outcomes among PWUD within rural communities.

Additionally, fewer participants in the benzodiazepine+opioid group reported experiencing homelessness in the past six months compared to the stimulant+opioid group, (35% vs 52%, respectively) (Table 1); however, both prevalences are high and likely reflect the stark relationship between substance use and homelessness generally. People experiencing homelessness are particularly impacted by substance use, and overdose is the leading cause of death among people experiencing homelessness^56,57. Further, opioid use disorder is the most common substance use disorder among people experiencing homelessness⁵⁸, however, methamphetamine is frequently used⁵⁹. Conversely, people experiencing homelessness are less likely to receive a benzodiazepine prescription than currently housed individuals, although those who do receive a prescription are more likely to be given a high risk dose⁶⁰. These findings suggest that benzodiazepine or stimulant use can affect the relationship between opioid use and homelessness. Similarly, closely related to the patterns we see in homelessness among our population, more of the benzodiazepine+opioid group received assisted income, which may reflect better connection to social services as well as a higher overall income compared to people who use stimulants+opioids⁶¹ (Table 2).

Our analysis compared two groups engaged in polysubstance use to ensure differences observed were not attributable to polysubstance use alone, which is associated with more severe substance use behaviors and outcomes, including disproportionate criminal legal system involvement^27,62,63 and higher mortality^64–67. Further by having past 30-day opioid use as the additional substance used, we ensure that our exposure of interest and comparison group is similar in both groups practice polysubstance use and both groups have used opioids in the past 30 days. This allows us to compare two polysubstance groups and describe the outcomes related to participants’ benzodiazepine use, rather than describing the outcomes that are being affected by increased likelihood of polysubstance use. Within our study, we observed no association between the number of substances used, substance use related health outcomes, or accessing addiction treatment characteristics between the two groups, indicating that polysubstance use behaviors and related outcomes were similar between the two groups.

In the overall ROI cohort, 47% of participants reported past 30-day benzodiazepine use²⁹, whereas only 10% did so in our study population. This discrepancy reflects our study design, which examines polysubstance use among people who use opioids. To maintain distinct comparison groups (benzodiazepine+opioid and stimulant+opioid), we excluded participants who used benzodiazepine with substances other than opioids or stimulants, as well as those who use all three substances.

Limitations & Strengths

This study has limitations. The cross-sectional design precludes inferences about causality and directionality. The phrasing of the benzodiazepine question (‘Have you ever used prescription anxiety drugs – like Xanax, Valium, Klonopin, etc. – to get high?’) names benzodiazepines drugs, but does not specifically use the term benzodiazepine. Similarly, this study lacked specificity as to which benzodiazepines each participant used, and each drug may have different demographic and use profiles. While RDS was used as a sampling frame for recruiting the participants, the short chains and high number of seeds in most sites might have introduced selection bias^31,68. Additionally, reliance on self-reported substance use responses introduces the possibility of recall bias or social desirability bias, however, is likely less biased and more complete ascertainment of substance use than interview-based approached⁶⁹. Lastly, data collection ended just as the COVID-19 pandemic was beginning, which we now know impacted substance use patterns across the US (e.g., increases in substance use during COVID have been reported)⁷⁰. Despite these limitations, these findings are strengthened by the large sample size and geographic diversity of the participants, which enhances the robustness and representation of rural PWUD. Additionally, this research addresses a critically important and understudied question of benzodiazepine and opioid co-use with an equally understudied population, namely, rural PWUD. This emphasis is particularly significant given the unique challenges faced by rural communities amid the opioid overdose epidemic, such as limited access to healthcare^71,72 and harm reduction services⁷³.

Conclusion

In conclusion, our analysis among rural people who used benzodiazepines+opioids provided new insights into the unique characteristics, health behaviors, and areas of need of this population. Several of the characteristics of the ROI cohort, particularly pertaining to non-benzodiazepine substance use behaviors, have not been previously documented in rural populations, and speak to areas of need to better understand benzodiazepine+opioid polysubstance use in rural areas. One notable finding was elevated heavy episodic drinking among people who use benzodiazepines+opioids, inviting questions about synergies between benzodiazepines and alcohol, overlap between benzodiazepines and alcohol misuse, and the use of benzodiazepines to relieve alcohol withdrawal. Meanwhile, the findings about gabapentin use among people who use benzodiazepines+opioids add to a growing body of literature documenting that gabapentin has abuse potential. We also illustrate how different polysubstance use typologies are associated with social outcomes including criminal legal system involvement and homelessness. More generally, our results bring attention to an understudied area of concern within the drug epidemic among these underserved communities.

Public health practitioners and healthcare providers can use these findings to implement targeted screening for opioid and benzodiazepine co-use and develop interventions, such as integrated care models addressing polysubstance use and mental health comorbidities, which is of particular importance due to the rising opioid overdose crisis and growing recognition of the dangers of polypharmacy. Additionally, public health professionals should prioritize advancing harm reduction services, specifically naloxone distribution and emphasizing safe prescribing practices for rural health care providers. Finally, these results elucidate the need for tailored interventions that address the specific needs of rural PWUD to reduce the risks associated with benzodiazepine and opioid co-use. Overall, our results point to the need for more awareness of the dangers of benzodiazepine+opioid misuse, particularly in rural areas. This descriptive analysis brings the opportunity to generate hypothesis-driven studies and interventions to address the dangers that benzodiazepines and opioid use pose to rural communities.