The impact of methamphetamine/amphetamine use on receipt and outcomes of medications for opioid use disorder: a systematic review

Background Methamphetamine/amphetamine use has sharply increased among people with opioid use disorder (OUD). It is therefore important to understand whether and how use of these substances may impact receipt of, and outcomes associated with, medications for OUD (MOUD). This systematic review identified studies that examined associations between methamphetamine/amphetamine use or use disorder and 3 classes of outcomes: (1) receipt of MOUD, (2) retention in MOUD, and (3) opioid abstinence during MOUD. Methods We searched 3 databases (PubMed/MEDLINE, PsycINFO, CINAHL Complete) from 1/1/2000 to 7/28/2020 using key words and subject headings, and hand-searched reference lists of included articles. English-language studies of people with documented OUD/opioid use that reported a quantitative association between methamphetamine/amphetamine use or use disorder and an outcome of interest were included. Study data were extracted using a standardized template, and risk of bias was assessed for each study. Screening, inclusion, data extraction and bias assessment were conducted independently by 2 authors. Study characteristics and findings were summarized for each class of outcomes. Results Thirty-nine studies met inclusion criteria. Studies generally found that methamphetamine/amphetamine use or use disorder was negatively associated with receiving methadone and buprenorphine; 2 studies suggested positive associations with receiving naltrexone. Studies generally found negative associations with retention; most studies finding no association had small samples, and these studies tended to examine shorter retention timeframes and describe provision of adjunctive services to address substance use. Studies generally found negative associations with opioid abstinence during treatment among patients receiving methadone or sustained-release naltrexone implants, though observed associations may have been confounded by other polysubstance use. Most studies examining opioid abstinence during other types of MOUD treatment had small samples. Conclusions Overall, existing research suggests people who use methamphetamine/amphetamines may have lower receipt of MOUD, retention in MOUD, and opioid abstinence during MOUD. Future research should examine how specific policies and treatment models impact MOUD outcomes for these patients, and seek to understand the perspectives of MOUD providers and people who use both opioids and methamphetamine/amphetamines. Efforts to improve MOUD care and overdose prevention strategies are needed for this population.


Introduction
Over 1.6 million people in the United States have opioid use disorder (OUD) [1]. Almost 50,000 people in the United States died of opioid overdose in 2019 [2], and overdose death has markedly increased during the COVID-19 pandemic [3][4][5][6][7]. Worldwide, OUD is one of the most prevalent drug use disorders and a notable source of global mortality and morbidity [8]. There are 3 US Food and Drug Administration (FDA)-approved treatment medications for OUD (MOUD), including methadone, buprenorphine and naltrexone [9]. Opioid agonist medications (methadone and buprenorphine) reduce risk of opioid overdose [10][11][12], and overdose risk has been observed to increase when patients exit agonist treatment demonstrating the importance of retention in treatment [10,12]. MOUD are considerably underused, and increasing access to and retention in MOUD treatment, particularly opioid agonist medications, is essential to addressing the opioid crisis and preventing overdose death [9]. In light of this goal, MOUD are increasingly being provided outside of specialty substance use treatment settings including in primary care [13,14] and community settings such as syringe services programs (SSPs) [15].
Multiple sources of data suggest that methamphetamine use is increasing among people with OUD. In the United States, a sharp increase in reported methamphetamine use has been documented among people entering OUD treatment-a nationwide survey found an 85% increase in prevalence between 2011 and 2018 [16,17], and an analysis of the national Treatment Episode Data Set found a 490% increase in prevalence from 2008 to 2017 [18]. An analysis of National Survey on Drug Use and Health data found that prevalence of recent illicit methamphetamine use more than tripled among people with recent heroin use or heroin use disorder from 2015 to 2017, and more than doubled among people with prescription OUD during the same period [19]. Amphetamine use is also growing globally-the United Nations reports that amphetamine seizures quadrupled worldwide from 2009 to 2018, and that methamphetamine/amphetamine use has increased across multiple regions [20].
Methamphetamine in particular is known to be highly addictive, and its use is often associated with multiple health and social problems [21].
Given the striking increase in methamphetamine/ amphetamine use both generally and among people with OUD specifically, as well as the highly addictive nature of methamphetamine and associated adverse effects, it is important to understand how use of these substances impacts receipt of and outcomes associated with MOUD. Blondino and colleagues published a systematic review of studies conducted in the United States and published before 11/28/2018 that examined associations between co-occurring substance use and retention in MOUD and opioid abstinence during MOUD, and summarized 7 articles assessing associations between amphetamine use and these 2 outcomes [22]. In order to more fully understand existing research and gaps in knowledge regarding the impact of methamphetamine/amphetamine use on MOUD-including its impact on the entire MOUD care continuum, potential trends reflecting changes in drug use patterns and MOUD provision, and potential variation across settings-an expanded review is needed that includes studies examining receipt of MOUD, studies published more recently, and studies conducted outside of the United States.
The objective of this systematic review was to identify studies that examine and report associations between methamphetamine/amphetamine use or use disorder and 3 classes of outcomes: (1) receipt of MOUD, (2) retention in MOUD, and (3) opioid abstinence during MOUD. We describe study characteristics and findings, as well as potential implications and key gaps in existing research.
the University of Washington. Boolean search queries were created using a combination of keywords and subject headings (complete search queries are included in Appendix 1). Reference lists of included studies were later hand-searched to identify additional studies meeting inclusion criteria.

Inclusion and exclusion criteria
Included studies met the following criteria: (1) the study sample was composed of people who use opioids and/or have documented OUD; (2) the study examined and reported on a quantitative association between methamphetamine/amphetamine use or use disorder and one of 3 types of MOUD outcomes of interest, with MOUD including methadone, buprenorphine and/or naltrexone; and (3) the study was published in English. We did not exclude studies if they did not limit their sample to people with diagnosed OUD, as many studies examining MOUD receipt do not assess OUD but examine samples likely to include many people who meet diagnostic criteria for OUD (e.g., people who inject heroin). MOUD outcomes of interest included (1) receipt of MOUD, which included initiation (i.e., newly starting MOUD during the study period) or any receipt (i.e., documentation of MOUD receipt during a specified period, which may or may not represent a new initiation); (2) retention in MOUD, which included both continuous measures of time in treatment (i.e., time from initiation until discontinuation) and categorical measures of time in treatment (i.e., remaining in treatment for a specified length of time); and (3) opioid abstinence during MOUD, measured through urine screens and/or self-report of opioid use. Studies were excluded if their sample was not restricted to people who use opioids and/or have documented OUD, if they examined any stimulant use (including cocaine and/or amphetamines) but did not separately examine the association of methamphetamine/amphetamine use with the outcome(s) of interest, and if they examined use of MOUD that was not prescribed. Studies were not excluded based on design (provided they included a quantitative analysis of the association of interest), geographic location, or clinical setting.

Study screening and selection
Abstracts were independently screened by 2 authors (MCF and HL) and excluded if they clearly did not meet inclusion criteria; disagreements were resolved through consensus between the 2 authors. Remaining full-text articles were independently reviewed for final inclusion/exclusion by the same 2 authors, and disagreements were resolved through consensus between the 2 authors or through consultation with the senior author (ECW) as needed. Reference lists of included articles were hand-searched by one author (MCF) to identify additional studies possibly meeting inclusion criteria, and inclusion or exclusion of these articles was independently confirmed by a second author (HL).

Data extraction and quality assessment
The same 2 authors independently extracted study data using a template developed by the study team to capture desired information; disagreements were resolved through consensus between the 2 authors. Extracted data included study design, dates, setting, population, adjunctive services to address substance use (i.e., psychosocial treatments or support groups, if the paper clearly described that these were provided or offered to study participants), average MOUD dose (if described), total number of subjects and number with methamphetamine/amphetamine use or use disorder, measure definitions, statistical analyses, covariates, and estimated association(s).
Risk of bias was assessed for each study using the Quality in Prognosis Studies tool [24], which assesses level of bias (low, moderate or high) in 6 domains: (1) participation, (2) attrition, (3) prognostic factor (i.e., methamphetamine/amphetamine use or use disorder) measurement; (4) outcome measurement, (5) confounding, (6) analysis and reporting. The level of bias for each domain was determined with respect to the specific association of interest for the present review-for example, if a study presented an unadjusted association for methamphetamine/amphetamine use and the outcome but did not include it in the multivariable model, the study was determined to have a high level of bias for confounding. The attrition domain was considered not applicable for cross-sectional studies and for longitudinal studies in which treatment retention/discontinuation was the only outcome of interest examined. Two authors (MCF and HL) independently conducted the risk of bias assessment; disagreements were resolved through consensus or through consultation with the senior author (ECW) as needed. Study screening, data extraction and quality assessment were performed using Covidence systematic review software [25].

Description of included studies
The database search returned 4852 records, and 1688 duplicates were removed. Seventeen additional articles were later identified through hand-searching reference lists of included articles. Three thousand one hundred eighty-one abstracts were screened, and 2604 were excluded. Five hundred seventy-seven full-text articles were reviewed and 538 were excluded, resulting in a total of 39 articles included for qualitative synthesis (Fig. 1). The 2 independent reviewers had "substantial agreement" at both phases of study selection based on a kappa statistic (kappa = 0.69 for abstract screening, kappa = 0.77 for full-text review) [26].

Receipt of MOUD treatment
Thirteen studies examined the association between methamphetamine/amphetamine use or use disorder and receipt of MOUD (Table 1). Eight used a crosssectional study design and 5 used a longitudinal study design. Time periods for data collection ranged from 1992 to 2018, with only 2 studies having collected data within the past 5 years (2016 or later). Eight studies were conducted in the United States; other studies were conducted in Canada, Thailand, Vietnam, Norway, and France. Study populations and settings included patients with OUD in general healthcare settings (4 studies; 1 limited to patients with both OUD and post-traumatic stress disorder), patients presenting for specialty substance use treatment for opioid use (3 studies), parents who used opioids enrolled in a child welfare-based substance use intervention program (1 study), people with OUD recruited through a community survey (1 study), and people who inject drugs (PWID) reporting opioid use recruited through SSPs or community surveys (4 studies; 1 limited to PWID with HIV). Four studies examined amphetamine use disorder, 4 examined amphetamine use, and 6 examined methamphetamine use (1 study separately examined both amphetamine and methamphetamine use). Amphetamine use disorder was measured using diagnostic codes for abuse or dependence, methamphetamine/amphetamine use was primarily measured by self-report of use during varying timeframes ranging from the past week to the past 6 months. Three studies examined receipt of any MOUD, 2 examined any agonist (methadone or buprenorphine), 1 examined buprenorphine or naltrexone, 6 examined methadone alone, 3 examined buprenorphine alone, and 1 examined naltrexone alone. Five studies adjusted for other substance use or use disorders.
Seven studies found a significant negative association between amphetamine use disorder or amphetamine/methamphetamine use and receipt of MOUD [18,[27][28][29][30][31][32]. Outcomes examined in these studies included receipt of any MOUD, any agonist, methadone alone, and buprenorphine alone. Two studies found a significant positive association; one between past 6 month methamphetamine use and lifetime receipt of injectable naltrexone among adults with OUD recruited through a community survey in a United States city [33], and the other between amphetamine use disorder and receipt of either buprenorphine or naltrexone (injectable or oral, measured through outpatient pharmacy claims) among commercially-insured adults with OUD in the United States [34]. Two studies found no significant association; one separately examined amphetamine and methamphetamine use and receipt of any MOUD within a child welfare-based substance use intervention program in Kentucky, United States [35], and the other examined "frequent" methamphetamine use and reporting current enrollment in methadone treatment among PWID with HIV recruited through a community survey in Vancouver, Canada, in which only 12 participants reported frequent methamphetamine use [36]. Two studies did not report tests of statistical significance [37,38]. There were no clear patterns in findings across studies with respect to study design, time period, geographic location, population/setting, predictor measurement or covariate  A detailed description of covariates is provided in Appendix 2: Table 5 c Crude measures of association are only presented alongside adjusted measures if there was a difference in statistical significance; p-values and/or 95% CIs are presented when they were reported    A detailed description of covariates is provided in Appendix 2: Table 5 c Crude measures of association are only presented alongside adjusted measures if there was a difference in statistical significance; p values and/or 95% CIs are presented when they were reported d Study also examined cumulative retention, however we were unable to interpret these results based on the description of these analyses

Retention in MOUD treatment
Twenty-one studies examined the association between methamphetamine/amphetamine use or use disorder and retention in MOUD ( Table 2). All studies used a longitudinal design; one was a secondary analysis of data collected for a randomized controlled trial. Time periods for data collection ranged from 1993 to 2018, with only 3 studies having collected data within the past 5 years (2016 or later). Thirteen studies were conducted in the United States, 2 in both Israel and the United States, 2 in Canada, and other studies were conducted in Israel, China, Norway and Ireland. All studies included patients receiving MOUD; study settings included methadone treatment programs (8 studies), buprenorphine treatment programs (5 studies), specialty opioid treatment programs providing both methadone and buprenorphine (3 studies; 1 youth treatment program), buprenorphine or naltrexone receipt assessed through medical records or insurance claims (3 studies), and community surveys of people who use opioids self-reporting methadone receipt (2 studies). Four studies examined amphetamine use disorder, 1 examined methamphetamine use disorder, 11 examined amphetamine use, and 5 examined methamphetamine use. Methamphetamine/amphetamine use disorder was measured using diagnostic codes or diagnostic criteria; methamphetamine/amphetamine use was measured either through urine drug screen (UDS) or self-report of use during varying timeframes either prior to intake or during treatment. Definitions of retention outcomes varied; some studies measured retention as a time-toevent variable, while others used binary or categorical measures of retention until various times ranging from 30 days to 3 years. Seven studies adjusted for other substance use. Nine studies found a significant negative association between methamphetamine/amphetamine use disorder or use and retention in MOUD [34,[39][40][41][42][43][44][45][46]. In one of these studies the association became nonsignificant after covariate adjustment [44], in 2 other studies the association was only significant in 1 of 2 populations that were examined (in both studies, the population with higher rates of amphetamine use had a significant negative association for amphetamine use and retention) [45,46]. One study conducted among patients receiving methadone treatment in Israel during 2004-2005 found a positive association between amphetamine use during treatment measured by UDS and retention over 13 months [47]. Eight studies found no significant association [36,[48][49][50][51][52][53][54]. Three studies did not report tests of statistical significance [55][56][57], with one noting that there were "too few patients to perform statistical comparison" for this association [57].
There were no clear patterns in findings across studies with respect to time period, geographic location, population/setting, predictor measurement, type of MOUD, or covariate adjustment including adjustment for other substance use. While most studies finding a significant negative association measured retention as a time-toevent variable or retention at 1 year, studies reporting non-significant associations generally looked at retention over shorter time periods (i.e., 6 months or less). Studies reporting non-significant associations generally had low numbers of participants with the predictor of interest, and many had wide confidence intervals around estimated associations suggesting low statistical power. Additionally, most studies that described provision of some type of adjunctive services for substance use (e.g., psychosocial treatment, support groups) to study participants reported non-significant associations, though it is possible these services were provided but not described in other papers. However, one study reporting a non-significant association did not align with these patterns [50]. Average MOUD dose was not consistently reported across studies, preventing assessment of potential patterns in findings across average dose.

Opioid abstinence during MOUD treatment
Eight studies examined the association between methamphetamine/amphetamine use or use disorder and opioid abstinence during MOUD (Table 3). Two used a cross-sectional study design and 6 used a longitudinal study design; 2 longitudinal studies were secondary analyses of data collected for randomized controlled trials. Time periods for data collection ranged from 2000 to 2016, with only 1 study having collected data within the past 5 years (2016 or later). Two studies were conducted in the United States, other studies were conducted in Taiwan, Vietnam, Norway, England, Ireland and Sweden. All studies included patients receiving MOUD; study settings included methadone treatment programs (3 studies), specialty opioid treatment programs providing both methadone and buprenorphine (2 studies; 1 youth treatment program), inpatient methadone treatment (1 study), an "interim" outpatient buprenorphine program (1 study) and people with OUD receiving sustained-release naltrexone implants as part of a clinical trial in inpatient treatment and prisons (1 study). One study examined amphetamine use disorder, 6 examined amphetamine use, and 1 examined methamphetamine use. Amphetamine use disorder was measured using diagnostic criteria, methamphetamine/amphetamine use was measured either through UDS or self-report of use during varying timeframes either prior to intake or during treatment. Opioid abstinence/use was measured as a binary variable, and definitions varied with respect to method of measurement (UDS or selfreport) and timeframe (e.g., at any point vs. at specific time points during treatment). No studies adjusted for other substance use or use disorders. Four studies found a significant negative association between amphetamine use disorder or methamphetamine/amphetamine use and opioid abstinence during MOUD treatment [58][59][60][61]. The other 4 studies found no significant association [54,[62][63][64]. There were no clear patterns in findings across studies with respect to study design, time period, geographic location, or definition of predictors/outcomes. Patients in studies finding a significant negative association were receiving methadone or sustained-release naltrexone implants, and patients in studies reporting non-significant associations were receiving methadone or buprenorphine. All but one of the studies finding a significant negative association adjusted for at least some covariates (though none adjusted for other substance use/use disorders), whereas all studies reporting non-significant associations presented unadjusted associations. Three of the 4 studies reporting non-significant associations had very low numbers of participants with the predictor of interest and wide confidence intervals, suggesting low statistical power. One study reporting a non-significant association that had a relatively higher number with the predictor of interest was the only study to examine diagnosed amphetamine use disorder as opposed to amphetamine/methamphetamine use during treatment [62]. Only one study described provision of any adjunctive services and average MOUD dose was not consistently reported across studies, preventing assessment of potential patterns in findings across these characteristics.

Risk of bias
Results from the risk of bias assessment are presented in Table 4. Most studies were found to have low risk of bias for participation; some were found to have moderate risk due to incomplete descriptions of recruitment methods/participation rates or higher refusal rates. The attrition bias domain was considered not applicable to cross-sectional studies and studies examining only retention/discontinuation from treatment as an outcome; most remaining studies were found to have low risk of bias for attrition, and some were found to have moderate or high risk due to higher levels of attrition. Risk of bias for prognostic factor measurement (i.e., measurement of methamphetamine/amphetamine use or use disorder) was found to be low for most studies; some were found to have moderate risk due to incomplete measurement definition or use of documented diagnostic codes to assess substance use disorder, which may be under-diagnosed or documented inconsistently. Risk of bias for outcome measurement was also found to be low for most studies; some were found to have moderate risk due to incomplete measurement definition, the outcome not having a consistent timeframe across all study participants, or use of pharmacy claims/prescription fill data which may not capture all receipt of MOUD. Most studies were found to have moderate or high risk of bias for confounding due to lack of adjustment for some or all potential confounding factors. Many studies were found to have moderate risk of bias for statistical analysis and reporting due to lack of conceptually driven model-building, or lack of clarity in description of analyses and/or results.

Discussion
This systematic review identified studies from multiple countries examining the association between methamphetamine/amphetamine use or use disorder and a range of MOUD care continuum outcomes. Overall, existing research suggests that methamphetamine/amphetamine use and use disorder negatively impact receipt of MOUD, retention in MOUD and opioid abstinence during treatment. No clear pattern in findings was observed across time periods or geographic locations, though potential patterns emerged across outcomes, including MOUD type, longer vs. shorter-term retention, and the provision of adjunctive services during MOUD. These patterns should be directly examined in future research.
Studies examining receipt of MOUD generally found that amphetamine use disorder or methamphetamine/  A detailed description of covariates is provided in Appendix 2: Table 5 c Crude measures of association are only presented alongside adjusted measures if there was a difference in statistical significance; p values and/or 95% CIs are presented when they were reported amphetamine use was negatively associated with receipt of opioid agonist medication. This finding appeared in studies spanning multiple time periods, geographic locations, clinical settings, and populations. It is possible that some observed associations are confounded by other substance use/use disorders, though 3 of the 7 studies finding a negative association adjusted for this. The 2 studies that found a positive association examined receipt of injectable naltrexone alone and naltrexone or buprenorphine [33,34]. It is possible that an apparent association between methamphetamine/amphetamine use and receipt of naltrexone is confounded by the presence of alcohol use disorder for which naltrexone is an indicated treatment [65]. The study by Morgan and colleagues adjusted for alcohol use disorder diagnoses while the study by Daniulaityte et al. did not. Naltrexone has been studied as a potential pharmacotherapy for amphetamine use disorder [66,67], however it is generally considered a second-line treatment for OUD [68], and may be less effective than agonist therapies in reducing risk of opioid overdose [11]. One study reporting a non-significant association likely had low power due to a very small number with the predictor of interest [36], and the other may have been the result of a unique study setting (a child welfare-based substance use intervention that aimed to facilitate linkage to MOUD) [35]. Overall, existing studies suggest that methamphetamine/amphetamine use may be a widespread barrier to receipt of opioid agonist medications among people with OUD, and further research is needed to determine whether receipt of naltrexone is more prevalent among people with OUD who use methamphetamine/amphetamines. Studies examining retention in MOUD generally found negative associations between methamphetamine/amphetamine use disorder or use and retention across multiple study time periods, geographic locations, clinical settings and populations, as well as across different types of MOUD. Some observed associations may be confounded by other substance use, though 5 of the 9 studies finding a negative association adjusted for this. As we do not expect methamphetamine/amphetamine use to positively impact retention relative to no use, we considered potential differences among studies finding a negative association compared to studies finding no association between methamphetamine/amphetamine use and retention.
Most studies reporting non-significant associations had relatively small numbers of participants with the predictor of interest, suggesting they may have been underpowered to detect associations. Besides the likely impact of low power, there were other potential differences among studies reporting negative associations compared to those reporting no association-most studies reporting no association examined retention over shorter periods of time than those that found negative associations, suggesting the possibility that methamphetamine/amphetamine use may have more of an impact on longer-term rather than shorter-term MOUD retention. Additionally, most studies reporting no association described some type of adjunctive services for substance use that were provided or offered to study participants, suggesting adjunctive services might improve retention for some people who use methamphetamine/amphetamines. However, provision of these services may not have been consistently reported across studies and low statistical power may be the primary factor driving non-significant results. Overall, existing studies suggest that methamphetamine/amphetamine use and use disorder negatively impacts MOUD retention.
Studies examining abstinence from opioid use during MOUD treatment generally found that methamphetamine/amphetamine use was negatively associated with opioid abstinence. However, as none of these studies adjusted for other substance use/use disorders, it is possible that observed associations are confounded by other substance use. Most studies finding significant negative associations were conducted in methadone clinics; one was a secondary analysis of randomized controlled trials testing sustained-release naltrexone implants [60]. Most studies reporting non-significant associations had very low numbers of participants with the predictor of interest and thus likely had low statistical power. One that may have had higher power was the only study to examine amphetamine use disorder [62], suggesting that only active use during treatment impacts opioid abstinence, however more research is needed to confirm this. Overall, existing studies suggest that methamphetamine/amphetamine use may negatively impact opioid abstinence during treatment among patients receiving methadone or sustained-release naltrexone implants, while the impact for patients receiving  buprenorphine or other types of naltrexone is unclear. However, further research is needed adjusting for other substance use.

Gaps in research and future directions
Research is needed to understand how varying characteristics of MOUD care influence the impact of methamphetamine/amphetamine use on MOUD outcomes. One study that did not meet inclusion criteria for this review (as it did not examine use of amphetamines/ methamphetamines specifically) found that removing a buprenorphine program's requirement that patients be abstinent from stimulants (cocaine or amphetamines) resulted in improved initiation, but decreased retention, for patients who used stimulants [69]. Future studies should similarly aim to understand the impact of specific clinical policies on MOUD receipt, retention, and treatment outcomes for people who use methamphetamine/amphetamines. Research is also needed to directly assess the impact of MOUD dose and receiving psychosocial treatments on MOUD retention and outcomes among people who use methamphetamine/amphetamines. Randomized controlled trials have found that providing contingency management and cognitive behavioral therapy to patients who used stimulants in MOUD reduced stimulant use, suggesting that offering concurrent, co-located treatments for multiple substance use disorders can benefit patients [70][71][72]. While there are currently no FDA-approved medications to treat amphetamine use disorder, ongoing work to advance pharmacologic treatment may also create opportunity for better simultaneous treatment [67]. However, treatment providers should recognize that requiring, rather than offering, additional treatment may create barriers to MOUD for some patients who use other substances, which could increase their risk of opioid overdose. Finally, most studies included in this review were conducted in specialty substance use treatment settings, though some were conducted in more general medical settings or involved community surveys. Studies are needed that examine outcomes for people who use methamphetamine/amphetamines in new settings where MOUD are increasingly being provided, such as emergency departments, prisons/jails, and community settings such as SSPs [15,73,74]. One study of SSP-based buprenorphine treatment found that stimulant use (cocaine or amphetamines) at enrollment was not associated with retention in bivariate analyses, suggesting MOUD outcomes for people who use methamphetamine/amphetamines might be improved in lower barrier settings [15]. Increased understanding of the perspectives of both MOUD providers and people who use drugs regarding co-occurring opioid and methamphetamine/amphetamine use is also needed. In surveys and qualitative studies buprenorphine providers have indicated they are less likely to prescribe for patients who use alcohol or benzodiazepines [75,76], however providers' thoughts on methamphetamine/amphetamine use are unclear. Some research suggests that people who use opioids/have OUD who also use methamphetamine/ amphetamines are less likely to express interest in receiving help for substance use [77,78]. Qualitative studies have found that people who use both opioids and methamphetamine describe a balancing effect of the drugs that increases functionality, which could be related to a lower perceived need for MOUD [17,79]. Another qualitative study found that methadone patients who used stimulants described several benefits they experienced from their stimulant use, including balancing sedating effects of methadone [80]. Future research should seek to further understand how people who have OUD and use methamphetamine/ amphetamine perceive their need for MOUD, whether they feel MOUD are accessible to and effective for them, and their recommendations to improve MOUD services.
Finally, evidence suggesting that methamphetamine/amphetamine use and use disorder is associated with reduced receipt of MOUD, reduced retention in MOUD, and opioid use during MOUD treatment highlights the necessity of maintaining and expanding evidence-based harm reduction strategies that prevent overdose death and reduce risk of other sequelae. Such strategies include widespread naloxone distribution, overdose prevention education, and supervised consumption facilities [81][82][83][84]. Harm reduction may play an increasingly important role in preventing overdose death if methamphetamine/amphetamine use continues to increase among people who use opioids, and efforts should be made to ensure that these services reach people who use multiple substances.

Limitations
While our search strategy identified a large number of studies for screening, it may have missed studies not included in searched databases. We addressed this limitation by performing a hand search of the reference lists of included articles. Additionally, the inclusion criterion that studies be published in English may have resulted in the exclusion of some relevant studies. Most included studies analyzed data collected prior to 2016, and patterns may be changing as methamphetamine use continues to increase among people who use opioids and MOUD delivery continues to evolve. Many included studies did not examine methamphetamine/amphetamine use or use disorder as a primary variable of interest, but rather as one of several variables of interest, and therefore many did not adjust for covariates based on hypothesized confounding specific to methamphetamine/amphetamine use or use disorder. As described above, several studies appeared underpowered to detect the association of interest based on small numbers of participants with the predictor of interest. One limitation specific to studies of MOUD receipt or opioid abstinence that did not clearly establish temporality between methamphetamine/amphetamine use and the outcome of interest is the possibility that findings reflect reverse causality (i.e., the impact of receiving MOUD or using opioids during MOUD on methamphetamine/ amphetamine use). However, in both outcome groups there were multiple studies that did clearly measure methamphetamine/amphetamine use prior to the outcome event that found a significant negative association. Finally, the scope of this review was limited to studies describing associations between methamphetamine/amphetamine use and MOUD-related outcomes. Future literature reviews should summarize existing research examining the impact of other specific substance use on MOUD, as well as the impact of methamphetamine/amphetamine use on treatment for other substance use disorders. Additionally, future reviews could summarize existing research examining the impact of methamphetamine/amphetamine and other substance use on sequelae of opioid use disorder among people receiving MOUD, including overdose.

Conclusions
Methamphetamine/amphetamine use has sharply increased among people with OUD. Findings from studies identified in this systematic literature review generally suggest that methamphetamine/amphetamine use negatively impacts MOUD receipt, MOUD retention, and opioid abstinence during MOUD. Future research should examine how specific aspects of MOUD care and low-barrier models of treatment impact MOUD outcomes for this population.
Research is also needed to better understand the perspectives of MOUD providers and people who use both opioids and methamphetamine/amphetamines. Continued efforts to expand and improve MOUD and overdose prevention strategies for this population are needed.

Database search queries
PubMed/MEDLINE (amphetamine OR amphetamines OR methamphetamine OR methamphetamines OR meth OR stimulant OR stimulants OR "other drug" OR "other drugs" OR "other substance" OR "other substances" OR polydrug OR polysubstance OR "multiple drug" OR "multiple drugs" OR "multiple substance" OR "multiple substances" OR "Methamphetamine"