Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Opioid Use Disorder Pharmacotherapy and Overdose Prevention

Abstract. Opioid use disorder pharmacotherapy and overdose prevention are linked intervention systems operating across receptor pharmacology, clinical initiation, long-term maintenance, acute toxicity, and population access to care. This fundamental critical narrative monograph reconstructs the history of maintenance treatment, compares methadone, buprenorphine, and naltrexone, evaluates randomized and observational evidence on retention, relapse, overdose and mortality, and analyzes naloxone distribution and post-overdose continuity. It integrates seminal trials, contemporary fentanyl-era studies, a 2026 randomized emergency-department trial, cohort mortality analyses, a 2026 network meta-analysis, and current clinical and regulatory sources. The synthesis identifies effective agonist maintenance and emergency opioid reversal as complementary interventions, clarifies the antagonist induction barrier, distinguishes study estimands and evidence certainty, and sets out unresolved questions in causal inference, equitable access, and sustained treatment engagement.

Keywords: opioid use disorder; methadone; buprenorphine; naltrexone; naloxone; overdose prevention; fentanyl; medication for opioid use disorder; opioid agonist treatment; treatment initiation; retention; mortality; precipitated withdrawal; emergency department; harm reduction; psychopharmacology

Scientific Scope, Research Questions, and Method

Opioid use disorder (OUD) pharmacotherapy comprises sustained treatments that alter opioid receptor signaling, withdrawal physiology, reinforcement, and exposure to lethal overdose, together with pharmacological reversal of acute opioid toxicity. This monograph examines the comparative effectiveness, mechanisms, implementation, and safety of methadone, buprenorphine, and naltrexone, and the critical interface of naloxone and other opioid-overdose reversal agents with ongoing care. The central scientific object is the intervention system: a pharmacological agent administered to a person within a changing clinical and social environment, evaluated through initiation, persistence, opioid use, functional outcomes, overdose, and mortality. The FDA (2026) identifies three FDA-approved medications for treating OUD: methadone, buprenorphine, and naltrexone.

The research questions are how pharmacological mechanisms relate to patient-important outcomes; how induction and discontinuation change the risk profile; which differences among treatments are substantiated by randomized trials; why observational survival estimates do not directly answer the same questions; how overdose-response systems connect people to durable treatment; and which barriers make effective therapy inaccessible. Evidence is organized by directness and design: original trials establish contrasts under assigned treatment strategies, cohorts assess time-varying treatment and rare outcomes, systematic syntheses appraise consistency, and regulatory sources define present legal rather than biological possibilities.

This is a critical integrative narrative monograph, based on purposively selected seminal studies, confirmatory trials, population-based research, peer-reviewed syntheses, and official clinical and regulatory material available through November 3, 2026. It is not a preregistered systematic review, and does not claim exhaustive search capture or an independent meta-analysis. The underlying studies retain their own inclusion criteria and bias structures. Clinical treatment decisions require local medical assessment and current product labeling; this scientific exposition does not supply individualized initiation instructions.

The Disorder: Diagnosis, Dependence, Exposure, and Harm

OUD is a behavioral and clinical syndrome involving a problematic pattern of opioid use with impairment or distress and characteristic domains such as impaired control, craving, harmful consequences, and persistent use. Physiological tolerance and withdrawal are pharmacological adaptations that may occur with prescribed opioid therapy; their presence alone does not establish compulsive substance use. Diagnostic assessment asks about behavior and functioning as well as exposure, while overdose prediction additionally asks about supply potency, co-exposures, abstinence-related tolerance loss, and access to rescue. The CDC (2024) recommends offering or arranging medication treatment rather than relying on detoxification alone.

The separation between disorder severity and overdose vulnerability matters. A person can experience a lethal opioid exposure without having a diagnosed OUD; a person with well-treated OUD can face intermittent exposure through contaminated drugs, interrupted medication access, or competing medical conditions. Thus overdose prevention has both a disorder-treatment pathway and a universal harm-reduction pathway. Longitudinal research should not silently substitute positive toxicology, possession of a medication, or a coded diagnosis for the entire clinical construct.

An adequate pharmacotherapy outcome system distinguishes engagement in care, actual drug exposure, craving, withdrawal burden, quality of life, social participation, overdose, and all-cause mortality. Each has different observation windows and different biases. Urine toxicology is an objective measure of recent exposure within a detection window, whereas sustained treatment retention can be favorable even when intermittent substance use occurs. Treating one outcome as a universal proxy for recovery makes studies difficult to compare.

Historical Emergence of Maintenance Treatment

The development of opioid maintenance treatment changed the clinical unit of analysis from an isolated episode of detoxification to a continuing condition in which pharmacological stabilization can support social and psychological functioning. Dole and Nyswander (1965) reported an early methadone clinical trial in 1965; its historical importance lies in showing that a sustained agonist intervention could alter the course of heroin use. Its small scale and historical treatment environment should not be read as modern randomized comparative evidence. Methadone subsequently became embedded in specialized treatment programs, where dispensing rules and service architecture shaped treatment exposure as powerfully as receptor pharmacology.

Buprenorphine introduced partial mu-opioid agonism with a distinctive respiratory-depression profile and expanded opportunities for office-based prescribing in the United States. Long-acting formulations subsequently shifted the question from daily adherence alone toward interval coverage, convenience, access, and continuity between doses. Antagonist treatment with naltrexone represented a different design: blocking opioid effects after adequate detoxification. The histories of these agents illuminate why access conditions, patient selection, and the transition onto medication must be included in comparative effectiveness.

Overdose-response systems developed along a parallel trajectory. Naloxone, once primarily a professional emergency medication, became an essential community intervention distributed to people likely to witness an overdose. The WHO (2014) formalized a community-management framework in 2014, recognizing the practical importance of bystanders, training, immediate airway support, and linkage to emergency services. These histories converge in the contemporary objective of preventing acute fatality while providing continued treatment after rescue.

Opioid Receptor Pharmacology and Behavioral Regulation

The mu-opioid receptor is central to opioid analgesia, reward-related signaling, physical dependence, and respiratory suppression. Pharmacological differences among a full agonist, partial agonist, and antagonist condition expected clinical effects but do not themselves specify a person's treatment trajectory. Methadone is a long-acting full agonist; its therapeutic effects include suppressing withdrawal and reducing the salience of additional opioid exposure during stable treatment. Its long and variable pharmacokinetics also create clinically consequential accumulation, interaction, and overdose considerations, particularly during treatment initiation and changes in tolerance.

Buprenorphine has high receptor affinity and partial agonist activity. Its receptor occupancy can suppress withdrawal and attenuate effects of additional full agonists, but sufficiently abrupt displacement of full agonists can produce precipitated withdrawal. The partial-agonist ceiling on respiratory depression is clinically important yet not an absolute shield against dangerous sedation, especially with alcohol, benzodiazepines, other sedatives, or medical instability. Clinical safety must account for formulation, co-exposures, environment, and monitoring; it cannot be derived from receptor labels alone.

Naltrexone competitively antagonizes opioid receptors. It provides a pharmacological blockade rather than agonist stabilization, does not itself treat opioid withdrawal, and requires an opioid-free transition to avoid precipitating withdrawal. Its clinical value depends on achieving induction and subsequent adherence or scheduled injections. Naloxone, in contrast, is a short-acting competitive opioid antagonist used for acute overdose reversal. These agents therefore occupy distinct positions in the causal chain: maintenance treatment changes the course of OUD, while emergency reversal interrupts an immediate physiological crisis.

Methadone: Maintenance, Retention, and Clinical Organization

Methadone has a long empirical history and remains a central treatment for moderate to severe OUD. Stable maintenance can reduce illicit opioid use and support retention, but outcomes depend on supervised clinical assessment, consistent medication access, responsive dose management, and the ability of the care setting to meet patients' needs. Because methadone is a full agonist with variable elimination and clinically important interactions, the earliest treatment phase requires particular attention to safety. No fixed dose or unsupervised schedule can stand in for individualized clinical management.

Treatment retention is a substantive therapeutic outcome. Leaving treatment may reflect clinical adverse effects, access costs, punitive program practices, housing and transportation barriers, competing responsibilities, or the person's preferences. A discontinuation date in administrative data is therefore both a potential change in pharmacological protection and a marker of altered access or health. Sordo et al. (2017) found substantially higher mortality rates out of treatment than during methadone maintenance across cohort studies, but cautioned that transitions and changing baseline risk complicate causal interpretation.

In the United States, methadone for OUD is largely delivered through regulated opioid treatment programs, unlike ordinary office-based prescribing of buprenorphine. Revisions to federal opioid treatment program regulations in 2024 sought greater flexibility, patient-centered decisions, and reduced barriers (SAMHSA (2024))). These are system-level interventions acting on continuity, dignity, and feasibility. Their health impact must ultimately be assessed in real-world cohorts and implementation studies, rather than assumed from the existence of a regulation.

Buprenorphine: Partial Agonism and Accessible Maintenance

Buprenorphine maintenance has evidence from controlled trials and systematic reviews for improving retention compared with placebo and reducing illicit opioid use under suitable treatment conditions. The Mattick et al. (2014) Cochrane synthesis compared buprenorphine with placebo and methadone across different regimens and settings; interpretation requires attention to comparator treatment adequacy, dose conditions, and study-era differences. In modern clinical practice, the comparison between buprenorphine and methadone includes supply accessibility, clinical preference, monitoring requirements, co-occurring medical issues, and continuation prospects.

Sublingual combination products typically include naloxone to deter some forms of non-prescribed administration; naloxone's presence in these medications is conceptually separate from community naloxone overdose reversal. Long-acting injectable buprenorphine allows therapeutic exposure between appointments and offers an alternative when daily medication storage, routine, or stigma create difficulties. These formulations can also introduce new access barriers through cost, authorization procedures, injection requirements, or disruption of follow-up.

A patient's response should be assessed through symptom stability, ongoing exposure, functioning, treatment preferences, and adverse effects. Statements that one formulation is uniformly superior to all others exceed the available evidence. Clinical decisions are contingent on whether patients can actually initiate a medication and remain supported by a treatment service; a theoretically effective pharmacological formulation can perform poorly when institutional design prevents receipt or continuation.

Extended-Release Buprenorphine and Fentanyl-Era Trials

A 2025 randomized, double-blind study by Shiwach et al. (2025) compared two maintenance regimens of monthly extended-release buprenorphine in a high-risk OUD sample. Among 435 analyzed participants, the primary response comparison yielded an absolute difference of 2.6 percentage points with a 95% confidence interval from −4.7 to 9.9 points, which did not establish overall superiority. Post hoc analyses suggested potentially greater benefit in subgroups reporting very frequent fentanyl use, but subgroup exploration cannot justify a universal dosing rule. The study's industry sponsorship and author affiliations are relevant to transparent appraisal, alongside its randomized design.

An especially instructive contrast appears in the 2026 multicenter emergency-department randomized trial by D’Onofrio et al. (2026). Across 29 US emergency departments and 1,994 participants, seven-day injectable and sublingual buprenorphine strategies had similar seven-day treatment engagement, 40.5% versus 38.5%. At 30 days, reported engagement was also similar. Because both arms received structured referrals, the trial tested a formulation strategy within a particular service environment rather than whether medication initiation was superior to withholding treatment.

These studies identify a recurring methodological distinction. The intervention assigned in a formulation trial can change pharmacological coverage, but the outcome that matters—remaining in care, reduced illicit use, or avoiding death—also depends on a sequence of appointments, insurance decisions, pharmacy access, and social support. The 2026 ED finding is evidence against assuming an automatic short-term engagement advantage for the tested injection, while still leaving open benefits in different populations, delivery systems, and longer outcomes.

Naltrexone: Antagonism and the Induction Barrier

Extended-release naltrexone provides opioid-receptor blockade without continuous agonist stimulation. The principal pharmacological requirement is successful transition after adequate opioid withdrawal, a condition that can be difficult to achieve in people with ongoing physiological dependence. Evidence restricted to participants who have already completed detoxification addresses effectiveness after initiation. Evidence assigning treatment before detoxification additionally captures how often the drug can be started at all.

The X:BOT trial by Lee et al. (2018) randomized 570 participants to extended-release naltrexone or buprenorphine-naloxone in community-based inpatient services. Successful induction occurred in 72% assigned to naltrexone compared with 94% assigned to buprenorphine-naloxone. Intention-to-treat comparisons favored buprenorphine-naloxone on relapse-free survival, substantially reflecting induction failures; among those who initiated assigned treatment, relapse outcomes were more similar, but that initiated subset had been selected after randomization. Neither result should be discarded. Together they show why a treatment's feasibility is an integral component of its effectiveness.

A prior placebo-controlled trial by Krupitsky et al. (2011) showed extended-release naltrexone efficacy among detoxified participants in a Russian treatment context. Its design supports pharmacological efficacy after successful initiation while limiting conclusions about the full initiation pathway in an unselected community population. Antagonist therapy is a real treatment option for appropriately selected individuals, but evidence that it reduces overdose mortality to the same extent as maintained agonist treatment is less certain; Larochelle et al. (2018) observed wide confidence intervals and relatively few naltrexone events in a post-overdose cohort.

A 2026 observational target-trial emulation by Hsu et al. (2026) followed 106,052 medically managed withdrawal discharge episodes in Massachusetts during a 24-week comparison. The estimated nonfatal opioid overdose risks were 11.6% after buprenorphine-naloxone initiation and 13.9% after extended-release naltrexone initiation; estimated all-cause mortality was 1.4% in both groups. These outcomes concern medication initiation after withdrawal treatment under modeled assumptions, not randomized assignment or the risk among everyone with untreated OUD. They provide a more recent counterpoint to X:BOT and reinforce that an apparently similar mortality estimate can coexist with a meaningful difference in nonfatal overdose.

Initiation, Withdrawal, and the Changing Drug Supply

Initiation is a causal transition in which previous opioid exposure, receptor occupancy, withdrawal state, treatment setting, and formulation jointly determine early tolerability. Delaying pharmacotherapy until all symptoms disappear can lose a critical engagement opportunity; initiating a high-affinity partial agonist under inappropriate physiological conditions can precipitate withdrawal. An objective withdrawal scale such as the Clinical Opiate Withdrawal Scale is a measurement aid, not a complete individualized decision rule. Clinical assessment additionally depends on timing, co-exposures, prior medication, and the patient's priorities.

The prevalence of illicitly manufactured fentanyl and analogues has increased clinical uncertainty about withdrawal timing. In a prospective multisite emergency-department trial report, precipitated withdrawal was uncommon; a separate retrospective hospital cohort by Thakrar et al. (2024) reported 26 events among 226 patients, or 11.5%, using a specified scale-change definition. These estimates should not be merged as if drawn from identical patient populations, induction regimens, ascertainment windows, or monitoring environments. Disagreement between rates is an empirical question about case mix and measurement rather than proof that one study is defective.

The Weimer et al. (2023) clinical considerations for high-potency synthetic opioids discuss adapted induction strategies as expert guidance in a changing practice landscape. Such considerations provide a framework for clinicians but should be distinguished from a high-certainty outcome trial establishing one optimum method across settings. Future randomized comparisons need to report precipitated withdrawal definitions, baseline fentanyl exposure, patient experience, sustained retention, and serious adverse events together. More flexible initiation pathways deserve evaluation because tolerability, rescue from withdrawal, and treatment engagement are jointly relevant.

Evidence Across Randomized and Observational Designs

Randomization protects a treatment comparison against measured and unmeasured baseline confounding in expectation, but actual treatment initiation, crossovers, attrition, masking limitations, and selection of settings can still constrain the estimand. A randomized trial that assigns medications after inpatient admission primarily estimates outcomes for people who reached that admission and accepted randomization. It cannot automatically answer what would happen if those same medications were offered to everyone after an emergency overdose in a geographically dispersed community.

Observational cohorts are indispensable for studying death, rare adverse events, long-term treatment episodes, and policy changes. Yet medication selection is influenced by severity, prior treatment, healthcare access, incarceration, employment, and unrecorded substance use. Person-time analyses can also misclassify changing exposure. Comparisons of deaths during and after treatment have an additional transition problem: the act of leaving therapy is associated with factors that influence baseline risk independently of any pharmacological withdrawal of protection.

The 2026 network meta-analysis by Rodrigues et al. (2026) synthesized 103 randomized trials enrolling 26,624 people across 13 interventions. It found stronger evidence for treatment retention and some abstinence outcomes than for mortality or overdose outcomes, for which the randomized evidence remained low certainty. That result is consistent with substantial observational mortality associations: the designs observe different event frequencies, comparison strategies, intervention duration, and populations. Absence of a conclusive mortality signal in underpowered trial networks is not affirmative evidence that effective medication has no survival benefit.

Treatment Retention as a Time-Varying Therapeutic State

Retention is often represented by the binary fact of being in treatment at a chosen endpoint. Yet OUD care may include episodic starts, interruptions, transitions between medications, and re-engagement following recurrence. Counting only final status can obscure the dose of exposure to a potentially protective treatment over time. Event-history models, recurrent-event approaches, and trajectory analyses better represent this pattern when records and assumptions support them.

Discontinuation is clinically and scientifically consequential. In the meta-analysis of 19 cohorts, Sordo et al. (2017) reported pooled all-cause mortality rates during and outside methadone treatment of 11.3 and 36.1 per 1,000 person-years; for buprenorphine the corresponding pooled rates were 4.3 and 9.5. These are aggregate rates from different observational populations, not a head-to-head randomized proof that buprenorphine is safer than methadone for every patient. The large heterogeneity among cohorts and differential eligibility are essential qualifiers.

The immediate period following medication discontinuation may concentrate risk because tolerance can be reduced while environmental exposure persists. A robust treatment system should therefore investigate the reasons for interruption, the availability of rapid re-entry, and continuing overdose prevention during gaps. In research, participants lost to follow-up should not be presumed abstinent, relapsed, or deceased; linkage to mortality and administrative data and sensitivity analyses are required to establish what happened.

Overdose and Mortality as Patient-Important Outcomes

Opioid overdose is an acute toxic syndrome in which opioid-induced respiratory depression may progress to hypoxemia, brain injury, or death. Fatality depends on the toxic exposure, other sedating substances, individual physiology, whether another person is present, access to immediate assistance, and the effectiveness of emergency response. Overdose-related mortality is therefore a particularly important endpoint but requires large populations or long follow-up to measure reliably in randomized treatment trials.

A Massachusetts retrospective cohort of 17,568 adult survivors of nonfatal overdose by Larochelle et al. (2018) found that methadone maintenance and buprenorphine receipt were associated with lower all-cause and opioid-related mortality relative to periods without medication. For all-cause mortality, adjusted hazard ratios were 0.47 for methadone and 0.63 for buprenorphine; naltrexone estimates were too imprecise for confident conclusions. The associations are clinically meaningful but may still reflect residual confounding, health-system contact, and classification of exposure.

The Wakeman et al. (2020) comparative effectiveness cohort of 40,885 people across six treatment pathways likewise found that buprenorphine or methadone was the pathway associated with reduced overdose and serious opioid-related acute-care use compared with no treatment at both three and twelve months. This convergence across sources strengthens confidence that treatment access matters. It cannot identify a single biological mechanism as the complete explanation, because pharmacological stabilization, structured contact, ancillary support, and selection may operate together.

A subsequent Ontario matched-cohort analysis by Kleinman et al. (2026) compared methadone with buprenorphine-naloxone among people with overdose in the preceding year. After propensity-score matching, the study included 2,941 people per treatment group and reported lower one-year death hazard following methadone initiation. Treatment episodes were short, and the investigators observed no significant mortality difference restricted to time receiving opioid agonist treatment. This pattern is consistent with the importance of retention and contextual access, but residual confounding and the Ontario care environment preclude a universal drug-ranking conclusion. The finding adds directly relevant 2026 evidence while preserving the distinction between initiation strategies and pharmacological exposure.

Acute Overdose Physiology and Clinical Recognition

The immediate physiological danger in an opioid overdose is inadequate ventilation. Reduced responsiveness, slow or absent breathing, cyanosis, and abnormal respiratory sounds are warning signs, though clinical presentation may be affected by co-ingestants. Respiratory support and rapid professional emergency response are crucial. Public-facing education should emphasize recognizing abnormal breathing, activating emergency services, giving an available opioid-overdose reversal medication according to product instructions, and following emergency dispatcher guidance about rescue breathing or CPR. Professional care remains important even if responsiveness temporarily returns.

The contemporary unregulated drug supply can contain fentanyl or other synthetic opioids, benzodiazepines, alcohol, stimulants, or non-opioid sedatives such as xylazine. Naloxone reverses the opioid component of toxicity but does not directly reverse a non-opioid sedative. A seemingly incomplete response may therefore require ongoing respiratory support and emergency medical treatment rather than concluding that naloxone has no value. Different opioid and antagonist durations also allow respiratory depression to recur after initial improvement.

Overdose recognition is a form of distributed behavioral vigilance: survival may depend on another person noticing clinical signs, having medication within reach, and acting under uncertainty. This is why education, peer networks, product availability, and emergency-service relationships are substantive components of an overdose-prevention system. They should be evaluated as real interventions with measurable implementation outcomes.

Naloxone: Mechanism, Delivery, and Community Access

Naloxone is a competitive opioid receptor antagonist used for emergency reversal of suspected opioid overdose. Nasal and injectable formulations have different product instructions, and emergency response can require repeated administration or continued observation because opioid effects may outlast naloxone's effect. The WHO (2014) recommends access to naloxone and training for people likely to witness an overdose, including people who use opioids and their relatives and peers. It also emphasizes basic life support and the continuing need for comprehensive medical care.

The scientific evaluation of community naloxone distribution spans physiological efficacy, reach, access, bystander competence, documented reversal, and population overdose mortality. Physiological reversal of opioid toxicity is well established. Whether a particular distribution program changes mortality in a region depends on coverage, targeting, training, people using alone, timing of response, and other contemporaneous changes in the drug supply and services.

In a Massachusetts interrupted time-series analysis, Walley et al. (2013) found community-level associations between implementation of overdose education and nasal naloxone distribution and lower opioid overdose death rates. This ecological evidence supports distribution strategies but cannot exclude all time-varying confounding or quantify the effect for a particular individual. It also highlights the importance of evaluating actual implementation intensity, not merely counting whether a community adopted a written policy.

Overdose Reversal Agents, Safety, and Regulatory Distinctions

Naloxone access has expanded through nonprescription products and public-health programs in the United States. Availability and reimbursement vary by location and product. Other reversal agents, including nalmefene products, have regulatory approvals with distinct pharmacological properties and labeling. A longer antagonist duration can affect observation and withdrawal experience but does not make all agents or delivery routes interchangeable; procurement policy should follow up-to-date FDA labeling and the evidence applicable to each preparation.

Emergency reversal may precipitate acute withdrawal in a person with physiological dependence. This possibility should be managed through supportive emergency care and does not outweigh the immediate imperative of restoring breathing in a suspected opioid overdose. Moreover, a return to consciousness after reversal is not a clinical endpoint equivalent to safety at discharge: subsequent toxicity, co-exposures, aspiration, trauma, or recurrent respiratory depression may require monitoring.

It is analytically useful to distinguish product approval, comparative clinical effectiveness, and public-health implementation. An FDA-approved indication establishes that a product met the applicable regulatory evidence standard for its labeled use. It does not prove that any distribution network has sufficient coverage or that every locally available product is equivalent in all real-world situations. The appropriate research design shifts as the question moves from receptor reversal to program effectiveness.

Prevention After a Nonfatal Overdose

Nonfatal overdose identifies a person with elevated future risk and a potential encounter with emergency, inpatient, or community services. The response should be understood as a linked sequence: stabilize the acute physiological event, address immediate safety, assess OUD and co-occurring needs when feasible, offer medication treatment, establish timely follow-up, and ensure future overdose-response resources. Investigating the sequence requires identifying where individuals are lost between each step.

The Larochelle et al. (2018) cohort demonstrates both the prognostic significance of the post-overdose period and the low observed uptake of maintenance medication in its historical Massachusetts sample. The findings cannot be generalized as current national rates because the cohort covers 2012–2014 and different health-service conditions. Their central implication is a measurable treatment gap among survivors; the strongest contemporary response is to study whether emergency-to-community transitions close that gap without merely transferring administrative responsibility.

Overdose rescue and maintenance treatment complement one another. Naloxone can reverse an acute opioid-mediated event, while methadone or buprenorphine can reduce ongoing withdrawal pressure and illicit opioid exposure over longer intervals. Medication alone cannot remove an adulterated supply or guarantee that someone will witness the next event. Conversely, distributing naloxone without an accessible treatment pathway forfeits opportunities for sustained risk reduction. Integrated strategies need both elements.

Emergency Department Initiation and Warm Handoffs

A randomized trial by D’Onofrio et al. (2015) demonstrated that emergency-department-initiated buprenorphine with facilitated follow-up could improve engagement compared with referral-oriented strategies in the study setting. This changed a crucial implementation assumption: emergency services can initiate the treatment trajectory rather than serving solely as sites of stabilization and advice. Its outcomes depend on a real receiving service, so reproducing the medication intervention without follow-up infrastructure may reproduce only part of the tested model.

The larger 2026 trial by D’Onofrio et al. (2026) provides a direct comparison of two buprenorphine initiation formulations while retaining referral support for all participants. Similar engagement outcomes at seven days illuminate the limits of formulation substitution as a stand-alone solution. Participants can receive effective initial medication and still encounter insurance, transportation, identification, scheduling, housing, or pharmacy barriers within days.

A warm handoff is more than the distribution of a telephone number. It involves making a receiving encounter likely, aligning records and consent, resolving immediate medication coverage, and identifying contingency pathways if the first appointment fails. The key endpoints include successful first contact, medication possession or administration, sustained retention, overdose after discharge, and patient-reported feasibility. These measures should be reported separately because program claims can otherwise conceal where attrition occurred.

Behavioral Treatment, Psychosocial Care, and Shared Decisions

Pharmacotherapy operates within a larger behavioral and interpersonal treatment environment. Motivational interviewing, contingency management, cognitive behavioral approaches, peer support, and psychiatric care may address different mechanisms such as adherence, reinforcement, coping, housing instability, and co-occurring symptoms. The evidence base does not justify withholding effective medication until a person completes counseling or achieves a predetermined psychological state. The CDC (2024) emphasizes access to evidence-based medication treatment for OUD.

Behavioral intervention research must specify its contrast. Counseling added to a well-supported pharmacotherapy program is a different trial from counseling offered as an alternative to medication, and different again from a complex service package compared with inadequate care. Improvements in retention may reflect transportation, staff continuity, legal support, and a non-punitive environment as much as an intervention's named psychotherapy component.

Shared decision-making is an empirical component of treatment acceptability. Patients may legitimately value the structure of a methadone program, the flexibility of buprenorphine, or the antagonist profile of naltrexone after detoxification. Respecting preferences requires presenting relative evidence, induction demands, monitoring needs, risks, and access realities accurately. Preferences can change after exposure to a formulation or shift in life circumstances; research should allow treatment switching without automatically treating it as failure.

Co-occurring Mental Disorders and Polysubstance Exposure

Depression, anxiety, trauma-related disorders, psychotic symptoms, chronic pain, and other substance use can complicate assessment and sustained treatment. The direction of causation between these conditions and OUD is often reciprocal or difficult to establish from cross-sectional data. Pharmacotherapy studies that exclude complex psychiatric presentations may overstate external applicability to real-world services, while studies using administrative diagnoses can conflate disorder presence with differential recognition and billing.

Concurrent alcohol or benzodiazepine exposure increases the clinical importance of sedation and respiratory safety. This does not create a universal inference that OUD treatment must be stopped whenever such exposure occurs; abrupt interruption can itself increase overdose risk. Medication decisions require clinician assessment of interacting risks and continued access to coordinated care. Psychosocial interventions should address additional exposure and distress while preserving the opportunity for effective OUD treatment.

Research programs should distinguish overdose involving multiple substances from toxicology detection of incidental or chronic exposure. Additional sedatives can change the physiological endpoint even if the primary disorder remains OUD. Screening and monitoring improve when they are clinically useful and transparently explained rather than experienced as a mechanism for arbitrary expulsion from treatment.

Pregnancy, Postpartum Care, and Neonatal Outcomes

Pregnancy creates a care context in which maternal stabilization, prenatal engagement, fetal exposure, delivery, and postpartum continuity must all be considered. Methadone and buprenorphine are established treatments for OUD in pregnancy. Medication choice should reflect prior stability, patient preferences, obstetric circumstances, available services, and specialized clinical judgment. Abrupt discontinuation or untreated OUD can expose both pregnant patients and infants to serious harms.

The Medicaid cohort study by Suarez et al. (2022) compared maternal and neonatal outcomes associated with buprenorphine and methadone exposure during pregnancy. Some neonatal outcomes favored buprenorphine in adjusted analyses, but observational medication selection, differences in clinical setting and severity, and residual confounding require careful interpretation. The study does not establish that every patient should switch a successful maintenance regimen when pregnancy begins.

The postpartum period can involve changes in sleep, social support, healthcare coverage, and treatment accessibility, with potential implications for overdose vulnerability. Studies concentrating only on birth outcomes omit the longer period of maternal functioning and survival. An adequate perinatal research architecture includes longitudinal linkage between addiction treatment, obstetrics, pediatric care, and postpartum support while protecting patient agency and confidentiality.

Adolescence, Aging, and Other Population Differences

Age can alter the clinical context of pharmacotherapy through differences in developmental capacity, family support, legal consent, comorbidity, metabolism, and service access. Evidence derived from predominantly middle-aged adult trial cohorts should not automatically be transferred to adolescents, older adults with polypharmacy, or individuals with significant organ impairment. These groups require careful attention to population-specific recommendations and contraindications established by the relevant product and clinical guidance.

For younger patients, family and peer environments shape access, stigma, treatment continuity, and overdose exposure. For older adults, sedative interactions, chronic pain, cognitive burden, falls, and complex medication regimens may shape safety and adherence. The same molecular drug can therefore produce different observed outcomes even when its fundamental receptor action is unchanged.

Sex and gender, racialized access disparities, poverty, rural geography, and disability are also modifiers of treatment opportunity and potentially of observed effectiveness. Researchers should report representation, differential loss to follow-up, and the degree to which measurement instruments behave similarly across groups. Stratified analyses require sufficient precision; post hoc claims should not be mistaken for established biological differences.

Correctional Settings, Reentry, and Treatment Continuity

Incarceration and release alter access to medication, exposure opportunity, social conditions, and physiological tolerance. Treatment interruption followed by re-exposure after release creates a particularly important prevention concern. Research in correctional settings should analyze initiation or continuation during custody, continuity on release, naloxone availability, and transitions into community care. The endpoint should extend beyond a single successful referral.

Institutional policy determines whether pharmacotherapy is accessible, whether existing regimens can continue, and whether transition arrangements survive changes of custody, residence, or insurance. A program may record clinical completion while the released person receives no medication. This difference between administrative completion and therapeutic continuity is central to the interpretation of implementation outcomes.

Ethical evaluation must attend to voluntariness, privacy, discrimination, informed decision-making, and the consequences of coercive conditions. Studies involving constrained choices may provide important evidence about systems, but they should not equate attendance with freely chosen engagement. Mortality and overdose after release require careful outcome surveillance and appropriate comparison groups.

Stigma, Structural Access, and Treatment Equity

Stigma acts on both individual behavior and institutions: patients may delay disclosing opioid use, clinicians may avoid prescribing medication, and policies may impose burdens that obstruct maintenance. The language of treatment can influence these mechanisms. Describing methadone or buprenorphine as merely substituting one addiction for another overlooks their evidence-based therapeutic function and confuses physiological dependence with the diagnostic syndrome of OUD.

Treatment access is a chain of practical conditions: workforce competence, legal authority, clinic hours, distance, transportation, insurance coverage, pharmacy stock, stable contact information, and continuity across systems. A medication can appear less effective in a population where patients repeatedly cannot obtain it, even if drug efficacy is unchanged. Comparative effectiveness therefore requires separating pharmacological response from differential access and analyzing both.

The 2024 US opioid treatment program reforms documented by SAMHSA (2024) are relevant because they modify this chain. Their existence does not quantify a mortality effect; that requires natural-experiment or interrupted-series designs with attention to state variation and contemporaneous interventions. Equity analysis should report who benefits, who remains outside care, and whether reforms change the distribution of protection rather than only its aggregate average.

Methods of Assessing Opioid Use and Treatment Response

Research instruments include structured clinical interviews, administrative diagnoses, opioid use self-report, toxicological testing, craving scales, withdrawal ratings, medication records, treatment-contact logs, overdose reports, and linkage to death registries. Each samples a different aspect of the underlying process. The temporal resolution of testing and the detection characteristics of assays constrain interpretation. A negative urine sample does not by itself demonstrate sustained abstinence, and a medication prescription does not demonstrate ingestion or durable receptor exposure.

Withdrawal instruments such as COWS operationalize a clinical state but are influenced by timing, observer assessment, and the presence of non-opioid symptoms. In fentanyl-exposed cohorts, different thresholds and post-initiation windows can create very different precipitated-withdrawal rates. The contrast between Thakrar et al. (2024) and prospective emergency-department trial reports makes harmonized definitions an important scientific goal.

Mortality registries provide hard outcomes but may miss the nuances of exposure timing, cause-of-death classification, and migration across jurisdictions. Self-report can describe the lived course of craving, stability, and access but is vulnerable to recall and social desirability; confidentiality and non-punitive practices can improve validity. Triangulation is strongest when researchers explicitly model the errors of each measure rather than assuming that several imperfect measures automatically equal a perfect outcome.

Causal Inference, Confounding, and Comparative Estimands

The causal question must specify a treatment strategy, eligibility criteria, time zero, outcome, and follow-up. Assigning treatment at baseline, receiving at least one medication dose, remaining on medication for twelve months, and re-initiating after overdose are different interventions and different estimands. A study of people who successfully complete detoxification estimates effects in a selected group; a study randomizing before induction includes the chance that a medication will never be started.

Observational analyses of maintenance treatment are vulnerable to immortal-time bias if the exposure group is defined by surviving long enough to receive medication. They are also vulnerable to time-varying confounding: recent substance use, withdrawal severity, housing, and healthcare contact can affect the next treatment decision and subsequent outcome. Target trial emulation, appropriate inverse-probability weighting, competing-risk consideration, and explicit missing-data analysis can improve inference when their assumptions are examined.

Confounding by indication can make an apparently safer medication look worse if it is preferentially offered to patients with more severe OUD; selection of stable patients can produce the opposite pattern. The cohort evidence from Larochelle et al. (2018) and Wakeman et al. (2020) should therefore be read as strong, policy-relevant associations, not randomized estimates of a fixed drug's isolated biological effect. Trials like Lee et al. (2018) and D’Onofrio et al. (2026) clarify specific allocation contrasts without resolving every long-term mortality question.

Implementation Science and Service-Level Evaluation

An intervention becomes a public-health treatment only when delivery systems reliably reach people with the condition. Implementation research separates reach, adoption, fidelity, acceptability, feasibility, maintenance, and equitable distribution. For OUD pharmacotherapy, these domains translate into whether a person is offered medication, can actually initiate it, continues to receive it, understands the care plan, and can return after interruption.

The emergency-department trials demonstrate why medication availability and follow-up function as connected interventions. D’Onofrio et al. (2026) did not find a significant difference in short-term engagement between the tested injectable and sublingual strategies; enhanced referral assistance across arms may have affected absolute engagement in both. An implementation replication should report staffing, reimbursement, appointment supply, patient transportation, and the actual proportion completing subsequent visits.

A good service-level evaluation avoids mistaking the number of prescriptions written for reduced population mortality. It tracks a cascade from diagnosis or overdose encounter through medication offer, initiation, continuous exposure, re-engagement, and survival. System changes may produce benefits at one stage while leaving later attrition unchanged; this is a substantive finding that indicates where the next intervention should act.

Economic Evaluation and Resource Allocation

Economic evaluations of opioid pharmacotherapy should consider medication costs, monitoring infrastructure, emergency visits, inpatient admissions, treatment persistence, criminal-legal interactions, patient travel, lost income, and preventable deaths. The relevant perspective—health system, payer, patient, or society—determines which costs appear in the analysis. Cost-effectiveness is therefore a conditional conclusion requiring an explicit analytic horizon, comparator, unit costs, and assumptions about treatment exposure.

Long-acting formulations may reduce the burden of daily dosing for some patients and increase per-unit drug costs for services. Office-based models may improve convenience but perform poorly where pharmacy supply or follow-up support is unreliable. More expensive medication can be economical if it meaningfully improves continuous protection in a suitable group; a cheap product can be ineffective if patients cannot receive it. Both claims require actual comparative data rather than a priori preference.

Because overdose deaths are relatively uncommon in short trials, models often extrapolate from retention or opioid use outcomes to survival. This is reasonable only if the assumed relationship is transparently justified and uncertainty is propagated. Sensitivity analyses should include discontinuation, re-entry, changing fentanyl exposure, intervention reach, and heterogeneous baseline risk. Economic estimates that assume uninterrupted lifelong treatment despite observed attrition may seriously misrepresent real-world resource needs.

International Comparison and Transferability

Opioid pharmacotherapy systems vary by regulatory arrangements, clinical workforce, pharmaceutical availability, drug-market composition, and cultural context. A trial from a specialized US inpatient detoxification setting is not a direct prediction for primary care in a low-resource country; a treatment program based in an urban region may not describe remote populations. International comparisons must separate medication effects from how national institutions arrange care.

The Krupitsky et al. (2011) naltrexone trial studied detoxified participants in Russia, whereas Lee et al. (2018) randomized across US inpatient services and follow-up programs. Both identify important pharmacological properties while sampling different stages of the care process. Meta-analyses combining jurisdictions should test heterogeneity and avoid treating the pooled result as a culturally unconditioned constant.

The WHO (2014) framework for community overdose management is broadly relevant, but implementation depends on naloxone availability, emergency-service response, training, and local drug patterns. Treatment research should therefore collect context variables in a reusable way: legal setting, access rules, predominant opioid exposures, co-use patterns, and the actual continuity of the receiving service. Such information is essential for determining whether an effect can be transported to another setting.

Current Research Frontiers and Therapeutic Innovation

Key frontiers include comparative trials of initiation in fentanyl-exposed populations, longer-acting formulations, service models that enable immediate treatment after overdose, strategies to re-engage people after interruptions, and precise evaluation of treatment in populations excluded from earlier trials. The Shiwach et al. (2025) and D’Onofrio et al. (2026) studies illustrate that an innovation can be pharmacologically plausible yet produce a small or statistically uncertain advantage on a selected primary outcome. Research quality depends on specifying the endpoint before investigating subgroups.

The 2026 network meta-analysis by Rodrigues et al. (2026) broadened comparative synthesis across 103 trials and 13 treatments. Its emphasis on limited certainty for death outcomes points toward a need for linkage of trial cohorts to registries, pragmatic randomization, longer follow-up, and stronger reporting of treatment discontinuation. Network rankings based on heterogeneous treatment structures should not be converted into universal first-line rules, especially where treatments require specialized settings or have distinct indications.

Artificial tools may contribute to appointment reminders, accessible information, service navigation, and analysis of longitudinal care pathways. These applications should be judged by measurable safety, access, fairness, privacy, and retention outcomes. Automatically generated clinical instructions or risk scores cannot replace clinician assessment, and apparently accurate prediction can still fail when it is implemented in a biased service system. The Artificial Era is relevant here as a possible transformation of care infrastructure and research methodology, not as a substitute for controlled clinical evidence.

Critical Scientific Discussion

The most consequential evidence distinction is between pharmacological efficacy conditional on treatment exposure and clinical effectiveness conditional on being able to receive treatment. The antagonist initiation problem in Lee et al. (2018) makes the distinction explicit. The same logic applies to inaccessible methadone programs, delayed buprenorphine appointments, and injectable therapies requiring insurance authorization. Analyses restricted to people successfully treated after selection can overstate the feasibility of the full strategy.

A second distinction concerns mortality. Randomized trial syntheses can be insufficiently powered to demonstrate effects on uncommon fatal events, as reported by Rodrigues et al. (2026), while observational cohorts and meta-analyses such as Sordo et al. (2017) identify important on-treatment survival associations. Scientific synthesis should consider agreement in direction, plausible mechanisms, dose and duration of exposure, bias, and the difference between conditional and population-level outcomes. A trial does not automatically answer every question more accurately than a cohort, and a cohort does not automatically identify causal effects.

A third problem is the instability of context. Fentanyl-era exposure, changes in emergency-service capacity, new product formulations, and revised prescribing rules alter the relevance of historical findings. Revisiting the same clinical question with a new population or service environment can be scientifically necessary. The strongest synthesis is time-indexed: it records what was studied, when it was studied, and what changes would limit transfer to present practice.

Finally, ethical quality and scientific quality are linked. Punitive discharge, inaccessible clinics, coerced treatment, stigmatizing terminology, and deficient follow-up can produce both harm and systematically distorted evidence. Patient participation in study design, clear consent, protection of confidentiality, and attention to lived experience are essential when the intervention requires years of contact with care systems.

Integrated Scientific Synthesis

The evidence supports a coherent model of OUD pharmacotherapy as sustained modification of risk through medication access, receptor-level effects, behavioral stabilization, and institutional continuity. Methadone and buprenorphine are strongly supported maintenance options, with convergent evidence for improved retention and associations with lower overdose and mortality; naltrexone has demonstrated efficacy after successful induction but an important initiation barrier and less certain comparative mortality evidence. These differences are grounded in pharmacology, randomized comparative trials, and real-world cohort studies (Mattick et al. (2014); Lee et al. (2018); Sordo et al. (2017); Larochelle et al. (2018))).

The overdose-prevention component supplies acute rescue to people who may or may not be in sustained treatment. Naloxone distribution, bystander training, emergency response, and post-overdose care form a complementary mechanism for reducing preventable death (WHO (2014); Walley et al. (2013))). The causal pathway continues beyond revival: engagement, medication initiation, continuity of care, and future overdose prevention should be studied as consecutive but distinct endpoints.

The resulting research architecture has three interacting time scales. At the immediate scale, respiratory support and opioid reversal prevent imminent toxicity. At the medium scale, treatment induction and stabilization modulate withdrawal, craving, and exposure. At the long scale, service retention, social context, re-engagement, and population access determine the persistence and distribution of protection. A treatment system succeeds when these scales are connected and measurable, while each retains its own evidence requirements.

Unresolved Problems and Research Agenda

Important unanswered questions include how to achieve safe, tolerable induction across diverse fentanyl exposures; which support models best preserve treatment after emergency visits; how long-acting preparations affect multi-year retention and mortality; how to identify high-risk treatment interruptions without penalizing patients; and how to integrate overdose response with care in rural, correctional, and perinatal settings. These questions call for pragmatic randomized trials, prospective cohorts, registry linkage, and qualitative research with people receiving care.

Measurement remains a central obstacle. Future studies should preregister meaningful outcomes, distinguish initiation from maintenance, harmonize precipitated-withdrawal definitions, report treatment transitions, and measure overdose and survival separately from short-term abstinence. Comparisons should include patient-valued outcomes such as pain, functioning, safety, stable relationships, and the burden of attending care. Adequate sample representation and transparent competing-risk methods matter whenever conclusions address mortality.

The field also requires implementation experiments examining relaxed attendance requirements, pharmacy access, telehealth, peer navigation, harm-reduction distribution, and rapid re-entry after missed care. Population evaluations should account for drug-market shifts and contemporaneous policy interventions. Independent replication, disclosure of commercial interests, public protocols, and access to deidentified analytic data where ethically appropriate will improve confidence in rapidly changing therapeutic claims.

Conclusion

Opioid use disorder pharmacotherapy and overdose prevention form an integrated field in which drug mechanisms, treatment initiation, long-term continuity, and emergency rescue must be evaluated together. Methadone and buprenorphine have substantial evidence as maintenance treatments, including systematic evidence for retention and robust real-world survival associations. Extended-release naltrexone provides a distinct antagonist option whose effectiveness depends on feasible induction. Naloxone and trained community response interrupt acute overdose while creating an opportunity for sustained care.

The central scientific conclusion is that clinical benefit is created through effective exposure to treatment across time. Randomized trials clarify specific intervention contrasts; observational survival studies illuminate longer and rarer outcomes while requiring explicit bias assessment; implementation studies establish whether a treatment reaches those who need it. Modern opioid science advances through their disciplined integration. The next generation of research must make continuity, safety, equity, and patient-valued recovery visible at the same precision as receptor pharmacology and trial endpoints.

References

  • Centers for Disease Control and Prevention. (2024). Opioid use disorder: Treating. https://www.cdc.gov/overdose-prevention/hcp/clinical-care/opioid-use-disorder-treating.html
  • D’Onofrio, G., Herring, A. A., Hawk, K. F., Perrone, J., Cowan, E., McCormack, R. P., Dziura, J., Matthews, A. G., Pantalon, M. V., Owens, P., Martel, S., Coupet, E., Jr., Lofwall, M. R., Walsh, S. L., Edelman, E. J., Carpenter, J. E., Strout, T. D., Baumann, M. R., Anderson, E.,. Fiellin, D. A. (2026). Emergency department–initiated buprenorphine for opioid use disorder: A randomized clinical trial. JAMA, 335(11), 948–960. https://doi.org/10.1001/jama.2025.27019
  • D’Onofrio, G., O’Connor, P. G., Pantalon, M. V., Chawarski, M. C., Busch, S. H., Owens, P. H., Bernstein, S. L., & Fiellin, D. A. (2015). Emergency department–initiated buprenorphine/naloxone treatment for opioid dependence: A randomized clinical trial. JAMA, 313(16), 1636–1644. https://doi.org/10.1001/jama.2015.3474
  • Dole, V. P., & Nyswander, M. E. (1965). A medical treatment for diacetylmorphine (heroin) addiction: A clinical trial with methadone hydrochloride. JAMA, 193(8), 646–650. https://doi.org/10.1001/jama.1965.03090080008002
  • Food and Drug Administration. (2026). Information about medications for opioid use disorder (MOUD). https://www.fda.gov/drugs/food-and-drug-administration-overdose-prevention-framework/information-about-medications-opioid-use-disorder-moud
  • Hsu, H. E., Lodi, S., Yan, S., Bovell-Ammon, B. J., Christine, P. J., Tilhou, A. S., Bernson, D., Novo, P., Lee, J. D., Rotrosen, J., Liebschutz, J. M., Walley, A. Y., & Larochelle, M. R. (2026). Buprenorphine-naloxone vs extended-release naltrexone following opioid withdrawal treatment. JAMA Network Open, 9(7), e2623280. https://doi.org/10.1001/jamanetworkopen.2026.23280
  • Kleinman, R. A., Larney, S., Sule, N. O., Ma, C., Hauck, T. S., & Kurdyak, P. (2026). Methadone vs buprenorphine-naloxone in opioid overdose survivors. JAMA Network Open, 9(8), e2629313. https://doi.org/10.1001/jamanetworkopen.2026.29313
  • Krupitsky, E., Nunes, E. V., Ling, W., Illeperuma, A., Gastfriend, D. R., & Silverman, B. L. (2011). Injectable extended-release naltrexone for opioid dependence: A double-blind, placebo-controlled, multicentre randomised trial. The Lancet, 377(9776), 1506–1513. https://doi.org/10.1016/S0140-6736(11)60358-9
  • Larochelle, M. R., Bernson, D., Land, T., Stopka, T. J., Wang, N., Xuan, Z., Bagley, S. M., Liebschutz, J. M., & Walley, A. Y. (2018). Medication for opioid use disorder after nonfatal opioid overdose and association with mortality: A cohort study. Annals of Internal Medicine, 169(3), 137–145. https://doi.org/10.7326/M17-3107
  • Lee, J. D., Nunes, E. V., Jr., Novo, P., Bachrach, K., Bailey, G. L., Bhatt, S., Farkas, S., Fishman, M., Gauthier, P., Hodgkins, C. C., King, J., Lindblad, R., Liu, D., Matthews, A. G., May, J., Peavy, K. M., Ross, S., Salazar, D., Schkolnik, P.,. Rotrosen, J. (2018). Comparative effectiveness of extended-release naltrexone versus buprenorphine-naloxone for opioid relapse prevention (X:BOT): A multicentre, open-label, randomised controlled trial. The Lancet, 391(10118), 309–318. https://doi.org/10.1016/S0140-6736(17)32812-X
  • Mattick, R. P., Breen, C., Kimber, J., & Davoli, M. (2014). Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence. Cochrane Database of Systematic Reviews, (2), CD002207. https://doi.org/10.1002/14651858.CD002207.pub4
  • National Center for Health Statistics. (2026, May 13). U.S. overdose deaths decrease for third consecutive year in 2025. Centers for Disease Control and Prevention. https://www.cdc.gov/nchs/pressroom/releases/20260513.html
  • Rodrigues, M., Dennis, B. B., Sadeghirad, B., Oprea, A., Malik, M., Naji, L. N., McEvoy, A., Malhotra, D., Parpia, S., Thabane, L., & Samaan, Z. (2026). Pharmacological management of opioid use disorder: A systematic review and network meta-analysis. The Canadian Journal of Psychiatry. Advance online publication. https://doi.org/10.1177/07067437261479071
  • Shiwach, R., Le Foll, B., Alho, H., Dunn, K. E., Strafford, S., Zhao, Y., & Dobbins, R. L. (2025). Comparison of extended-release buprenorphine doses for treating high-risk opioid use: A randomized clinical trial. JAMA Network Open, 8(12), e2548043. https://doi.org/10.1001/jamanetworkopen.2025.48043
  • Sordo, L., Barrio, G., Bravo, M. J., Indave, B. I., Degenhardt, L., Wiessing, L., Ferri, M., & Pastor-Barriuso, R. (2017). Mortality risk during and after opioid substitution treatment: Systematic review and meta-analysis of cohort studies. BMJ, 357, j1550. https://doi.org/10.1136/bmj.j1550
  • Suarez, E. A., Huybrechts, K. F., Straub, L., Hernández-Díaz, S., Jones, H. E., Connery, H. S., Davis, J. M., Gray, K. J., Lester, B., Terplan, M., Mogun, H., & Bateman, B. T. (2022). Buprenorphine versus methadone for opioid use disorder in pregnancy. The New England Journal of Medicine, 387(22), 2033–2044. https://doi.org/10.1056/NEJMoa2203318
  • Substance Abuse and Mental Health Services Administration. (2024). 42 CFR Part 8 final rule. https://www.samhsa.gov/substance-use/treatment/opioid-treatment-program/42-cfr-part-8
  • Thakrar, A. P., Christine, P. J., Siaw-Asamoah, A., Spadaro, A., Faude, S., Snider, C. K., Delgado, M. K., Lowenstein, M., Kampman, K., Perrone, J., Nelson, L. S., & Kilaru, A. S. (2024). Buprenorphine-precipitated withdrawal among hospitalized patients using fentanyl. JAMA Network Open, 7(9), e2435895. https://doi.org/10.1001/jamanetworkopen.2024.35895
  • Wakeman, S. E., Larochelle, M. R., Ameli, O., Chaisson, C. E., McPheeters, J. T., Crown, W. H., Azocar, F., & Sanghavi, D. M. (2020). Comparative effectiveness of different treatment pathways for opioid use disorder. JAMA Network Open, 3(2), e1920622. https://doi.org/10.1001/jamanetworkopen.2019.20622
  • Walley, A. Y., Xuan, Z., Hackman, H. H., Quinn, E., Doe-Simkins, M., Sorensen-Alawad, A., Ruiz, S., & Ozonoff, A. (2013). Opioid overdose rates and implementation of overdose education and nasal naloxone distribution in Massachusetts: Interrupted time series analysis. BMJ, 346, f174. https://doi.org/10.1136/bmj.f174
  • Weimer, M. B., Herring, A. A., Kawasaki, S. S., Meyer, M., Kleykamp, B. A., & Ramsey, K. S. (2023). ASAM clinical considerations: Buprenorphine treatment of opioid use disorder for individuals using high-potency synthetic opioids. Journal of Addiction Medicine, 17(6), 632–639. https://doi.org/10.1097/ADM.0000000000001202
  • World Health Organization. (2014). Community management of opioid overdose. https://www.who.int/publications/i/item/9789241548816