Retention and dropout from sublingual and extended-release buprenorphine treatment: A comparative analysis of data from a nationally representative sample of commercially-insured people with opioid use disorder in the United States

Abstract

Background and aims:

Maintenance on medications for opioid use disorder, particularly buprenorphine, is critical for reducing overdose risk and improving health outcomes in the United States. This study evaluates retention and dropout probabilities between sublingual buprenorphine (SL-BUP) and extended-release buprenorphine (XR-BUP) among commercially-insured individuals with opioid use disorder (OUD).

Design and setting:

A retrospective cohort study using Meretive^™ Markeskan^® claims data from 2019 to 2020. A multi-state Markov model assessed transitions between treatment states over 12 months.

Participants:

The study included 58,933 individuals aged 18–64 years with OUD, initiating SL-BUP (n = 57,520) or XR-BUP (n = 1,413). XR-BUP patients were divided into XR-BUP only (n = 684; 49 %) and XR-BUP with supplemental SL-BUP (XR-BUP+sSL; n = 729; 51 %).

Measurements:

Primary outcomes included probabilities of remaining in treatment or transitioning between states at 1, 3, 6, and 12 months. The impact of dosage and days of supply on retention was also examined.

Results:

The probability of permanent treatment dropout at 6 months was similar for SL-BUP (38.59 %, 95 % CI: 37.9 %-39.4 %) and XR-BUP (41.3 % 95 %CI: 36.8 %-46.1 %), yet the probability of remaining in treatment was significantly higher for SL-BUP than XR-BUP (49.5 %; 95 %CI: 48.8 %-50.1 % vs. 13.5 % 95 % CI: 10.5 %-16.5 %). The high proportion of individuals initially prescribed XR-BUP later transitioned to SL-BUP. Higher doses and longer days supplied of SL-BUP reduced dropout rates. Among patients receiving ≥16 mg/day and ≥28 days, dropout probabilities were 5.7 % (95 % CI: 5.4 %-6.0 %) at 1 month, 15.4 % (95 % CI: 14.8 %-16.2 %) at 3 months, 28.0 % (95 % CI: 26.9 %-29.2 %) at 6 months, and 47.8 % (95 %CI: 45.2 %-49.5 %) at 12 months. In contrast, patients prescribed <16 mg/day and <28 days had a 46.3 % (95 %CI: 45.0 %-47.6 %) dropout rate by 6 months.

Conclusion:

SL-BUP demonstrates higher retention rates and lower dropout compared to XR-BUP in real-world settings. Optimizing SL-BUP dosing and providing extended supplies can improve retention and reduce treatment discontinuation.

Introduction

Opioid overdose mortality in the United States had been rising steadily before a recent decline, yet it remained high, with 107,543 deaths recorded in 2023, driven by the overprescription of opioid painkillers and the transition to more potent synthetic opioids like fentanyl (U.S. Overdose Deaths Decrease, 2024; Hébert & Hill, 2023; Jalal et al., 2018). Currently, only 41 % of the 2.1 million people in the U.S. who have opioid use disorder (OUD) are commercially-insured (Gupta et al., 2024; Davenport & Matthews, 2018). In 2022, 109,000 people died of drug overdoses in the United States, with nearly 70 % involving synthetic fentanyl (Control CfD, 2024). The current opioid epidemic in the U.S. negatively impacts public health, contributing to decreased life expectancy and the transmission of blood-borne viruses (Hébert & Hill, 2023; Investigation FBo, 2019; Hodder et al., 2021). Additionally, mass incarceration exacerbates the structural vulnerabilities that heighten the risk of overdose, while a punitive approach to drug policy contributes to the structural stigma that prevents people from accessing treatment (Grella et al., 2020; Lopez et al., 2022).

Maintenance on medications for opioid use disorder (MOUD), especially opioid agonist therapies (OAT) like methadone or buprenorphine, is the most effective treatment for OUD (Volkow et al., 2014; Degenhardt et al., 2019). Taking prescribed MOUD and retention in it are key measures of treatment effectiveness for OUD, where longer treatment duration is associated with improved outcomes, including reduced illicit opioid use and lower mortality (Schwartz et al., 2013; Kresina & Lubran, 2011; Mancher & Leshner, 2019; Stone et al., 2021). In contrast, dropping out from treatment substantially increase the risk of death (Sordo et al., 2017). In the North America, Europe, and Australia, among the various MOUD options available, treatment retention is highest for methadone, followed by buprenorphine, while opioid antagonists like long-acting injectable naltrexone have substantially lower retention rates (Morgan et al., 2021; Burns et al., 2015; Hickman et al., 2018; Tucker et al., 2004; Nosyk et al., 2024). SL-BUP’s partial agonist properties and broader accessibility make it a more effective overdose prevention strategy, and it does not contribute to buprenorphine-involved overdose mortality, even after its widespread expansion during the COVID-19 pandemic (Larochelle et al., 2018; Tanz et al., 2023). While only a quarter of people diagnosed with OUD initiate MOUD, buprenorphine is by far the most commonly prescribed MOUD in the United States, despite methadone being the least expensive and the most effective in terms of treament retention (Krawczyk et al., 2022; Hadland et al., 2017; Connock et al., 2007; Farnum et al., 2021). U.S. nationwide, six-month retention for buprenorphine varies between 22 % and 49 % according to claims data, with some studies reporting lower retention rates, particularly among Medicaid enrollees and those treated in primary care or substance use disorder facilities (Xu et al., 2024; Chua et al., 2023; Ivasiy et al., 2024).

In 2017, an extended-release formulation of buprenorphine (XR-BUP) was approved for treatment due to perceived benefits, such as higher tolerability (i.e., no dysgeusia), lower diversion potential, and potentially higher adherence (Altice et al., 2022; Marsden et al., 2023). XR-BUP, however, remains extremely expensive with limited access for most affected communities (Choi et al., 2024; Morgan & Assoumou, 2023). A prospective trial comparing XR-BUP to SL-BUP found similar treatment response (negative urine drug testing), (Lofwall et al., 2018) including among baseline fentanyl users, (Lofwall et al., 2018) but did not report retention and dropout data. No other head-to-head real-world trial has compared these two formulations.

We analyzed a nationally representative sample of commercially-insured patients diagnosed with OUD who initiated treatment with any formulation of buprenorphine. Our goal was to compare the probability of dropout from any type of buprenorphine treatment and the likelihood of remaining on buprenorphine over 12 months based on their original prescription.

Methods

Data source and sampling

The nationally representative Meretive^™ Markeskan^® claims data for commercially-insured individuals in the U.S. were used to assess enrollees aged 18–64 years who were diagnosed with OUD and prescribed any formulation of buprenorphine for chronic maintenance. The claims data consisted of pharmacy, outpatient, and inpatient claims, with each enrollee assigned a unique identifier. All claims included ICD-10 diagnosis and procedure codes, while each pharmacy claim corresponded to a single prescription, detailing the days of buprenorphine supplied and dosing of medication, identifiable using national drug codes (NDC) (See Appendix 2). From the ICD-10 codes, we constructed a cohort of enrollees newly diagnosed with OUD in 2019, using the 2018 calendar year sample as a washout period, to ensure newly diagnosed OUD, patients who received MOUD in 2018 were excluded from the analysis. Among those diagnosed with OUD, we identified individuals initiating and MOUD using procedure codes for methadone, XR-BUP, and extended-release naltrexone (XR-NTX), and NDC for SL-BUP, XR-BUP, and XR-NTX. The sampling methodology and approach to MOUD identification have been detailed in a previously published study (Ivasiy et al., 2024). Further analyses were restricted to individuals prescribed any formulation of buprenorphine. These enrollees were followed for another calendar year until the end of 2020, ensuring at least one year of follow-up for each enrollee in the sample. A small proportion (<0.03 %) of enrollees who transferred between different types of MOUD during the observation period were not included in the analysis.

Measures

Standard SL-BUP is taken daily, with maintenance doses typically ranging from 8 to 24 mg daily, and can be provided as take-home dosing for up to 28 days (Substance Abuse & Mental Health Services Administration). XR-BUP is administered as a monthly subcutaneous injection (300 mg for the first two months, then 100 mg monthly) (Indivior Inc, 2023). During each follow-up period after the initiation of buprenorphine, enrollees could be in one of five mutually exclusive states: interim off-treatment followed by re-enrollment into treatment, permanent off-treatment, sublingual buprenorphine only (SL-BUP only), XR-BUP only, or receiving both formulations simultaneously. In the latter case, sublingual buprenorphine was considered supplemental to XR-BUP (XR-BUP + sSL) if it was dispensed five or more days before the last day of XR-BUP supply. If the period between the last day of any supply and the next prescription was 10 days or more, it was coded as an interim off-treatment period followed by the new treatment episode. Episodes with interruptions of <10 days were considered continuous, uninterrupted treatment. Permanent dropout was defined as the absence of return throughout the entire follow-up period. Each newly enrolled individual was tracked for 12 months, and those who remained in care beyond 365 days were censored. For each enrollee, we calculated the baseline days of supply and baseline buprenorphine dosage, categorizing them as either <28 days or 28 days and more, and <16 mg or 16 mg and more, respectively. The prescribing information package insert states the average steady-state concentration of both the 100 mg and 300 mg doses of XR-BUP is higher than 16 mg of SL-BUP (Indivior Inc, 2023; Rutgers Medication-Assisted Treatment Center of Excellence). We compared states based on demographic and clinical characteristics in claims data, including sex, mean age, dichotomized age distribution by lowest quartile (≤29 vs. >29 years), regional classification (Northeast, Midwest, South, West), and the prevalence of comorbid HIV and HCV diagnoses.

Statistical analysis

A time-homogeneous multi-state transition Markov model using five mutually exclusive states was constructed. A multi-state model captures complex treatment pathways by modeling transitions between different treatment states (Boucherie et al., 2015). This approach allows for a more nuanced understanding of the treatment process, as it can handle complex scenarios where individuals move through various states, offering greater flexibility and insight compared to simpler models (Laird et al., 2013). The analysis examined 14 potential transitions, including reversals, between different states. Results are presented using the probability of transition (PT) and probabilities of remaining (PR) in each state at 1, 3, 6 and 12 month periods, along with their 95 % confidence intervals (95 % CI) calculated by usuing bootstrap resampling method, which generates an empirical distribution of the estimates (Jackson, 2011). The impact of dosage and days of supply on these transitions was assessed using a proportional intensities model. The covariates in the final model were selected via forward selection, comparing −2 log-likelihood values (likelihood ratio test). Model fit was assessed by visually comparing the observed and predicted state prevalences over designated time intervals. Analyses were conducted using SAS 9.4 (SAS Institute, Inc., Cary, NC, USA) and R version 2022.02.1 package ‘msm’ version 1.7.1 (Jackson, 2011).

We weighted the MarketScan sample to align with the national Employer-Sponsored Insurance population using demographic strata from the American Community Survey Public Use Microdata Sample (United States Census Bureau, 2023). This adjustment ensures that healthcare utilization estimates more accurately reflect the broader ESI population. During the weighting process, we identified up to 5 % of missing records for the weights. Since these missing values appeared to be missing at random, we addressed them using multiple imputation techniques to maintain the integrity and reliability of the weighted estimates. More details on the weighting methodology approach are described elsewhere (Ivasiy et al., 2024).

Results

Table 1 presents the weighted sample characteristics stratified by buprenorphine formulation (SL or XR). The sample comprises 58,933 individuals, categorized into two primary groups: SL-BUP only (N = 57,520; 97.6 %) and XR-BUP (N = 1733; 2.4 %). Of those on XR-BUP, they were further categorized into those who received only XR-BUP (N = 684; 39.5 %) and those with supplemental sublingual (XR-BUP+sSL): N = 729; 60.5 %). The sample was comprised mostly of men (64.9 %) in their late 30 s (mean=39±11 years). Regionally, the largest proportion of enrollees were in the South region (38.5 %), followed by the West (24.0 %), the North Central (19.4 %), and the Northeast (18.1 %). In terms of comorbid diagnoses, the prevalence of HCV was 4.4 %, which was similar across all prescribed BUP formulations, while HIV was 0.2 % overall and highest in the XR-BUP + sSL group (0.8 %). HIV and HCV co-infection was 0.1 % and similarly distributed in all those prescribed BUN. For the initial prescription, over half (52.7 %) of the sample was prescribed <16 mg of buprenorphine, and similarly, 59.1 % was prescribed <28 days of supply.

Table 1.

Sample characteristics stratified by type of BPN-based MOUD (weighted).

		Total		By Tretment Modality
				SL-BUP		XR-BUP
						Total		By Suplemetal Intake
								XR only		XR+sSL
		58,933 (100 %)		57,520 (97.6 %)		1413 (2.4 %)		684 (48.4 %)		729 (51.6 %)
Mean Age, years ±SD		39 ± 11		38 ± 11		35 ± 10		34 ± 10		36 ± 10
		N	%	N	%	N	%	N	%	N	%
Sex	Male	38,240	64.9	37,181	64.6	1059	74.9	537	78.5	522	71.7
	Female	20,693	35.1	20,339	35.4	354	25.1	147	21.5	206	28.3
Age, years	≤29	13,898	23.6	13,416	23.3	483	34.2	286	41.8	197	27.0
	>29	45,034	76.4	44,104	76.7	930	65.8	398	58.2	532	73.0
Region	Northeast	10,660	18.1	10,403	18.1	257	18.2	154	22.5	103	14.1
	North Central	11,442	19.4	11,185	19.4	258	18.3	90	13.2	168	23
	South	22,673	38.5	22,216	38.6	456	32.3	192	28.1	264	36.3
	West	14,158	24	13,715	23.8	442	31.3	248	36.3	194	26.6
Comorbid Diagnoses	HCV	2599	4.4	2567	4.5	32	2.3	0	0	32	4.4
	HIV	130	0.2	124	0.2	6	0.4	0	0	6	0.8
	HIV&HCV	53	0.1	48	0.1	5	0.4	5	0.8	0	0
Baseline Dose,mg	<16	30,984	52.7	30,303	52.7	772	54.6	393	57.0	379	52.0
	≥16	27,857	47.3	27,217	47.3	640	45.3	290	42.4	350	48
Baseline Days of Supply	<28	34,798	59.1	34,091	59.3	799	56.5	384	56.1	415	56.9
	≥28	24,042	40.9	23,428	40.7	614	43.5	300	43.9	314	43.1

BUP: Buprenorphine; SL: Sublingual; XR: Extended-Release; sSL: Supplemental sublingual.

Fig. 1 (see also Appendix 1 for table format) illustrates the estimated probabilities of transitioning between treatment states and remaining in a given state over 1, 3, 6, and 12 months, with 95 % confidence intervals adjusted for sex, age, dosing, and days of supply. The figure presents transitions between XR-BUP, SL-BUP, XR-BUP + SL-BUP, Interim Off-Treatment, and Permanent Off-Treatment, highlighting the dynamic nature of treatment retention. Notably, XR-BUP has a high probability of treatment retention at 1 month (69.8 % [66.4–72.6]) but declines significantly over time, reaching 13.5 % [10.5–16.5] by 6 months. In contrast, SL-BUP demonstrates greater adherence beyond 1 month (82.1 % [81.8–82.3]) and at 12 months (30.9 % [30.1–31.6]). Transition probabilities to Interim Off-Treatment are higher for XR-BUP than for SL-BUP.

Fig. 1.

Estimated probabilities of transition (PT) and probabilities of remaining on treatment (PR) for a 1, 3, 6 and 12-month period and their 95 % confidence intervals. PT and PR presented in this model were adjusted to baseline BUP dose and Days of Supply, Age and Sex.

Fig. 2 compares treatment adherence by evaluating the probability of remaining on the prescribed treatment (SL-BUP vs. XR-BUP). At all time points, SL-BUP retention is significantly higher: 1 month (82.1 % [81.8–82.3] vs. 69.8 % [66.4–72.6]), 3 months (64.2 % [63.7–64.6] vs. 35.7 % [30.1–40.5]), 6 months (49.5 % [48.8–50.1] vs. 13.5 % [10.5–16.5]), and 12 months (30.9 % [30.1–31.6] vs. 2.2 % [1.3–3.2]).

Fig. 2.

Probability remaining on the buprenorphine treatment (sublingual vs long-acting injectable) at various time points over 12 months.

Fig. 3 reflects a real-world finding that patients transition from one treatment to another due to a multitude of factors. This figure shows the probabilities of permanent dropout from any type of buprenorphine-based maintenance treatment at 1, 3, 6, and 12 months controlled by dosing and days of supply. The Part A of the figure pertains to BUP formulation factors while in the Part B are the estimates stratified by factors related to prescription of SL-BUP (dosage and days supplied). Enrollees prescribed XR-BUP with sSL have a significantly lower probability of dropout at 1 month but are similar between BUP formulations thereafter. When enrollees are simultaneously prescribed both formulations, the probability of dropout increases over time: 1 month (P [95 % CI]: 5.0 [4.3–6.0]), 3 months (19.9 [17.6–22.9]), 6 months (37.4 [34.3–40.6]), and 12 months (60.9 [58.5–63.7])..

Fig. 3.

Probabilities of permanent dropout from different types of buprenorphine-based medication-assisted treatment at 1, 3, 6 and 12 months.

The probability of permanent dropout is similar for SL-BUP and XR-BUP: 1 month (8.6 [8.4–8.8] vs. 10.0 [8.3–12.3]), 3 months (22.4 [21.9–22.9] vs. 24.9 [21.1–29.6]), 6 months (38.6 [37.9–39.4] vs. 41.3 [36.8–46.1]), and 12 months (61.5 [60.6–62.4] vs. 63.5 [59.9–67.4]), with no statistically significant differences.

For sublingual buprenorphine, factors controlled by prescribers (i.e., dosage and days supplied) significantly impact dropout probability. Longer days supplied and higher dosages reduce dropout risk the most (see Fig. 3, Part B). Among patients receiving ≥16 mg of SL-BUP with a ≥ 28-day supply, the probability of dropout is 5.7 % [5.4–6.0] at 1 month, increasing to 15.4 % [14.8–16.2] at 3 months, 28.0 % [26.9–29.2] at 6 months, and 47.8 % [46.2–49.5] at 12 months. Notably, the 12-month dropout probability for this group is similar to the 6-month probability for those prescribed <16 mg/day with a < 28-day supply (46.3 % [45.0–47.6]).

Discussion

To our knowledge, this is the first real-world study to compare the probability of remaining on the originally prescribed treatment as well as dropout probabilities in patients with OUD after initiating various formulations of buprenorphine, the most prescribed MOUD in the United States. This study provides a comprehensive comparative analysis of the retention and dropout probabilities associated with different formulations of buprenorphine among a nationally representative sample of commercially-insured people diagnosed with OUD. In this study, we use methodologies for identifying real-world dropout from any type of buprenorphine treatment, given that patients may transition to differing treatment modalities for a variety of reasons, and identify the probabilities of dropout from the treatment option they were initially prescribed. The findings contribute to the ongoing discourse on the effectiveness of XR-BUP compared to the more commonly and substantially less expensive prescribed sublingual buprenorphine.

Our results reveal significant differences in treatment adherence and dropout rates between XR-BUP and SL-BUP. While XR-BUP is designed to overcome barriers to daily medication adherence through reduced administration frequency, it demonstrates lower retention rates compared to SL-BUP. The 3-month adherence data for XR-BUP (36 %) in our study was substantially lower than the 50 % retention rates reported from 2018 data in commercially-insured patients, (Morgan et al., 2021) though the 3-month adherence in our study was similar for SL-BUP retention rates reported before (64 %). The 2018 sample, however, likely had better retention than our sample as these were early adopters of XR-BUP during the first year after its approval, and the later data reflects less of a “honeymoon effect”. Early adopters are often more enthusiastic about new treatments, driven by a sense of novelty and optimism regarding the innovation’s potential to address unmet clinical needs (Knudsen & Roman, 2015). This early enthusiasm can lead to greater adherence and higher retention rates, as both patients and providers invest more effort into ensuring the success of the new treatment.

Though reasons for discontinuing XR-BUP are not known, lessons from other extended-release medications for psychiatric or substance use disorders may be informative. For example in extended-release antipsychotic medications, the main drivers associated with discontinuation include adverse side effects and suboptimal efficacy (Auxilia et al., 2023). In the case of XR-naltrexone, an opioid antagonist used to treat OUD, treatment discontinuation was high, primarily due to patients’ unfulfilled expectations about its treatment efficacy (Brenna et al., 2022). Patients who were initially prescribed XR-naltrexone, however, switched to an opioid agonist like methadone or buprenorphine. In the current study of XR-BUP, many patients who discontinued it switched to another formulation of buprenorphine – SL-BUP.

Opioid use disorder is a chronic relapsing disease and requires sustained treatment. In the case of XR-BUP, the probability of remaining on treatment is negligible (2 %) by 12 months. Moreover, among those who initiated XR-BUP, over 50 % were simultaneously prescribed supplemental sublingual buprenorphine. While the standard formulation of XR-BUP translates to 16 mg per day of SL-BUP, the supplemental treatment likely reflects the need for higher doses as many initiates may have been using fentanyl before initiating treatment with buprenorphine. This suggests that despite the theoretical advantages of XR-BUP, such as decreased outpatient visits and lower risk of misuse, (Chappuy et al., 2020) these benefits may not translate into long-term treatment adherence in a real-world setting. In contrast, SL-BUP shows higher retention rates than XR-BUP, even at 12 months, indicating that patients are more likely to continue treatment when prescribed this formulation. Additionally, the intake of supplemental sublingual buprenorphine during XR-BUP administration suggests that XR-BUP levels may be insufficient to fully alleviate withdrawal symptoms in some patient groups or, as in the case of XR-naltrexone, have unfulfilled expectations of efficacy.

The findings of this study align with previous research highlighting the higher retention rates of SL-BUP compared to XR-BUP (Morgan et al., 2021; Brenna et al., 2022; Chappuy et al., 2020; Morgan et al., 2021). The dropout rates from XR-BUP observed in our study align with those reported in clinical trials focused on patient-centered outcomes for individuals receiving XR-BUP (Lofwall et al., 2018). Unlike retention analyses in clinical trials that typically use an intention-to-treat approach without accounting for state transitions, this study provides a more detailed examination of these transitions. Furthermore, given that randomized controlled trial (RCT) conditions may not fully represent real-world outcomes, comparing RCT findings with results from a large claims data sample offers valuable insights, while acknowledging the limitations associated with this type of analysis. In addition to XR-BUP not improving retention in care, it is extremely costly and increasingly recognized as a less cost-effective strategy relative to sublingual buprenorphine (Flam-Ross et al., 2023). Real-world studies analyzing commercial and public insurance claims datasets have found that 6-month retention rates for buprenorphine treatment vary widely, ranging from 22 % to 50 %, influenced by insurance type, co-occurring disorders, and demographic factors (Morgan et al., 2021; Xu et al., 2024; Chua et al., 2023; Ivasiy et al., 2024; Ijioma & Chilcoat, 2023; Xu et al., 2023). In contrast, our cohort focuses exclusively on the commercially-insured population, where the overall buprenorphine 6-month retention rate is 56 %, as reported elsewhere (Ivasiy et al., 2024).

A more nuanced understanding of prescribed dosage and number of days supplied of sublingual buprenorphine is crucial. Importantly, dosages exceeding 16 mg per day and providing a supply of over 28 days of medications were associated with lower dropout rates over the course of an entire year. Relative to XR-BUP, supplying medication beyond 28 days conferred a retention benefit at 6 months irrespective of the dose prescribed. One potential explanation is that attending frequent appointments undermines patients’ ability to succeed in their treatment by placing unnecessary and cumbersome demands on their time. Transportation time, taking time off from work, or leaving caregiving responsibilities often compete with frequent medical appointments required by clinicians (Godersky et al., 2019). Our study’s follow-up period includes the COVID-19 pandemic in 2020, a time when changes in healthcare delivery, increased stressors, and the expansion of telehealth services significantly impacted buprenorphine retention rates in the U.S. While potential job losses during the early pandemic threatened healthcare coverage and access to medications for opioid use disorder, evidence suggests that new federal- and state-level policies helped stabilize buprenorphine use among those with employer-based insurance (Cantor et al., 2021). Additionally, some studies indicate that telehealth-based buprenorphine treatment improved early retention by reducing barriers related to transportation, stigma, and clinic availability (Al Faysal et al.; Lewis et al., 2024). These findings together point to the need to change the way that treatment is delivered that optimizes treatment dosage, especially when highly potent fentanyl is common, and minimizes clinical demands on patients, many of whom are able to self-manage and remain engaged in care. Many prescribers believe that in order to continue treatment with buprenorphine patients must be fully accountable. To balance this accountability, some have advocated for video-observed therapy (VOT) to enhance trust by providers (Hallgren et al., 2022). However, some argue that VOT may be too cumbersome and time-consuming for clinicians, potentially eroding trust from patients who may find it intrusive, while clinicians might also feel that the lack of urine drug testing limits their ability to fully trust the results (Godersky et al., 2019). Pilot studies of VOT for SL-BUP, however, have not resulted in improvements in illicit drug use and treatment engagement, making it a non-feasible and costly strategy (Tsui et al., 2021).

The benefits of optimal dosing and extended take-home dosing have been reported in several global settings, including both high and low- and middle-income settings, (Farnum et al., 2021; Ivasiy et al., 2024; Ivasiy et al., 2022; Fareed et al., 2012; Amram et al., 2021) yet have not translated to major shifts in clinical practices. Our findings suggest that while optimizing dosing strategies could be crucial for improving treatment retention and reducing the risk of discontinuation, a supply of 28 days or more had an even stronger association with retention, particularly when patient visits were minimized.

These results have important implications for clinicians and policymakers. While the study cohort included patients treated with both generic and original formulations of sublingual buprenorphine (see Appendix 2 for details), generic SL-BUP costs approximately $864 to $920 annually, whereas Sublocade, the extended-release formulation, costs about $6,600 per year (Canadian Agency for D, Technologies in H. Pharmacoeconomic Review Report: Buprenorphine extended-release injection (Sublocade): (Indivior Canada, Ltd.) 2019). Given the high cost of XR-BUP and its relatively lower retention rates, it may be necessary to reconsider its widespread use, especially in populations where long-term adherence is critical. XR-BUP, however, can be beneficial for certain patient sub-populations, particularly those nearing release from correctional facilities, as it can improve the flexibility of engaging with patients during the transition from incarceration (Lee et al., 2021). Its use at the time of release, however, is unlikely to have long-term benefits unless a plan is set up to allow flexible shifts from XR-BUP to SL-BUP, which is common. The findings also support the continued use of SL-BUP as a primary treatment option for OUD, particularly in settings where extended-release formulations may not be feasible or cost-effective.

This study has several limitations. First, the use of insurance claims data restricts the generalizability of findings to the broader population of OUD cases, particularly those who are uninsured or covered by Medicaid. However, these data still offer a comprehensive real-world perspective on individuals with employer-based insurance. Second, insurance claims data capture prescription fills but do not account for missed doses, meaning actual buprenorphine use may differ from what is recorded. Nonetheless, prescription fills remain a widely used and reliable proxy for treatment engagement in real-world settings. Third, the lack of detailed patient-level data limits the ability to assess factors such as socioeconomic status and the motivations of both patients and prescribers in choosing XR-BUP versus sublingual buprenorphine. Fourth, individuals obtaining buprenorphine outside the prescription drug benefit, such as through cash payments, were not included, potentially underestimating MOUD uptake. However, insurance claims data still capture the majority of prescribed treatments, reflecting key trends in clinical practice. Finally, although formulations intended solely for pain management were excluded, the analysis may still capture individuals prescribed buprenorphine for OUD but using it off-label for pain treatment. Despite these limitations, this study provides valuable real-world insights into the comparative use of sublingual and extended-release buprenorphine formulations in the U.S., contributing to a broader understanding of treatment patterns in routine clinical care.

Conclusion

Extended-release buprenorphine offers certain advantages in terms of reduced administration frequency and lower misuse potential, but its substantially lower retention rates compared to SL-BUP raises questions about its long-term effectiveness in real-world settings. Moreover, optimizing dosage and providing an extended amount of SL-BUP is substantially superior to XR-BUP and to lower dosages and fewer days supplied, pointing to the need for practice transformation to reduce clinical demands on patients. Future research should explore the underlying factors contributing to these differences and examine the potential benefits of combined or tailored treatment approaches to improve retention and overall treatment outcomes for individuals with OUD.