Associations Among Sleep, Pain, and Medications for Opioid Use Disorder: a Scoping Review

Abstract

Purpose of Review

We present current evidence on the associations among sleep, pain, and medications for opioid use disorder (MOUD) among individuals with opioid use disorder (OUD).

Recent Findings

We searched MEDLINE, Embase, PsycInfo, Web of Science, and Cochrane Library from inception until September 2023 for original research studies examining sleep, pain, and MOUD. We identified 19 manuscripts (14 were cross-sectional studies, four were prospective cohort studies, and one was a randomized controlled trial). Measures of sleep and pain varied. Sleep disturbance and pain were highly prevalent and associated. However, the associations between MOUD treatment characteristics (e.g., initiation, type, dose, and prior MOUD) and a) sleep and b) pain were mixed or unclear. Limited sample sizes and covariates such as opioid use disorder severity sometimes complicated the examination or interpretation of these associations. Few studies examined possible mediators underlying these associations.

Summary

While sleep and pain were consistently associated, it is unclear whether sleep and pain are associated with MOUD treatment characteristics or other covariates such as opioid use disorder severity. Future research on the associations among sleep, pain, and MOUD among individuals with OUD should consider a) comparing different MOUD treatments including formulations and dose schedules, b) qualitative and mixed methods studies to assess patient and provider preferences for the treatment of sleep and pain in OUD treatment settings, c) longitudinal studies that employ reliable and valid measures with sufficiently powered sample sizes to examine mediation and moderation, and d) testing whether interventions addressing pain or sleep among patients receiving MOUD improve pain, sleep, and MOUD outcomes.

Introduction

Over 6 million people in the United States are estimated to have an opioid use disorder (OUD), and opioid-involved drug overdose remains a leading preventable cause of mortality [1]. Despite the effectiveness of FDA-approved medications for OUD (MOUD) in preventing mortality, treatment initiation and retention remain low and highly variable [2–6]. Pain has received considerable attention since the first wave of the opioid epidemic, which was fueled by the surge in opioid analgesic prescribing for the treatment of chronic non-cancer pain [7–9]. Resultant policy and practice changes have led to reductions in opioid analgesic prescribing as well as debates concerning the undertreatment of pain and its impact on the opioid crisis [10–13]. Indeed, chronic pain (i.e., pain lasting three months or more) is common among patients receiving MOUD, with prevalence estimates exceeding 40% and double that of the general population [14, 15]. In recent years, sleep disturbance (i.e., sleep that is unrefreshing or insufficient in duration, continuity, or timing) has received increasing attention as a highly prevalent problem common to OUD and chronic pain [16–21]. A bidirectional relationship between sleep and OUD has been proposed [22–24], although increasing evidence suggests that sleep may have a greater influence on pain perception rather than the reverse [25, 26]. Pain and sleep disturbance are potential targets for improving MOUD treatment outcomes.

Both sleep and pain have important roles in neurobiological and physiological processes that are dysregulated in substance use disorders [27–29]. Opioids disrupt sleep architecture and increase risk for sleep-disordered breathing [30]. Long-term opioid use, whether in the setting of OUD or chronic opioid therapy, is associated with cellular and molecular adaptations leading to increased pain sensitivity [31]. Sleep disturbance and pain perception are often intensified during opioid withdrawal, and opioid use is subsequently negatively reinforced. While prior reviews have summarized associations between a) pain and sleep [16, 17], b) pain and OUD [10–12], c) sleep and OUD [19], and d) pain, sleep, and chronic opioid therapy [21, 32, 33], to our knowledge, the associations among sleep, pain, and MOUD among individuals with OUD have not yet been examined concurrently. The purpose of this scoping review is to summarize these associations in the existing literature and identify gaps for future research.

Methods

Search Strategy

We conducted this scoping review using the framework described by the Joanna Briggs Institute and in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) checklist [34, 35]. We elected to conduct a scoping review to map the relevant literature on this broad topic. The protocol for this scoping review was registered in Open Science Framework (https://doi.org/https://doi.org/10.17605/OSF.IO/VT2DZ).

A medical librarian (MF) guided the search methodology and performed the literature search.

To maximize comprehensiveness, we first identified keywords and related terms to broadly capture concepts and conditions related to “opioid use disorder”, “pain”, and “sleep.” A medical subject heading (MeSH) analysis of known key articles was also performed. Through an iterative process, the search strategy was developed in the Ovid MEDLINE database (Table 1) and peer-reviewed by an independent librarian using the Peer Review Electronic Search Strategies (PRESS) guidelines [36]. The following databases were formally searched on September 12, 2023: Ovid MEDLINE, Embase (Ovid), PsycInfo (Ovid), Web of Science, and Cochrane Library.

Table 1.

Search strategy on Ovid MEDLINE(R)

1	exp Sleep wake disorders/ or exp Sleep/ or (Arousal or Apnea* or circadian or daytime or day time or Dream* or Dyssomnia* or Hypersomn* or Obesity hypoventilation or Narcolep* or Night* or Noctur* or Parasomnia* or Rapid eye movement* or REM or Restless Leg* or Sleep or Bruxism or Somn* or Wake or Wakefulness).mp
2	exp pain/ or pain.mp
3	exp Opiate Substitution Treatment/ or exp methadone/ or exp buprenorphine/ or naltrexone/
4	((opiate* or opioid* or narcotic* or heroin or fentanyl or carfentanil or tramadol or hydrocodone or hydromorphone or oxycodone or oxycontin or oxymorphone or morphine or opium or narcotic or diacetylmorphine) adj5 medicat*).tw,kw
5	((opiate* or opioid* or narcotic* or heroin or fentanyl or carfentanil or tramadol or hydrocodone or hydromorphone or oxycodone or oxycontin or oxymorphone or morphine or opium or narcotic or diacetylmorphine) adj2 (agonist or substitut* or replac* or maint* or program or treatment or therap*)).tw,kw
6	(methadone or buprenorphine or naltrexone or vivitrol or subutex or suboxone or levomethadyl or methadyl acetate or moud or mouds).tw,kf
7	(medication assisted adj2 (treatment or therap*)).tw,kw
8	(Substance-Related Disorders/ and (Fentanyl/ or exp narcotics/ or morphine/)) or exp Opioid-Related Disorders/ or OUD.tw,kw. or ((drug or substance or heroin or morfin* or morphin* or narcotic* or opiate* or opioid* or opium or oxycodone or oxycontin or oxymorphone or fentanyl) adj5 (habit* or abus* or dependen* or disorder* or addict* or misus* or “use” or “user” or “users” or “using” or overus* or withdrawal or abstinence)).tw,kw
9	3 or 4 or 5 or 6 or 7 or 8
10	1 and 2 and 9
11	10 not (Animals/ not (Animals/ and Humans/))

Study Selection and Inclusion and Exclusion Criteria

We included original research studies published in peer-reviewed journals (i.e., we did not search the grey literature, and we excluded review articles, protocols, editorials, case reports, conference abstracts, etc.). We did not limit the search by year or language. Articles were eligible for full-text review if they investigated both sleep and pain in adult participants (18 years or older) who were engaged in treatment for OUD. By limiting participants to those who were treatment-seeking or in treatment for OUD eliminated the possibility where a diagnosis of OUD was questionable (e.g., opioid “misuse” or “nonmedical use”). We excluded animal/non-human, basic science, and pre-clinical studies.

Search results were pooled in EndNote, deduplicated and uploaded to Covidence for screening. Two independent reviewers (KD and CH) screened the titles and abstracts. Disagreements were resolved through discussion and review by a third reviewer (OJ and DB).

Of the 7,509 articles that were identified, 3,422 duplicates were removed. A total of 4,087 titles/abstracts were screened, and 81 studies were selected for full-text screening. Upon reviewing discrepancies and expert consultation, the research team decided further modification of our research protocol to exclude a) studies where sleep and pain were described only as symptoms of opioid withdrawal and other conditions (e.g., menopause) or as components of broader constructs (e.g., quality of life) and b) studies that did not include participants receiving MOUD or where treatment with MOUD was not clear. Overall, 19 studies were included for extraction (Fig. 1).

Fig. 1.

PRISMA flow diagram of study selection

Data Charting and Extraction

Extracted information included study type, sample size and characteristics, MOUD type and dose, study aim, sleep and pain measures, other variables, analytic methods, and findings. We organized the results from the selected studies in the following manner. First, we summarized the study characteristics and the measures used to assess sleep and pain. Second, we summarized the findings related to the associations among different MOUD characteristics in relation to sleep and pain. Third, we summarized the findings related to the associations between different sleep and pain characteristics and MOUD. Fourth, we summarized the findings related to the effects of sleep or pain interventions among patients receiving MOUD. Finally, we summarized the findings related to covariates and associated outcomes.

Results

Study Characteristics

Descriptive characteristics of all 19 included studies are summarized in Table 2. All studies were quantitative studies: one randomized controlled trial [37], four prospective cohort studies [38–41], and the remaining were cross-sectional studies (n = 14; [42–55]) (see Table 2). Five studies were conducted at a single institution [40, 42–45], and two studies used the same data set [46, 47]. Sample sizes ranged from 23 to 603. Fourteen studies included participants receiving methadone [39–47, 50, 51, 53–55], one study examined participants receiving buprenorphine [37], and four studies included participants receiving either buprenorphine or methadone [38, 48, 49, 52]. Two studies included participants with former but not current receipt of MOUD treatment [45, 48]; one of those studies also included participants with no history of OUD or opioid use as a comparison group [48]. Pooled together, 2,176 individual participants received methadone and 225 received buprenorphine. No included studies examined extended-release naltrexone.

Table 2.

Characteristics of included studies

Study type	Authors, year, [ref], country	Setting	Sleep measures	Pain measures	Summary
RCT	Goyal et al. 2023, [34], India	100 male participants with OUD treated with buprenorphine, PSQI > 5, and no psychiatric comorbidity randomized to receive trazodone (n = 51) or placebo (n = 49) for 6 weeks	PSQI Sleep-50 ESS	BPI	* Trazodone improved sleep (PSQI < 6) in 82.4% vs. 17.6% (placebo) at 6 weeks * Improved sleep was not associated with changes in BPI, ESS, opioid withdrawal, opioid craving, depression, or anxiety
Prospective cohort	Finan et al. 2020, [35], USA	55 participants (25% female) with OUD treated with either methadone (n = 26, 38% female) or buprenorphine (n = 29, 14% female) and followed for up to 16 weeks	Sleep diary Home EEG	BPI	* No significant difference in sleep measures between methadone and buprenorphine when controlling for pain severity and gender * Pain severity was inversely associated with subjective sleep quality
	Nordmann et al. 2016, [36], France	173 participants (16% female) with OUD initiating methadone and followed for 1 year	2 items about sleep disturbance severity	BPI	* Methadone treatment and dose had no significant effect on sleep disturbance * Bodily pain was significantly associated with severe, but not moderate sleep disturbance
	Peles et al. 2011, [37], Israel	23 participants (17.4% female) with OUD initiating methadone and followed for 1 year	PSQI ESS Polysomnography	Peles questionnaire	* No change in chronic pain prevalence, PSQI scores, ESS scores, total sleep time, sleep efficiency, % REM, or % deep sleep after 1 year of methadone treatment * Prevalence of absent deep sleep decreased * New OSA was significantly associated with weight gain, but not methadone dose. No CSA was observed
	Torrens et al. 1997, [38], Spain	135 participants (31.1% female) with OUD initiating methadone and followed for 1 year	NHP - sleep domain	NHP - pain domain	* Sleep and pain significantly improved after 1 year of methadone treatment, with the most significant improvements in the first month * Treatment retention after 1 year (60.7%) was significantly associated with amount and route of heroin use at baseline, but not sleep or pain
Cross-sectional	Peles et al. 2006, [39], Israel	101 participants (21.8% female) with OUD treated with methadone	PSQI	Peles questionnaire	* PSQI scores were significantly associated with chronic pain, methadone dose, duration of OUD, use of benzodiazepines, and psychiatric diagnoses * Methadone dose was significantly higher in chronic pain * PSQI scores and methadone dose were highest among patients who had symptoms of sleep apnea and/or leg restlessness
	Peles et al. 2009, [40], Israel	44 participants with OUD treated with either “high-dose” methadone > 150 mg (n = 25, 12% female) or “low-dose” methadone < 80 mg (n = 19, 27.8% female)	PSQI ESS Polysomnography	Peles questionnaire	* High-dose group had significantly less % deep sleep compared to the low-dose group, with no difference in PSQI scores, ESS scores, sleep efficiency, total sleep time, % wake time, % REM sleep, respiratory events, or chronic pain * Chronic pain was significantly associated with higher PSQI scores, lower sleep efficiency, lower total sleep time, and greater % wake time, with no difference in ESS scores, % deep sleep, or respiratory events
	Peles et al. 2014, [41], Israel	123 participants (27.6% female) with OUD treated with methadone	PSQI ESS	Peles questionnaire	* Chronic pain was significantly associated with higher PSQI scores, but not ESS scores
	Peles et al. 2015, [42], Israel	154 participants with OUD in sustained remission with either current treatment with methadone (n = 55, 34.5% female) or prior (> 10 years ago), but not current treatment with methadone (n = 99, 19.2% female)	PSQI ESS	Peles questionnaire	* Current compared to prior treatment with methadone had significantly higher PSQI scores and greater pain severity, with no difference in ESS scores * Among those without chronic pain, PSQI scores were not different between current and prior treatment with methadone
	Baldassarri et al. 2020, [43], USA	164 participants (40.9% female) with OUD treated with methadone	PSQI ESS Berlin Questionnaire	BPI	* PSQI scores were significantly associated with pain interference with sleep * ESS scores were not associated with chronic pain * PSQI and ESS scores were not associated with methadone dose
	Ponce Martinez et al. 2020, [44], USA	89 participants (37.1% female) with OUD treated with methadone and chronic pain	PSQI	BPI PCS	* Pain intensity, pain catastrophizing, and sleep disturbance were significantly correlated * Pain intensity had a direct effect on pain catastrophizing, but not sleep disturbance * Mediation analyses found that sleep disturbance exacerbated pain intensity indirectly through pain catastrophizing
	Frers et al. 2021, [45], USA	120 participants (60.8% female) with chronic non-cancer pain and divided into 4 groups: 1) OUD treated with methadone (n = 30); 2) OUD treated with buprenorphine (n = 30); 3) OUD with prior (> 6 months ago), but not current treatment with MOUD (n = 30); 4) “ opioid-naïve,” i.e., no history of OUD and no lifetime receipt of opioids > 1 month (n = 30)	PSQI	2 items about pain intensity and interference	* No significant difference in sleep between methadone, buprenorphine, and prior treatment with MOUD (groups 1–3) * Both current and prior treatment with MOUD (groups 1–3) compared to opioid-naïve (group 4) had significantly worse sleep duration, sleep quality, sleep disturbances, and daytime dysfunction * Among participants with prior treatment with MOUD (group 3), there was a significant negative association between weeks from MOUD discontinuation and sleep disturbance
	Dunn et al. 2018, [46], USA	185 participants (52.2% female) with OUD treated with either methadone (n = 125) or buprenorphine (n = 60)	MOS Sleep Scale	BPI	* Methadone compared to buprenorphine was significantly associated with daytime drowsiness, but there was no significant difference in other sleep subscales, pain intensity or interference * Sleep impairment was significantly associated with pain intensity and pain interference
	Stein et al. 2004, [47], USA	255 participants (45.9% female) with OUD treated with methadone	PSQI	1 item about pain interference	* Higher PSQI scores were significantly associated with bodily pain, methadone dose> 100 mg, greater nicotine dependence, depression, anxiety, and unemployment
	Zhong et al. 2019, [48], China	603 participants (30.2% female) with OUD treated with methadone	3 items about frequency of insomnia 1 item about frequency of nightmares	1 question about pain intensity	* Higher methadone dose (> 70 mg) was significantly associated with more frequent nightmares * Frequent nightmares were significantly associated with the use of hypnotics, insomnia, at least moderate pain intensity, and worse functional impairment
	Huffman et al. 2022, [49], USA	38 participants (28.9% female) with OUD treated with either methadone (n = 31), buprenorphine (n = 6), or unknown MOUD (n = l)	PSQI	PDI	* Sleep disturbance significantly positively correlated with pain disability, self-assessed disability, symptom catastrophizing, injustice experience, and negatively correlated with quality of life
	Han et al. 2022, [50], USA	47 participants (23.4% female) older than 50 years old (mean age = 58.8 years, no SD) with OUD treated with methadone	4 items from HRS (insomnia)	2 items from HRS (chronic pain limiting function)	* Older adults receiving methadone for OUD compared to matched 2016 HRS cohort had a significantly higher prevalence of chronic pain, insomnia, and other “geriatric conditions” (e.g., falls, medical multimorbidity, urinary incontinence, hearing and visual impairment) except for perceived functional impairment
	Pud et al. 2012, [51], Israel	73 participants (17.8% female) with OUD treated with methadone	GSDS	3 items about pain intensity and interference	* Forced cluster membership by pain, depression, and sleep disturbance scores into clusters of low, medium, and high symptom burden found that pain most distinguished the three clusters
	Zahari et al. 2015, [52], Malaysia	168 male participants with OUD treated with methadone and no acute or chronic pain, divided into “pain-sensitive” (n = 132) and “pain-tolerant” (n = 36) groups using results from the cold-pressor test	PSQI	Cold-pressor test	* Pain-sensitive group had significantly higher PSQI scores compared to the pain-tolerant group, but there was no significant difference in methadone dose

OUD Opioid use disorder; PSQI Pittsburgh sleep quality index; ESS Epworth sleepiness scale; BPI Brief pain inventory; EEG Electroencephalography; REM Rapid eye movement; OSA Obstructive sleep apnea; CSA Central sleep apnea; NHP Nottingham health profile; MOUD Medication for opioid use disorder; PCS Pain catastrophizing scale; MOS Medical outcomes study; PDI Pain disability index; HRS Health and retirement survey; GSDS Global sleep disturbance scale

Bold indicates primary measures

Sleep Measures

Studies found a high prevalence of subjective sleep disturbance among participants receiving MOUD, ranging from 51.3—90% [39, 42, 44–46, 49–53]. The most frequently used sleep measure was the Pittsburgh Sleep Quality Index (PSQI) (n = 12; [37, 40, 42–48, 50, 52, 55]), a 19-item self-report questionnaire evaluating past-month sleep quality and disturbances, where a global score ranges from 0 – 21, and a cutoff of > 5 indicates poor sleep quality [56]. Additional instruments used to measure subjective sleep quality and disturbances included the Medical Outcomes Survey (MOS) Sleep Scale [57] (n = 1 [49];) and the General Sleep Disturbance Scale (GSDS) [58] (n = 1; [54]). Other sleep instruments used included the Epworth Sleepiness Scale (ESS) to measure excessive daytime sleepiness [59] (n = 6; [37, 40, 43–46]), the Berlin Questionnaire to screen for obstructive sleep apnea [60] (n = 1; [46]), and the Sleep-50 Questionnaire insomnia subscale [61] (n = 1; [37]). One study collected sleep diary estimates [38], while four studies obtained primary sleep data by asking questions or using items from other questionnaires assessing the presence of insomnia (n = 1; [53]), frequency of nightmares and insomnia (n = 1; [51]), and presence (n = 1; [41]) or severity (n = 1; [39]) of sleep disturbance. Only three included studies evaluated objective sleep data (sleep continuity, architecture, and disordered breathing) with polysomnography [40, 43] or electroencephalography (EEG) [38].

Pain Measures

Overall, studies reported the prevalence of pain (i.e., chronic pain, or pain during the time of data collection) among participants receiving MOUD ranged from 37.7% to 65.8% [38–40, 42–46, 51, 53, 54]. Six studies [37–39, 46, 47, 49] used the Brief Pain Inventory (BPI), an 11-item self-rated questionnaire assessing pain intensity and pain interference (i.e., interference in functioning attributable to pain) [62]. One institution used their own questionnaire assessing chronic pain duration and severity [40, 42–45]. Other self-reported measures included the Pain Disability Index (PDI) to assess the extent to which pain interferes with participation in various life activities [63] (n = 1; [52]), the Pain Catastrophizing Scale (PCS) to measure pain-related rumination, magnification, and helplessness [64] (n = 1; [47]), and items taken from other questionnaires assessing pain presence [41], intensity [48, 51, 54], or interference [48, 50, 53, 54]. One study [55] evaluated objective pain sensitivity using the cold pressor test [65].

Associations Among Different MOUD Treatment Characteristics, Sleep, and Pain

The MOUD treatment characteristics examined in the included studies comprised MOUD initiation, MOUD type, MOUD dose, and prior MOUD treatment.

MOUD Initiation and Treatment

Three cohort studies followed patients initiating methadone treatment (MT)¹ for one year [39–41]. Torrens et al. found significant improvements in subjective sleep and pain scores after one year of MT, with the most pronounced improvements occurring by the end of the first month (n = 135) [41]. The two other studies found no significant change in sleep or pain [39, 40]. Peles et al. observed minor improvements in sleep architecture (e.g., significant reduction in the proportion of absent deep sleep) on polysomnography; they otherwise found no change in sleep latency, sleep efficiency, total sleep time, perceived sleep quality or daytime sleepiness, or pain severity after one year of MT (n = 23) [40]. They also observed an increase in obstructive sleep apnea (OSA) incidence that was attributed to weight gain, not MT. Nordmann et al. also found no significant change in subjective sleep disturbance after one year of MT (n = 173); they did not assess the association between pain and MT [39]. None of the included studies investigated buprenorphine initiation.

MOUD Type

Three studies compared sleep and pain between participants receiving methadone or buprenorphine and largely found no significant differences [38, 48, 49]. Specifically, Frers et al. found no significant difference in PSQI subdomains: sleep quality, sleep duration, sleep disturbances, and daytime dysfunction or pain between participants receiving MOUD with methadone (n = 30) and buprenorphine (n = 30) [48]. Finan et al. found no significant differences in sleep diary estimates or home sleep EEG between those receiving methadone (n = 26) and buprenorphine (n = 29) when controlling for pain severity and gender [38]. Dunn et al. found no significant differences in BPI or total MOS Sleep Scale scores between those receiving methadone (n = 125) and buprenorphine (n = 60); however, significantly more methadone recipients reported daytime drowsiness, and there was a trend towards higher pain intensity among those receiving methadone compared to buprenorphine [49]. In the latter two studies, pain intensity was significantly and positively associated with sleep impairment regardless of MOUD type.

MOUD Dose

Multiple studies evaluated the effects of methadone dose on sleep and pain, and the results were mixed or inconclusive [40–43, 46, 50, 51, 55]. An early study by Peles et al. found a linear correlation between methadone dose and PSQI scores (n = 101), and both were significantly higher among patients with chronic pain compared to those without chronic pain [42]. They performed a follow-up study comparing patients receiving high-dose (> 150 mg, n = 25) versus low-dose (< 80 mg, n = 19) methadone and found significantly lower % deep sleep on polysomnography among the high-dose group, but no significant differences in sleep continuity, sleep-disordered breathing, PSQI, ESS, or pain severity scores [43]. They concluded that poor sleep among patients receiving MT was related to chronic pain and other covariates (e.g., benzodiazepine use, psychiatric disorders, and longer duration of OUD), but not methadone dose.

Among patients receiving MT, Baldassarri et al. found no significant bivariate associations between methadone dose and PSQI or ESS scores (n = 164) [46], while Stein et al. found significant associations between methadone dose > 100 mg and PSQI scores in bivariate but not multivariate analyses (n = 255) [50]. Meanwhile, Zhong et al. found significant multivariate associations between methadone dose > 70 mg and frequent nightmares among Chinese patients receiving MT (n = 603) [51]. These studies all found significant multivariate associations between sleep disturbance and pain; they did not assess the association between methadone dose and pain.

Using the cold pressor test [65], Zahari et al. found no significant difference in mean methadone dose between “pain-tolerant” (i.e., able to withstand the cold pressor test by > 37.53 s) and “pain-sensitive” (i.e., cold pressor test < 37.53 s) Malaysian males receiving MT; all participants had no acute or chronic pain, and no regular use of other substances (except nicotine) at baseline (n = 168) [55]. Pain sensitivity was significantly associated with PSQI scores; they did not assess the association between methadone dose and PSQI scores. None of the included studies investigated buprenorphine dose.

Prior (But Not Current) MOUD Treatment

Two studies included participants with a history of MOUD treatment (i.e., prior but not current receipt of MOUD) [45, 48]. Peles et al. found that among participants in sustained remission from OUD, current receipt of MT (n = 55) compared to prior, but not current receipt of MT or other MOUD (n = 99) was associated with significantly higher PSQI scores and pain severity; however, in the absence of chronic pain, sleep disturbance did not differ significantly between the two groups [45]. Among participants with similar levels of chronic pain, Frers et al. found no significant difference in PSQI subdomains between current (n = 60) and prior (n = 30) treatment with methadone or buprenorphine; however, sleep disturbance was significantly negatively associated with weeks following MOUD discontinuation [48]. Both current and past receipt of methadone or buprenorphine (n = 90) was significantly associated with greater subjective sleep disturbance compared to “opioid-naïve” (i.e., no history of OUD and no lifetime receipt of opioids > one month) individuals with chronic pain (n = 30).

Interim Summary

These findings indicate limited evidence supporting associations between MOUD treatment characteristics (initiation, type, dose, and prior MOUD treatment) and either sleep disturbance or pain; observed associations may be due to confounding factors. Sleep and pain appear associated regardless of MOUD treatment characteristics.

Associations Among Different Sleep Characteristics, Pain, and MOUD

As described above, studies examined multiple aspects of sleep (e.g., sleep quality, insomnia, daytime sleepiness, and objective data) with mixed results regarding the associations between sleep and MOUD (Table 3). Poor sleep quality [38, 39, 42–46, 49–52, 55], but not daytime sleepiness [43, 45, 46] appeared consistently associated with pain. On polysomnography, Peles et al. found that patients with chronic pain (n = 18) compared to those without chronic pain (n = 25) had significantly disrupted sleep continuity (e.g., lower sleep efficiency, lower total sleep time, and higher wake time), but exhibited no significant difference in sleep architecture or disordered breathing [43]. Finan et al. found a nonsignificant trend in the associations between lower sleep efficiency, total sleep time, and % deep sleep measured on home sleep EEG and higher pain intensity (n = 49) [38].

Table 3.

Associations among sleep, pain, and MOUD

Author, year, [ref], country	Are MOUD and sleep disturbance associated?	Are MOUD and pain associated?	Are sleep disturbance and pain associated?	Covariates
Goyal et al. 2023, [34], India	Did not measure	Did not measure	No*	N/A
Finan et al. 2020, [35], USA	No	Did not measure	Yes	Gender
Nordmann et al. 2016, [36], France	No	Did not measure	Yes	Nicotine, age, suicidal risk, employment, alcohol, gender
Peles et al. 2011, [37], Israel	No	No	Did not measure	Body mass index, benzodiazepine use
Torrens et al. 1997, [38], Spain	Yes	Yes	Did not measure	Amount and route (intravenous) of heroin use at baseline
Peles et al. 2006, [39], Israel	Yes	Yes	Yes	Benzodiazepine use, psychiatric diagnoses, OUD duration, age, gender, cocaine use, cannabis use, non-prescribed opioid use
Peles et al. 2009, [40], Israel	No	No	Yes	Benzodiazepine use, age, OUD duration, body mass index
Peles et al. 2014, [41], Israel	Did not measure	Did not measure	Yes	Gender, methadone “privileges,” OUD duration, duration of MT, benzodiazepine use, cocaine use, non-prescribed opioid use, cognitive impairment
Peles et al. 2015, [42], Israel	No	Yes	Yes	Psychiatric diagnoses, cognitive impairment, gender, duration of MT, current tobacco use, years of abstinence from non-prescribed opioids
Baldassarri et al. 2020, [43], USA	No	Did not measure	Yes	Somatization, employment, body mass index, use of additional pain medications, age, gender, race, education, history of homelessness, sexual desire
Ponce Martinez et al. 2020, [44], USA	Did not measure	Did not measure	Yes	Depression
Frers et al. 2021, [45], USA	No	Did not measure	Did not measure	Use of sleep medications
Dunn et al. 2018, [46], USA	No	No	Yes	Psychiatric impairment, current opioid withdrawal, negative affect, gender, age, race, employment,
Stein et al. 2004, [47], USA	Yes	Did not measure	Yes	Nicotine dependence severity, depression, anxiety, employment, sedative use, gender, age, race
Zhong et al. 2019, [48], China	Yes	Did not measure	Yes	Hypnotic use, history of compulsory treatment, history of injecting heroin, education, employment, hepatitis B, anxiety, age, gender
Huffman et al. 2022, [49], USA	Did not measure	Did not measure	Yes	N/A
Han et al. 2022, [50], USA	Yes	Yes	Did not measure	N/A
Pud et al. 2012, [51], Israel	Did not measure	Did not measure	Did not measure	Gender, depression
Zahari et al. 2015, [52], Malaysia	Did not measure	No	Yes	OUD duration

MOUD Medication for opioid use disorder; OUD Opioid use disorder; MT Treatment with methadone; ADHD Attention-deficit hyperactivity disorder

Bold indicates an association (bolded covariates are associated with at least one of the variables: sleep, pain, MOUD treatment characteristics)

^*Sample lacked significant pain

Effects of Sleep Interventions

Only one study evaluated the effects of sleep interventions. Goyal et al. found that among male patients with poor sleep quality receiving buprenorphine for OUD in India, flexible-dose trazodone (n = 51) significantly improved sleep compared to placebo (n = 49) [37]. Improved sleep had no effect on pain; however, the mean total BPI intensity and interference scores were only 3.2 (SD = 2.9; range: 0—40) and 0.4 (SD = 0.7; range: 0—70) respectively in this otherwise healthy population with no medical, psychiatric, or other substance use disorders (except nicotine).

Associations Among Different Pain Characteristics, Sleep, and MOUD

As described above, studies were mixed regarding the associations between pain and MOUD treatment characteristics, however, there were consistent associations between different pain characteristics (e.g., intensity, interference, and objective pain sensitivity) and sleep among individuals receiving MOUD (Table 3). Ponce Martinez et al. measured pain catastrophizing and found significant positive associations with depressive symptoms, sleep disturbance, and pain intensity among patients receiving MT with co-occurring chronic pain (n = 89) [47]. Huffman et al. measured pain-related disability and found a significant association with poor sleep quality among employment-seeking participants receiving methadone or buprenorphine (n = 38) [52].

No included studies examined the effects of pain interventions on sleep or other variables among patients receiving MOUD.

Variables Associated with Sleep, Pain, and MOUD

Covariates and Confounders

Some studies drew attention to covariates such as nicotine dependence severity [39, 45, 50], use of benzodiazepines [42, 43], history of injection drug use [41, 51], employment status [46, 50], cognitive status [44, 45], medical and psychiatric comorbidities [42, 45, 49–51], weight [40, 46], age [39, 43], and gender [38, 45, 54]. Covariates were not consistently associated (Table 3). One possible confounder (often unmeasured) is OUD severity. For example, some studies reported associations between duration of OUD [42–44, 55] or take-home “privileges” [44] and sleep, pain, and MOUD. The authors postulated that the apparent effects on sleep, pain, and MOUD may be due to OUD severity, rather than duration of OUD or take-home “privileges.”

Moderators and Mediators

None of the studies specifically examined moderators when examining associations among sleep, pain, and MOUD. Two studies tested variables for mediation [47, 48]. As mentioned previously, Frers et al. found that among participants with chronic pain (n = 120), group (methadone, buprenorphine, “prolonged abstinence,” or “opioid-naïve”) significantly predicted four PSQI subdomains (sleep quality, sleep duration, sleep disturbance, and daytime dysfunction) [48]. Using the Short Form Health Survey (SF-36) [66], they found that general health mediated the relationship between group and both daytime dysfunction and sleep disturbance, while emotional role functioning also mediated the relationship between group and sleep disturbance. Both physical functioning and physical role functioning did not mediate any relationship between group and sleep subdomains.

Ponce Martinez et al. tested mediation pathways between pain intensity, pain catastrophizing, and sleep disturbance in patients receiving MT with co-occurring chronic pain (n = 89) [47]. Pain catastrophizing mediated the relationship between sleep disturbance (exposure) and pain intensity (outcome). Pain intensity did not mediate the relationship between sleep disturbance and pain catastrophizing, nor did sleep disturbance mediate the relationship between pain intensity and pain catastrophizing. Pain intensity was directly associated with pain catastrophizing.

Associated Outcomes: Disability, Functional Impairment, and Quality of Life

While examining associations among MOUD, sleep, and pain, some studies examined associated outcomes, such as disability, functional impairment, and quality of life. Han et al. found that older adults receiving MT (n = 47) had a greater prevalence of medical and psychiatric comorbidities, insomnia, chronic pain, and other “geriatric conditions” compared to matched participants from the Health and Retirement Survey (n = 470) [67], but no difference emerged on perceived functional impairment [53]. Huffman et al. found that among employment-seeking participants receiving methadone or buprenorphine (n = 38), poor sleep quality was associated with greater pain-related disability, symptom catastrophizing, injustice experience, overall disability, and poorer quality of life [52]. Pud et al. performed a forced cluster membership into high, moderate, and low symptom severity by pain, sleep, and depression [54]. Pain, sleep, depression, and quality of life measures were worst in the high symptom cluster, while only pain severity differentiated moderate and low symptom clusters, suggesting pain as the major distinguishing factor in symptom severity and quality of life (n = 73). In sum, patients receiving MOUD have a high burden of sleep disturbance, pain, medical and psychiatric comorbidities, disability, and poor quality of life. The degree of functional impairment may become less disproportionate with aging. The impact that symptom burden has on quality of life might be distinguished by pain severity.

Discussion

This scoping review aimed to identify what is known from the current literature on the associations among sleep, pain, and different MOUD characteristics (type, dose, duration of use, current or past treatment). To our knowledge, this is the first review to have examined these three components concurrently. We identified 19 studies that met criteria for inclusion in this review. Results of our review showed that sleep and pain were highly prevalent among people receiving MOUD. Whether sleep and pain are related to MOUD treatment characteristics remains unclear. The effects of several covariates are also unclear. Regardless of the effects of covariates, there seems to be consistent associations between sleep and pain.

Sleep and Pain Measures

Most studies obtained subjective sleep measures that varied from single items to validated questionnaires and daily sleep diaries. Only one study included bed partner information and found that participants taking the highest methadone doses had symptoms of sleep apnea or sleep-related movement disorders [42]. The few studies that obtained objective sleep data found that some individuals had deep sleep suppression or sleep-disordered breathing [40, 43], which may lead to less restorative sleep. Only one study compared subjective and objective sleep measures and found that patients receiving methadone or buprenorphine perceived their sleep to be better than objective measurements [38]. However, a prior study by Sharkey et al. found good concordance between sleep diaries and home polysomnography [68]. Nonetheless, subjective, objective, and bed partner data offer valuable information in determining the nature of sleep disturbance to guide interventions.

Pain measures also varied across studies. Several studies measured pain intensity; some included pain interference and duration, and there was a lack of other pain characteristics such as location and type (e.g., nociceptive, neuropathic, musculoskeletal, abdominal, etc.). Duration for “chronic pain” also varied or was not clearly defined. Different pain conditions (e.g., fibromyalgia, trigeminal neuralgia) may respond differently to interventions for sleep or pain. A systematic review on positive airway pressure (PAP) therapy in co-occurring OSA and chronic pain found that PAP may reduce chronic headache pain, but there was less evidence for reducing non-headache chronic pain [69].

MOUD Treatment Characteristics

Most studies included participants receiving methadone, and few studies compared buprenorphine and methadone. Importantly, we did not find any studies that examined sleep and pain in participants receiving extended-release naltrexone (XR-NTX) for OUD. One randomized clinical trial in Norway compared sublingual buprenorphine/naloxone (BP-NLX) (n = 79) to XR-NTX (n = 80) and found similar levels of anxiety and depression, but the XR-NTX group had significantly less insomnia compared to the BP-NLX group [70]. Additionally, switching from BP-NLX to XR-NTX did not worsen mild-to-moderate chronic pain [71]. Studies that included participants with prior, but not current MOUD treatment support findings that complete abstinence from opioid agonists improves sleep over time without exacerbating pain [45, 48, 70]. Chrobok et al. found in a cross-sectional study that “recently detoxified” individuals who were at least 14 days abstinent from opioids and not receiving MOUD (n = 27) had significantly fewer sleep disturbances than patients maintained on methadone (n = 84), diacetylmorphine (n = 65), or buprenorphine (n = 23) [72]. While discontinuing opioid agonist treatment may improve sleep without exacerbating pain, this comes with risk for overdose and death and may not be appropriate for many patients who benefit from the treatment. Whether dose reduction improves sleep or pain remains inconclusive; however, observed suppression of REM and deep sleep [40, 43], which has been observed in other studies [73–75], would support restoration of sleep architecture with dose reduction. We did not find any studies that examined the effects of flexible dosing (e.g., divided, nocturnal) of methadone or buprenorphine, which is typically recommended as a pain management strategy [76, 77]. Flexible dosing and continuous formulations may also have varying effects on sleep disturbance, daytime sedation, and pain.

Sleep and Pain Interventions

Our review included only one RCT that evaluated the use of trazodone as a sleep intervention in male patients receiving buprenorphine with no medical, psychiatric, or other substance use disorders, and found that found that trazodone was effective for reducing PSQI scores, but as described above, this sample lacked significant pain [37]. A prior study by Stein et al. found that trazodone did not significantly improve subjective or objective sleep measures in patients receiving methadone, but they did not restrict participation by gender or medical, psychiatric (except psychotic and cognitive disorders), and other substance use disorders (n = 137); pain was not measured [78]. No included studies evaluated pain interventions on sleep and MOUD.

No included studies evaluated non-pharmacologic interventions for sleep or pain. This is a noteworthy omission since behavioral interventions such as cognitive behavioral therapy (CBT) for insomnia [79, 80], CBT for pain [81], and mindfulness-oriented recovery enhancement for chronic pain [82], and adjuvant therapies such as exercise [83–86] and acupuncture [87] have shown potential for patients with substance use disorders, including OUD.

Covariates

Finally, studies summarized significant covariates and other outcomes associated with sleep, pain, and MOUD. Studies concluded that there were significant confounding effects that warrant further exploration. Studies that attempted to minimize the effects of confounding had interesting findings. For example, after excluding participants who were female or had medical or psychiatric comorbidities, including other substance use disorders (except nicotine), the mean BPI scores were low among male patients with sleep disturbance receiving buprenorphine for OUD in India (n = 100) [37].

Only two studies evaluated mediation effects (e.g., pain catastrophizing). Pain catastrophizing may be a potential target for intervention to improve pain intensity [47]. A recent study by Baime et al. found that pain catastrophizing mediated the relationship between chronic pain and insomnia severity, suggesting that targeting pain catastrophizing may also improve insomnia severity [88]. There is some evidence suggesting that pain catastrophizing is associated with risk for opioid misuse [89, 90]. It remains unclear whether interventions for sleep or pain affects MOUD treatment characteristics or other OUD outcomes.

Limitations

There are several limitations to this scoping review. Generally, studies did not examine sleep, pain, and MOUD treatment characteristics equally. Sleep was a primary objective for most articles, and there was a lack of MOUD treatment characteristics as dependent variables, making it difficult to conclude how sleep and pain influence MOUD treatment characteristics or outcomes. The majority of studies had small samples and may not have been sufficiently powered to find associations. Loss of significance after controlling for covariates or performing subgroup analyses may have been due to loss of power due to small sample sizes rather than the confounding effects of covariates. Most studies were conducted in one or few study centers, which limits generalizability of study findings. Seven countries were represented, and there may be political, racial/ethnic, and cultural differences influencing results. Most studies were cross-sectional, and caution must be taken when interpreting the strength or directionality of associations and causal inferences. Furthermore, the variability in measures used for sleep, pain, and other covariates makes it difficult to determine associations.

We excluded studies that examined sleep and pain only during opioid withdrawal. Sleep disturbance and pain are exacerbated and often severe during opioid withdrawal, and the management of opioid withdrawal symptoms has been studied extensively over the past several decades [91–95]. We also excluded adolescents, which is an important and unique population beyond the scope of this review.

Additionally, we excluded studies where a diagnosis of OUD or the effects of MOUD were not clear: studies that described substance use disorders more broadly, identified participants from large health databases using diagnostic codes, or recruited participants with opioid “misuse” or self-reported OUD that could not be confirmed. We reasoned that limiting our inclusion criteria to studies that included participants engaged in treatment and receiving MOUD would provide the most clarity, however, we acknowledge that people with OUD not engaged in traditional treatment settings are an important, large, and understudied population [96, 97]. Additionally, people with opioid “misuse” and “pre-OUD” may offer significant insights into risk factors associated with sleep and pain in developing OUD [98–101].

As is typical in scoping reviews, we did not perform critical appraisals of included studies; the limitations listed above would likely confer a high risk of bias in at least some of the studies.

Future Directions

High-quality studies investigating the influence of both sleep and pain on treatment outcomes among persons receiving MOUD would help advance the field and improve treatment for people with OUD. Studies should consider using reliable and valid measures of pain and sleep; a standardized battery of measures would allow comparisons across studies. At a minimum, pain measures should include a definition for chronic pain, pain duration, location, and type, in addition to pain intensity and interference. Additional measures concerning the management of pain (e.g., pain coping, resilience), psychiatric distress (including anxiety and depression), and quality of life and functioning would also be useful if time permits [102–104]. More objective sleep data is needed, and potentially becoming increasingly accessible with wearable technology. Sleep measures should consider more specific sleep dimensions (i.e., not only sleep quality and insomnia) to assess for sleep-disordered breathing, sleep-related movement disorders, and circadian rhythm disturbances. Assessing for central disorders of hypersomnolence (e.g., narcolepsy) in addition to daytime impairment will be increasingly important as orexin receptor antagonists are currently being studied in this population [105].

Further studies comparing the effects of opioid agonist, partial agonist, and antagonist medications for OUD on sleep and pain are needed. There is also a dearth of literature exploring whether sleep and pain influence MOUD treatment preferences, i.e., methadone vs. buprenorphine vs. extended-release naltrexone vs. other MOUD available outside of the United States. MOUD measures should include not only dose and formulation, but also schedule and frequency of dosing. Qualitative and mixed methods studies are needed to assess patient and provider knowledge, attitudes, and preferences towards the treatment of sleep and pain conditions. Experience sampling methods might be useful to explore the directionality between sleep, pain, and opioid use over a brief duration. Longitudinal studies are needed to explore the natural progression of sleep, pain, and OUD, and changes associated with aging. Patients who achieve prolonged abstinence from nonmedical opioid use may offer insights into protective factors and mechanisms. Larger studies are needed to explore the influence of covariates including sociodemographic characteristics, objective measures (body weight, stress and inflammatory markers), the use of prescribed and non-prescribed psychoactive substances, and medical and psychiatric comorbidities. Finally, studies investigating pharmacologic and non-pharmacologic interventions for sleep and pain are needed. Distress and catastrophizing are potential targets for intervention.

Data Availability

No datasets were generated or analysed during the current study.