BACKGROUND

Managing acute pain (such as from surgery or injury) in patients taking medications for opioid use disorder (OUD) can be difficult due to patients’ increased pain sensitivity and higher opioid tolerance, as well as clinicians’ fear that the use of an opioid during an acute pain episode could trigger an OUD relapse.1,2,3 The potential for serious adverse events when methadone and buprenorphine/naloxone (referred to as buprenorphine for brevity) are changed or combined with additional opioids adds to the complexity of acute pain management. Methadone (a full opioid agonist) can cause serious, unpredictable effects such as respiratory depression or overdose when the dose is changed or when other opioids are added.4, 5 Buprenorphine (a partial opioid agonist) partially activates these receptors and reduces the effect of other opioids. Both medications can cause withdrawal when discontinued.4, 6 Naltrexone (an opioid antagonist) blocks the effect of other opioids and is often administered in an extended-release injectable format which cannot be withdrawn or reversed in cases of unexpected acute pain.7, 8

Guidance from professional societies such as the American Society of Addiction Medicine (ASAM) and the Perioperative Pain and Addiction Interdisciplinary Network (PAIN) suggests possible approaches to managing pain in OUD patients taking medications; however, these approaches are primarily based on expert consensus due to the paucity of available research.9, 10 For those taking methadone, the 2015 ASAM guidelines comment that higher doses of full opioid agonists in addition to methadone may be needed to manage acute pain and that short-acting opioids may be needed for those undergoing surgery, citing two observational studies.11, 12 For those taking buprenorphine, ASAM comments that temporarily increasing buprenorphine dosing may be effective for treating mild acute pain, that in cases of severe acute pain buprenorphine may need to be discontinued and replaced with a high-potency opioid, and that if buprenorphine is discontinued before an elective surgery, it should occur 24–36 h beforehand and be restarted when the need for full-agonist opioids has passed. The authors do not cite any specific studies supporting these statements but do cite 2 observational studies indicating that high doses of opioids in addition to buprenorphine may be needed to overcome the blocking effects of buprenorphine.13, 14 By contrast, the 2019 guidelines from PAIN which are exclusively focused on buprenorphine comment that it is rarely appropriate to reduce the dose of buprenorphine before surgery and suggest continuing buprenorphine and adding a full-agonist opioid postoperatively if pain is not being effectively managed, then considering reducing buprenorphine dosage if pain persists. The PAIN guideline,10 which was based on a systematic review,15 acknowledged that there is weak evidence to support these statements. For those taking naltrexone, the ASAM guideline suggests using nonopioid pain management strategies for acute unexpected pain or discontinuing naltrexone prior to surgery. The authors do not cite any studies supporting these statements.

Given the high prevalence of opioid use worldwide (16 million people worldwide use opioids illicitly;16 2 million people in the USA have OUD17), and the number of people taking medications for OUD is increasing every year,18 there is an urgent need to identify evidence-based strategies to manage acute pain in patients taking medications for OUD. In August 2019, the Department of Veterans Affairs Evidence Synthesis Program (VA ESP) conducted a rapid evidence review to inform a VA Health Services Research & Development conference on strategies to improve opioid safety. This manuscript summarizes evidence from that report19 supplemented by an updated search conducted in April 2020. The primary objective of our review was to assess the available evidence on the benefits and harms of different acute pain management strategies for patients taking medications (methadone, buprenorphine, or naltrexone) for OUD, as well as whether benefits and harms vary by the type of medication or type of acute pain (emergency condition vs planned surgery). Because we identified a lack of rigorous evidence on this topic, our secondary objective was to describe the gaps in literature and provide recommendations for future research.

METHODS

The complete description of our methods can be found on the PROSPERO international prospective register of systematic reviews (http://www.crd.york.ac.uk/PROSPERO/; registration number CRD42019132924).

Search Strategy

An information specialist searched MEDLINE, PsycINFO, CINAHL, and Cochrane Central Register of Controlled Trials (CENTRAL) using terms for OUD, medication-assisted treatment, and acute pain from database inception to April 2019 and updated the search in April 2020 (see ESM Appendix A for complete search strategies). Additional citations were identified from hand-searching reference lists and consultation with content experts. We limited the search to published and indexed articles involving human subjects available in the English language.

Study Selection

Study selection was based on the eligibility criteria described below (Table 1). Titles, abstracts, and full-text articles were reviewed by 1 investigator and checked by another. All disagreements were resolved by consensus.

Table 1 Eligibility Criteria
Full size table

Quality Assessment and Data Abstraction

For observational studies with control groups, we used Cochrane’s ROBINS-I tool20 to evaluate the potential for bias from participant selection, classification of interventions, departure from intended interventions, measurement of outcomes, confounding, and missing/unreported data. Overall bias ratings range from low, unclear, to high risk of bias. For observational studies without control groups, we adapted the criteria from ROBINS-I as well as the CARE Checklist21 and focused on the quality of reporting, rather than the potential for bias. Overall quality of reporting ratings ranged from not reported, partly reported, mostly reported, to well reported.

We abstracted data using piloted forms from all studies, including study characteristics, populations, comparators, intervention, and results. All data abstraction and internal validity/quality of reporting ratings were first completed by 1 reviewer and then checked by another. All disagreements were resolved by consensus.

Strength of Evidence and Data Synthesis

We graded the strength of the evidence based on the AHRQ Methods Guide for Comparative Effectiveness Reviews.22 This approach incorporates 4 key domains: risk of bias (includes study design and aggregate quality), consistency, directness, and precision of the evidence. Strength of evidence is graded for each key outcome measure and ratings range from high to insufficient, reflecting our confidence that the evidence reflects the true effect. Due to limited data and heterogeneity, we synthesized the evidence qualitatively. We focused our discussion on the studies that provided the best available evidence (i.e., observational studies with control groups), and discussed lower-tier evidence (i.e., observational studies without control groups) only when it addressed gaps in higher-tier evidence.

RESULTS

Our search identified 324 unique articles, of which 12 articles11, 13, 14, 23,24,25,26,27,28,29,30,31 met our inclusion criteria (Fig. 1). All studies were retrospective (3 studies with a control group and 9 without): 4 studies examined patients with OUD taking buprenorphine,14, 24, 28, 30 5 examined those taking methadone,11, 25,26,27, 31 1 examined a patient taking naltrexone,29 and 2 examined a mixed group of medications.13, 23 Four studies examined those with emergency conditions,14, 24,25,26 7 examined those undergoing planned surgery,13, 23, 27,28,29,30,31 and 1 examined a mixed group of emergency and surgical patients.11 Study size ranged from 1 to 134 participants, and studies were conducted in hospitals/tertiary care centers, specialized pain centers, and outpatient addiction treatment clinics. Follow-up ranged from 1 day to over 2 years. Detailed data abstraction is available in ESM Appendix B and highlighted findings appear in Tables 2 and 3. Detailed quality assessment and quality of reporting ratings are available in ESM Appendix C and overall ratings appear in Tables 2 and 3.

Figure 1
figure 1

Literature flowchart.

Full size image
Table 2 Findings from Studies with Control Groups
Full size table
Table 3 Findings from Studies Without Control Groups
Full size table

Studies with Comparison Groups

Overall, the best available evidence comes from 3 observational studies with comparison groups (Table 2). None of these studies directly compared different pain management strategies; instead, they provide indirect evidence by comparing the same or similar pain management strategies in different populations. One study found that continuing the use of both methadone and buprenorphine in patients after surgery may reduce total doses of opioids and a second found that discontinuing methadone can lead to disengagement from care. A third study confirms that patients taking medication for OUD are opioid-tolerant and may need higher doses of opioid agonists for effective pain control; however, because the study did not report whether medications for OUD were continued or discontinued, it is hard to know in what circumstances high doses of opioids should be prescribed. Our confidence in the findings of these studies is low as studies provided only indirect evidence to address our review question, lacked detailed descriptions of pain management strategies, rarely used validated outcome measurements, and did not look at long-term harms of pain management strategies such as relapse or overdose.

The first observational study with a control group13 compared surgical patients (orthopedic, abdominal, orofacial, thoracic, and other) taking methadone to those taking buprenorphine 1 day after surgery. The study found similar pain management strategies (high doses of morphine-equivalent opioids in the intraoperative period) and outcomes (use of patient-controlled analgesia [PCA], pain severity, and adverse events like nausea, vomiting, and sedation) in both groups. However, only half of the patients taking buprenorphine and three-quarters of patients taking methadone received their dose of OUD medication the day after surgery. Those taking buprenorphine who missed their dose used more PCA for longer than those that did not, and similar trends were found in patients taking methadone. The authors of the study did not know why some patients missed their dose.

The second observational study with a control group23 compared patients undergoing hip or knee joint replacement surgery who were taking medication for OUD (methadone or buprenorphine/naloxone) to non-OUD patients. Patients taking medication for OUD received 8 times the dose of opioids at discharge as non-OUD patients. Both groups had similar pain, functionality, and quality of life outcomes at 6 weeks and 1 year. It was unclear whether OUD medications were continued or discontinued in these patients, as the authors of the study only reported overall doses of morphine-equivalent doses. It was also unclear whether these high opioid doses were titrated down over time.

The third study with a control group11 compared patients with OUD taking methadone with acute pain from surgery or injury (type of surgery or injury not reported) to patients not taking methadone (presumably without OUD but the study does not specify). This study found that when similar doses of opioids were used to manage pain, patients taking methadone had similar number of pain reports but higher rates of behavioral problems and were more likely to discharge against medical advice. Due to limitations in how data on patients’ pain was collected, the authors of the study could not determine why patients discharged against medical advice and the study did not report any follow-up information on these patients (for example, whether they continued OUD treatment in outpatient settings). Previous research has indicated that those with substance use disorders (SUDs) are up to 3 times as likely to discharge against medical advice as those without SUDs, with untreated withdrawal, uncontrolled pain, perception of stigma, and hospital restrictions as potential contributors.32 It is plausible then that patients taking methadone in this study experienced uncontrolled pain or withdrawal symptoms which resulted in them discharging against medical advice; however, other factors (including perceived stigma and hospital restrictions) may have also played a role.

Studies Without Comparison Groups

Additional evidence from 9 observational studies that lacked control groups14, 24,25,26,27,28,29,30,31 (Table 3) address some gaps in evidence, as they looked at additional causes of acute pain (especially emergency conditions), examined naltrexone, and provided more detailed descriptions of the timing, dosage, and sequence of acute pain management strategies. The most notable finding among these studies is that tramadol adequately managed pain in a patient on extended-release naltrexone undergoing a planned surgery.29 However, because all these studies examined small numbers of patients, rarely used measurement tools to assess outcomes, and are by design at high risk of both selection and reporting bias, they do not provide a strong foundation on which to guide clinical decision-making.

Differences by Patient and Intervention Characteristics

It was not possible to determine whether benefits and harms of acute pain management strategies vary by patient characteristics or type of acute pain due to insufficient descriptions of patient populations (especially those with emergency conditions) as well the acute pain management strategies, including adjuvant analgesics.

DISCUSSION

This rapid evidence review synthesizes the available evidence on the benefits and harms of strategies for managing acute pain in patients taking medications for OUD. Given that there are increasing numbers of patients taking medications for OUD, it is critical to identify effective ways to manage acute pain in these patients. Unfortunately, we identified limited research on this topic, with the best available evidence suggesting that continuing buprenorphine or methadone for patients undergoing major surgery may reduce overall opioid doses,13 that discontinuing methadone can lead to disengagement from care,11 and that high doses of opioids may be required to control pain for some (although it is unclear which) patients. Our overall confidence in the findings is low as the 3 best-conducted studies provided only indirect evidence comparing the same or similar pain management strategy in different populations and had substantial methodological limitations, most notably failing to report if OUD medications were continued in one study, rarely using validated outcome measurements, and not reporting on long-term, patient-important outcomes such as OUD relapse.

Our findings on buprenorphine align with a recent systematic review by Goel and colleagues on managing perioperative pain in patients on buprenorphine that informed the 2019 PAIN guidelines,15 although our review identified an additional case study24 that reported buprenorphine needed to be discontinued in a patient with injuries from a motorcycle accident in order to achieve adequate pain control with full-agonist opioids. Although this study was not included in the review by Goel and colleagues, it aligns with the 2019 PAIN guidelines and ASAM guidelines that buprenorphine may need to be discontinued in cases of severe acute pain. Similarly, our findings on methadone align with a recent systematic review by Taveros and colleagues.33 In addition to confirming the results of these reviews, our review adds to the evidence base by looking for studies evaluating naltrexone. Unfortunately, we only identified a single case study, which found that tramadol can be used to manage acute pain in patients taking naltrexone.29

There were limitations of both our rapid review methods as well as limitations of the primary studies we identified. Our rapid review literature search required that a study include the term “acute pain,” which may have missed studies in which it is assumed that the patient is in acute pain but is not described that way (such as management of pain in the “perioperative” period). Additionally, our use of first reviewer inclusion and data abstraction with second reviewer checking may have resulted in missing eligible studies or study data. However, given our results align with two recent systematic reviews on this topic, we believe we identified most of the important data. Additionally, there were significant limitations in both the design and reporting of primary studies that limit our confidence in the findings. First, there were no prospectively designed studies, so evidence is limited to information that researchers collected from data sources not originally designed to address these questions. This was especially problematic in assessments of pain—some studies assessed pain through a scale, but others just noted how often a patient mentioned the word pain or gave a physicians’ overall impression of a patient’s pain. Second, most studies did not thoroughly describe their pain management approach—for example, 2 studies with control groups11, 23 lacked detailed information on whether OUD medication was continued, whether the dosage stayed the same, increased, or decreased, and why those decisions were made.

There are several important gaps in the available literature that should be addressed by future research:

  1. 1.

    There is a need for well-described, prospective studies (such as randomized controlled trials [RCTs] or cohort studies). We did not identify any studies that adequately described a specific pain management strategy and its effects on patient outcomes. Prospective studies with deliberate reporting of intervention elements (such dosage and timing of OUD, dosage and timing of opioid or nonopioid analgesics, and a rationale for why any changes are made) can be more informative than retrospective studies where intervention reporting is often less complete. Randomizing patients into different groups (i.e., an RCT) would provide the most rigorous, defensible answers to what pain management approaches are safe and effective; however, even a well-described, controlled cohort study that adjusts for differences in patient groups at baseline would be a useful step forward.

  2. 2.

    There is a need for studies evaluating approaches to acute pain management when methadone and buprenorphine are continued. Given the best evidence suggests continuing methadone and buprenorphine during episode of acute pain, future research should compare protocols that vary the timing (e.g., dividing doses) and/or doses (e.g., increasing the dose) and evaluate adjunctive opioid and nonopioid pain treatments. Slow-release oral morphine is being used for OUD treatment in some settings as an alternative to methadone and buprenorphine.34 The use of slow-release oral morphine in acute pain management for patients with OUD is another strategy that could be explored.

  3. 3.

    There is a need for studies on acute pain management in patients taking naltrexone. We only identified a single case study on the management of acute pain in a patient taking naltrexone. Research on management of acute pain in these patients is urgently needed.

  4. 4.

    Studies should use validated measurements and measure long-term, patient-important outcomes. Of the best available evidence, only 1 study23 used validated tools to measure quality of life or functionality. All other studies provided limited information on how outcomes were measured, with many commenting that outcomes were retrospectively collected from medical records. We identified no studies that rigorously evaluated patient satisfaction, healthcare utilization (other than length of hospitalization), opioid withdrawal symptoms, substance use relapse, opioid overdose, or suicide ideation or suicidal self-directed violence. Furthermore, with few exceptions, studies ended when patients were discharged, so it is impossible to determine what the long-term effects of pain management strategies were on patients’ health, especially the impact of administering opioids on patients’ likelihood of relapse or overdose.

This review confirmed that there is a lack of rigorous evidence on the management of acute pain in patients taking methadone, buprenorphine, or naltrexone, although the best available evidence suggests that continuing methadone and buprenorphine during episodes of acute pain is preferable to discontinuing these medications. More research is needed that evaluates patient outcomes following well-characterized acute pain management interventions including OUD medication dose and schedule adjustments and use of adjunctive nonopioid pain management strategies.