Long-term opioids for chronic musculoskeletal pain may not improve pain-related outcomes compared to a mostly non-opioid medication strategy
EBM Focus - Volume 13, Issue 12
- Evidence evaluating long-term opioid use for chronic musculoskeletal pain is lacking.
- The SPACE trial randomized 240 adults with chronic musculoskeletal pain to opioid vs. mostly non-opioid multi-step medication strategies for 1 year.
- The opioid strategy did not improve pain-related outcomes compared to the mostly non-opioid strategy.
Opioids are often considered to manage chronic musculoskeletal pain in the hopes that their analgesic effects outweigh their adverse effects, but most randomized trials evaluating their efficacy in these patients are ≤ 6 weeks in duration and none are ≥ 1 year (JAMA 2016, Pain Physician 2017). The recent SPACE pragmatic randomized trial evaluated opioid vs. mostly non-opioid medication strategies for 1 year in 240 US Veterans (mean 58 years old, 87% men) who had moderate-to-severe chronic back pain (in 65%) or hip or knee osteoarthritis pain (in 35%) despite analgesic use. Each strategy consisted of 3 steps, with a choice of medications within most steps and more intense medications in each successive step. The steps in the opioid group were 1) morphine immediate release (IR), hydrocodone/acetaminophen, or oxycodone IR; 2) morphine sustained-action (SA) or oxycodone SA; and 3) transdermal fentanyl. The steps in the mostly non-opioid group were 1) acetaminophen or nonsteroidal anti-inflammatory drugs; 2) adjuvant oral medication (such as nortriptyline, amitriptyline, or gabapentin) or topical capsaicin or lidocaine; and 3) pregabalin, duloxetine, or tramadol. Medication adjustments and step changes were based on shared decision-making incorporating patient response, patient preference, and the clinical pharmacist’s judgement. Concurrent non-pharmacological therapies were allowed. Patients and caregivers were not blinded to treatment allocation, but outcome assessors were.
At baseline, fewer patients in the opioid group expressed a preference for opioid therapy (21%) than patients in the mostly non-opioid group (37%) (statistical comparison not reported). Also, more patients in the opioid group dropped out of the trial (19% vs. 8%, statistical comparison not reported); all but 2 patients were included in the intention-to-treat analyses. The authors did not report the distributions of step levels or specific medications used. However, they did report that, during the final 3 months of the trial, 12.6% of patients in the opioid group were taking doses ≥ 50 mg/day morphine equivalent and 11% of patients in the mostly non-opioid group were taking tramadol. Pain-related efficacy outcomes were evaluated with the Brief Pain Inventory (BPI) interference scale assessing the impact of pain on patient function and the BPI severity scale assessing pain intensity. For each scale, a clinically meaningful improvement was defined as a ≥ 30% reduction from baseline.
The opioid medication strategy did not increase the likelihoods of a clinically meaningful improvement in function (in 59% with opioids vs. 60.7% with mostly non-opioids, not significant, 95% CI -14.4% to +11%) or a clinically meaningful reduction in pain severity (41% vs. 53.9%, p = 0.05, 95% CI -25.6% to 0%). Opioids also did not significantly improve most secondary outcomes relating to physical and mental health and quality of life. Opioids were associated with slightly more medication-related adverse symptoms (mean 1.8 vs. 0.9 symptoms on the 19-point Medication-Related Symptom Checklist) and a non-significantly increased risk of visiting the emergency department (50% vs. 39%, 95% CI -2% to +24%), but not significantly increased risks of measures of potential opioid misuse including unexplained positive urine drug tests, clinician-assessed behaviors, or patient-reported substance use.
The lack of patient blinding introduces a potential source of bias, especially considering that more patients in the mostly non-opioid group had expressed a preference to be in the opioid group at the beginning of the trial. Also, although the use of flexible treatment strategies in each group is laudable, the mostly non-opioid strategy in this trial was more flexible in that it allowed the choice of or concurrent use of several classes of medications, which enabled it to target several pathogenic mechanisms, whereas the opioid strategy did not. It is possible that an opioid-based strategy that also allows medications of other classes would be more likely to improve pain-related outcomes than either of the strategies evaluated in this trial. Nevertheless, long-term opioid use has potentially severe adverse effects and are generally not recommended, except as a last resort, for patients with chronic pain of any type (JAMA 2016, Pain Physician 2017). The SPACE trial supports this recommendation by providing high-quality evidence that a long-term opioid strategy may not be more likely to improve pain-related outcomes compared to a mostly non-opioid strategy in patients with chronic back pain or hip or knee osteoarthritis.