Patient-friendly summary
If you read nothing else
Bottom line
Opioids do not broadly improve quality of life in chronic non-cancer pain; the physical benefit is statistically real but clinically small and the mental-health benefit is absent, so opioid use should be individually assessed rather than routinely recommended.
Moderate evidencePublished
Evidence hierarchy
Study participants
Adults with chronic non-malignant pain, including osteoarthritis, chronic low back pain, neuropathic pain, and other conditions such as fibromyalgia
Study Summary
This systematic review and meta-analysis pooled 35 double-blind, placebo-controlled randomized trials (representing 11,057 patients) to ask whether opioid therapy improves health-related quality of life (HRQL) in people with chronic non-cancer pain. The authors found a statistically significant but small advantage for opioids on physical HRQL (SF-36 physical component score mean difference 1.82; EQ-5D mean difference 0.06), which they judged to be below the threshold for clinical relevance. No benefit was seen on the mental dimensions of HRQL (MCS mean difference 0.65, not significant), and in sensitivity analyses excluding designs that favored opioids the physical and EQ-5D benefits weakened or disappeared. The body of evidence was graded low to medium, and the findings apply mainly to the initial months of treatment.
Key Findings
| Finding | Detail | Impact |
|---|---|---|
| Small, clinically unimportant gain in physical quality of life | The meta-analysis of the SF-36 physical component score (18 studies, 7391 patients) showed a statistically significant mean difference of 1.82 (95% CI 1.32 to 2.32, p < .001) favoring opioids, but the authors judged this below the minimal clinically important difference, so it was not clinically relevant. | High |
| No effect on mental quality of life | The SF-36 mental component score (17 studies, 7237 patients) showed no significant difference between opioid and placebo (mean difference 0.65, 95% CI -0.43 to 1.73, p = .24). In a sensitivity analysis excluding designs favoring opioids the effect shifted toward zero (MD 0.07, p = .88), and the osteoarthritis subgroup actually worsened with opioids (p = .048). | High |
| EQ-5D benefit disappears in sensitivity analysis | The EQ-5D meta-analysis (5 studies, 3634 patients) showed a significant mean difference of 0.06 (95% CI 0.00 to 0.12, p = .045) favoring opioids, but excluding withdrawal or enriched designs eliminated the effect (MD 0.03, p = .43). | Medium |
| More dropouts from adverse events on opioids, more from lack of efficacy on placebo | Withdrawals due to adverse events were significantly higher with opioids (risk difference 0.15, p < .001), while withdrawals due to lack of efficacy were higher with placebo (RD -0.11, p < .001). Overall premature withdrawal was slightly higher with opioids (RD 0.04, p = .07) and significantly higher in the osteoarthritis subgroup (RD 0.09, p = .006). Overall dropout was about 41%. | High |
| Vote counting favored opioids on physical measures but was split on mental health | Of 26 studies assessing the SF-36 PCS, 22 showed a positive direction for opioids; of the EQ-5D studies 4 of 7 were positive; and 6 of 7 BPI pain-interference studies were positive. For the MCS, results were nearly evenly split (10 positive, 9 negative, 6 neutral). | Medium |
The meta-analysis of the SF-36 physical component score (18 studies, 7391 patients) showed a statistically significant mean difference of 1.82 (95% CI 1.32 to 2.32, p < .001) favoring opioids, but the authors judged this below the minimal clinically important difference, so it was not clinically relevant.
The SF-36 mental component score (17 studies, 7237 patients) showed no significant difference between opioid and placebo (mean difference 0.65, 95% CI -0.43 to 1.73, p = .24). In a sensitivity analysis excluding designs favoring opioids the effect shifted toward zero (MD 0.07, p = .88), and the osteoarthritis subgroup actually worsened with opioids (p = .048).
The EQ-5D meta-analysis (5 studies, 3634 patients) showed a significant mean difference of 0.06 (95% CI 0.00 to 0.12, p = .045) favoring opioids, but excluding withdrawal or enriched designs eliminated the effect (MD 0.03, p = .43).
Withdrawals due to adverse events were significantly higher with opioids (risk difference 0.15, p < .001), while withdrawals due to lack of efficacy were higher with placebo (RD -0.11, p < .001). Overall premature withdrawal was slightly higher with opioids (RD 0.04, p = .07) and significantly higher in the osteoarthritis subgroup (RD 0.09, p = .006). Overall dropout was about 41%.
Of 26 studies assessing the SF-36 PCS, 22 showed a positive direction for opioids; of the EQ-5D studies 4 of 7 were positive; and 6 of 7 BPI pain-interference studies were positive. For the MCS, results were nearly evenly split (10 positive, 9 negative, 6 neutral).
Strengths
- Included only placebo-controlled, double-blind RCTs, the highest grade of evidence, and pooled a large sample of 11,057 patients across 35 trials
- Prospectively registered (PROSPERO) and reported following PRISMA, with two independent reviewers screening studies
- Combined quantitative meta-analysis with a qualitative vote-counting approach to capture studies that could not be pooled
- Conducted sensitivity analyses excluding enriched/withdrawal designs, funnel-plot and Egger's test for publication bias, and trim-and-fill correction
- Assessed clinical relevance against the minimal clinically important difference rather than relying on statistical significance alone
Limitations
- The body of evidence was graded only low to medium
- High risk of attrition bias, with an overall dropout rate of about 41%, plus frequent use of enriched study designs that can overestimate benefit and underestimate harms
- High clinical heterogeneity across pain conditions and across different opioids (weak to strong), limiting precision of pooled estimates
- Findings apply mainly to the initial months of treatment (median 9.3 weeks); long-term effects on HRQL were not established
- Multi-morbid patients were largely excluded from the included trials, limiting how well results generalize to real-world chronic-pain patients
- Standard deviation of mean change had to be estimated for some studies, and definitions of the minimal clinically important difference varied
Key Takeaways for Patients
What This Means for You
- 01For long-term non-cancer pain, opioids appear to offer only a small improvement in physical quality of life that the researchers judged too small to be clinically meaningful, and no improvement in mental or emotional quality of life.
- 02When researchers removed studies whose design tended to favor opioids, even the small physical and overall quality-of-life benefits faded.
- 03People taking opioids were more likely to stop the medication because of side effects, while people on placebo were more likely to stop because it wasn't helping their pain.
- 04These findings mostly reflect the first few months of treatment, so they do not tell us much about the long-term effects of staying on opioids.
- 05The authors conclude that opioids should not be routinely recommended for chronic non-cancer pain; the decision should be individualized and monitored closely.
Read the Full Paper
Access the complete peer-reviewed study from British Journal of Pain
View Full Study