Objective: For chronic pain treatment many health care authorities consider morphine to be the reference standard for strategic decisions in pain therapy. Although morphine's effectiveness is clear and its cost is low, it's unclear whether morphine should remain the first choice or reference treatment.
Research design and methods: We performed a systematic review to evaluate the evidence available to support the position of morphine as the reference standard for step III opioids based on efficacy and tolerability outcomes.
Results: The search yielded 5,675 titles and 56 studies were included. Considerable heterogeneity precluded pair-wise meta-analysis on change of pain intensity and no difference between morphine and other opioids were found for tolerability outcomes. The network meta-analysis showed no statistically significant difference in change of pain intensity between morphine and oxycodone, methadone and oxymorphone. Compared to morphine, patients using buprenorphine are more likely to discontinue treatment due to lack of effect (OR 2.32, 95% CI 1.37 to 3.95). Patients using methadone are more likely to discontinue due to adverse events (OR 3.09, 95% CI 1.14 to 8.36), whereas this risk is decreased for patients using fentanyl (OR 0.29, 95% CI 0.17 to 0.50) or buprenorphine (OR 0.30, 95% CI 0.16 to 0.53). The most important limitation of this review is that the included studies are heterogeneous with regard to study population and intervention, which may affect the pooled effect estimates. The main strength is that we only included parallel RCTs, the strongest design for intervention studies.
Conclusions: The current evidence is moderate, both in respect to the number of directly comparative studies and in the quality of reporting of these studies. No clear superiority in efficacy and tolerability of morphine over other opioids was found in pair-wise and network analyses. Based on these results, a justification for the placement of morphine as the reference standard for the treatment of severe chronic pain cannot be supported.