Fifteen double-blind RCTs (n=1,145) were included. Eleven RCTs (6 crossover and 5 parallel group) used oral opioids and four used intravenous opioids.
The studies scored high for quality (mean 4 out of 5, range: 3 to 5) and validity (mean 14 out of 16, range: 10 to 18). Only 6 studies reported a power calculation.
Intravenous opioids (4 crossover RCTs, 120 patients randomised, 115 completed).
All 4 studies found average pain relief of 30 to 60% with opioids, compared with a 5% increase to a 25% decrease with placebo. The results were similar for different types of neuropathic pain (results presented graphically).
Most of the patients reported adverse effects. Problems included a 90% rate of infusion termination before 5 hours with fentanyl (1 RCT) and the exclusion of 14% of patients who were unable to tolerate a morphine infusion (1 RCT); in the latter study, 60% of those included reported adverse events. One RCT did not report any adverse effects with a morphine infusion. Adverse effects included vomiting in 37% of patients receiving fentanyl (1 RCT).
Oral opioids (11 RCTs conducted over 4 days to 8 weeks, 1,025 patients randomised, 698 evaluable).
Mean pain relief with opioids was about 30% for neuropathic and nociceptive pain. Two RCTs found that oxycodone significantly reduced steady pain, brief pain and dynamic allodynia compared with placebo in patients with allodynia.
All 7 studies assessing the quality of sleep found a significant improvement with opioids.
Five RCTs found no significant difference with either an opioid or placebo for various measures of physical activity. Two studies reported lower disability scores with oxycodone compared with placebo.
One of 3 RCTs assessing quality of life using a validated tool found that oxycodone improved most quality of life domains.
There was no statistically significant difference in the rates of discontinuation for opioids and placebo (30% versus 26%; RR 1.0, 95% CI: 0.1, 1.2). The most common reasons for discontinuation were adverse effects with opioids and lack of efficacy with placebo. Patients taking opioids were more likely to report at least one adverse effect than patients taking placebo (80% versus 56%; NNH 4.2, 95% CI: 3.1, 6.4). The most common adverse effects with opioids were constipation (41% versus 11% with placebo; NNH for 3.4, 95% CI: 2.9, 4.0), somnolence (29% versus 10% with placebo) and nausea (32% versus 12% with placebo). Opioids also significantly increased vomiting (15% versus 3%), dizziness (20% versus 7%) and itching (15% versus 7%) compared with placebo.
Two of the 5 studies reporting withdrawal symptoms found none. The others reported withdrawal symptoms in 1 or 2 patients each. In one RCT, drug craving was reported in 8.7% (4 patients) with morphine compared with 4.3% (2 patients) with placebo. Two other studies reported no drug abuse or aberrant drug-related behaviour.
Two of the 3 studies assessing tolerance found pain intensity rose after the 4-week titration period, while the other found no evidence of tolerance.
Follow-up studies (8 RCTs with open-label follow-up).
The studies found that about 44% of patients remained on opioids after 7 months to 2 years. The most common reasons for discontinuation were adverse effects or lack of efficacy. All 3 studies assessing tolerance found no development of tolerance in the majority of patients: tolerance was 6% in one study, 1 of 9 remaining patients followed up over 2 years in a second study, and not reported in the third study. Two studies reported no signs of addiction (in 3 patients in 1 study). Severe withdrawal was reported in a total of 3 patients in 2 studies. One study reported that 3 out of 106 patients took more drug than prescribed.