Antidepressants in neuropathic pain |
||
Clinical bottom line: Antidepressants are effective in reducing neuropathic pain. The overall NNT for at least 50% pain relief for antidepressant compared with placebo in diabetic neuropathy was 3.0 (2.4 to 4.0) and was similar across pain conditions. 30% of patients will obtain more than 50% pain relief, 30% will have minor adverse reactions and 4% will have to stop treatment because of major adverse effects. SSRIs may be less effective, but are associated with a 50% reduction in major adverse reactions. |
- Systematic review
- Placebo-controlled comparisons
- Diabetic neuropathy
- Figure: Antidepressant compared with placebo in diabetic neuropathy
- Postherpetic neuralgia
- Atypical facial pain
- Central pain
- Active-controlled comparisons
- Adverse effects
- Related topics
- Further reading
Antidepressants in neuropathic pain
Antidepressants have been used for over 30 years to manage neuropathic pain, but in the UK no antidepressant has a product licence for this indication. Although many studies have been carried out, it remains difficult to determine, for example, which antidepressant is most effective, or that antidepressants are better than anticonvulsants. What is clear is that antidepressants have an analgesic effect on top of any mood effect.
Systematic review
HJ McQuay, RA Moore. An evidence-based resource for pain relief. Oxford University Press March 1998 ISBN 0-19-262718-X.
Date review completed: 1994
Number of trials included: 18 (21 placebo-controlled arms / 11 active controls)
Number of patients: 400 active versus 373 placebo controls.
Control group: placebo or active (antidepressant / analgesic / anticonvulsants / tranquillisers)
Main outcomes: Relative benefit and NNT to achieve at least 50% pain relief (with 95% confidence intervals).
Inclusion criteria were randomised controlled trials of antidepressants in neuropathic pain (including diabetic neuropathy, postherpetic neuralgia, atypical facial pain and central pain); placebo or antidepressant or 'other intervention' control group; full journal publication; group size at least 10; pain outcome.
A clinically relevant outcome was defined as a measure equivalent to at least 50% pain relief, and outcomes of the longest duration were selected.
Placebo-controlled comparisons
Diabetic neuropathy
Six of 13 comparisons in diabetic neuropathy showed significant improvement over placebo (covering nine different antidepressants, Figure). Desipramine and tricyclics produced the best NNTs. The overall NNT was 3.0 (2.4 to 4.0).
Figure: At least 50% pain relief with antidepressant compared with placebo in diabetic neuropathy
|
| Condition | Number of trials | Antidepressant improved/total | Placebo improved/total | NNT (95%CI) |
| Diabetic neuropathy | 13 | 180/260 | 73/205 | 3.0 (2.4 to 4.0) |
| Postherpetic neuralgia | 3 | 43/77 | 8/68 | 2.3 (1.7 to 3.3) |
| Atypical facial pain | 2 | 62/88 | 30/85 | 2.8 (2.0 to 4.7) |
| Central pain | 1 | 10/15 | 1/15 | 1.7 (1.1 to 3.0) |
Postherpetic neuralgiaTwo of three comparisons in post-herpetic neuralgia showed significant benefit. The combined NNT was 2.3 (1.7 to 3.3). Atypical facial painTwo of two comparisons in atypical facial pain showed significant benefit. The combined NNT was 2.8 (2.0 to 4.7) Central painOnly one of three trials had extractable data, with an NNT of 1.7 (1.1 to 3.0) in a small number of patients. Active-controlled comparisonsIn three of three reports tricyclics were significantly more effective than benzodiazepines. Two of two reports showed no differences between various tricyclics. Many of the trials demonstrated analgesic benefit without significant changes in mood measures. There was no difference in efficacy across different pain conditions. Although only one trial compared antidepressants with anticonvulsants directly. This showed greater benefit at lower risk with antidepressant. Adverse effectsThe number needed to harm (NNH) for minor adverse effects was 3.7 (2.9 to 5.2) based on 11 reports, combining across pain syndromes. For major effects the NNH was 22 (14 to 58), based on 19 reports. Effects were lower for SSRIs (fluoxetine and paroxetine) than with tricyclics. Related topicsAnticonvulsants in chronic pain Topical capsaicin NNT Relative benefit/risk Further readingThe current review is an expansion of:
|