Twenty-seven RCTs were identified (2,082 participants, range 11 to 338). Twenty-one trials had a crossover design. Six trials were of parallel design. Five RCTS had a Jadad score of 3, 11 RCTs scored 4, 10 RCTs scored 5 and the Jadad score was not reported for one RCT. Nineteen studies included a placebo control.
Pain relief: There was no significant benefit over other analgesics in reducing neuropathic pain for benzodiazepine lorazepam (one RCT), phenothiazine fluphenazine (one RCT), sodium channel-blocking agents, mexiletine (one RCT) and carbamazepine (one RCT). Gabapentinoids pregabalin (one RCT) and gabapentin (two RCTs) were not significantly more effective than tricyclic antidepressants (amitriptyline (AT)) and significantly less effective in two RCTs. Two studies found non-significant trends for opioids (morphine, methadone) to be superior to tricyclic antidepressants (amitriptyline) and gabapentinoids (gabapentin).
Comparisons of antidepressants: Tricyclic antidepressants did not significantly differ in their pain-reducing efficacy (amitriptyline, nortriptyline, desipramine and imipramine). Noradrenergic desipramine was at least as effective as amitriptyline in reducing pain (two RCTs). Tetracyclic maprotiline was significantly less effective in reducing pain than amitriptyline (two RCTs). Mianserin did not significantly reduce pain compared to placebo in one RCT where impramine was significantly effective. Three small lower-quality studies found that the serotonergic drug clomipramine was significantly more effective in reducing pain than impramine and desipramine, and impramine was significantly more effective than the SSRI paroxetine. There was poor evidence for the effectiveness of the SSRI fluoxetine (one RCT with no significant effect and one RCT with a low but significant effect compared to desipramine and amitriptyline, which were more effective). Impramine was more effective than venlafaxine in reducing pain in one RCT, but the difference was not significant. There was evidence for individual variability in responsiveness when individual antidepressants were compared in three RCTs. The authors concluded that noradrenergic and nonadrenergic/serotonergic tricyclic antidepressants were more effective for neuropathic pain.
There appeared to be no greater benefit for the use of newer drugs (gabapentin, venlafaxine, pregabalin and lamotrigine). Results for other outcomes were reported: patient’s global indicator of change (PGIC), present pain index (PPI), overall activity, vitality scores, pain-related interference in daily activity, short form 36 health survey (SF-36), quality of life and adverse events. Little data were reported for adverse events.