Tuesday, November 4, 2008
Antipsychotics for pain?
As much as we've enjoyed, um, politely complaining about the helpfulness of Cochrane reviews of palliative care-related topics, one was just released which got me a little worked-up. It's a systematic review of antipsychotics for acute and chronic pain control. They included randomized controlled trials of adult patients who were treated with an antipsychotics, in which the major outcome of interest was pain (any control was allowed - e.g. placebo, active, no treatment, etc. - and any duration or severity or type of pain was allowed). All the typical Cochrane search and quality strategies were employed.
They found 11 studies (1972-2006), generally with relatively small numbers ranging from 29-326 patients; 8 studies were placebo-controlled. Pain syndromes studied ran the gamut: trigeminal neuralgia, postherpetic neuralgia, acute myocardial infarction, acute postoperative, etc. One study looked at 'terminal' cancer patients (1972; thioridazine - they note 'pain control was not reported to be superior to placebo' in this trial). Anyway - it goes without saying that the trials were heterogeneous, had mixed results (some positive some negative), used different assessment methods, and several weren't amenable to pooled analysis. In their pooled analyses (which only involved a few trials) they conclude that antipsychotics might in fact be helpful for pain, but looking at those analyses a good proportion of the patients were from a trial of antipsychotics for acute migraine headaches, the one pain syndrome in which dopamine blockers have a relatively well established track records!
All of this raises the question to me of why even do such an analysis which involves numerous pain syndromes with different etiologies, courses, and therapies; numerous different drugs which while they might all be 'antipsychotics' they are still drugs with distinct pharmacodynamics (many involving serotonin, norepinephrine, and acetylcholine; some with tricyclic antidepressant properties, etc.); and in which a strong positive treatment effect from one trial looking at one disease state (acute migraine) may in fact muddy the results, not clarify them? Meta-analysis was developed to pool results from many different trials to help clarify questions that individual trials have not been able to answer; this study seems to be doing just the opposite. Just because you can combine data doesn't mean you should.
Anyway - I was trained that antipsychotics are ineffective as analgesics (migraine excepted) and that's that: does anyone use them for pain with any success? I don't want my therapeutic nihilism to go unchallenged if people have had good experience with this....
Vocabulary builder bonus from the article: "Neuroleptanalgesia (i.e. a state of quiescence, altered awareness, and analgesia produced by a combination of taking an opioid analgesic and an antipsychotic)."
