Wednesday, April 22, 2009
I have accumulated a handful of pain & opioid-related papers that I hadn't planned on blogging extensively about, and now that I've got several stacked up I thought I'd blog them briefly. (Image via Wikipedia which supposedly means this is fair use/public domain; the album itself is pretty good if you don't mind your music simultaneously gloomy and groovy.)
Pain Medicine has a review of opioid-induced nausea and vomiting. It's an indepth review, heavy on the physiology, and is a good one for the teaching file especially those interested in 'deeper' reading (ie fellows). It makes the fascinating (and very tentative in its clinical relevance...ok probably clinically irrelevant) claim that higher doses of opioids may actually be preventive of nausea (long story, due to variable effects at the vomiting center vs. chemoreceptor trigger zone; I can't say that I've noticed this at all personally). The review promotes metoclopramide as the antiemetic of choice for opioid-related nausea, although this is all based in physiological and pharmacologic assumptions and not any real clinical data.
I got a little excited when they began talking about novel approaches, hoping there'd be something in the pipeline clinically or research-wise (they toss around some ideas about opioid receptor antagonists), and then mention tapentadol:
A novel approach is to combine more than one analgesic principle in one molecule so that both mechanisms are pharmacologically engaged. This concept was realized in tapentadol, a compound purposely designed to combine two analgesic actions, mu opioid receptor agonist activity with norepinephrine reuptake inhibition. In this case, the analgesic effect is not reliant solely on agonist activity at the mu receptor that is also responsible for side effects. Data from preclinical studies indicate that the combination of these two analgesic actions is less likely to produce opioid-mediated side effects; for example, in one of the most commonly used emesis model species, the ferret, tapentadol may produce fewer episodes of retching and vomiting as compared with morphine . This suggests a potential therapeutic advantage over the currently available classical opioid analgesics, offering improved tolerability and equivalent analgesic efficacy.So yes, looking good in a ferret model...sigh.
Clinical Journal of Pain has an industry-funded randomized, placebo controlled trial of extended release gabapentin for post-herpetic neuralgia. The ER gabapentin is delightfully described as 'gastric retentive,' which does in fact describe its mechanism: it 'swells' in the stomach and stays there for up to 8 hours. This was a 4-week study of ~150 patients with PHN who had at least 4/10 pain; their other analgesics (including adjuvants) were discontinued and they were randomized to 1800 mg once daily of ER gabapentin or divided as 600-1200 mg (there was a 2 week titration phase as patients got up to these doses, then were followed for another 2 weeks).
Studies like this are very common in the industry funded research literature, and one assumes this was done with an aim to getting this novel formulation of gabapentin approved for the now generic drug. Patients were taken off all other analgesics and gabapentin is well established as a moderately effective treatment for PHN (moderately like all other treatments), so unless the new formulation was very poorly tolerated (which it wasn't) the results were likely to be positive (which they were). There's no reason to think this formulation offers anything new to our patients, except increased cost and I guess convenience. For me it's always interesting to see in these very clean (no other analgesics at baseline) placebo controlled trials just what the actual magnitude of benefit of the drug is.
ER gabapentin was effective, modestly so. Both regimens decreased pain by about 2/10 points (twice daily a little more; which achieved statistical significance); placebo decreased by ~1/10 points. The number-needed-to-treat to achieve a 50% pain reduction in the twice daily gabapentin group compared to placebo was 6 after 4 weeks.
CJP also has a small placebo controlled trial of tramadol for persistent neuropathic pain after spinal cord injury (started on 50 mg 3x a day, able to titrate to 400 mg daily). 20% of subjects were on concomitant opioids at the time of the study. Tramadol was (guess what?) modestly effective compared to placebo, although a high number of significant adverse effects occurred.
What was most interesting to me was the breakdown of the global impression of change data as it is illustrative of a major clinical and research issue in pain management (and other disease states of course): about 2/3 of those on tramadol said they were the same or minimally improved; about 1/3 said they were much improved. This sort of finding, while not surprising, is one of the reasons one gets frustrated with a lot of pain research (it's not the researchers fault). Assuming for the sake of argument that this is a real effect/finding: what it means is that while most patients who you'd start on tramadol wouldn't benefit; there are a handful however who would benefit tremendously. That is of course what we see in clinical practice, and why particularly for complicated pain syndromes most of us cycle through a variety of different agents and drug classes, hoping one, eventually, works. And sometimes it does.
To state this another way, while these studies of populations demonstrate some average benefit of a drug applied to a population, in reality (and especially for studies involving such highly variable and idiosyncratic syndromes like neuropathic pain syndromes) many patients don't benefit at all, a few are frankly hurt (all side effects, no benefit), and a few benefit a lot. Average it out, and the drug shows a 'statistically significant' benefit. The problem is that we have zero clue as to where an individual patient will fit into that. Our patients, in pain, leery of these drugs with terrifying package inserts, have to live through this while hoping we stumble blindly into the agent which happens to work decently for them (if any do).
I don't mean to whine, and say this with the humility of one who doesn't do research, but what would be most helpful as a clinician are ways to help predict who will respond to what, such that our patients can be spared the expense, toxicity, and delay of us cycling through multiple agents until we establish some drug/regimen which is helpful for them. I assume the answer ultimately will be a genetic one (receptor, enzyme polymorphisms and whatnot), and this is what the whole 21st century 'personalized medicine' movement is supposed to be about....
Ok, I know I used to words 'not extensively about' earlier so I'll quickly exit the post....
CJP also has a tidy review on the endocrine effects of opioids: teaching file/fellows.
Pain has a review on the effectiveness of patient educational interventions for cancer pain management: not too many studies, 'heterogeneity,' some evidence pointing to such interventions can in fact improve analgesia ~1/10 points.