Due to word limitations in these posts, I'm going to jump right in. Those who champion liberal chronic opioid therapy (COT) frequently claim that the phenomenon of opioid-induced hyperalgesia (OIH) doesn't exist and/or is "clinically insignificant." I'd like to disabuse the reader of both of these fallacies.
We've talked about OIH a lot more in the perioperative setting over the past three decades or so, and I'm not really going to appeal to those data but will rather focus on the nonsurgical COT-OIH association, which is where the imaginary controversy is being promoted.
Establishing Causality in the Real World
Philosophers often make the point that we can never really be certain that Event A causes Event B. We rely on sequential patterns (I flip the switch and the light comes on) But people far smarter than me can show that at some level there's no difference between the assumption that my motor action results in illumination, and the assumption that the hands on my clock turning 10:01 will make the sun rise tomorrow (yes, it's January in Anchorage.) Statisticians (another arcane discipline above my pay grade), as I understand it, get people to the moon and back based on probability. No matter how tight your machining or computing specs are, there's always going to be some wiggle room.
In the era of evidence-based medicine, we're frequently encouraged to disregard (or at least turn up our nose at) any information that isn't backed up by Level 1 evidence. It's safe to say, though, that there will never be a prospective trial of a high-risk intervention such as opioid therapy looking for an adverse outcome such as OIH outside the context of initial clinical trials of course. This holds for constipation as well as OIH, endocrinopathy, and immunosuppression as well as addiction. Or again, any adverse outcome. We're just not going to experimentally arrange for bad things to happen to people.
So let's look at this whole causality issue from the classic epidemiologic Bradford-Hill criteria, which I'm rearranging/collapsing a little bit to keep it brief:
- Temporality, consistency, and specificity of association: Obviously, if Event A doesn't reliably precede Event B, and if we can't adjust out potential (real) confounders, we're on shaky ground. Summarizing three decades of supportive evidence isn't possible in this forum; let me just refer the reader to the .
- Strength and gradient: Not a deal breaker by any means, but we all get that the greater the effect, and certainly if there's a dose-response curve, it's more likely to be real. (See the link above)
- Analogy, plausibility, and coherence: There are ample analogies from other well-established neurobiologic phenomenon, such as alcohol-lowered seizure threshold, etc. -- we'll talk about this more in Part 2 of this post. I want to briefly focus on plausibility and, in particular, the elegant neurobiologic (gliabiologic?) discoveries of the past decade, convincingly showing that (a) opioids activate the microglial inflammatory pathway (by the toll-like receptor [TLR]-4 receptor); and (b) microglial TLR-4 activation results in downstream neural inflammation and sensitization. In plain English, COT ticks off microglia, the immune cells of the brain and spinal cord, and a cascade of cytokines including tumor necrosis factor α, and a bunch of interleukins and interferons ensues
We can, of course, invoke a syllogism here: if a = b (opioids inflame microglia) and b = c (inflamed microglia cause hyperalgesia), then a = c (therefore opioids cause hyperalgesia.) But we don't have to resort to that kind of mathematical or philosophical hand-waving. A = c has been shown in mice and men, which basically fulfills the coherence criterion, which we can think of as translational findings -- clinical data lines up very nicely with the bench research -- among other recent broad-brush reviews, see ., and among systematic reviews, see , and , both from 2019.
All Right, Enough with the Data -- How is This Relevant Clinically?
No one asks how acetaminophen-related hepatotoxicity or NSAID-related GI bleeds are relevant, and by the way, there's as much preclinical and clinical evidence for OIH as for these adverse drug effects. There's something unique about people's attachment to opioids and the pressure they put on prescribers to continue prescribing them. But I digress.
There are one (or two) more criteria we haven't covered yet. One is experiment, and the other is reversibility. There actually have been a number of experimental studies exposing volunteers to varying degrees of opioids and tallying subjective pain ratings, and more objective outcomes -- such as cold pressor and other quantitative sensory testing. Most telling, however, is the reverse evidence (not part of Sir Hill's original criteria) -- corollary reduction in the dependent variable following elimination of the proposed causal agent/independent variable.
OIH doesn't mean the patient has become tolerant to opioids (they're not mutually exclusive, though). The line between OIH and withdrawal is a very difficult one to draw clearly as well, and they may in fact represent the same phenomenon (they certainly share neurobiology.)
The point, though, is that OIH means the drug is making the pain worse. And along those lines, improvement if not resolution of pain after withdrawing the offending agent (in this case COT) basically clinches the diagnosis. We see this improvement every week, if not every month, in our practice, but I'm not going to ask you to buy a bunch of "n-of-1" trial data from Alaska. Dr. Fishbain has recently provided a (including 20 studies and over 2,100 patients) showing no worsening of pain in 15% of the patients undergoing COT tapering, and improvement in pain in over 80%.
I'm not saying there isn't a time and a place for COT, but clearly we've gotten it wrong a lot over the past two decades.
Let Me Rephrase That -- How Is This Relevant to Me?
Sure. "It's out there, but I don't see it in my practice." (Kind of like the industry-paid spokesperson/pain physician I heard at an in-service recently claiming that he had no opioid-dependent COT patients in his practice.) If you don't think of it, you won't diagnose it, they taught me 20 years ago in medical school. While hardly excusable, an even worse offense would be failure to diagnose due to refusing the evidence.
Think honestly about the patients you prescribe COT to; putting a face to the condition ought to make it all the more concerning. (To paraphrase Stalin, "the hyperalgesia of millions is a statistic; the hyperalgesia of one is a tragedy.") Granted, OIH isn't a life-threatening condition, but opioid dose-escalation sure can be, right?
Next time we'll look at how to navigate patients through/out of this labyrinth, and away from the minotaur of OIH.
Heath McAnally, MD, MSPH, is a board-certified anesthesiologist, pain physician, and addictionologist practicing in Alaska (the military sent him there and he decided to stay). If he wasn't trying to guide people in improving their own lives, teaching medical students to do the same, or writing about it, he'd probably be outdoors right now slogging up a mountain with a good friend or two.