The trouble with the OPAL trial of opioids for back pain
When the OPAL trial of opioids for back pain was published by The Lancet in mid-2023, it was big, bad news that “opioids don’t work for acute back pain.”
Skeptics love negative trials. It tickles our cynical biases to point to a big study in a fancy journal and smugly declare, “See? X doesn’t work!”
But the truth is more complicated.
And OPAL may not have told us enough for that kind of conclusion.
And mere smugness is the least of our worries here: more serious political agendas may explain the “celebration” of this bad news like it’s good news. We’re in the midst of a hot war over opioid prescribing, with polarized interests and opinions, and patients stuck in the middle.
OPAL arguably studied the wrong thing, for the wrong people, at the wrong times, with statistically insignificant results … and then the authors rode a wave of media hype and made bold public statements about opioids that reached past what their evidence could support. The result was like pouring gasoline on the fire of “opioid phobia.”
Meanwhile, The Lancet failed to publish critical letters for a full year. Those letters are the inspiration for this post.
Ripening seed head of an opium poppy… perhaps the most challenging plant in history.
OPAL in a nutshell
The OPAL study was a 2023 Australian trial of a short course of slow-release opioids for subacute low back and neck pain of moderate severity. Three hundred people got either a placebo or an oxycodone/naloxone combo, plus standard care, for up to six weeks. There was no major difference in pain relief at the end of the test. The opioid group got more side effects, like constipation. The paper concludes:
Opioids should not be recommended for acute non-specific low back pain or neck pain.
Or as put by one of the authors on social media:
The OPAL trial findings provide strong evidence that opioids should not be prescribed to people with acute back and/or neck pain.
Is it really “strong” evidence for that, though? Many experts disagree.
The other side of the story
OPAL wasn’t all bad, of course, but there are two categories of legitimate concern about it:
- The science itself. There are legitimate concerns about OPAL’s design, especially whether they studied the right drug for the right patients. These were ignored by The Lancet for a year before they finally published three response letters.
- How the science was interpreted and presented. Namely, as if the results were much more broadly applicable than the evidence could actually support (even if they could be trusted, which they can’t, see concern 1.)
The role of opioids in managing severe acute pain is not simple: they are a powerful tool with complex pros and cons, so it’s probably unwise demonize them based on one flawed study.
Highlights of the criticisms of OPAL
The papers’ conclusion “does not match nor reflect the trial’s question, design, and scope” (from the first of the response letters, by Weisman, Eubanks, Masharawi). Given that Jones et al. did not test what current guidelines actually recommend, it’s somewhat strange and galling that the authors seem to believe that clinical guidelines should be changed based on this one trial. Here are some of the issues related to this…
- This was not a study of the opioids you’d get if you went to the ER with 10/10 pain. Only one type of slow-release opioid was tested — a drug that is already officially not recommended for acute pain because of its sluggish pharmacokinetics. (So what exactly should be changed in guidelines that already tell doctors not to do this?)
- It seems odd that the OPAL authors concluded that opioids shouldn’t be prescribed for “acute” pain, when what they actually studied was moderate subacute pain. Managing severe acute pain with opioids is is just as much about the psychoactive effects — the euphoria! — as any analgesic effect, and the goal is to help patients right away. What do pain levels several weeks into an episode have to do with what physicians do with opioids in the ER?
- Rather than adjusting the dosage to patient function (again, known best practice), “pain scores of 0–1 were targeted, which might not be considered judicious use” (Wahba and Macintyre).
- Other pain medications aren’t without their own limitations and risks, and sometimes cannot be used at all (e.g. you can’t give NSAIDs to someone who’s had a recent gut surgery). And so blanket statements like “opioids should not be prescribed” are glaring oversimplifications.
Wahba and Macintyre:
this study can only conclude that modified-release opioids used for the management of subacute low back pain, with or without neuropathic pain, were no better than placebo. No change to current guidelines can be based on this conclusion.
Unfortunately, OPAL also cannot be trusted to tell us anything even within its scope…
- The results had low statistical significance (p=.051). Although very close to the traditional threshold for statistical significance, this result is on the wrong side of it (and while somewhat arbitrary, that line is drawn there for a reason). Such a result cannot inspire confidence, even if there were no other problems. We wouldn’t take a positive trial of a drug seriously if the results weren’t statistically significant, so why should we take a negative one seriously?
- The study mixed specific spine pathology with undiagnosed back pain, which is problematic.
- Neuropathic pain is less opioid-responsive, but was not excluded from the study. If you test opioids on patients known to get less benefit from opioids … you’re going to see less benefit.
- Non-pharmacological interventions were not standardised or recorded, so there’s some potential confusion there.
- Almost a quarter of people in the placebo group also used non-tested opioids independently, a potential confounder.
- There was a high dropout rate, too.
Yoda et al., perhaps understating the implications:
Considering these points, the results could be underpowered. We should be careful when drawing decisive conclusions based on this trial.
Better late than never? Not much better!
The publication of the critical letters in The Lancet was much too late in terms of the “public relations” damage. Almost all of the demonizing damage it could possibly have done has already been done, and many people have been hurt by “opioid phobia.” Weisman, Eubanks, and Masharawi:
We are afraid that in its current form, the paper and its conclusion are a disservice to patients needing opioids and to practising clinicians, and could contribute to promoting anti-opioid agendas, demonisation, phobia, and further unwarranted policies.
And the last word goes to Dr. Jim Eubanks (via private correspondence):
There are differences between our use of opioids for acute pain vs chronic. I think everyone who is in academic medicine at this point appreciates chronic opioids are not a viable long term option for most, but as opioids remain critical to the management of acute pain, and are cheap and generally available, I’m frustrated that the OPAL team left nuance out of their initial discussion.
And much of the discussion in the aftermath, too.
The authors’ dismissed most or all of the criticisms … and blocked and insulted critics who expressed these concerns on social media. For instance, Dr. Chris Maher, probably the most “famous” of the OPAL authors wrote on ex-Twitter:
“I am hoping that the arm chair critics get in the trenches and do some research rather than sit on the sidelines and create noise. I suspect I will be waiting a long time. … I don’t think we should let a small bunch of research Karens hijack the narrative for their amusement.”
He’s referring to critics who do indeed “do some research,” and definitely do not just “sit on the sidelines.” Replying elsewhere, Dr. Maher writes:
I have blocked several people to hopefully create a twitter sub-world that is enjoyable to participate in. Let the others live in the sewer of their own making.
That’s a noteworthy degree of contempt shown for earnest criticisms from serious professionals — criticisms published in major medical journal.