Anticonvulsants break hips

Pain patients are being prescribed ever more drugs like Lyrica and Neurontin (the anticonvulsants, AKA gabapentinoids). Their use is way up, as previously discussed, and so is their abuse. And now a new study links those drugs to … 30% more hip fractures!

That is a much worse outcome for most of those patients than whatever kind of pain they were hoping to treat.

The likely mechanism? The risk of falling goes way up with side effects like dizziness. That risk was almost double in “frail” people, and well over double with kidney disease (which also makes people wobbly).

“The exponential increase in gabapentinoid use has been particularly pronounced among older adults, especially after the expansion of the indication to include various painful syndromes (eg, neuropathic pain and fibromyalgia).”

And falls are likelier in the same older adults most likely to need pain control.

To justify the risks, I sure hope the evidence for using anticonvulsants for pain is really robust! 🤞🏻 (Spoiler alert: it ain’t.)

This post is about a 5-minute read. I’ll cover some of the background, the nerdy details of the study, the risk math measured in grandmas 👵 👵 👵, the importance of sedation versus analgesia, and a risk/reward analysis

The evidence for using anticonvulsants for pain is not robust (except for neuropathy)

The evidence is not even “promising” for many of the kinds of pain these drugs are being prescribed for. Quick review: anticonvulsants can be effective for some neuropathy, although “your mileage may vary” in a big way. Their value for other kinds of pain is low to nil. For instance, the NICE guidelines specifically advise doctors not to offer them sciatica: “there is no overall evidence of benefit and there is evidence of harm.” And yet that is one of the most common reasons to prescribing! And they generally overprescribed for many other dubious purposes, infamously driven by two major factors:

1. Huge, illegal marketing campaigns encouraging doctors to prescribe them for conditions they aren’t approved for.
2. Medical desperation to prescribe something, anything other than opioids. Which is not an “excuse” for bullshit over-prescription, but it is certainly an “explanation.”

And so some cynicism about these drugs is highly justified.

This study would not be particularly newsworthy without the context of over-prescription — it would just be delivering evidence of an important side effect for patient's to know about. But that side effect is much more ominous if it’s being imposed on millions of people who never had any realistic hope of benefitting. Which it is. This has already happened on a truly epic scale, and it is continuing to happen.

Nerdy details about the study

Leung et al. looked at data on about twenty-eight thousand older Australians with hip fractures between 2013 and 2018, whittling that down to 2600 who had consumed at least one prescription for a gabapentinoid (mostly pregabalin), and another thirteen thousand that didn’t as controls. They also studied two other major known risks for hip fracture, kidney disease (30% of the subjects) and general frailty (45%).

This experiment had a lot of self-control. That is, subjects acted as their own controls in multiple ways: a fancy “case-case-time-control” design that compared subjects to themselves at different times (“case-crossover”), but also to other patients across time (“case-time-control”). It’s too nerdy a design even for the salamander to ramble on about! It was fancy and interesting, and that’s enough said here (but try your luck with a good technical description).

The design means that they could be much more sure that the results weren’t a statistical illusion, namely that gabapentinoids are quite powerfully linked to a risk of hip fracture:

• 30% more breaks even in the healthiest subjects (1.3×)
• 75% more in the frail (1.75×)
• 140% with kidney disease! (2.4×)

Okay, but how many grandmas is that?

An “odds ratio” of 2.4 is a bit abstract. Let’s make the risk math more a bit more real by counting grandmothers. I’ve only ever really known a handful of grandmothers. Imagine two hundred! So many cookies.

And now imagine them all in rough shape: all frail, all with kidney disease.

One hundred of them never took a gabapentinoid, and twenty of those grandmas are fated to break a hip. But the other hundred did get that prescription … and so forty-eight of them break a hip.

But it’s not just frail grandmas with bad kidneys with lousy odds. For every hundred healthier people who break a hip sans gabapentinoids, one hundred thirty will break a hip if they do take gabapentinoids. That’s a big deal.

Also, hip fractures are not the only way to get hurt from feeling wibbly-wobbly!

Anticonvulsants can make you dizzy … and dizzyness can make you fall… and falling can break your hip … among many other things. It’s hard for younger people to relate to, but gravity really is a major health hazard that escalates with age. My wife’s recent fall on ice (severely aggravating a major old spinal injury) was a Major Life Event, and has resulted in months of hypervigilant caution to prevent another fall while she’s recovering, which really could be catastrophic. How much greater would the risk be if she was a little “wobbly” from being on anticonvulsants? And how much more tragic would it be if they weren’t even helping her? Which is likely?

Anticonvulsant comments, both for and against

There were many good quality comments about this when I shared it on Facebook, Twitter, Threads early this year. (I’m only making a meal of it just now, three months later.)

The response was roughly split between pro and anti-anticonvulsants. Some people have gotten no benefit from them, and are appalled by the overprescription and rampant off-label usage; but many others are happy with their prescription, and justifiably concerned about a good thing being demonized.

• “This is the worst medication I’ve ever been prescribed. Fun fact, did you know that drug addicts also use it to enhance the high they get from their drugs?” (Yep. I knew about the abuse of gabapentinoids … and the spectacular irony that the huge surge in over-prescription is actually driven largely by the fear of prescribing opioids and the desperation to prescribe something, anything else to pain patients. And so, weirdly, now they prescribe … both!)

• “For neuropathic pain? It’s a game changer for many, myself included. Most of the problems with drugs either being prescribed and not working for xyz type pain, or being abused, can be laid right at the feet of the people with the most power in the medical relationship. Hint: it ain’t the patients.”

• “The GABA’s are dished out like lollies in Australia, as a kind of ‘throw it against the wall and see if it sticks,’ despite the fact that the kind of pain they can be effective for is clearly not present. Maybe this is desperation on the part of doctors who have exhausted their options and are turning themselves inside out to compassionately provide relief, but when it makes no difference the prescription must be withdrawn.”

• “Pregabalin was enormously successful in treating my pain, which had a lot of neuropathic overtones though also diagnosed as fibromyalgia. If someone said they were taking it away I would have been terrified.”

• “Nothing helped my fibro pain like going on anticonvulsants.”

• “Wildly overused on the basis of even sketchier evidence for all kinds of pain in dogs and cats. So easy to mistake sedation for pain relief, and frustrating to see pain untreated because of caregiver placebo effects and NSAID-phobia.” [That’s from Brennen McKenzie’s perspective as a skeptical veterinarian! More on this important idea below.]

• “Gabapentin is the only thing I’ve found that can make my idiopathic peripheral neuropathy remotely tolerable. But I definitely notice the extra wobbliness, which is not a good combination with the lack of proprioception in my feet.”

• “Often used for headache migraine sufferers and studies show it has little to no effect. Tragic really, god knows why it is used.”

• “Not to mention all the other unbearably terrible side effects of anticonvulsants. But some tolerate them well and the meds provide those folks with a new lease to life. So cons and pros, for sure.” [A rare case there of someone commenting from both the pro and con perspectives in one shot!]

• “Also common for patients to be started on gabapentinoids peri-operatively as part of a ‘multi-modal analgesia’ approach, and these meds are often inappropriately continued at discharge & long term. Among the drugs I find myself stopping most often in my hospitalized patients.”

Anticonvulsants have great customer “sat,” but why?

Tim Cook of Apple loves to talk about customer “sat,” a corporate-speak abbreviation of “satisfaction.” The customer sat for anticonvulsants is quite high, independently of the evidence. Why?

A real but small percentage of that is probably that anticonvulsants actually do help some people for unclear reasons. For instance, some cases of unexplained widespread pain may have an unknown neuropathic component that is successfully addressed by these drugs. Yahtzee!

But a great many people are probably mistaking the woozles for analgesia. Anticonvulsants are sedating. And this kind confusion is a strong theme in the science of pain. This is quite important, so I’m really going to emphasize it (as nicely phrased by Brennen McKenzie above, who pointed out that it’s especially a problem with pets):

It’s easy to mistake sedation for pain relief.

It’s easy to mistake sedation for pain relief.

It’s easy to mistake sedation for pain relief.

And so, not only are anticonvulsants over-prescribed, but a lot of patients are essentially too buzzed to care. This is a great example of how we can get a lot of “it works for me” anecdotes for a well-studied medication that definitely cannot pass a rigorous test.

Is the risk of hip fracture acceptable if the drug is actually helping?

If anticonvulsants are truly useful to someone now, then perhaps it’s acceptable to have a significantly elevated risk of falling and fracturing later. Would I take them myself if they were the only thing that worked for my pain?

Tough one. Truly. Short term relief versus long-term risk is one of the Great Problems in health, and there is no right answer.

Many people self-medicate pain with alcohol, and that is definitely not an effective pain killer — sedation versus analgesia again! But the psychoactive effects can certainly provide a form of real relief for people. And that relief, modest as it may be, is something that people choose for the short term over and over and over again despite the fact that most people now do understand that alcohol is a poison with serious long-term consequences. Most people will in fact choose minor short-term benefits over even serious long-term harms. It’s a very human thing to do.

And it’s a thing I do, so I’d have to be quite the hypocrite to say that people “should” pass on gabapentinoids just because they could make you slip in the shower someday.

But hip fractures are really terrible! And you know what’s worse than absolutely any problem? That same problem … plus a hip fracture! So it really is a hard call, even when gabapentinoids are indeed “the only thing that works.”

I have heard some reports about the side effect backing off after a while, but I guess a key lesson from Leung et al. is that it doesn't back off far enough, soon enough for a lot of people. ☹️

Is the risk of hip fracture acceptable if the drug is not actually helping?

Rhetorical question, there. 🙂 The premise of short-term benefit definitely does not apply to huge numbers of people taking these drugs. The elephant in the room here — the whole reason this new risk data matters — is that these drugs are being dramatically over-prescribed for conditions they cannot actually help.

And that is very clearly not worth the risk of harm. Duh.