Not a good day for exercise-as-medicine

Some smartypants researchers did a study which showed that exercise is no better than saline injections for knee arthritis, in about 200 patients over a couple months. 😬

You’re familiar with “exercise,” I’m sure: the stuff that’s “the closest thing we have to a miracle cure”? Yeah, well, not for arthritic knees apparently. This pushes that annoying “nothing works” vibe in pain and rehab medicine to a new level, and challenges my pro-exercise bias. But, try as I might, I cannot find a serious flaw here. I cannot just wave this paper away, much as I might like to.

Exercise versus fake exercise, sorta

This experiment was rather cleverly controlled. This is always a challenge with exercise trials, because there’s no way to fool someone into thinking that they are exercising when they really aren’t (though it’s fun to try). So Bandak et al used a kind of “hack” to use saline injections as a good-enough proxy for an exercise sham.

Say what now?

Saline injections are obviously a fine sham when comparing to a medicinal injection. The recipient has no idea! But injections are very different from exercise. So the goal is to get to “equipoise” despite the difference. Equipoise is a tricky concept if there ever was one. What it means here is that everyone involved has to take both treatments seriously, for whatever reason. In this case, saline can be taken seriously — not as a “sham” for exercise, because no one is going to mistake it for exercise — but because it has its own credible effects on knee pain (via the placebo effect) as established in other kinds of trials.

That is a bit of a brain teaser, but a valid trial design. I wrestled with it for a good half hour to make sure I understood what I was endorsing.

The exasperating result

Once the validity of this weird comparison was established, Bandak et al were able to proceed to pit exercise against something like exercise … and, dammit, there just was no real difference. Exercise was not significantly better than saline injections. Meaning it was no better than placebo. Meaning it doesn’t work any more than a drug that cannot beat a placebo. Journal Watch:

“That doesn’t necessarily mean that we should totally abandon exercise interventions, for at least two reasons: First, average outcomes in clinical trials sometimes obscure benefits for individuals. Second, for some patients, exercise might confer general well-being that isn’t easily measured.”

Indeed, there are still many reasons to exercise. But this is really not a good day for exercise-as-medicine.

UPDATE: More evidence that “Exercise for arthritis is underwhelming.”