Chiropractic for Low Back Pain: The Evidence, Aggregated

For the video version, click here.

Spinal manipulation therapy is one of those treatments that just thrives on anecdote. If you’re like me, you have patients who swear by it. Also, if you’re like me, you’re skeptical about some of the claims certain providers make about spinal manipulation. Mostly because a lot of these claims lack biologic plausibility.

But for low back pain, spinal manipulation has a bit more face validity than spinal manipulation for, say, gallstones.

So I was intrigued to see this meta-analysis appearing in the Journal of the American Medical Association:

Would spinal manipulation prove the panacea for painful posteriors? Or would the anecdotal evidence wash away amidst a torrent of hard data?

The crack researchers examined randomized trials of spinal manipulation therapy for acute lower back pain from 2011 to 2017.

26 randomized trials were found. Most of them…weren’t great.  No studies blinded the provider (though I admit that would be tough).  And only 4 blinded the patients by using sham procedures.

Let’s look at the effect of spinal manipulation on pain scores.

This forest plot shows you that, broadly speaking, spinal manipulation reduced pain scores by about 10 points out of a 100 point scale. Is that meaningful?  We don’t really know, because the minimum clinically important difference for low back pain hasn’t been established yet. But that 10 point difference is about what you see when you use NSAIDs for low back pain.
 

But look at those top two trials.  Those were the trials where spinal manipulation was compared to a “sham” procedure – just pushing randomly on someone’s back. In both those well-controlled trials, we don’t see any effect.

Functional outcomes mirror the pain outcomes, a modest benefit, but unclear clinical significance.

The meta-analysis also found high heterogeneity of effect.  What that means is that the results of the trials don’t hang together very well – they are all over the map. While this doesn’t tell us exactly what the studies are doing wrong, it tells us we can’t be all that confident in the combined treatment estimate.

My theory on the heterogeneity? Spinal manipulation is by no means a uniform treatment. It’s not like a chemical compound, where everyone gets exactly the same thing. Some providers use a “thrust” technique. Some don’t.  Some have soft music playing in the room, some don’t.  It may be any of these factors that actually help the back pain. 

We need to chase biologic plausibility. And, as I say, there is at least a whiff of that with spinal manipulation for low back pain. But for a study to be reliable, it needs to be rigorous. And the most rigorous studies in this meta-analysis, those that randomized patients to spinal manipulation or sham spinal manipulation, showed no effect. A few more large studies of that type though, and we’re back in business.