In 18 sciatica injection trials, the people who got salt water kept getting better

Salt water is supposed to do nothing. That is the entire reason it sits in the control arm of an injection trial: you give one group the drug, you give the other group sterile saline, and whatever separates them is the drug working.

A review published on 24 August in JBJS Reviews went back and looked at what happened to the salt water group (JBJS Reviews, 2026). Across 18 randomized trials and 759 patients, those people got better. Not slightly, and not briefly.

A team at Brown University's Warren Alpert Medical School gathered every randomized trial published between 2000 and 2025 that included a saline injection arm for lumbar radiculopathy, which is the nerve pain most people call sciatica, or for disc related low back pain. Then they asked a question trials almost never report on its own. Instead of comparing the active injectate against saline, they looked only inside the saline arm, before and after.

For radiculopathy, leg pain on a 10 point visual analogue scale fell by 1.79 points at one month and by 3.85 points at one year. Disability on the Oswestry index dropped 7.30 points at one month and 17.83 points by a year. Those later figures sit well past the threshold researchers use to call a change clinically meaningful rather than merely detectable, and the improvement was largest at the final time point, not the first. The pattern held most consistently when the saline went in epidurally.

Back pain coming from a disc behaved differently. Those patients changed less, and most of the pooled estimates never reached the meaningful threshold at all. Same inert fluid, two different problems, two different stories.

Here is the part I'd want anyone reading a headline about this to hold onto. A change measured inside one arm is not a placebo effect, and the authors don't claim it is. It bundles together everything that happens to a person across twelve months: sciatica that settles on its own in a large share of cases, regression toward the mean because people enrol in trials at the point when they feel worst, the medication and hands on care they receive alongside, and the plain fact of being examined, taken seriously, and followed by someone for a year. The conclusion the authors draw is about study design. Saline is not an inert comparator, so any active injectate has to clear a baseline that is already climbing, which makes its own additional benefit harder to see. Their recommendation is for the people who run trials, not for the people who sit in the waiting room.

Three questions worth taking to your next appointment

I'm a Feldenkrais® practitioner, not a physician, and whether to have an injection is a decision for you and your clinician. What this paper offers is a sharper set of questions to bring into that room.

If we wait another three months and do nothing invasive, what does the usual course look like for someone with my findings? This is the question the review licenses most directly, because a large share of what those saline arms recorded is time. Radicular pain has a natural history. Knowing roughly what yours looks like changes how you read every number you are quoted afterwards.

When you tell me this injection helps, what was it compared against? If the comparator was saline, then some portion of the improvement in both arms belongs to the baseline the paper just described. That doesn't make the injection useless. It makes the honest question "how much of this is the drug" rather than "does anything happen."

Whatever we decide about the injection, what should I be doing with my body between now and then, and what would make you want me to stop? Waiting is not the same as resting. Most people, left without an instruction, quietly stop using the affected leg, and by the time three months have passed they have lost more than the nerve ever cost them.

What I notice in practice

A woman in her sixties told me last spring that she had stopped putting weight through her left leg without ever deciding to. She'd had one epidural and was scheduled for a second. When we looked at how she stood up from a chair, she was pushing almost entirely through the right side, and had been for months. Nothing about that was a decision. Her nervous system had made a sensible arrangement during the worst weeks and then simply kept it.

That arrangement is what I see far more often than the disc itself. Pain teaches a fast, intelligent lesson about which movements to avoid, and the lesson outlasts the injury that taught it. You haven't lost movement. You've lost trust that movement will help, and trust comes back slower than tissue does.

Awareness Through Movement® works on exactly that layer. The lessons are slow and small, you only go where it feels easy, pleasant, and comfortable, and the point is not to strengthen anything but to give the brain enough information to stop guarding a leg it no longer needs to guard. This sits alongside what your physiotherapist or your spine clinician gives you rather than in place of it. If you want something concrete to try while you decide, our page on Feldenkrais for sciatica uses a ball as soft support so the leg can move without carrying its own weight.

Low back pain is the single leading cause of disability worldwide, and it affected an estimated 619 million people in 2020 (WHO, 2023), and a great deal of what gets tried for it is tried in good faith with thin evidence behind it. This paper isn't an argument against injections. It's a measurement of how much recovery was already underway in the room.

The 759 people in those saline arms were never the story anyone set out to tell. They enrolled hoping to be in the other group. A year later a good number of them were walking further than they had in a long time, and nobody has a clean account of why.