What nearly twenty years of sciatica trials say about waiting for surgery

A panel of spine surgeons published a set of clinical recommendations this week on the question that sends people into internet rabbit holes at two in the morning. When should a herniated disc be operated on?

The recommendations are careful, and worth reading. What caught me was something sitting underneath them, visible only once you line up the trials they assembled and look at the arc instead of the answer. The arc is not really about operating sooner or later. It is about how many people, given enough months, quietly changed their minds.

What they concluded

AO Spine gathered the evidence base for surgery versus conservative treatment in lumbar disc herniation. Two randomised trials, two more with observational cohorts running alongside them, one prospective study, one retrospective. Each one was summarised, assessed for how well it was done, and given a clinical recommendation (Global Spine Journal, 2026).

The conclusion splits cleanly in two.

If there is real muscle weakness, timing carries most of the weight. For moderate to severe weakness the panel recommends that a clinical team aim for decompression within three days, because that window appears to matter for whether the nerve comes all the way back. New or worsening weakness in a leg is the one place in this whole literature where the panel is unambiguous, and it belongs in front of a doctor rather than in front of a search engine.

If there is no weakness, which describes most people living with sciatica, the recommendation reads quite differently. How long the symptoms have gone on, how severe they are, and what the patient prefers should drive the choice, with the benefit of surgery growing as the months accumulate. Patient preference. Sitting there in a surgical guideline, listed as a criterion.

Nineteen years, one shape

Go back to 2007. The Dutch Sciatica Trial randomised people to early surgery or to six months of conservative care, and at one year the two groups had arrived in roughly the same place. The advantage of surgery was speed, not destination. Five years on, with 231 of the original 283 people still being followed, the picture had not changed: no meaningful difference between the groups in pain, in disability, or in how recovered people said they felt (BMJ Open, 2013).

Then SPORT, the large American one, reporting eight year outcomes in 2014 from 501 randomised patients plus 743 more in a parallel observational cohort, across thirteen spine clinics in eleven states. Analysed by what people were assigned to, the differences were small and could comfortably have been chance. Analysed by what people actually received, surgery came out ahead. And neither group slid backwards between year four and year eight (Spine, 2014).

Then the sharpener. In 2020 a single centre trial took 128 people whose sciatica had already been running for somewhere between four and twelve months, which is a different population from every trial that came before it, and randomised them to microdiscectomy or to six further months of structured conservative care. At six months the surgical group's leg pain had come down to under three on a ten point scale while the other group was still sitting above five (New England Journal of Medicine, 2020). Nine people had adverse events connected to the operation.

Laid end to end the line is legible. For sciatica that started recently, the two roads converge and surgery is the faster one. For sciatica that has already run for most of a year, surgery brought faster relief at six months, at least in the single selected group anyone has studied so far. That is close to the sentence about benefit increasing with duration that turned up in this week's recommendations, arrived at over nineteen years and several hundred randomised patients.

The part that rarely gets mentioned

Look at how many of them changed roads.

In the Dutch trial, 46% of the people allocated to conservative care had surgery anyway. In SPORT, 48% of those assigned to conservative care received surgery, and only 60% of those assigned to surgery went through with it. In the 2020 trial, 34% of the conservative group crossed over, at a median of eleven months.

That is why the assigned numbers and the actually received numbers keep disagreeing with each other, and it is also why a clean answer never quite arrives. These are not clean trials of two treatments. They are records of what several hundred people decided, month by month, while their symptoms went about their own business.

The five year Dutch data holds one more thing. Three factors predicted an unsatisfactory result half a decade later: being over forty, severe leg pain at the outset, and how emotionally distressing the pain felt at the start. Of the three, that last one was the largest. About 21% still reported an unsatisfactory recovery at the five year mark, and 31% had at least one unsatisfactory outcome at some point during follow up, whichever treatment they had.

The months nobody randomises

Whatever a person chooses, there are months.

That is the stretch I keep meeting. People arrive having already done the reading, sometimes with a date in the diary, sometimes having decided firmly against one, and what they want to talk about is not the disc. It is that they have stopped trusting their own leg. Bending to the dishwasher has become a negotiation. Rolling over in bed gets rehearsed before it gets done. They aren't weak and they aren't imagining it, and the guarding was a sensible idea when it started, but it has outlived the moment it was invented for.

You haven't lost movement. You've lost trust that movement will help.

This is the stretch of time where gentle movement work can have a practical role. Not because it changes what a disc is doing, but because it gives you a way to explore movement again without repeatedly provoking the symptoms. In an Awareness Through Movement® lesson you only go where it feels easy, pleasant, and comfortable, which is about the only condition under which a guarded body will lend you anything at all. It sits alongside whatever your surgeon or your physiotherapist has you doing, never in place of it, and back pain that has settled in for the long run tends to need something to occupy that middle stretch regardless.

When people ask what they can actually do during those months, I point them to the sort of movement a Feldenkrais Method® lesson is made of. You lie on your back, bend your knees and stand your feet about pelvis width apart, and let your pelvis roll slowly and gently, in small circles or a little way from side to side, going only as far as stays easy and comfortable, and resting whenever you like. The sore leg is asked for almost nothing. If you want somewhere to begin, these gentle sciatica exercises follow that pattern.

Nineteen years of trials have given us a clearer picture of when surgery brings faster relief, and when waiting can still lead to a similar long term result. What none of them found a way to measure is the person standing at the junction for eleven months, working out how to get out of bed.