Spinemathema podcast · Episode 13 · About 9 minutes · October 2026

Episodes are narrated using an AI model of Mr Sadek’s own voice, from scripts he has written and approved.

Read the full article with references: Does Vertebroplasty Work? What the Sham Controlled Trials Found

Transcript

Hello and welcome to the Spinemathema podcast. I’m Ahmed Sadek, a neurosurgeon. In this episode I’m looking at a set of landmark trials, and they teach us something important, not just about one treatment, but about how we test treatments in general.

The treatment is called vertebroplasty. And the trials are what we call sham controlled trials. Two were published side by side in the New England Journal of Medicine in two thousand and nine, and a larger one, called VERTOS four, was published in the British Medical Journal in two thousand and eighteen.

Let me set the scene. In people with osteoporosis, the bones become thinner and weaker. A vertebra, which is one of the bones of the spine, can collapse with little or no injury. These are called compression fractures, and they can be very painful. In vertebroplasty, a doctor uses X-ray guidance to pass a needle into the broken bone and inject a special bone cement. The idea is to stabilise the fracture and relieve the pain.

Before two thousand and nine, vertebroplasty had become a routine treatment. But much of the evidence came from case series and small studies where patients knew they’d had the procedure. And that’s a problem. Pain from these fractures often settles by itself as the bone heals. And simply having a procedure, with all the attention and expectation that comes with it, can make people feel better. So how do you tell whether the cement itself is doing the work?

That’s where a sham controlled trial comes in. One group has the real procedure. The other group has a pretend version that looks and feels the same, but leaves out the key step. In this case, the cement. Neither the patients nor the people assessing them know who had which. If both groups improve by the same amount, the improvement probably isn’t coming from the cement.

So let’s look at the three trials.

The first was led by Buchbinder and colleagues. It included seventy eight people with one or two painful osteoporotic fractures, less than a year old and still unhealed on a magnetic resonance imaging scan, or MRI. Thirty eight had vertebroplasty and forty had a sham procedure. The main result was pain at three months.

The second, led by Kallmes and colleagues, was called INVEST. It included one hundred and thirty one people aged fifty or over, from eleven centres in the United States, the United Kingdom and Australia. Sixty eight had vertebroplasty and sixty three had a simulated procedure. The main results were disability and pain at one month, and after that people were allowed to switch groups.

The sham in INVEST was quite elaborate. Everyone had local anaesthetic. The cement ingredient was opened in the room so that patients could smell it. And the doctor gave the verbal and physical cues of an injection, pressing on the back for example. But in the sham group, no needle went into the bone and no cement was injected.

The third trial, VERTOS four, was led by Firanescu and colleagues in the Netherlands. It was designed to improve on the earlier trials. It included one hundred and eighty people aged fifty or over with recent fractures, from four community hospitals. Ninety one were assigned to vertebroplasty and eighty nine to a sham. Everyone had local anaesthetic and needles placed, and the cement was mixed nearby so it could be heard and smelt. Patients were followed for a year, and nobody was allowed to switch. The day after treatment, more than eight in ten people in both groups believed they’d had cement.

So what did they find? In all three trials, the story was the same. Pain improved a lot in both groups, and vertebroplasty didn’t do significantly better than the sham.

In the Buchbinder trial, pain on a scale of nought to ten had fallen by an average of two point six points after vertebroplasty, and one point nine points after the sham, at three months. That difference wasn’t significant, and vertebroplasty showed no significant advantage on any measure at any time.

In INVEST, at one month, the groups differed by less than one point on both pain and disability, and again that wasn’t significant. Both groups improved straight away. Interestingly, by three months, about half of the sham group had switched to vertebroplasty, compared with about one in eight of the vertebroplasty group switching the other way.

In VERTOS four, pain fell in both groups from the very first day. At twelve months, the average pain score had dropped by five points after vertebroplasty and by four point seven five points after the sham. The groups didn’t differ significantly at any time point.

These trials were then brought together in a Cochrane review, updated in two thousand and eighteen. A Cochrane review is a careful, independent summary of all the good quality trials on a question. It pooled five sham controlled trials with five hundred and forty one people. On average, pain was about seven tenths of a point lower after vertebroplasty, on a scale of nought to ten. That’s well below the size of change that patients usually notice as important. The reviewers concluded there was high to moderate quality evidence of no important benefit. They also noted rare but serious side effects after vertebroplasty, including bone infection and pressure on the spinal cord, although the numbers were too small to say exactly how often these happen.

Now, I want to be fair about the limitations. The two thousand and nine trials were fairly small, and INVEST allowed switching after a month, which makes the later results harder to read. Not every trial agrees, either. Another sham controlled trial, called VAPOUR, published in two thousand and sixteen, did report better pain relief with vertebroplasty. It mainly included hospital inpatients with very recent fractures and high pain scores. And in VERTOS four, some extra analyses that weren’t planned at the start suggested that, at a year, more people in the sham group still had moderate or severe pain, and more had further collapse of the fractured bone. Findings like that need confirming in new trials before we can rely on them.

So what does this mean for you? If you have a painful osteoporotic fracture in your spine, the best evidence suggests that most people improve a great deal over the following weeks and months, whether or not cement is injected. Good pain relief, gentle activity, physiotherapy and treatment of the underlying osteoporosis usually form the core of care.

The Cochrane reviewers advised that patients should be told about both the lack of important benefit and the potential for harm. The VERTOS four authors suggested that cement might still have a place for a selected group whose pain doesn’t settle, but that hasn’t been proven. If vertebroplasty is suggested to you, it’s perfectly reasonable to ask why, what the evidence shows for someone in your situation, and what the alternatives are.

And the wider lesson? Sometimes a treatment seems to work, and patients really do get better. But a fair test, with a convincing sham, can show that the improvement would have happened anyway. That’s why these trials are landmarks.

One safety point. If back pain comes with new weakness, numbness, or loss of control of your bladder or bowels, or if you also feel unwell with a fever or unexplained weight loss, please get medical help promptly.

Thank you for listening. Please remember that this podcast is general information and not a substitute for advice about your own situation, so do speak to your own doctor. You’ll find the full article, the diagrams and all the references at spinemathema dot com. And if you found this helpful, please subscribe to the Spinemathema channel on YouTube.

This podcast is general information and is not a substitute for advice about your own situation. Please speak to your GP or specialist.