Vertebroplasty - Spinal Fracture Repair

Five sham controlled trials. One of them says yes.

Vertebroplasty is the most tested and most disputed procedure in spine care. This page sets out what each of the blinded trials measured, why four found nothing and one found a real effect, and the narrow group of patients the disagreement leaves room for.

What a vertebral compression fracture is

A vertebra is a block of bone with a hard shell and a spongy interior. Osteoporosis thins the spongy part until the block can no longer carry the load it carries every day, and at some point it collapses, usually at the front, so the block becomes a wedge and the spine bends a little further forward at that level.

Nothing has to happen to cause it.

People describe lifting a kettle, coughing, stepping off a curb, or simply waking with a band of pain across the back that was not there the day before, and that pain sits at one level, runs far worse standing than lying down, and can be severe enough to keep somebody in bed for a week. On an MRI the collapsed block glows, which is how a radiologist distinguishes a fracture that happened recently from one that healed years ago, and that distinction matters more on this page than on almost any other, because everything the evidence has to say on treating these fractures turns on how old the fracture is. Most of them heal on their own over six to twelve weeks whether anybody intervenes or not. That fact sits underneath the entire argument that follows. It also makes this condition unusually easy to appear to treat successfully, since almost any intervention delivered in week two will be followed by improvement in week six, and a clinician who does ten of these and sees nine patients improve has learned nothing at all about whether the ten treatments helped.

The operation everyone argues about

Most pages describing a procedure summarize the evidence in a paragraph and move on. This one cannot, because vertebroplasty has been tested against a dummy version of itself five separate times, which almost no operation in any specialty has, and the results did not agree.

Four said no. One said yes.

That single sentence is the honest state of the field, and it explains why two hospitals in the same city can give you opposite advice on the same scan without either of them lying to you. The rest of this page walks through the trials one at a time, because a patient deciding whether to fly somewhere for an injection of cement deserves to see the actual disagreement rather than a confident summary of whichever half of it a department happens to believe. We perform this procedure. We also decline it more often than we perform it, and the sections below explain exactly where we draw that line and why. Nothing on this page is written to talk you out of treatment or into it. It is written so that when somebody quotes a trial at you, in this clinic or any other, you can ask which trial, how old the fractures were, and what the comparison group received.

Who gets these fractures

Across the trials on this page the participants were mostly women, and the average age in the various studies ran between the early sixties and the early eighties. Those numbers describe the population, and they shape every decision taken for it.

The two other groups

Cancer that has spread to a vertebra causes the same collapse for a different reason, and the evidence discussed here does not apply to those patients, who are assessed alongside an oncologist and often treated with radiotherapy as well as or instead of cement. Long term steroid treatment thins bone in the same way osteoporosis does, sometimes faster, and produces fractures in people who are considerably younger than the usual picture. Both of those groups need a different conversation from the one on this page, and the first job of any assessment is to work out which of the three you are in, because a wedge shaped vertebra on an X-ray looks broadly the same in all three. Age helps, history helps more. A collapsed vertebra in a woman of seventy nine with a previous wrist fracture and no cancer history reads one way, and the identical picture in a man of fifty two on long term steroids for a lung condition reads another, and the same picture again in somebody with treated breast cancer reads a third way and needs a bone scan before anybody discusses cement.

What happens without a procedure

Knowing what conservative treatment actually involves is necessary before any comparison makes sense, and it is more than being told to go home and rest.

• Pain relief on a schedule instead of on demand, stepped up until the person can stand and walk, which frequently means opioids for a short period and a plan for stopping them.
• Early mobilization, since bed rest beyond a few days costs muscle and bone in exactly the population that can least afford to lose either.
• A brace in some cases, worn for a few weeks, which lets some people stand upright and which others abandon within two weeks because they cannot bear it.
• Treatment of the osteoporosis itself, which is the part that changes what happens over the following years and the part most often forgotten.
• Time. Most of these fractures settle over six to twelve weeks as the bone knits, and the pain curve falls steeply in the first month whatever anybody does, which is the single fact every comparison in this field has to be read against.

That last point is what makes this field so hard to read. Everybody gets better. The question the blinded trials set out to answer was whether the cement makes people better faster than they were getting better anyway.

Conservative treatment has a real cost of its own, and pretending otherwise would be dishonest in the other direction. Strong painkillers in somebody of eighty bring confusion, constipation and falls, and a week spent flat in bed costs muscle that takes months to rebuild and sometimes never comes back. So the case for intervening has never rested on the cement being wonderful. It rests on the alternative being expensive in an elderly body, which is exactly the argument the positive trial was built around.

What the procedure is, and what the day looks like

Vertebroplasty counts as an injection more than an operation. You lie face down, either sedated and awake or under a general anesthetic depending on the case and on how much pain lying flat causes you.

Needle, cement, and two hours

Two large needles are passed through the skin of the back and driven through the pedicles, the two bony columns at the back of each vertebra, into the collapsed block itself, with a live X-ray showing the position at every stage. Bone cement mixed to the consistency of toothpaste then goes in slowly under continuous imaging, so the operator watches it fill the bone and stops the moment it approaches an edge. The cement sets hard within minutes. Count on roughly an hour per vertebra treated. Afterward you lie flat for an hour while it finishes hardening, sit up at an hour, walk at two, and most people stay a single night for observation and go home the next morning. Bending, lifting and strenuous activity wait two weeks. The needles leave two puncture marks and no stitches. There is no wound to dress, no drain, and nothing to remove later, which is part of why the procedure reads as low stakes to patients and to the doctors offering it.

Nothing in that description is difficult, which is part of the problem. A procedure this quick and this tolerable is easy to offer to anybody with a painful back and a wedge on their scan, and the temptation to do so is precisely what the trials in the next sections were built to test.

Kyphoplasty, and whether the balloon changes anything

Kyphoplasty adds one step to the same procedure. A balloon goes in through the needle and is inflated inside the collapsed vertebra first, making a cavity and partly restoring the height of the block, and the cement then goes into that cavity at lower pressure.

The argument for it is mechanical and reasonable. Lower injection pressure ought to mean less cement escaping where it should not go, and restoring some height ought to matter for a spine that is bending forward.

Whether it delivers is another matter. Seven randomized trials covering 968 patients have compared the two directly, and they sit inside the same pooled review that found no clinically important benefit for vertebroplasty against a dummy procedure in the first place. The honest framing, then, is that kyphoplasty may be a better engineered version of a procedure whose core benefit is disputed, and a patient being steered toward the more expensive of the two should ask what that upgrade is buying in their particular case rather than in general. The height restored runs to a few millimeters, and nobody has shown that those millimeters translate into a straighter spine ten years later. Posture after these fractures is set by the whole column and not by one block. Somebody with four old wedged vertebrae and a new fifth one will stand no straighter for having the fifth one propped up, and the stooped posture they came in with belongs to the four that healed before anybody saw them.

The two trials that stopped the field

In August 2009 the same issue of one journal carried two independent sham controlled trials, one from Australia and one from the United States, and both reported the same thing. Vertebroplasty performed no better than a procedure that looked and felt identical and involved no cement at all.

What the two 2009 trials reported

The Australian trial randomized 78 patients with one or two fractures under twelve months old, confirmed unhealed on MRI, and found no significant advantage in any measured outcome at one week or at one, three or six months, with a between group difference in pain at three months of 0.6 points on a ten point scale and a confidence interval that crossed zero. The American trial randomized 131 patients to cement or to a simulated procedure and found a difference of 0.7 points in disability and 0.7 points in pain at one month, neither of them significant. Both groups improved immediately after their intervention. Neither improved more than the other.

One detail from the American trial deserves its own line, because it captures how uncomfortable these results were. At three months, 51 percent of the patients in the sham group crossed over to have real cement, against 13 percent of those who had already received it, so the people who had been given nothing were unhappy enough to want the procedure while the measurements still showed no difference between the groups. Both of those things are true at once, and holding them together is the hard part. Patients wanted it. The instruments could not find it. Whenever a treatment produces that pattern, the usual explanation is that the act of being treated carries a good deal of the effect, and the specific ingredient carries less than anybody expected.

The trial with the honest sham

A standing objection to the 2009 trials was that their sham procedures were not really nothing. Needles touched bone, local anesthetic went in, and patients felt something happen. So a Dutch group ran the comparison again in four community hospitals, with the design pushed further.

Practice moved after those two papers, and it moved sharply. Rates of vertebroplasty fell across several health systems, some insurers stopped paying for it, and professional bodies split into those who revised their guidance and those who argued the trials had enrolled the wrong patients. That argument is still running, and the next three sections are what it consists of.

That trial carries more weight against routine use than any other, because it removed the objection raised against its predecessors and produced the same answer over a full year of follow up.

The trial that went the other way

Then comes the one that found an effect, and it explains why anybody who reads this literature carefully still offers the procedure at all. An Australian group argued that the earlier trials had enrolled the wrong patients, and designed a trial around a much narrower group.

What the positive trial did differently

It enrolled 120 patients, all with fractures under six weeks old and all with back pain of at least 7 out of 10, meaning pain that medication was failing to control. It used a deliberate technique aimed at filling the vertebra properly rather than simply placing cement in it. At fourteen days, 24 of 61 patients in the cement group had pain below 4 out of 10, against 12 of 59 in the placebo group, a gap of 23 percentage points with a confidence interval running from 6 to 39 and a probability value of 0.011. The authors concluded that vertebroplasty is superior to placebo for pain reduction in acute fractures of less than six weeks.

Read the entry criteria again, because they are the whole finding. Fractures under six weeks. Pain of seven or more out of ten. Patients whose medication was not holding them. Most people who write to us about vertebroplasty meet none of those three. They have back pain that has run for months, a wedge on an old X-ray, and painkillers that take the edge off, and somewhere they have read that a simple injection will fix it, which is a sentence written by somebody selling injections. That describes a far smaller group than the one being offered cement in clinics around the world, and a far smaller group than most hospital websites list as candidates.

Why the trials disagree

Three explanations circulate, and they do not exclude one another. Which of them you find convincing determines what you think should be done, so they are set out side by side below.

Note that the pooled review tested the first explanation directly and rejected it, reporting that the effects did not differ according to whether the pain had lasted six weeks or longer. That analysis pooled patients across trials rather than running a fresh one, which is a weaker way to answer the question than designing a study around it, and the people who ran the positive trial say exactly that. So the argument is live rather than settled. That finding deals a serious blow to the acute fracture argument, and any department that quotes the positive trial without mentioning it has handed you one side of a two sided result.

The pooled verdict

Twenty one randomized trials have now been gathered into one analysis, five of them against a placebo procedure with 541 participants between them. The pooling produced the figures below, measured at one month.

This table scrolls sideways on a narrow screen. Swipe or drag to see every column.

Who should still have it

Having laid all that out, here is our actual practice, stated plainly.

The narrow group, and the wide one we turn down

We consider cement for the patient the positive trial described, and for nobody outside that description. A fracture under six weeks old with bright signal on MRI, pain at seven out of ten or above, and medication that is failing to hold it, usually in somebody who cannot get out of bed and is losing muscle by the day. In that situation the potential gain is real enough to weigh against a small procedural risk, and the alternative is a further three weeks flat in bed with everything that costs an eighty year old. Outside it we say no. Fractures older than six weeks, pain that analgesia is controlling, back pain with an old healed wedge on the scan and no bright signal, and multiple chronic fractures in somebody who wants their posture corrected are all cases where the blinded evidence gives us nothing to offer and where saying so is the service. A department that agrees to treat all of those has stopped reading the trials. The line is uncomfortable to hold in a clinic room. Somebody has flown a long way, they are in real pain, the scan shows a real fracture, and the procedure is quick and available, and every incentive in private medicine points toward doing it. What the evidence supports and what a patient in front of you wants are frequently different things, and this is one of the places where the gap between them is measurable.

If you send us a scan and the answer comes back no, the letter says so and says why, because we would rather lose a booking than take payment for a procedure the evidence says will leave you where it found you. That has cost us patients. It has also never cost us one who came back angry.

Risks, and the osteoporosis nobody treats

Serious complications stay uncommon, with reported rates under 2 percent for osteoporotic fractures and higher when the vertebra has been destroyed by a tumor. Uncommon and absent are different words, and the specific risks follow from what is being done.

What can go wrong, and the thing that goes wrong afterward

• Cement escaping the vertebra, which is common on imaging and almost always harmless, and which occasionally reaches a nerve root or the canal and causes pain or weakness.
• Cement entering a vein and traveling to the lungs, which is the mechanism behind the respiratory failure reported in the trials, and the reason injection is slow and imaged throughout.
• Infection in the treated bone, rare and serious, since cement cannot be washed out the way a wound can.
• A new fracture at the level above or below, reported more often after cement in the pooled data at 48 of 418 against 31 of 422, with a confidence interval too wide to settle whether the cement causes it.
• Bleeding, bruising and a sore back for a few days, which is the ordinary outcome and needs nothing.

Now the part that matters more than any of that. A vertebral fracture from minor force is a diagnosis of established osteoporosis, and it is the single strongest predictor that somebody will break another bone. Roughly half of the people who break a vertebra this way will break another within a few years if nothing is done for the bone itself. Cement treats one block of bone for a few weeks. It does nothing for the disease that broke it, and the patient who goes home with a treated vertebra and no bone protection has been handed a local repair and left with the condition intact. So every one of these patients needs a bone density measurement, blood tests for vitamin D and calcium and the treatable causes of bone loss, and a drug treatment started and continued for years. That part of the plan does more for your future than the injection does, and a hospital cannot sell it as a package, and that is why it goes missing. Ask for it by name. A written bone plan naming the drug, the dose, the duration and the date of the next review is a deliverable you can hold a hospital to, it costs almost nothing to produce, and a department that treats your vertebra without producing one has treated the symptom and walked past the disease. The drugs themselves are unglamorous and they work. A weekly tablet, a six monthly injection or a yearly infusion, chosen according to what your kidneys and your dentist allow, cuts the chance of the next spinal fracture substantially, and unlike the cement that effect has never been in dispute.

Cost and having it done in Istanbul

No price appears on this page, partly because it would be out of date by the time you read it and partly because the first question here is whether the procedure should happen at all. For this treatment more than any other on our site, the assessment is the product.

This table scrolls sideways on a narrow screen. Swipe or drag to see every column.

Plan on four to six nights in the country. That covers the assessment, the procedure, the overnight observation, the bone density study and the endocrinology appointment that sets up your treatment for the years that follow, which is the appointment worth flying for even when the answer on the cement is no. Back home, your own doctor continues that treatment, and we send them the procedure report, the images, the bone density result and the drug plan with its review dates. Send that packet on at your first appointment rather than describing it, since a bone drug started abroad and never written down tends to stop quietly at the six month mark, which is the point at which it was starting to do something.

Send the MRI and the date the pain started. Those two settle most of it.

Vertebroplasty FAQ

The questions below arrive in most first messages from abroad.

References

1. Buchbinder R, Johnston RV, Rischin KJ, Homik J, Jones CA, Golmohammadi K, et al. Percutaneous vertebroplasty for osteoporotic vertebral compression fracture. Cochrane Database of Systematic Reviews. 2018;11(11):CD006349.
2. Buchbinder R, Osborne RH, Ebeling PR, Wark JD, Mitchell P, Wriedt C, et al. A randomized trial of vertebroplasty for painful osteoporotic vertebral fractures. New England Journal of Medicine. 2009;361(6):557-568.
3. Kallmes DF, Comstock BA, Heagerty PJ, Turner JA, Wilson DJ, Diamond TH, et al. A randomized trial of vertebroplasty for osteoporotic spinal fractures. New England Journal of Medicine. 2009;361(6):569-579.
4. Firanescu CE, de Vries J, Lodder P, Venmans A, Schoemaker MC, Smeets AJ, et al. Vertebroplasty versus sham procedure for painful acute osteoporotic vertebral compression fractures (VERTOS IV). Randomised sham controlled clinical trial. BMJ. 2018;361:k1551.
5. Clark W, Bird P, Gonski P, Diamond TH, Smerdely P, McNeil HP, et al. Safety and efficacy of vertebroplasty for acute painful osteoporotic fractures (VAPOUR). A multicentre, randomised, double-blind, placebo-controlled trial. Lancet. 2016;388(10052):1408-1416.