August 7, 2009
Vertebroplasties: Not Very Valuable
Osteoporosis, the demineralization and weakening of bones, is common in older patients. A potentially incapacitating consequence of osteoporosis is a vertebral fracture, in which one of the vertebrae in the spine collapses and breaks under the weight it’s carrying. Like other broken bones, this is frequently very painful. Sometimes the fractured vertebra heals and the pain resolves after some time, but other times the pain can be incapacitating and prolonged.
A few years ago a procedure called vertebroplasty was developed to stabilize fractured vertebrae and provide some pain relief. In it, a radiologist numbs the skin over the broken vertebra with a local anesthetic, then inserts a needle into the broken vertebra and injects some surgical cement. The thought is that as the cement hardens it fuses the broken fragments of the vertebra and thereby gets rid of the pain. Last year this minor surgery was done about 100,000 times in the U.S. It is occasionally spectacularly successful. Some patients who are initially bed-bound in pain are walking comfortably a day later.
We physicians want to help patients and need to believe we are helping patients. So it’s perhaps not surprising that this new procedure which was generally assumed to be helpful was never rigorously tested, until now. Two studies in this week’s New England Journal of Medicine tested the effectiveness of vertebroplasty for vertebral fractures.
The designs of the studies were ingenious. Patients with vertebral fractures were randomized to vertebroplasty or sham surgery. The patients agreed at enrolment that they would not know which procedure they received. The sham surgery consisted of the application of the local anesthetic, and in one study even the insertion of the needle into the broken vertebra, but without the infusion of the cement. Because the cement has a strong scent, the radiologist even opened a container of cement during the sham surgeries to let the odor fill the room.
Both studies showed the same surprising result: patients receiving the sham surgery had as much pain relief as patients receiving vertebroplasty. Both the sham and vertebroplasty groups improved, both immediately and months later. But there was no benefit of vertebroplasty over sham surgery.
How can this be? How can we have done hundreds of thousands of procedures which are no better than placebo? Asked another way: how can the placebo be so good?
One explanation is that the natural history of vertebral fractures is very favorable. Fractures tend to heal naturally. So just as with colds, anything you do for a vertebral fracture will appear effective since you’re intervening in a problem that is likely to improve anyway.
Another explanation is what statisticians call regression to the mean. Illnesses tend come to medical attention when symptoms are at their worst, so on average symptoms for stable illnesses will improve after medical attention no matter what is done.
The final explanation is the power of the placebo effect. Patients want to get better, and they know that the physician expects them to improve. For subjective outcomes such as pain, expectations are a powerful treatment. Many studies have shown the surprising efficacy of placebos, and some have shown that an invasive procedure has an even stronger placebo effect than a sugar pill.
The lesson for doctors is that we need to keep reminding ourselves to test our assumptions. Just because we mean well doesn’t mean we’re helping. The lesson for patients is that just because you’re better doesn’t mean we helped.
New England Journal of Medicine articles and editorial:
A Randomized Trial of Vertebroplasty for Painful Osteoporotic Vertebral Fractures A Randomized Trial of Vertebroplasty for Osteoporotic Spinal Fractures Balancing Science and Informed Choice in Decisions about Vertebroplasty
Wall Street Journal article: Spine Surgery Found No Better Than Placebo
Important legal mumbo jumbo: