Surgery for Knees May Be Useless
Every year, 650,000 people undergo painful and costly knee surgery to relieve the pain of arthritis in a procedure that may actually be useless, medical researchers said Wednesday.
In the first systematic assessment of arthroscopic surgery to treat osteoarthritis, the researchers discovered that people who underwent a sham operation felt just as much pain relief, and sometimes even more, than those who underwent the real knee surgery, according to a study published in the New England Journal of Medicine.
By challenging one of the most popular procedures for the most common form of arthritis, the findings cast doubt on the billion-dollar industry of arthroscopic knee surgery. More than a third of all such surgeries in the United States are performed to relieve osteoarthritis pain.
“We have shown the entire driving force behind this ... industry is the placebo effect,” said Dr. Nelda P. Wray, a health services researcher at the Houston VA Medical Center and Baylor College of Medicine. Wray was the senior scientist on the study.
The procedure first became popular in the 1980s and early 1990s but was never rigorously tested. Unlike new drugs or medical devices, which must undergo extensive clinical trials to prove their effectiveness before they can be used, most surgical procedures aren’t clinically tested.
The only way to rigorously assess surgery is to compare the real procedure with a fake without the patient knowing which would be applied. Such unorthodox clinical trials using placebo surgeries have been notoriously difficult to do because of ethical concerns and the difficulty of recruiting patients.
The study conducted by Wray and her colleagues is only the second placebo surgery trial in the last 40 years, experts said.
The researchers tracked 180 patients for two years. Some of the patients received the surgery while others received placebo surgery, in which a small incision was made but nothing else was done.
Both groups of patients reported virtually the same moderate improvements in pain and joint function, the researchers reported.
Dr. Bruce Moseley, lead author of the study and a clinical associate professor of orthopedic surgery at Baylor, said that there is no reason to perform arthroscopy for arthritis knee pain, although the procedure still helps people with tears to the ligament or meniscus.
Despite the findings, some doctors defend the procedure. Seaid Rezaian, a Beverly Hills physician who performs these surgeries, said that if the patient is younger and the destruction doesn’t involve cartilage of the joint, the surgery will be very effective.
In the study, patients received one of three procedures: an arthroscopic lavage, in which fluid is washed through the knee join; a lavage and debridement, in which torn or loose cartilage is removed from the knee; or a sham surgery.
During the two years of follow-up after the surgeries, researchers measured changes in the patients’ pain and knee function, assessing if they could lift groceries better and climb stairs faster.
All groups felt the same degree of improvement, leading researchers to conclude that the benefit could be attributed entirely to the placebo effect. At several points in the trial, the placebo group actually reported better outcomes.
Some doctors aren’t surprised by these findings. Dr. David McAllister, assistant professor of orthopedic surgery at the David Geffen School of Medicine at UCLA, said, “The treatment never made any sense. The proposed advantage is to wash away debris, but that doesn’t change arthritis.”
Moseley, who is also an orthopedic surgeon for the Houston Rockets and Houston Comets basketball teams, had his own skepticism regarding the procedure. During his early medical training, Moseley was “struck by how little visual change there was after surgery.”
Wanting to determine exactly how the surgery might help arthritis, Moseley went to Wray, who heads the Houston Center for Quality of Care and has expertise in designing clinical trials. Wray convinced Moseley that if they truly wanted to understand what the effect the procedure had on knee pain, they would need to include a placebo surgery into the trial, a very difficult thing to do in the ethical climate of today’s medical community.
There have only been two previous placebo controlled surgery trials in this country, the first in the 1950s on a commonly used procedure for angina. Doctors found the sham surgery worked as well as the real one, and the procedure went the way of the dinosaurs.
Placebo surgery fell out of favor for the next 40 years. It was considered too unethical. Then in 1999, researchers tested the effect of fetal cell injections into brains of Parkinson’s patients. Half of the 40 trial participants underwent sham surgery. Three of those reported feeling better.
But critics of the Parkinson’s study argued that patients were undergoing surgical risk without the possibility of reward, an argument that has made it difficult to conduct placebo surgery trials.
Dr. Baruch Brody, director of Baylor’s Center for Medical Ethics and Health Policies, saw the issue another way. “This surgery has been done in thousands of people without any indication that it helps. What’s unethical is doing these procedures without a clinical trial.”
Wray and Moseley worked closely with Brody to design a trial that met strict ethical standards. They made sure the placebo surgery carried little risk: Patients were sedated rather than given the general anesthesia administered to those undergoing the real surgery.
Researchers also made sure that patients understood they might undergo fake surgery. Almost half refused to participate, which is considered relatively high, indicating to Brody that participants clearly understood the issue.
Wray emphasized the study’s wider implications for health policy. She hopes the findings will convince the medical community that “the placebo effect is important in health care, and it should be understood and used ethically.”
Brody, the study’s ethics collaborator, concurs. “There are lots of other surgeries that have never been validated,” and this study provides an example of how to test them.
The best candidates for placebo trials are surgeries for relief of subjective symptoms like pain, he said.
Paul Shekelle, a physician at the VA Greater Los Angeles Healthcare System and associate professor of medicine at the UCLA medical school, said the “implications are potentially profound.”
“What is the appropriate thing to do for these patients? No one thinks it is ethical to lie to a patient or to do a useless procedure,” said Shekelle. “Should we give them a procedure that isn’t better than placebo?”
