Knee arthroscopy/placebo study coming under fire

HOUSTON — A knee arthroscopy/placebo study recently published in the New England Journal of Medicine and widely reported in the mass media is facing tough criticism by several orthopedic surgeons.

The research results were not news to any of the knee surgeons who spoke with ORTHOPEDICS TODAY. The inclusion criteria, however, were flawed, they said. Perhaps more damaging was the way that the media covered the article, essentially leading readers to incorrectly conclude that all knee arthroscopy is ineffective.

Conducted by investigators at Baylor College of Medicine and the Department of Veterans Affairs, the trial randomized patients with painful osteoarthritis (OA) to arthroscopic debridement, arthroscopic lavage or a sham procedure. Patients in all three groups had only moderate improvements in pain and function at two-year follow-up.

J. Bruce Moseley, MD, who performed all the procedures in the trial, and colleague Nelda P. Wray, MD, MPH, told ORTHOPEDICS TODAY that they stand by the results of their study and that many criticisms are based on limited information available in the NEJM article. Additional articles providing more details and subgroup analysis have been submitted for publication or are in preparation, they said.

“We feel strongly that the results and conclusions of our study are accurate and true, and that there is a significant placebo effect from arthroscopy for OA of the knee. … We feel that arthroscopy for OA of the knee is beneficial only for a placebo effect, and there are no exceptions.”

Selecting the right candidate

Thomas P. Schmalzried, MD, questioned the patient selection criteria. “The candidate for arthroscopy is not a patient who meets pain and radiographic criteria. It’s a patient who has OA and then has some internal derangement superimposed on the degenerative joint disease. Nobody does arthroscopy just for pain.” Schmalzried is Joint Reconstruction section editor for ORTHOPEDICS TODAY.

Moseley and Wray said that trial participants were representative of typical patients who would be considered for arthroscopy for knee OA after undergoing nonsurgical treatments. Many patients complained of one or more mechanical symptoms or effusion, had pain significant enough to seek treatment, and had exhausted nonsurgical treatments.

Freddie Fu, MD, of the University of Pittsburgh Medical School, said that for endstage arthritis, arthroscopy is less effective yet it remains an essential part of the treatment plan for patients with knee OA. “There are many stages of arthritis before patients require a total knee replacement. That is when arthroscopy is called for, such as staging, diagnosis, to treat a torn meniscus, loose bodies and for microfracture. I probably agree that if you have true arthritis … maybe scoping is not that meaningful,” Fu said.

J. Whit Ewing, MD, executive vice president of the Arthroscopy Association of North America (AANA), said that the “patient selection was flawed.” Elements critical to diagnosing and treating patients with knee OA were omitted or overlooked, such as PA weight-bearing flexion radiographs and consideration of alignment.

AP standing, flexion weight-bearing PA (the “Rosenberg View”), standing hip-ankle AP, lateral and sunrise radiographs were done to determine a numeric OA grade in the medial, lateral and patellofemoral areas of the knee, Moseley and Wray said.

The three grades were totaled for an overall arthritis severity score indicating mild, moderate or severe OA. “We had roughly equal representation for all three groups in our study and the statement that we had only ‘very bad arthritic knee(s)’ is not true,” they said.

William A. Grana, MD, said: “I would ask the authors how they validated the criteria that they used to determine success and failure. To do an outcome study like this you have to validate the instruments you’re using.” Grana is secretary of AANA and is the Sports Medicine section editor for ORTHOPEDICS TODAY.

The researchers said that “most of our measurement instrument was “taken from instruments with proven reliability and validity, such as the SF-36 general health measure, the AIMS-2 pain scale and Wisconsin brief pain questionnaire.

“We also developed some questions … specifically for this study,” which were extensively analyzed and found to have high reliability and validity.

Media coverage disturbing

Media reports of the trial are being criticized. “The article and particularly the Associated Press report disturbs me very much,” said ORTHOPEDICS TODAY Chief Medical Editor John B. McGinty, MD.

Moseley and Wray said they also have concerns about how the media handled the story. “We are in agreement that the media has, in many instances, condemned all arthroscopic surgery on the basis of our study. This has occurred in spite of our efforts to the contrary.”

Kenneth E. DeHaven, MD, of Rochester, N.Y., said his “biggest concern about this study is that it may not be kept in perspective, and that it is automatically extrapolated to all types of arthroscopic surgery.”

Moseley and Wray said that the media have failed to report that arthroscopic surgery is appropriate in unarthritic knees for several indications. “We have emphasized this consistently in our dealings with the media, but this is one of the unexciting details that seems to get edited out of most stories.”

AANA’s response

The negative press has not helped the longtime public relations campaign that members of AANA have been informally conducting concerning knee arthroscopy. “This is a damaging paper in terms of patient confidence,” Ewing said.

He said that members of AANA have tried to educate surgeons that knee arthroscopy is not indicated for patients with arthritic knee pain and swelling alone, and that it should be reserved for those presenting with pain in conjunction with mechanical knee problems, like locking or catching.

In mid-July, Ewing was preparing AANA’s response to the NEJM study, which the organization planned to publish in a major newspaper or medical journal. The document stated: “The article … leads the casual reader to think that arthroscopic surgery is useless for any type of knee surgery involving arthritis. This is most unfortunate.”

AANA called for evaluation of patients at a longer follow-up than two years. “There is too much chance that the placebo effect will still be in effect if the result is determined in less than two years of follow-up.” In response, Moseley and Wray said that two years is the standard for publication in most orthopedic journals and this was a “more than adequate amount of time to evaluate the effectiveness of an arthroscopy surgery.”

The AAOS perspective

The AAOS issued a response, praising the study’s use of evidence-based medicine, a large number of patients, an excellent control group and long-term assessment.

“The weaknesses include a potential selection bias caused by the number of patients who decided not to participate and the nonspecific indications for arthroscopy,” AAOS executive vice president William W. Tipton Jr., MD, reported. Of the 324 patients who met the inclusion criteria, 144 declined to participate.

While the AAOS found the study ethically sound, the organization said the results should be reproduced in a non-VA setting with selection for pre-existing malalignment, mechanical symptoms, body weight and other factors.

“We would also agree with the authors that an expert panel be convened to review the findings of this study and make recommendations as to the selective use of arthroscopy in the treatment of OA of the knee.”

Ethics of placebo

“This method, a randomized controlled study with a control treatment, a placebo method, is the gold standard for all clinical studies, but it is not always ethical. Some people would question whether this was ethical or not,” Grana said.

Two complications were reported in the 165 patients who completed the trial. One patient developed incisional erythema, which was treated with antibiotics. Another had calf swelling in the operated leg. “In no case did a complication necessitate the breaking of the randomization code,” investigators said.

Even though an editorial in the NEJM said the trial “exemplifies the ethically justified use of placebo surgery,” Schmalzried disagreed. “I don’t like the idea of having a placebo procedure,” he said.

The heart of the ethical issues surrounding this study lies in those patients who were treated. “The real editorial should have been on the ethics of operating on people that shouldn’t have been operated on at all,” Schmalzried said.

“We feel very strongly about the ethics of our study,” Moseley and Wray said. “We had a nationally recognized ethicist involved in the study from its onset.” Once it was determined that placebo arthroscopy could be safely performed, “[Baruch A. Brody, PhD] felt it would be unethical to not perform the study.”

For your information:

• Moseley JB, O’Malley K, Petersen NJ, et al. A controlled trial of arthroscopic surgery for osteoarthritis of the knee. N Engl J Med. 2002;347:81-88.
• Horng S, Miller FG. Is placebo surgery unethical? N Engl J Med. 2002;347:137-139.