Abstract & Commentary
Synopsis: A prospective, randomized, placebo-controlled trial found arthroscopy no better than sham surgery for treatment of arthritis of the knee.
Source: Moseley JB, et al. N Engl J Med. 2002;347:81-88.
Although arthroscopic debridement is a common procedure for knees with degenerative arthrosis, most of the evidence regarding its efficacy is based on retrospective, noncontrolled trials. Moseley and colleagues at the Houston Veterans Affairs Medical Center developed a carefully controlled trial to specifically examine the effectiveness of arthroscopy for arthritic knees.
Approximately 60 patients were prospectively randomized into 1 of 3 groups. These included: 1) simple arthroscopic lavage with at least 10 liters of fluid; 2) the same arthroscopic lavage plus debridement of any meniscal, impinging chondral, or synovial lesions; or 3) a sham operation in which incisions were made, the patient was anesthetized, and the OR environment made to mimic a surgical experience from the patient’s perspective but no instruments were introduced into the knee. For the lavage-only group, if the surgeon encountered an entrapped bucket handle tear it was resected. For the debridement group, no abrasion chondroplasty or microfracture was performed, but only shaving of rough articular cartilage and meniscal tears. Impinging osteophytes that blocked extension in the intercondylar notch were removed. All observers and patients were blinded as to treatment group. They were assessed at multiple time points ranging from 2 weeks to 2 years with the use of 5 different self-reported scores, 3 of which involved pain and 2 of which involved function. Additionally, they were timed for maximal speed walking and stair climbing.
At no time point did either of the arthroscopic treatment groups show any improved outcomes over the placebo group that had sham surgery. Furthermore, at some of the earlier time points the arthroscopic treatment groups had even worse scores than the placebo groups for some of the variables. Through a careful statistical analysis Moseley et al conclude that their study methods and rating scales had sufficient power to detect any differences that would have been evident. They conclude that arthroscopic treatment for arthritic knees does not show any clinical benefit and should not be performed.
Comment by David R. Diduch, MS, MD
This paper has already gotten a lot of press and attention from the public, and like other papers we have seen published in nonorthopaedic journals, the orthopaedic issues have already been misrepresented. To its credit, this was a well-conducted study with careful statistical analysis. There was a pilot study to determine statistical power and size of treatment groups needed, patients and observers were all blinded, it was prospective, it was controlled, it was randomized, and very few patients were lost in follow-up. One concern is the validity of the scoring systems used to evaluate differences among patients. Most scoring systems were components of other larger scales and involved just a few questions per scale. None of the conventional orthopaedic scoring systems were used, such as HSS scores or knee society scores. Nevertheless, I do believe that their assessment and, certainly, their sample size, was sufficient to detect any major differences among treatment groups. In fact, there were no differences.
So should this change what we do? I think it should just reinforce what we do but perhaps change the way we present things to patients. Arthroscopy to treat mechanically impinging meniscal, chondral, or loose body lesions is well established. This paper does nothing to affect our recommendations of arthroscopy for these proper indications. We have known from other retrospective studies that the outcome for arthroscopic debridement for an arthritic knee is not very predictable. Indeed, we need to get weightbearing (preferably flexion) views of the knee to carefully assess for arthrosis before going straight to arthroscopy. I know I see a problem in the managed care environment with the primary care physician going straight to a MRI that will always show meniscal and chondral pathology in an arthritic knee. The focus then becomes the meniscal pathology and the patient gets arthroscopy before a proper assessment on the amount of arthrosis is done. That is the patient that does not get better and the type of patient that this study analyzed.
There do remain appropriate indications for arthroscopy in patients with arthritic knees, however. The patient too young for joint replacement, and without malalignment to require osteotomy, and with mechanical type symptoms of catching, giving way, locking and effusions, who has failed other nonoperative measures to include NSAIDs, exercises, and injections is certainly an appropriate candidate. These patients should be properly counseled as to the unpredictable nature of the outcome, but when everything else has failed arthroscopy can be of benefit. Unfortunately, the lay press has misrepresented this study and people are questioning the value of arthroscopy in general. It is important to remember that this study only looked at arthritic knees and really is information that we already knew and should have been presenting to our patients in honest fashion all along.
Dr. Diduch, Associate Professor, Department of Orthopaedic Surgery, University of Virginia School of Medicine, Charlottesville, VA, is Editor of Sports Medicine Reports.