Correspondence
Arthroscopic Surgery for Osteoarthritis of the Knee
N Engl J Med 2002; 347:1717-1719November 21, 2002
- Article
To the Editor:
Although the study by Moseley et al. (July 11 issue)1 was based on sound scientific principles and included a dramatic control (a sham operation), it was flawed because of the selection of patients. The study group comprised veterans, 75 years of age or younger, almost all of whom were men. This cohort of patients does not truly represent the general population. Osteoarthritis does not suddenly appear late in life, when increasing symptoms are accompanied by radiologic changes. Osteoarthritis is a degenerative process that starts with the earliest changes in the articular cartilage seen at arthroscopy and progresses to the secondary changes that can finally be identified by plain-film radiography.
Although I am in full agreement with the authors that arthroscopy is of little value in the more advanced stages of osteoarthritic degeneration, it is my experience that arthroscopic lavage and débridement can dramatically improve symptoms in the early stages of osteoarthritis and can do so for several years.
To cast doubt on all arthroscopic surgery for osteoarthritis, with the implication that it is worthless, does a disservice to the increasing number of people in our population who have early stages of arthritis and might deprive them of this treatment.
1 ReferencesRobert W. Jackson, M.D.
Baylor University Medical Center, Dallas, TX 75246-99901
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
Full Text | Web of Science | Medline
To the Editor:
No responsible orthopedic surgeon will claim that arthroscopic surgery is a help to all patients with arthritis of the knee. That would be untrue and inappropriate. It is just as inappropriate to state that arthroscopic surgery is useless in any case of arthritis of the knee. There is a subgroup of patients with arthritis of the knee that can be substantially helped with appropriate arthroscopic surgery. This point is alluded to in the editorial1 that accompanies the report by Moseley et al.
The selection of patients is all important in arthroscopy of the knee in those with pain and early degenerative arthritis. Plain-film radiography during posterior–anterior flexion in a weight-bearing position may be the most important study in the evaluation of patients with knee pain. These evaluations were apparently not done in the group selected for the reported study. Minor alignment problems were apparently also disregarded. One could therefore predict, in advance of the study, the results that were obtained in a very elegant fashion.
1 ReferencesWhit Ewing, M.D.
John W. Ewing, M.D.
Arthroscopy Association of North America, Rosemont, IL 600181
Felson DT ,Buckwalter J . Débridement and lavage for osteoarthritis of the knee. N Engl J Med 2002;347:132-133
Full Text | Web of Science | Medline
To the Editor:
Moseley et al. used the pain subscale of the Medical Outcomes Study 36-item Short-Form General Health Survey to determine the power of their study but then used the score on their own Knee-Specific Pain Scale as the primary end point. There was no link between the Knee-Specific Pain Scale as an outcome measure and the design of the study with respect to sample size. This instrument is a nonvalidated measurement and was “created for this study.” Though designed as a superiority trial, the study was converted into an equivalence trial. The authors defined post hoc equivalence bounds, or “minimal important differences,” which they “calculated on the basis of the trial data.”
To avoid data-dependent bias, equivalence bounds should be determined in advance.1,2 Furthermore, it is well established that equivalence trials generally require greater power than superiority trials. Piaggio and Pinol3 state, “The use of conventional superiority approach to design equivalence trials has led underpowered trials to show equivalence within clinical relevant margins.” Application of the standard power formula for equivalence studies4 to the data presented by Moseley et al. reveals a power range of 14 to 70 percent for different variables (Table 1Table 1
Power to Demonstrate Equivalence.), with all the values in the range below accepted levels for claiming equivalence. Most of the values are so low that any conclusions about equivalence are meaningless. It is disturbing that the outcome with the greatest post hoc power was the one chosen as the primary end point, especially because it was created for the study and may have lacked validity as an outcome measure.4 ReferencesKeith G. Chambers, M.D.
Michael Schulzer, Ph.D.
Vancouver Hospital, Vancouver, BC V5Z 1L8, Canada
kchamber@vanhosp.bc. ca 1
Jones B ,Jarvis P ,Lewis JA ,Ebbutt AF . Trials to assess equivalence: the importance of rigorous methods. BMJ 1996;313:36-39[Erratum, BMJ 1996;313:550.]
CrossRef | Web of Science | Medline2
Ebbutt AF ,Frith L . Practical issues in equivalence trials. Stat Med 1998;17:1691-1701
CrossRef | Web of Science | Medline3
Piaggio G ,Pinol APY . Use of the equivalence approach in reproductive health clinical trials. Stat Med 2001;20:3571-3577
CrossRef | Web of Science | Medline4
Fleiss JL . General design issues in efficacy, equivalency and superiority trials. J Periodontal Res 1992;27:306-313
CrossRef | Web of Science | Medline
To the Editor:
Moseley et al. note that about 13 percent of the patients in both the placebo and the intervention groups believed that their procedures were placebos. Do the authors have any data about these two subgroups? Did these patients have the same rate of relief of symptoms as the overall group, or did the mind-set of believing the procedure was a placebo influence the outcome?
Richard S. Blacher, M.D.
Tufts University School of Medicine, Boston, MA 02111
rblacher@massmed.org To the Editor:
In a Sounding Board article accompanying the report by Moseley et al., Horng and Miller ask, “Is placebo surgery unethical?”1 The Council on Ethical and Judicial Affairs of the American Medical Association, which I chair, has published an ethical opinion2 that addresses the same question. The appropriateness of surgical placebo controls should be based on the following points. Surgical placebo controls should be used only when no other trial design will yield the requisite data; informed consent should receive particular attention, with disclosure of the risks of the operation and a description of the differences between the study groups in the trial; the use of surgical placebo controls is not justified when testing the effectiveness of an innovative surgical technique that represents a minor modification of existing procedures; surgical placebo controls may be justified when a new surgical procedure is developed to treat a condition for which no surgical procedure exists but must be weighed against the benefits, risks, and side effects of the current standard of care, including nonsurgical treatment; standard treatment must be offered as part of the study design, if it is efficacious and acceptable to the patient and if forgoing it would result in injury; when standard treatment is not fully efficacious or is unacceptable to the patient, surgical placebo controls may be used and standard treatment forgone if the informed-consent process includes adequate safeguards.3
The fact that 44 percent of the patients in the study by Moseley et al. declined to participate in this controlled trial of arthroscopic surgery speaks highly of the authors' emphasis on informed consent.
3 ReferencesLeonard J. Morse, M.D.
Department of Public Health, Worcester, MA 01610-27151
Horng S ,Miller FG . Is placebo surgery unethical? N Engl J Med 2002;347:137-139
Full Text | Web of Science | Medline2
Surgical “placebo” controls. In: Council on Ethical and Judicial Affairs. Code of medical ethics: current opinions with annotations: 2002–2003 edition. Chicago: American Medical Association, 2002:28-9.
3
Tenery R ,Rakatansky H ,Riddick FA Jr , et al. Surgical “placebo“ controls. Ann Surg 2002;235:303-307
CrossRef | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Jackson and Ewing and Ewing suggest that, although our results are valid in the patients we studied, there may be subgroups of patients — those with early stages of arthritis, those with normal alignment, and (as noted in the accompanying editorial) those with mechanical symptoms — in whom arthroscopy would be more efficacious. With regard to variation in outcomes related to the severity of arthritis or alignment, as measured by plain-film radiography during posterior–anterior flexion in a weight-bearing position: we have performed an extensive subgroup analysis that does not find differences to suggest any bias. Is this procedure effective in patients with mechanical symptoms? Of the 180 patients enrolled in this study, 172 had one or more mechanical symptoms.
Most of the patients in our study were men; however, as Felson and Buckwalter1 state: “There is no reason to believe that the response to surgery would vary according to sex.” We believe that the group of patients we studied is highly representative of the spectrum of patients in whom this procedure is currently being performed. As suggested by Jackson and Ewing and Ewing, these patients do have improvement after the procedure, but we showed that the benefit of arthroscopy is not greater than the benefit of the placebo effect. Given our results, we would argue the following: if someone questions whether arthroscopic surgery would be efficacious in a specific subpopulation of patients, then the ethical way to proceed would be to test the hypothesis by conducting a placebo-controlled trial in that specific subgroup.
When our study was designed, no psychometrically valid scale to assess knee pain was available. Thus, the Knee-Specific Pain Scale was developed for this study by a research team with two psychometricians. Subsequently, we have conducted rigorous psychometric testing that demonstrates that this scale has strong psychometric properties.2 This study was designed as a superiority trial. However, when we failed to show that arthroscopy was superior to placebo, we did not assume equivalence; we statistically tested for equivalence.3,4 As with all power calculations, the power approach to assessing the results of a negative trial addresses only the potential for the trial to detect certain differences (e.g., equivalence). However, the approach we used tested directly for equivalence. The analyses demonstrated equivalence; hence, concern about a priori power then becomes irrelevant.
Many of the outcomes in patients who guessed that they had undergone the placebo procedure were worse than those who guessed they had undergone one of the arthroscopic procedures.
4 ReferencesNelda P. Wray, M.D., M.P.H.
J. Bruce Moseley, M.D.
Kimberly O'Malley, Ph.D.
Houston Veterans Affairs Medical Center, Houston, TX 77030
nwray@bcm.tmc. edu 1
Felson DT ,Buckwalter J . Débridement and lavage for osteoarthritis of the knee. N Engl J Med 2002;347:132-133
Full Text | Web of Science | Medline2
O'Malley KJ, Suarez-Almazor M, Aniol J, et al. Joint specific multidimensional assessment of pain (J-MAP): factor structure, reliability, validity, and responsiveness in patients with knee osteoarthritis. J Rheumatol (in press).
3
Hauck WW ,Anderson S . A proposal for interpreting and reporting negative studies. Stat Med 1986;5:203-209
CrossRef | Web of Science | Medline4
Blackwelder WC . “Proving the null hypothesis“ in clinical trials. Control Clin Trials 1982;3:345-353
CrossRef | Medline
Author/Editor Response
The American Medical Association's ethical opinion on placebo surgery should encourage investigators to plan, and institutional review boards to approve, clinical trials of surgical procedures involving the use of placebo surgery when methodologically warranted and ethically justified.
Sam Horng, B.A.
Franklin G. Miller, Ph.D.
National Institutes of Health, Bethesda, MD 20892-1156
fmiller@cc.nih. gov - Citing Articles (7)
Citing Articles
1
Richard W Nutton. (2009) Is arthroscopic surgery a beneficial treatment for knee osteoarthritis?. Nature Clinical Practice Rheumatology 5:3, 122-123
CrossRef2
John C. Richmond. (2009) Surgery for Osteoarthritis of the Knee. Medical Clinics of North America 93:1, 213-222
CrossRef3
Kirkley, Alexandra, Birmingham, Trevor B., Litchfield, Robert B., Giffin, J. Robert, Willits, Kevin R., Wong, Cindy J., Feagan, Brian G., Donner, Allan, Griffin, Sharon H., D'Ascanio, Linda M., Pope, Janet E., Fowler, Peter J., . (2008) A Randomized Trial of Arthroscopic Surgery for Osteoarthritis of the Knee. New England Journal of Medicine 359:11, 1097-1107
Full Text4
John C. Richmond. (2008) Surgery for Osteoarthritis of the Knee. Rheumatic Disease Clinics of North America 34:3, 815-825
CrossRef5
M. Maravic, P. Landais. (2005) Arthroscopie pour gonarthrose en France en 2001. Revue de Chirurgie Orthopédique et Réparatrice de l'Appareil Moteur 91:8, 768-772
CrossRef6
Thomas J. Errico, Robert J. Gatchel, Jerome Schofferman, Edward C. Benzel, Thomas Faciszewski, Marjorie Eskay-Auerbach, Jeffrey C. Wang. (2004) A fair and balanced view of spine fusion surgery. The Spine Journal 4:5, S129-S138
CrossRef7
Milka Maravic, Paul Landais. (2003) Arthroscopy for knee osteoarthritis. Joint Bone Spine 70:6, 404-406
CrossRef
