Response to Placebo Knee Surgery Study: An Expert Interview With Don Johnson, MD

Peter Jhon


July 19, 2002


The following interview with Don Johnson, MD, took place on July 12, 2002. Dr. Johnson is Director of the Carleton Sports Medicine Clinic in Ottawa, Ontario, Canada, and is a member of the Medscape Orthopaedics and Sports Medicine Editorial Board. These are his comments on the recently published article from The New England Journal of Medicine that reported no difference between arthroscopic knee surgery and placebo surgery for the treatment of knee osteoarthritis (OA).

Medscape: Can you provide a summary of the controversial study that was released in The New England Journal of Medicine?

Dr. Johnson: The study[1] took patients from the VA hospital and broke them down into 3 categories: (1) the traditional arthroscopic debridement group, which is really just a clean-up of the joint; (2) a group who had saline fluid flushed through the joint; and (3) a group who had a knife stuck in their skin to create "portals." This last group was the placebo or "sham" arm of the study, as no one actually underwent a surgical procedure. The results that came out were that all of the groups were essentially the same after 2 years in terms of pain relief. The 2 major [pieces of information] that came out of this paper were that surgery has a high percentage of placebo effect, which a number of people didn't think could happen, but it does; and that unfortunately a lot of people will generalize that arthroscopy for knee surgery is not effective.

We have to go back and analyze what the problems are; we have realized for 20 years that arthroscopy for certain types of arthritic knee conditions is not effective. We've been teaching people this for years -- there are simply certain categories of patients that are not suitable for knee arthroscopy. We advise patients and try to select those patients who we feel will do well with surgery. Generally, those who do well have mechanical symptoms -- patients whose knees catch, lock, or give way. This is due to either a loose piece of meniscus or cartilage, and that can be fixed fairly easily.

Medscape: What are the flaws of this study?

Dr. Johnson: The main flaw lies with the selection of patients. The patient selection criterion was arthritis, and it wasn't any narrower than that. We had an opportunity to speak to some of the coauthors when this paper was first presented, and both admitted that the patients picked for this study were not ideal. At the VA hospital, the residents picked off patients who had mechanical problems, the ones who looked like suitable candidates for surgery. The patients who were picked for this study were probably not ones who would have benefited from surgery, and would not be appropriate candidates in my own practice. It wasn't surprising to me that patients didn't get any better from the placebo operation, because you wouldn't expect them to. If they had a category of patients who had mechanical problems, then this group would have probably shown considerable improvement over the other 3 groups.

So the 2 points to be gleaned are that (1) there is a strong placebo effect, but that (2) the study is flawed because of this selection bias of this patient population.

Medscape: What sort of beneficial data came out of this study?

Dr. Johnson: It was surprising that so many patients reported pain relief just from a sham operation. I think this is very interesting, but this doesn't really help the way I practice orthopaedic medicine.

Medscape: How is this information different from the data presented at AAOS 2001?

Dr. Johnson: It's the same paper.

Medscape: What was the reaction of orthopods when this paper was presented in 2001?

Dr. Johnson: About the same as it is now: they disregarded it because of the selection bias in the patient population. The patients in the study were not going to get better with an operation, so of course the results of the operative group would be the same as that of the sham group.

Medscape: Will this study change the way orthopods practice medicine?

Dr. Johnson: No, unless insurance providers decide not to pay for this procedure. It is interesting because this has raised a huge controversy for the states, payers, and insurance providers. If you label arthritis as not amenable to arthroscopic surgery, you are going to do a huge disservice to a number of people. Most of these patients are in the Medicare/Medicaid population. Most of the people who have an arthroscopic debridement are in the older age group.

Medscape: What are the specific indications for patients with OA that would make them candidates for arthroscopic knee surgery?

Dr. Johnson: The indications are pain, swelling, and the mechanical symptoms of catching, locking, or giving way. This means that something is getting in the way of the joint surfaces, and this is a fairly easy thing to fix. They will still have the underlying arthritic condition, so they will have pain-related activity, but their function will be better. These are the ideal patients; once the arthritis becomes too advanced and the leg becomes deformed or bowed, arthroscopic procedures have no role. There are a couple of different spectrums that you don't operate on: (1) those patients who have pain alone, who you can manage conservatively, and (2) those patients whose OA is too severe to surgically manage.

Medscape: What are the nonsurgical options for these patients?

Dr. Johnson: Almost all of our patients first get treated conservatively: rest, activity modification, anti-inflammatory drugs, physiotherapy, etc. If they fail this conservative routine, and if they have mechanical symptoms, these patients may be candidates for surgery. In our practice, this could take a year, so a lot of people are weeded out. In the US, this conservative time frame may be cut back, but most folks agree that one should have between 3-4 months of conservative treatment to see if there is improvement.

Medscape: Was this study ethical?

Dr. Johnson: Yes, I think that the methodology of this study was okay, except for the patient inclusion criteria. They had an outcome measure we don't normally associate with knee OA -- mostly patient satisfaction, and not any other objective measurements, which is questionable, but acceptable. But is it ethical to do a sham operation? No, but I don't think that we should slam the study due to methodology. It was fine; it's just too bad that the study wasn't more clearly defined.

Medscape: Are there additional studies that need to be done in order to refute this study?

Dr. Johnson: I think we have got several studies in the literature to show which patients you should operate on and the efficacy of long follow-up is, such as Bob Jackson's as well as several others.[2,3] I'm not sure we need another study; it's just that this particular study has sort of thrown a wrench into things.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.