Arthroscopic Debridement of the Knee - Sam Church 30/9/2003

A Controlled Trial of Arthroscopic Surgery for Osteoarthritis of the Knee

N Engl J Med, Vol 347, No.2 – July 11, 2002

Moseley et al (Houston)

One of the most talked about papers since I became a SpR. High impact paper in a high impact journal. Even made the national press in this country. Did it deserve all that attention?

They start by saying that although 50% of patients with OA report improvement following arthroscopy, there is no physiological explanation for this.

They took 180 patients <75y.o. (mean 52.3), with >3/10 pain on VAS despite 6 months of maximum medical treatment. None had serious medical problems, severe deformity or >8/12 OA on radiographic assessment. They then randomised them to:

• Lavage: arthroscopy + 10L washout only. However, if an unstable meniscal tear was encountered, it was excised to stability. No other debridement was performed.
• Debridement: chondral surfaces, menisci & tibial spine spurs were shaved smooth + 10L washout.
• Placebo: patients were heavily sedated (no GA), three 1cm incisions were made in the knee and an arthroscopy was simulated (in case the patient remembered anything) – including the sound of splashing water! No instruments entered the knee.

The groups were well matched for all pre-op variables and one surgeon performed all the operations.

Patients were followed up for 2 years with 3 self-reported pain scales, 2 self-reported function scales and one objective function scale (amount of time it took to walk 30m then climb up and down a flight of stairs).

At no point did either arthroscopic-intervention group have greater pain relief or improvement in function than the placebo group. In fact, the only significant difference they found in the whole study was that objective function was better in the placebo group at 2 weeks (p=0.02) and at one year (P=0.04).

In their discussion, they are very careful not to say that arthroscopy for OA is a complete waste of time. They simply state that it is no better than placebo. They accept the fact that these results could have come about because the surgeon was crap at arthroscopy but list his credentials (including surgeon to the NBA & US Olympic basketball team) which are pretty impressive. They also state that only 56% of eligible patients agreed to take part in the study; they may have been so convinced that arthroscopy would help that they were willing to take the chance of a placebo procedure and this may have rendered them more susceptible to the placebo effect (funny thing to say bearing in mind that if they declined they definitely could have had an arthroscopy). They end by saying that the billions of dollars spent on arthroscopy for OA ‘might be put to better use’ and that we should not underestimate the placebo effect, regardless of its mechanism.

CRITIQUE

·          Double-blind, prospective, randomised controlled trial. Solid methods & stats leaving little room for criticism.

·          My main criticism is that they talk about the effect (placebo or otherwise) of arthroscopy but they never quantify it. All they do is compare the outcomes of the 3 different interventions with each other but they do not compare the outcomes with pre-op scores. From a patient’s point of view, the cause of the effect is not as important as the significance of the effect and they do not analyse that at all (despite having collected all the pre-op data). To put it another way, if the effect is significant, does it matter whether or not we understand the cause?

·          Since it’s publication, there have been numerous comments on it in other journals. There seems to be a general feeling that their conclusions were a bit wishy-washy and that a correct interpretation of their results would be that the benefits of arthroscopy for the treatment of OA of the knee is to provide subjective pain relief, and that the means by which it provides this benefit is via the placebo effect.

Please log in to view the content of this page.
If you are having problems logging in, please refer to the login help page.