I have previously written about the (non) role of arthroscopy for osteoarthritis or degenerative meniscus tears in the knee (here, here and here). Surgeons have continued to operate, based on a belief that (now) centres on the presence of mechanical symptoms. An analysis of the recent sham surgery trial of arthroscopic partial meniscectomy (APM), which showed APM to be no better than sham for patients with meniscus tears without arthritis, has shown that this procedure is no better than sham surgery for patients with mechanical symptoms.
I spend a lot of my time complaining that we don’t have enough high level evidence for many of the procedures we do. For knee arthroscopy, however, we have multiple high level studies, including the recent randomized trial of APM versus sham surgery for degenerative meniscus tears – a study that has been rated as having the least risk of bias of any study in this area, or just about any area. Yet knee arthroscopy is still one of the commonest orthopaedic procedures in the world, with about a million performed each year in the US alone, mostly in middle aged or older patients and patients with degenerative changes in the knee.
Surgeons used to perform this procedure for arthritic knees, and many still do. After the sham surgery study by Mosely in 2002 (and other studies), surgeons changed their tune. They continued to operate, but said it was for the meniscus tear, not the arthritis. Then the study by Sihvonen et al showed that it was not effective for degenerative meniscus tears without arthritis. Surgeons continued to operate, arguing that they were operating for ‘mechanical symptoms’ – a group of symptoms difficult to define and, depending on how you define it, present in most arthritic knees. In fact, nearly all of the patients in the first sham surgery study (by Mosely) had mechanical symptoms.
So Sihvonen and colleagues re-analysed their famous APM sham surgery study in patients with meniscus tears without arthritis, looking at who had mechanical symptoms and who did not, and whether surgery or sham treatment changed that (article is here). They decided to use catching or locking as the definition of mechanical symptoms, which is what most people would use, and this was present in nearly half of the patients. Patients with a knee that would not straighten (a true ‘locked’ knee) were excluded from the initial study.
For those having the real APM surgery, the proportion of patients with mechanical symptoms ROSE from 46% before surgery to 49% afterwards.
For those having sham surgery, the proportion of patients with mechanical symptoms FELL from 49% before, to 43% afterwards.
This means that APM is worse for mechanical symptoms than sham surgery, but the difference is not statistically significant so I cannot conclude that it is worse. Importantly, however, NO ONE can conclude from this that APM makes people with degenerative meniscus tears and mechanical symptoms better.