Intended for healthcare professionals


Arthroscopic surgery for degenerative knee

BMJ 2015; 350 doi: (Published 16 June 2015) Cite this as: BMJ 2015;350:h2983
  1. Andy Carr, professor and director
  1. 1Botnar Research Centre, Oxford University Institute of Musculoskeletal Sciences, NIHR Oxford Musculoskeletal Biomedical Research Unit, Oxford OX3 7LD, UK
  1. andrew.carr{at}

Overused, ineffective, and potentially harmful

The most frequent indication for knee arthroscopy is degenerative joint disease in middle aged and older patients. Each year, more than 700 000 knee arthroscopies are done in the United States and 150 000 in the United Kingdom.1 Magnetic resonance imaging evidence of meniscal abnormality, osteophytes, cartilage damage, and bone marrow lesions is often present. All these imaging abnormalities are common in the general population and are often asymptomatic.2 The evidence base for arthroscopic surgery is known to be weak, and a pressing need exists for more high quality multicentre randomised controlled trials, systematic reviews, and meta-analyses to inform clinicians and improve care for patients.3 Researchers have already reported that trials of arthroscopic surgery find no benefit over control interventions ranging from exercises to placebo surgery.4

A linked paper by Thorlund and colleagues (doi:10.1136/bmj.h2747) adds substantially to the debate by systematically reviewing all the evidence on the benefits and harms of arthroscopic knee surgery for middle aged and older adults with knee pain and degenerative knee disease.5 The authors report that the small benefit seen after arthroscopic surgery of the knee is short lived and disappears within one to two years.

In the light of this evidence, why is arthroscopy still so common? It even seems to be increasing in both North America and Europe.6 7 Is the published evidence flawed? This is certainly the view of some surgeons, including the editors of the journal Arthroscopy who believe that “the New England Journal of Medicine is biased against knee surgery.”8 In the journal’s defence, the available evidence is certainly of low quality in places. Only two of the nine trials reviewed by Thorlund and colleagues were adequately blinded, and many of the other trials had a high risk of bias. In five of the nine trials, the comparator was exercise therapy that was poorly described and given at a suboptimal dose.

Another possibility is that surgeons are falling prey to confirmation or myside bias,9 whereby robust and high quality evidence is contested and ignored in favour of deeply held convictions or entrenched attitudes. Such bias is not new and was well described by Leo Tolstoy in 1899: “I know that most men not only those considered clever, but even those who are very clever, and capable of understanding most difficult scientific, mathematical, or philosophic problems can very seldom discern even the simplest and most obvious truth if it be such as to oblige them to admit the falsity of conclusions they have formed, perhaps with much difficulty conclusions of which they are proud, which they have taught to others, and on which they have built their lives.”10

One thing is clear from all randomised trials: patients improve after arthroscopy. This is in line with surgeons’ own observations and with evidence from uncontrolled observational studies. However, in robust and bias-free trials that use placebo controls, active treatment works no better than control treatment. In response, leaders of the arthroscopic surgery community have asserted that patients who participate in placebo controlled trials “may not be of entirely sound mind” and that “ethically, sham surgery is a questionable research method, which may be harmful.”8

A recent systematic review of the use of placebo in surgical trials shows that in more than half of these studies surgery had no greater effect than a placebo.11 This review also reported that very few harms occurred after placebo surgery. Placebo surgery was safer than the treatment under investigation. These findings make a strong case for the use of placebo controls when a placebo effect may be present and for the discontinuation of procedures that offer patients no discernible benefit. The treatment effect associated with arthroscopic surgery of the knee may well have a placebo component. Outcomes are mostly subjective—improvement in pain is the main justification for the procedure. Placebo effects can be modified and substantially enhanced by a variety of factors that alter beliefs and expectations.12

Importantly, Thorlund and colleagues also review the harms associated with arthroscopic knee surgery. They were unable to identify harm from randomised trials alone because the trials were too small, so they did a wider review including observational studies. These studies were heterogeneous and inconsistent, but the risks associated with non-surgical treatment including exercises are clearly rare and minor. Harms associated with arthroscopic surgery are also rare but include serious adverse events such as deep venous thrombosis, infection, pulmonary embolus, and death.

Supporting or justifying a procedure with the potential for serious harm, even if this is rare, is difficult when that procedure offers patients no more benefit than a placebo. If, as reported, the mortality associated with arthroscopic knee surgery is 0.96 (95% confidence interval 0.04 to 23.9) per 1000 cases and the rate of deep venous thrombosis is 4.13 (1.78 to 9.60) per 1000 cases then, with rates of surgery at their current level, a substantial number of lives could be saved and deep venous thromboses prevented each year if this treatment were to be discontinued or diminished.

We may be close to a tipping point where the weight of evidence against arthroscopic knee surgery for pain is enough to overcome concerns about the quality of the studies, confirmation bias, and vested interests. When that point is reached, we should anticipate a swift reversal of established practice.


Cite this as: BMJ 2015;350:h2983


  • Research, doi:10.1136/bmj.h2747
  • Competing interests: I have read and understood the BMJ policy on declaration of interests and declare the following interests: I am supported by the NIHR Oxford Biomedical Research Unit and have received research grants from NIHR and Arthritis Research UK.

  • Provenance: Commissioned; not externally peer reviewed.


View Abstract