Knee MRI – are we getting TMI (Too Much Information)?

A recent BMJ study is an eye-opener about problems interpreting knee MRI’s in older adults. The bottom-line? A picture may not be worth a thousand words.

Let’s look briefly at the study. The authors assembled a large group of “older” adults in Framingham, Mass. (I chose not to take offense at their definition of older, which is 50 years old and above.) They took all comers, regardless of knee pain, and did routine xrays on their knees. Anyone with radiographic evidence of osteoarthritis was eliminated, leaving 710 individuals. The authors then did knee MRI’s on all 710 folks, and the findings were interesting.

Overall, 89% of all MRI’s were abnormal, with findings consistent with OA. Drilling down, people with no knee pain had just about as many abnormalities as those with knee pain. The authors concluded several things:

  • MRI’s of the knee are frequently abnormal, and the older a person gets, the more likely it is that an MRI of their knee will show changes of OA.
  • There is little correlation between knee MRI findings and a person’s symptoms.
  • For an older person with knee pain, abnormalities found on MRI may have nothing to do with why the knee hurts.
  • Treatment decisions in older adults should not be based on knee MRI findings alone.

It should be noted that this study did not look at the utility of MRI’s in knee pain due to trauma, or when used as part of an evaluation of other knee problems. Also, the authors did not look at the role played by routine knee radiographs.

However, this study does have immediate practical value for clinicians. Several years ago, we learned that MRI’s of the lumbo-sacral spine are often abnormal in people with and without back pain, making it difficult to know whether MRI-identified problems were actually causing back pain. Now, the same appears to be true for knees. When evaluating an older patient with knee pain, if their xray is normal, pause before ordering an MRI. Remember there is a 9 in 10 chance their knee MRI will show abnormalities, and these findings may be incidentalomas.

Sometimes the best way to evaluate a problem is to use one of those tried-and-true approaches of the past, as “older” doctors learned in medical school before MRI scanning appeared on the scene.

A constant challenge we clinicians face is separating the wheat from the chaff.

Richard Fleming, MD

Comments are closed.