I was lecturing at a recent medical conference where another lecturer showed a picture of a colleague’s thumb with the caption, “MRI-My Regional Investigator”. The point was that we have become too dependant on advanced imaging in musculoskeletal diagnosis and too often, what’s seen on the MRI doesn’t represent what’s causing the patent’s pain. Since we see this all day every day in the office, I thought it was a point well taken.
Studies of structural imaging and it’s correlation to pain have been dismally disappointing. Several low back studies show that patients with severe problems on MRI are often pain free, while other patients with severe pain often have limited structural changes on MRI of the spine. I wrote a paper on this issue published in the Annals of Rheumatology in 2002, posted here. More surprising was a recent study published in the New England Jouranl of Medicine showing that about 60% of patients without a history of active knee pain have meniscus tears on MRI. This study calls into question the reasoning behind likely hundreds of thousands of knee surgeries performed over the past two decades. Since many of these tears may have been asymptomatic, why are we operating on them?
We see this occur everyday in our clinical practice. Let me provide some concrete examples. A late 30’s mother of twins told she needed a hip replacement because her MRI showed mild to moderate hip arthritis and a shallow hip joint socket (dysplasia), who really turns out to have an irritated nerve in her back. We inject the irritated nerve and take her hip pain away, she’s now back to her passion of rock climbing, no hip replacement. An elderly male told that he needs a high risk hip replacement (because of his advanced age the surgery would be very risky) based on his MRI, who turns out really to have SI joint pain (the next joint up). We inject the SI joint, no pain, and he returns to full activities. A middle aged woman told she needs a hip replacement based solely on her MRI. She can no longer get in and out of cars and exam shows that she really has myofascial trigger points in her psoas muscle. We inject the psoas muscle trigger points a few times and the patient no longer has hip pain or problems getting in and out of cars. A middle aged woman with severe knee arthritis who was told she needed a knee replacement. Turns out treating the pinched nerve in her back resolved her knee pain (meaning the severe arthritis wasn’t the cause of the pain). A photographer who injured his knee in the antarctic while filming glaciers who was told he needed micro fracture surgery or a knee replacement based on a knee MRI, who really had trigger points in his quadriceps. We got rid of the trigger points and he was able to go biking and hiking for the first time in months. I could literally go on all day on the list of patients we have seen where their MRI imaging only did more to confuse the picture than to diagnose the source of their pain.
So why does this happen? Why does looking at structure rarely predict what hurts? As a simple example, we’d all accept at face value that taking a picture of our car engine when the car won’t start and e-mailing that to a mechanic would likely result in a mechanic hanging up the phone. While there are instances when that picture might help, most of what can go wrong with a car wouldn’t show on such a picture. The mechanic needs to take a history of how the car got that way, what lead up to the problem, poke and prod at the engine, test various electrical signals, etc… Only then can he or she give an accurate “diagnosis” about what may be causing the problem. Another interesting and more scientific example is a recent paper entitled, “Pain Causes Arthritis, Arthritis doesn’t Cause Pain”. The issue is that activated pain sensors lead to bad chemicals being dumped into the joint which the causes the arthritis. The spine is involved in this dance. This is likely why we see patients with severe arthritis who have little pain, they simply don’t have activated pain sensors in the joint.
This is not to say that we don’t use imaging, we do. We use very sophisticated 3.0T MRI with obscure sequences like T2 mapping and others to show various tissues like cartilage. However, it’s always ultilized as a clue to determine what might be causing pain, not a definitive test to rule in or out disease that is causing pain. When the time spent reviewing the MRI exceeds the time spent actually talking to or examining the patient, we increase the likelihood of a bad diagnosis being made based solely on the imaging. We could save billions in health care dollars if every doctor starting using “My Regional Investigator” along with “Magnetic Resonance Imaging”.