• A L

~30% of us have a torn rotator cuff even though we have no pain – why you can’t always trust an MRI

The Scenario: 42 years old. I felt a sharp pain in my shoulder during the last game of a baseball tournament – next day could barely lift my arm. 2 months later and it’s gotten significantly better. MRI says partially torn rotator cuff which must’ve been from the injury… right?

In the field of physiotherapy, the use of magnetic resonance imaging (MRI) for shoulder injuries is not that old of a practice. The first published article looking at using MRIs to investigate shoulders came from physician Dominik Huber from the Department of Radiology at the Long Island Jewish Medical Center in New York in 1986.1 He was the first to try and improve MRI techniques as they applied to the shoulder, and later concluded in his study that MRI use to investigate suspected rotator cuff tears could be a warranted practice in the future.

Fast forward 30+ years and the use of MRI on the shoulder has become quite commonplace in the rehab world. But unless you’ve had an MRI prior to your injury to compare a post-injury MRI to, MRI interpretation can be difficult. How do we know that what shows up is actually causing you your symptoms?

The analogy that I like to use is like the idea of how intact was the body of your car prior to a car crash? Maybe it was pristine before and every scratch we see now is due to the crash. Or maybe your car was like my 1991 Ford Taurus station wagon in 2007 and 90% of those scratches were already there.

An MRI of your rotator cuff post-injury shows a tear – but how much of that existed before?

Not so easy to know.

Most of the answers to this question come from studies investigating asymptomatic people – that is, people with no history of shoulder injuries or shoulder pain. These studies are often fairly straightforward in design – grab a group of people with no history of shoulder injuries or shoulder pain, and perform an MRI. To make it even less biased, give the MR image to a radiologist and don’t tell them a single thing about these people to keep them honest in their interpretation. Then, analyze the stats.

Let’s take a look at 4 of the earliest studies to do this.

One of the first to investigate MRIs in asymptomatic people was Scott A. Mirowitz from St. Louis, Missouri in 1991.2 Scott looked at 15 people (8 men, 7 women) who were between 22 – 54 years old and with no history of shoulder injury or pain. In these 15 asymptomatic people with no history of injury or pain, 11 people were found to have some degree of impingement on MR imaging.

1992: Christine Neumann of the San Francisco Magnetic Resonance Center did the same study with 32 people (23 men, 9 women) between the ages of 22 – 44 years old, again with no history of shoulder injury or pain.3 In this study, 89% of their population showed some degree of degeneration.

1995: Anthony Miniaci from Toronto, Ontario, Canada took 20 people ranging from 17 – 49 years old with no history of shoulder injury or pain.4 This time, not a single shoulder was without some type of degenerative change – that is, 100% of the study group showed some degree of degeneration with 7 shoulders (23%) showing up with a partial-tear.

Finally, also in 1995, Jerry Sher of the University of Miami looked at 96 people between the ages of 19 – 88 years old who were all again free of current or previous shoulder injury or pain.5 Overall 33 (34%) people had tears within their rotator cuff – with 19 (20%) of them being partial tears and 14 (15%) being full tears of the rotator cuff.

Benign rotator cuff findings on MRI are more common than you think – and that’s just fine. My station wagon, with no working locks, a radiator that needed a top up of fluid every other week, and a broken-key stuck within the ignition took me nearly 4000 kilometres from Ontario to Newfoundland and back without a hitch – what I mean to say is, it was still perfectly functional. Maybe not ideal, but with some diligent maintenance, perfectly functional nonetheless.

Some degeneration within your rotator cuff even without any history of injury or pain is most likely. A partial tear on an MRI? There’s nearly a 1 in 3 chance that it could have been there before the pain ever started. MRIs serve a purpose, and not all partial rotator tears are as benign as this post may make it seem to be. But not all of them are as sinister as they sound or require a surgery. If there is progress in the rehab, don’t let that MRI get you down or dictate the outcome – work hard and trust the process. Be proud of that wagon.


  1. Huber DJ, Sauter R, Mueller E, Requardt H, Weber H. “MR imaging of the normal shoulder.” Radiology 1986;158(2):405-408.

  2. Mirowitz SA. “Normal rotator cuff: MR imaging with conventional and fat-suppression techniques.” Radiology 1991;180(3):735-740.

  3. Neumann CH, Holt RG, Steinbach LS, Jahnke Jr AH, Petersen SA. “MR imaging of the shoulder: appearance of the supraspinatus tendon in asymptomatic volunteers.” American Journal of Roentgenology 1992;158(6):1281-1287.

  4. Miniaci A, Dowdy PA, Willits KR, Vellet AD. “Magnetic resonance imaging evaluation of the rotator cuff tendons in the asymptomatic shoulder.” The American Journal of Sports Medicine 1995;23(2):142-145.

  5. Sher JS, Uribe JW, Posada A, Murphy BJ, Zlatkin MB. “Abnormal findings on magnetic resonance images of asymptomatic shoulders.” The Journal of Bone and Joint Surgery 1995;77(1):10-15.

1 view0 comments