Partial Rotator Cuff Tears: What is a Surgeon to Do?

By: William B. Stetson, M.D.

Member, AANA Communications and Technology Committee

Akin Akinwumi, B.S.

Pre-Med Intern, Stetson Lee Orthopaedics and Sports Medicine

Patial thickness rotator cuff tears of the supraspinatus and infraspinatus tendons can be interstitial or involve the articular side, bursal side or both sides of the rotator cuff tendons. They can of course be asymptomatic or can be a potential source of shoulder dysfunction. 

The occurrence of abnormal MRI signal changes in these rotator cuff tendons does not always correlate with symptomatology. In a prospective, randomized study of 96 asymptomatic individuals, 20% had partial thickness tears. With those over the age of 60, 26% had a partial thickness cuff tendon tear which was asymptomatic¹. In overhead athletes such as baseball players² or volleyball players³, partial rotator cuff tears by MRI are seen in 43%-65% of asymptomatic athletes. 

What we do know is that partial rotator cuff tears are not always symptomatic, the frequency increases with age and activity level and their occurrence in asymptomatic individuals makes it important to fully evaluate the patient. The evaluation should include a complete history and thorough physical examination before concluding that the MRI findings are the cause of the patient’s complaints.

The typical patient with a partial supraspinatus (most common) or infraspinatus rotator cuff tendon tear has signs and symptoms of subacromial impingements which are non-specific and typically have a painful arc of motion between 60-120 degrees, which is also commonly seen in subacromial impingement^4,5, or also referred to as rotator cuff tendinitis. Subacromial impingement and other shoulder issues such as scapular dyskinesis and biceps tendinitis might be causing the symptoms and should be addressed first nonoperatively as the partial rotator cuff tear might not be the problem. Indeed, a partial rotator cuff tear by MRI is not an automatic indication for surgery. 

Classification Challenge: The next challenge we have as Orthopaedic Surgeons when discussing partial rotator cuff tears is the classification of these tears. In 1990, Ellman’s study in Clinical Orthopaedics and Related Research (CORR) was a review article about incomplete rotator cuff tears⁶ and the arthroscopic classification of these tears. He described these tears as articular, bursal or intratendinous and by the depth of the tear. A grade I was < 3 millimeters (mm) deep, a grade II was 3-6 mm deep and a grade III was > 6 mm in depth. Ellman’s arthroscopic classification system was based on the assumption that the normal rotator cuff tendon is 10-12 mm thick. However, there have never been any studies to date that have validated Ellman’s classification system, and not all rotator cuff tendons have the same thickness.  Kuhn et al.⁷ conducted a prospective study with 12 fellowship-trained Orthopaedic Surgeons on the interobserver agreement in the classification of partial rotator cuff tendon tears using the Ellman classification system. Using this system, viewing arthroscopic videos of 30 shoulder surgeries, there was very poor interobserver agreement when classifying the depth of the tear (Kappa coefficient = 0.19) (no to slight agreement).

If we can’t even agree how badly a tendon is partially torn, then how can we then agree on the best way to treat these tears? That is the dilemma that confronts us today as we discuss the best way to manage these partial rotator cuff tears, both nonsurgically and surgically. 

“The 50% Rule:” As Orthopaedic Surgeons, we like guidelines to help us in our decision making when addressing partial ACL tears, flexor tendon tears or lacerations, partial tears of the biceps tendon, partial thickness rotator cuff tears and even other ligamentous or tendonous tears.  The rule appears to have evolved from the hand literature toward a somewhat arbitrary application for other orthopaedic conditions. Little scientific evidence is available to support the 50% rule⁸ and that goes true for “the 50% rule” in arthroscopic shoulder surgery and partial rotator cuff tears. In a systematic review of the literature, Pedowitz et al. reviewed 291 articles pertaining to partial-thickness rotator cuff pathology, 23 articles mentioned the 50% rule and most of these articles cited the Ellman study as a common reference⁸. However, there has only been one biomechanical study by Mazzocca et al.⁹ to validate the effect of the depth of a partial rotator cuff tear on function, noting a gradual change in in vitro biomechanical parameters with greater than 50% tearing. Although an excellent study, at the time of arthroscopic shoulder surgery, are we always able to determine whether the tear is greater or less than 50% of the tendon? There has been no study to date that has been able to reliably verify the “50% rule” for partial rotator cuff tears. This is very important as this is often the criteria used at the time of arthroscopic shoulder surgery to determine whether a partial tear should be debrided or repaired.

The Snyder Classification System: Snyder and colleagues proposed a different classification system in 1991¹⁰, grading partial tears as articular-sided (A), bursal-sided (B) and complete tears (C). A grading system from 0 to 4 (IV) was proposed, depending on the severity of the tear. A normal articular side or bursal side is 0, a grade I is a small tear or fraying less than one centimeter, a grade II is partial tearing of 1-2 cm in size, a grade III has tearing of 2-3 cm and a grade IV has tearing greater than 3 cm in size, often times with a large flap component. Snyder’s classification system is more specific in that it classifies both the articular side and the bursal side of the rotator cuff and does not try to use the greater or less than “50% rule.” For example, an A-I, B-IV partial rotator cuff tear using Snyder’s classification system means that there is less than one centimeter tearing of the articular side of the cuff (A-I) but there is greater than 3 centimeters of tearing on the bursal side (B-IV). This can give a more accurate description of the partial cuff tearing at the time of diagnostic arthroscopy and hopefully help guide treatment recommendations.

In 2016, we published our results of the “Interobserver Agreement in the Classification of Partial-Thickness Rotator Cuff Tears Using the Snyder Classification System” and found a Kappa coefficient of 0.512 indicating moderate reliability and “very good”¹¹. We concluded that the Snyder classification system was reproducible and recommended that it be used for future research in studies analyzing the treatment options of partial thickness rotator cuff tears. 

The Ellman vs. The Snyder Classification Systems: Bi and Verma recently published an eloquent review article on the topic of the Ellman and Snyder classifications systems for partial-thickness rotator cuff tears and the limitations of each¹². The authors make the valid point that both classification systems were developed from an arthroscopic diagnosis and by definition, cannot help in the preoperative planning or indications for surgery. We agree 100% but it is important to point out that the purpose of the Snyder classification system is to help classify and guide the surgeon with intraoperative decision making and not preoperative planning. Because of the lack of sensitivity and specificity of using MRI, ultrasound or other imaging modalities, there is really no practical preoperative classification system of partial-thickness rotator cuff tears at this time. Even with the use of MR arthrography in detecting partial thickness articular-sided rotator cuff tears¹³, one cannot fully determine the extent of the partial tearing of the rotator cuff.  Diagnostic shoulder arthroscopy remains the gold standard for detecting and classifying partial thickness and full thickness rotator cuff tears. The Snyder classification system allows the surgeon to classify these tears at the time of surgery and helps guide us in intraoperative decision making as to the best treatment options at the time of surgery: debridement vs. repair and then if repair, what is the best technique?

Treatment Algorithm: A treatment algorithm can be based on the Snyder classification system depending on whether the partial tear involves only the articular side (A), bursal side (B) or both sides (A and B) of the rotator cuff tendon. We use the rule of “3’s” which has not been validated by any study but does give the arthroscopic shoulder surgeon some guidance of whether to debride or repair. For example, any articular or bursal side tear that is 3 cm or greater (A-III for articular sided and B-III for bursal sided), we feel that there is enough damage to the rotator cuff tear to warrant repair. What type of repair is controversial as to whether to do an in-situ repair (e.g. PASTA repair) or to debride the tear completely to convert to a full thickness tear, and then to repair. We have had good success with both techniques depending on the age and the activity level of the patient, doing PASTA repairs for younger patients and conversion to complete tears with repair in older patients. Both types of techniques can yield good results, but the PASTA repair technique can be more challenging¹⁴.

If there is only minimal tearing on the articular side (e.g. A-I or A-II) with no damage to the bursal side of the rotator cuff, it seems wise to debride the cuff and leave the remainder, making sure that we are also treating other pathology including the superior labrum, biceps tendon, subacromial space and acromioclavicular (AC) joint, that may be the source of the patient’s symptoms. For partial bursal sided tears (e.g. B-I or B-II) with a normal articular side (A-0), some surgeons would argue that these types of bursal sided tears should be treated more aggressively as they may be more symptomatic than partial articular sided tears. Further studies need to be done for addressing these types of tears, but I tend to treat them more aggressively with repair, having had some early failures in my career, treating these patients with only debridement, and then having them return with similar symptoms and asking me why I did not “fix their problem!” Ouch, not what a surgeon wants to hear! With that being said, further studies need to be done with these types of tears, but no one will fault the surgeon for making logical, well-thought-out decisions at the time of surgery.

In general, if I find tearing both on the articular side and the bursal side that add up to “3 or greater” (e.g. A-2, B-1 or greater), it tells me there is significant enough damage to the rotator cuff that needs to be addressed with repair. For most tears, I convert them to a full thickness tear with debridement and then repair them as if I am repairing a full thickness tendon tear. We have had very good, long-term results with this technique with greater than 80% of patients returning to a pre-injury level of function greater than seven years after surgery¹⁵.

Conclusions: As a specialty, we still have a long way to go before we fully understand and learn how to properly treat partial-thickness rotator cuff tears. We do know now that partial rotator cuff tears often present in asymptomatic individuals and a partial tear by MRI is not an automatic indication for surgery. What is the contribution to symptoms in a painful shoulder from the partial rotator cuff tear or what else can be the pain generators¹⁶?  These are clinical diagnostic dilemmas we face as surgeons which are not easily answered. But we do know that conservative management is the initial step in treating these patients and nonoperative treatment can often lead to a full recovery. Further, prospective, randomized studies are needed before fully understanding the natural history and progression of these partial rotator cuff tears, the best reproducible classification system and how to properly diagnose and treat these partial cuff tears. 

References