The Tortoise Versus the Hare—It’s a Photo Finish
Commentary on the article by Jay D. Keener, MD, et al.: “Rehabilitation Following Arthroscopic Rotator Cuff Repair. A Prospective Randomized Trial of Immobilization Compared with Early Motion”
Stephen F. Brockmeier, MD


There are a number of known factors that can affect clinical outcomes and structural healing after rotator cuff repair. Some of these are fixed variables, such as patient age and comorbidities, tear size, and tissue quality, while others, such as patient-related factors like smoking, repair construct, and postoperative management, are modifiable and may allow for improvement in patient outcomes. The traditional school of thought regarding the initiation of postoperative rehabilitation following rotator cuff repair has favored early passive shoulder motion to prevent stiffness. Until recently, this approach has been largely uniform. However, over time, data suggesting poorer tendon healing in certain subsets of patients have accumulated, leading some surgeons to advocate a delayed rehabilitation protocol with an initial phase of strict immobilization. The theoretical “trade-off” of this approach would be to improve the likelihood of tendon healing at the potential expense of an increased risk of postoperative stiffness and a delayed recovery timeline.

The current study by Keener and coauthors from Washington University analyzes this question and does so in a superb fashion. In this Level-I, prospective, randomized trial, the authors identified and enrolled 124 patients under the age of sixty-five years who underwent arthroscopic repair of a small or medium tear (<30 mm in anterior-to-posterior dimension) of the supraspinatus tendon. The surgical technique and repair construct were standardized to a double-row transosseous equivalent technique. Patients with preoperative stiffness, a subscapularis tear, arthritis, or a concomitant labral repair were excluded; however, patients who underwent concomitant biceps tenodesis or tenotomy or distal clavicular excision were not. Patients were randomized into either a “traditional” rehabilitation protocol, which consisted of six weeks of simple sling immobilization and initiation of therapy at the first postoperative visit, or a delayed rehabilitation protocol, with simple sling immobilization and initiation of physical therapy at six weeks postoperatively.

Their results essentially prove the null hypothesis as there were no significant differences found in functional outcomes scores (American Shoulder and Elbow Surgeons, visual analog pain scale, Simple Shoulder Test, and Constant scores), active shoulder motion, or strength between the two groups at twelve or twenty-four months. They did note better mean active elevation and external rotation in the traditional rehabilitation group at three months, but this difference normalized at later time points. Additionally, ultrasound evaluation at a minimum of twelve months was carried out in 116 patients and, interestingly, demonstrated no differences in healing rates between the two groups.

The attributes of this paper are obvious. The research methodology is of the highest caliber, and the question they seek to answer is of clear clinical relevance. They tightly controlled a number of potentially impactful variables such as patient age (by including only patients under the age of sixty-five years), tear location and size, and repair construct. Additionally, the study retention at twelve and twenty-four months was >80%, which is impressive. While there have been a few recent studies1,2 that are similar, the numbers in the current study are substantially higher and appear to be appropriately powered to support the findings and conclusions. On the basis of these features, I have no reservation at all in believing their findings and using them to support clinical judgment.

However, a few points merit emphasis when interpreting these results as they relate to the general population of patients with rotator cuff pathology. First, the authors’ inclusion criteria limited this study to patients who were younger than average for rotator cuff pathology, with a mean age of approximately fifty-five years in both cohorts. These patients all had single tendon tears of <30 mm that were able to be managed with direct repair without complex releases or side-to-side sutures. They additionally made no mention of smoking status as it related to either cohort. So this is clearly an “optimized” group of patients, which is further supported by their overall rate of tendon healing of 92% at the final ultrasound evaluation. Second, the authors admit that patient compliance was not directly monitored in either group.

In the end, the authors astutely conclude that there appears to be no “apparent advantage or disadvantage of early passive range of motion compared with immobilization with regard to healing or functional outcome.” This statement serves to offer the reader the option to use these results to support either approach, depending on his or her interpretation, preference, or preexisting bias. One could argue, however, that the group of patients in whom a delayed rehabilitation approach might be most advantageous would be those with a less optimal set of confounding variables, such as increased patient age, large or massive tears, or a revision repair. In my current practice, I use a delayed rehabilitation protocol, which is identical to that outlined in this study for these patients, as the current data suggest that such patients have a substantially inferior likelihood of tendon healing. As I critically analyze the current study, I find myself wondering the most about the applicability of their findings to these scenarios. While the finding of no disadvantage of a delayed rehabilitation approach with regard to final shoulder motion and strength would more than likely still apply to this more challenging group of patients, would the lack of an advantage regarding rates of tendon healing also translate? This is the question that needs to be addressed next and, in the end, I suspect that one final variable, the biology of the repair site, will prove most critical in the ultimate outcome.


  • * The author did not receive payments or services, either directly or indirectly (i.e., via his institution), from a third party in support of any aspect of this work. He, or his institution, has had a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. The author has not had any other relationships, or engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.


  1. 1.
  2. 2.