Burkhart first introduced the concept of a partial rotator cuff repair, in which the torn rotator cuff is partially repaired, without complete coverage of the native footprint [19]. In this technique, the anterior and posterior margins of the tear are repaired, leaving the retracted central portion of the tear in place. Burkhart named this a “force couple repair ,” as it allows adequate balancing of the anterior and posterior forces acting on the proximal humerus [20]. Burkhart first noticed that some patients, despite having a massive rotator cuff tear, were able to maintain normal or near-normal shoulder strength and range of motion. He proposed that this was possible because the coronal and axial forces acting on the glenohumeral joint were still balanced, despite the defect within the rotator cuff. The balancing forces in the frontal plane are provided by the deltoid and subscapularis. In the axial plane, the balancing forces are from the subscapularis and infraspinatus/teres minor [21]. Therefore, complete repair of the supraspinatus might not be required for sufficient balancing of the forces. In fact, even with complete supraspinatus paralysis, 75% of abduction strength and 85–90% of external rotation strength are preserved [22]. Thus, a large defect might weaken the anterior and posterior translational forces in equal amounts, such that the resulting forces continue to maintain the humeral head in an anatomic location on glenoid surface throughout a functional range of motion [19].

Burkhart proposed that in patients with massive rotator cuff tears, achieving a partial repair could restore function by rebalancing the anterior and posterior moments acting on the glenohumeral joint [23]. Thus, this type of repair is often called a “functional repair .” This also reduces the strain at the margins of the tear so that fixation will be adequate, despite being weaker than that of a complete repair. It protects the tendon-bone repair interface during healing, and allows a functional recovery despite a less-than-complete repair. Burkhart found an increase in active elevation of 90–150° and mean 2.3 grade improvement of elevation strength in 14 patients using an open approach to achieve a partial repair [19]. He found that once the margin of the tear is fixed, it is no longer impinged under the acromion or between the rotator cable and the humeral head. Thus, shoulder motion can be achieved without pain.

Burkhart also introduced the concept of “margin convergence ,” also known as a “side-to-side” repair. Unlike a partial repair, in which cuff tissue is brought to bone, margin convergence involves fixation of cuff tissue to cuff tissue. It is usually performed by bringing together the retracted margins at the apex of a retracted U-shaped tear, thereby converting a massive U-shaped tear into a smaller tear. Following a margin convergence procedure, the remaining cuff tissue may or may not be fixed to bone, depending on the tension created by the repair.

Margin convergence or side-to-side repair is labeled a nonanatomic repair, since rotator cuff tissue is being fixed to other cuff tissue, rather than being brought to its anatomic location at the bony footprint on the proximal humerus. Several studies have found good clinical results and high patient satisfaction, with recurrence rates from 19 to 48% [24–26]. Bukhart et al. found excellent clinical results with massive tears and 3.5-year follow-up. In his series, he used an arthroscopic approach, and found that margin convergence of large U-shaped tears was similar to complete (tendon to bone) repair of medium-sized crescent-shaped tears [12]. In massive cuff tears, Burkhart was able to achieve an increase in forward flexion by 42–132 degrees and improvements in strength and functional scores as well.

There is evidence of improved clinical outcomes with partial rotator cuff repair at 3–5-year follow-up [1, 21, 27, 28]. In these studies, the partial rotator cuff repair was performed with the addition of margin convergence. Jones and Savoie found good-to-excellent results in 88% of patients. Fixation of cuff tissue onto the bony footprint did not improve long-term clinical results [29]. Van der Zwaal et al. reported good clinical results with a 19% re-tear rate at 26.5-month follow-up [24]. They used a running suture technique for their side-to-side repair [30]. Rousseau et al. found good clinical results with a 44% re-tear rate at 38-month follow-up in an elderly population (mean age 66) [25]. Nove-Josserand found an 85% satisfaction rate at 5-year follow-up with this technique [31]. They included only patients with subacromial space >7 mm and Goutallier stage 1 or 2 fatty infiltration. The recurrence rate was 17%. Duralde and Bair reported 67% of patients with good-to-excellent ASES scores with 92% satisfied with the results of partial repair [32].

Quite a few studies have suggested that a partial repair of massive rotator cuff tear results in equivalent outcomes to a complete repair. Godneneche et al. found that both partial and complete repairs gained equivalent improvements in Constant scores. However, rotator cuff strength was found to be significantly lower in the partial repairs. In addition, the re-tear rate was significantly higher in partial repairs (48.8% versus 20%). Heuberer et al. confirmed this (53% versus 29%).

A partial rotator cuff repair may result in greater functional loads on the remaining rotator cuff, thus resulting in muscular fatigue of the intact tendons [33]. Heuberer et al. prospectively analyzed 68 patients with massive rotator cuff tears and found that complete repair, partial repair, and debridement alone without repair all led to significant improvements in clinical and subjective outcomes. However with respect to strength and functional scores, complete repairs were significantly superior to both partial repairs and debridement alone. The re-tear rate for partial repairs was 29% versus 53% for complete repairs, as evaluated by MRI. Interestingly, there was no difference between debridement and partial repairs with regard to functional and pain scores. Thus, the authors questioned whether partial repair or debridement is the preferred arthroscopic procedure in the case where a complete repair cannot be achieved. However, there is data that shoulder function deteriorates over time following a debridement [34]. Thus in the long term, partial repair may still be the preferred option, although there is limited data to support this conclusion.

Moser et al. [35] retrospectively evaluated 38 patients who underwent either complete repair, partial repair, or debridement. They found a trend towards improved motion and strength in complete and partial repairs, when compared with debridement alone. When comparing complete to partial repairs, there was a significant difference in external rotation strength and a trend towards improved functional scores, favoring complete repair.

Francesci et al. compared partial rotator cuff repair to debridement and acromioplasty and found superior functional outcomes and quality-of-life scores for partial repair [8]. Although patients who underwent debridement experienced less pain in the first postoperative month, partial repair of the rotator cuff resulted in significantly increased strength and range of motion.

Shon et al. demonstrated that despite initial improvement in clinical and radiographic parameters, about half of the patients demonstrated worsening of outcomes, including increasing patient dissatisfaction, over a 2-year period of follow-up after partial rotator cuff repair [1]. The authors suggested multiple possibilities to account for this decline, including re-tearing of the cuff and deterioration of the positive effect of subacromial decompression and biceps tenotomy/tenodesis. However, unlike in other studies, the patients in the study by Shon et al. did not receive margin convergence procedures in addition to the partial rotator cuff repair.

Several preoperative factors have been identified as contributing to a poor outcome following rotator cuff repair. These include older age, female sex, smoking, diabetes, longer duration of symptoms, preoperative disability, pseudoparalysis, size and shape of the tear, retraction of the tear, subscapularis tears, and atrophy and fatty infiltration of the muscle [36–40]. One must consider that the factors which portend a poor outcome for complete rotator cuff repairs may also be important determinants of outcomes in partial cuff repairs. Shon et al. identified fatty infiltration of the teres minor muscle, in particular, as a factor associated with worse outcomes following a partial rotator cuff repair [1]. This was confirmed by Godeneche and Barth [41], who found lower functional scores with higher stages of fatty infiltration (greater than Goutallier stage 1). These observations suggest that tears should be treated promptly, and may explain why repair of acute traumatic tears results in better outcomes than chronic degenerative tears [41–43].

Franceschi et al. identified acromiohumeral index (AHI) and patients’ daily activity levels as factors predicting a poor outcome for partial repair [8]. They found that for patients with high activity levels or a preoperative AHI grade 2 or worse, that is, acromiohumerus interval of 5 mm or less, partial repair did not have any advantage when compared with debridement and acromioplasty.

Holtby et al. found that although advanced age is correlated with larger tear size, poorer tendon quality, and more U-shaped tears, age was not a predictor of whether a massive tear was amenable to complete or partial repair [44]. In that analysis, men and women were found to have an equal incidence of U-shaped tears, although men had more crescent-shaped tears while women had more L-shaped tears. Crescent-shaped tears had the best outcomes. One factor that did predict reparability was external rotation strength. Patients with stronger external rotation at neutral prior to surgery were significantly more likely to have a complete repair. Interestingly, neither active flexion nor abduction was a significant predictor. High subjective disability, lack of external rotation, poor tendon quality, larger tear size, and U-shaped tears were all associated with inability to achieve a full repair, thus resulting in a partial repair of the cuff. Both patients with partial repairs and full repair achieved significant gains in strength, ROM, and improvement in disability levels. However, those with partial repairs did have slightly higher levels of disability and had less improvement in functional scores at 2 years post-op [44].