subscapularis) rotator cuff tears

Arthroscopic Repair of Anterosuperior (Supraspinatus/Subscapularis) Rotator Cuff Tears: A Prospective Cohort With 2- to 4-Year Follow-up. Classification of Biceps Subluxation/Instability William F. Bennett, M.D.

Purpose: The purpose of this study was to evaluate the outcome of patients who underwent arthroscopic repair of anterosuperior rotator cuff tears. The null hypothesis, that there was no difference between preoperative scores and postoperative scores, was tested statistically. Type of Study: A cohort study. Methods: The preoperative and postoperative status of patients with anterosuperior rotator cuff tears was analyzed using the Constant score, American Shoulder and Elbow Society Index (ASES Index), a visual analog pain scale (VAS), a single question of percent function compared with the opposite unaffected extremity, and a single question reflecting satisfaction, “would you undergo the surgery and the postoperative rehabilitation to achieve the result you have today.” There were also 2 groups compared: 1 that had a “tac” used for repair of the subscapularis tendon, and the other that used a “tie” technique for subscapularis repair. All supraspinatus tendon tears were complete and were repaired using a soft-tissue fixation device. Results: There was a statistically significant difference for all outcome measures except for the objective Constant score of the tie group, P ⫽ .58. Follow-up was 2 to 4 years. There were no differences based on sex or type of fixation device used for repair of the subscapularis tendon. There were no reruptures, clinically. Conclusions: The arthroscopic repair of anterosuperior rotator cuff tears provides reliable expectation for improvement in function, decreases in pain, decreases in clinical findings of biceps subluxation and inflammation, improvement in shoulder scores, and the improvement of clinical findings of subscapularis insufficiency. Key Words: Anterosuperior rotator cuff tear—Arthroscopic repair—Biceps subluxation/instability classification—Subscapularis—Hidden rotator interval lesion.

T

he term anterosuperior rotator cuff tear is defined1 as a lesion of the cuff that includes fullthickness tears of the supraspinatus that extend anterior to its border involving the rotator interval structures and potentially involving the subscapularis tendon. A lesion of the long head of the biceps tendon is considered an anterior lesion. Warner et al.2 used this term to differentiate massive rotator cuff tears,

From Private Practice, Sarasota, Florida, U.S.A. Address correspondence and reprint requests to William F. Bennett, M.D., 5741 Bee Ridge Rd, Suite 470, Sarasota, Fl 34233, U.S.A. © 2003 by the Arthroscopy Association of North America 0749-8063/03/1901-3304$35.00/0 doi:10.1053/jars.2003.50023

supraspinatus and infraspinatus tears that included subscapularis tendon tears, from massive rotator cuff tears that included the supraspinatus and infraspinatus alone, or the posterosuperior rotator cuff tears. The anterosuperior rotator cuff tear is thought to be less common than and responded poorer to open operative repair than the posterosuperior rotator cuff tear.1,3 For the purpose of this report, the term anterosuperior rotator cuff tear is defined as a tear involving only the supraspinatus and subscapularis, with or without involvement of the biceps pulley. Other nomenclature for the biceps pulley include reflection pulley, superior glenohumeral/medial head coracohumeral ligament complex (SGHL/MCHL complex), and medial wall bicipital sheath. There are few studies that have

Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 19, No 1 (January), 2003: pp 21-33

21

22

W. F. BENNETT

reported on the repair of the subscapularis tendon and anterosuperior rotator cuff tears.2-9 Walch et al.6,7 have described the hidden rotator interval. They have described partial tears of the subscapularis and coracohumeral ligament tears. Combinations of these lesions have been previously thought to be unclassifiable.1 The arthroscope has allowed for the further delineation of subtle differences in rotator interval lesions10,11 and arthroscopic identification of these combinations has allowed for a biceps subluxation/instability classification11 to be developed. A hidden rotator interval is not just a subscapularis tear; lesions may include the heads of the coracohumeral ligament. While not specifically in the rotator interval, lesions of the lateral wall of the bicipital sheath or lateral head of the coracohumeral ligament complex can be found in anterosuperior rotator cuff tears. Anterosuperior rotator cuff tears may include clinical findings of biceps inflammation and/or subluxation and arthroscopic findings of inflammation (tenosynovitis) and/or tendonitis/fraying, flattening, and/or subluxation or avulsion, especially when the medial and/or lateral wall of the bicipital sheath is torn or plastically deformed.11 It has been noted that the long head of the biceps can be enlarged and/or flattened with these lesions.1 The purpose of this study was to evaluate the outcome of patients who underwent arthroscopic repair of anterosuperior rotator cuff tears. The null hypothesis, that there was no difference between preoperative scores and postoperative scores, was tested statistically. Additionally, the arthroscopic findings of the long head of the biceps are noted, the arthroscopic technique of sheath reconstruction is outlined, and a discussion as to how this reconstruction prevents biceps subluxation is included.

the rotator cuff or partial tears of the supraspinatus tendon and auto accidents. Workers’ Compensation patients were not excluded. Previous reports on isolated subscapularis tears have included partial-thickness tears of the supraspinatus tendon.8 This study excluded partial-thickness tears of the supraspinatus tendon. Additionally, if patients had an intra-articular lesion (chondromalacia of the humeral head or glenoid, SLAP lesions, or Bankart lesions) found at the time of arthroscopy, they were excluded from this group. All new patients filled out a shoulder form questionnaire, were tested for strength using a tensiometer without lidocaine injection of the subacromial space, and had a preoperative radiograph and magnetic resonance imaging (MRI) or magnetic resonance arthrography (MRA). The shoulder form includes questions from the ASES Index, Constant score, percent function evaluation, a 10-point visual analog pain scale (VAS), and a satisfaction question: “Hypothetically, would you undergo the surgery and postoperative rehabilitation to achieve the present status of your shoulder?” Additionally, the group was chosen to have a minimum 2-year follow-up and because of this criterion the study group was cut off at December 31, 1999. Thus, the study group had a 2- to 4-year followup. Patients were evaluated with the scoring systems at 3, 6, 9, and 12 weeks and then at 6, 9, 12, 24, 36, and 48 months, respectively. At each office visit with the intervals previously stated, the patients were asked to fill out their shoulder form. A continuum of their progress was thus recorded in the data registry with the most recent form replacing the data from the previous visits

METHODS

Specific examinations directed at determining subscapularis lesions and/or biceps tendon inflammation and subluxation were routinely included. These include the Belly press, Gerber lift-off, lag signs, Anguin’s test, Yergason’s test, and a new biceps subluxation test.11 With this new passive test, the shoulder is flexed to 90° and brought to full external rotation and cradled in the crux of the examiner’s elbow. The patient must be completely relaxed and the external rotation should not create a painful situation. From here the arm is brought passively (the patient must be completely relaxed and not use any active motion while the examiner is moving the arm) to cross-body adduction, while simultaneously decreasing flexion of the shoul-

Between 1995 and 1999, 139 complete arthroscopic rotator cuff repairs were performed. Of these repairs, there were 35 repairs that involved the supraspinatus tendon and the subscapularis tendon and any affected heads of the coracohumeral ligament. The inclusion criteria were any patients who had a tear of the subscapularis tendon (intra-articular subscapularis tendon [IASS]) and supraspinatus lesions. The supraspinatus lesions were all complete tears and the subscapularis tears ranged from partial-thickness intra-articular subscapularis tendon (IASS) lesions to full-thickness IASS lesions with or without retraction. Exclusion criteria included involvement of any other tendon of

Preoperative Assessment

CLASSIFICATION OF BICEPS SUBLUXATION der and passively internally rotating the arm. Pain, the feeling of something slipping, catching, or popping during the passive range may indicate increased or abnormal biceps medial/lateral motion. These findings do not indicate a positive test if they are found at the extremes of the range only. The Hawkins and Neer impingement tests were also included as well as the Jobe supraspinatus test. Dynamic anterosuperiorization was evaluated.1 Passive range of motion measurements were performed for all planes: glenohumeral flexion, glenohumeral abduction, glenohumeral external rotation with the arm at neutral, glenohumeral external rotation with the arm at 90° of abduction, and glenohumeral internal rotation with the arm at 90° of abduction. Additionally, internal rotation by level was recorded, as this information was needed to score the Constant score. These motions were compared with the contralateral extremity. Strength testing was performed with a tensiometer and measured in pounds. Operative Technique Patients were placed in the beach-chair position and routine portals were developed. A supraspinatus portal was used for shuttling sutures between the anterior portal. The glenohumeral joint was inspected for intraarticular pathology, such as SLAP lesions, Bankart lesions, and chondromalacia. There were 2 study groups. The first group had repair of the subscapularis tendon using two 8-mm bioabsorbable PGA tacs, (Suretac; Accufex, Mansfield MA). The second group had repair of the tendon using routine suture tying techniques with two No. 2 Tevdeks per one metal 5-mm screw (Metal Corkscrew; Arthrex, Naples, FL). This subdivision was not a randomized trial but was based on the specific findings at the time of arthroscopy. Subscapularis tears that had significant involvement of the SGHL/MCHL complex often did not lend itself to tac repair. As noted in another article,6 if the medial head of the coracohumeral ligament is torn with a tear of the subscapularis tendon, the biceps can sublux into the joint. If the MCHL is torn without the subscapularis tendon, the biceps can sublux between the outer surface of the subscapularis tendon and the coracohumeral ligament. So if the MCHL was robust, it was repaired with a tac because this created a stable reflection pulley. But if the MCHL was flimsy, torn or significantly frayed, in order to reconstruct the medial sheath reflection pulley adequately, a tie technique was used.

23

The subscapularis tendon tears are repaired from an intra-articular approach as opposed to a subacromial approach. The anterior portal is placed as medial to the lateral edge of the coracoid as possible. The extraarticular coracohumeral ligament is partially resected leaving the medial and lateral heads intact so as not to develop a condition in which biceps subluxation may occur. Most of the coracohumeral ligament is extraarticular except for the medial and lateral heads that contribute fibers to the medial and lateral wall of the bicipital sheath and superior capsule respectively. This area is best observed with the shoulder at 90° of flexion and in internal rotation, with some adduction (Fig 1). When the shoulder is flexed to 90° and the arthroscope is placed in the posterior portal, the view is like looking down the arm from above. The insertion area of the intra-articular subscapularis tendon (IASS) when torn is usually identified by a tuft of tissue. The articular cartilage immediately superior to the insertional trough is denuded for better visualization of the insertion site and for access to the inferior portions of the subscapularis footprint. The fibrous scar tissue can be removed from the anterior portal or the supraspinatus portal. A supraspinatus portal is developed, which is important for passing the sutures from the anterior portal to this portal. The arm is brought into external rotation and the tendon is visualized. Typically, with tears of the subscapularis tendon, the IASS has pulled off and the continuity of the subscapularis to the lateral side of the bicipital sheath remains by way of the lateral head of the coracohumeral ligament and some fibers of the lower portions of the subscapularis insertion. The lower portion of the subscapularis tendon is mainly capsular-ligamentous-muscular.12 Figure 2 shows the normal appearance of the intraarticular subscapularis tendon in an arthroscopic photograph superimposed on an artist’s rendition of the normal anatomy. Tears of the IASS essentially peel off from the insertion site from a cephalad to caudad direction in degenerative conditions, and tear or rupture in traumatic conditions. The thickened portion of the tendon belongs attached to the bone (Fig 3). A small ball-burr is placed through the supraspinatus portal (sometimes directly through the anterior portal or an accessory anterolateral portal) and can be brought anterior to the superior glenohumeral/medial head coracohumeral ligament complex (SGHL/ MCHL) and then through the opening in the rotator interval. A bed of bleeding bone is creating at the trough from where the IASS has pulled off. To reduce the tendon to the insertion area of the IASS, a soft-

24

W. F. BENNETT

FIGURE 1. Arm Positioning: from the position visualized, if the arm is brought into cross-body adduction with some internal rotation, the lesions of the rotator interval and intra-articular subscapularis lesions are best visualized. The black arrow has been placed to indicate that when the arthroscope is placed into the posterior portal, the view will be that of looking down the arm. All of the following arthroscopic views are the same (right shoulder) except for Fig 7.

tissue grasper is placed through the supraspinatus portal or through an accessory anterolateral portal and the thickened area is grasped. The reduction is facilitated by placing the shoulder in the position of adduction

FIGURE 2. The arthroscopic view is from the posterior portal of a right shoulder. The view is like looking down the arm from above. The photograph is superimposed on a rendition of this anatomy to show what a normal subscapularis insertion looks like. (CHL, coracohumeral ligament; SS, subscapularis tendon; HH, humeral head; MCHL, medial head coracohumeral ligament; LCHL, lateral head coracohumeral ligament.)

FIGURE 3. The tear of the intra-articular subscapularis insertion (IASS) is represented by the thickened portion of the tendon. It is pulled off the bone. Notice how the lower or caudal fibers are still attached. The insert in the upper right represents a normal insertion site for reference. This is a type-1 biceps subluxation/instability lesion (see Fig 9). (IASS, intra-articular subscapularis; SS, subscapularis tendon.)

and internal rotation. A Mayo stand helps to hold the arm in position. This reduction does not compromise external rotation to any degree further than external rotation of the normal side. Next, a 5-mm Corkscrew was introduced through the anterior portal and this portion of the tendon was penetrated with the screw. The first screw was placed slightly inferior to the most thickened part of the subscapularis tendon as preparations for a second

FIGURE 4. The soft-tissue fixation device has been passed through the tendon and is secured in the bone. The suture retriever is retrieving sutures that have been brought out through the supraspinatus portal. The following step is to bring these sutures out through the anterior portal and tie them down.

CLASSIFICATION OF BICEPS SUBLUXATION

FIGURE 5. The first of the knots is shown in this photo. Notice how the thickened portion of the subscapularis tendon has been brought into approximation with its footprint on the humeral head. The insert in the upper area represents a normal insertion site for reference. (SS, subscapularis tendon.)

25

reduction has been performed, the guidewire is removed and it is then reintroduced to impale the medial head of the coracohumeral ligament. Then by aiming medially it can be further introduced to impale the subscapularis tendon and the entire complex can be reduced to the footprint. Following this, the drilling is performed and the tac is inserted, usually blindly until the complex is soundly reduced. The arthroscope can be moved forward from the posterior portal into the rotator interval and the tac can be visualized in the cannula if needed. In bigger tears that extend inferiorly, preliminary tacs are used to secure the lower portions of the subscapularis tendon. The final tac should secure the medial head of the coracohumeral ligament (Fig 7). If the biceps was frayed and it was less than 50% (arbitrarily chosen) the width of the tendon, the fraying was debrided. If the tendon was severely frayed, it was tenodesed into the subscapularis repair. Postoperative Management

screw were made. The screw was not advanced through the inner surface. The tendon was then reduced using the post of the screw handle and the screw was introduced into the bone. A second screw was introduced in similar manner through the most thickened part and placed through the tendon into bone. The sutures were shuttled by using a suture retriever and running the sutures through the supraspinatus portal (Fig 4). Then the sutures were retrieved through the anterior portal and a modified Revo knot was used for tying (Fig 5). Two knots were tied per screw. The second suture was passed not only through the subscapularis tendon but the medial head of the coracohumeral ligament, allowing for a nice reconstruction of the medial wall of the bicipital sheath (Fig 6). Alternatively, the screws can be placed first, and then the shuttle technique is used to pass the suture through the tendon. If the tac technique is used for the medial sheath reconstruction and/or subscapularis repair, a supraspinatus portal is not needed for shuttling the sutures but may be needed for passing a soft-tissue grasper down to the subscapularis tendon and MCHL and reducing it into place. Or as an alternative, this can be achieved by creating a stab incision at the anterolateral corner of the acromion. However, the easiest way to perform this reduction is to preliminarily take the guidewire on the drill and run it through the anterior portal and impale the subscapularis tendon at an area that would correspond to the torn intra-articular subscapularis tendon IASS and reduce it to its footprint. Once this

The patients were instructed to sleep with the sling for 6 weeks. They were advised to stay in the sling for the first 3 weeks but could take their arm out for gentle Codman-Pendulum exercises. At 3 weeks, they began formal physical therapy regaining passive range of motion, working with a therapist. Passive external rotation with the arm at the side and with the shoulder at 90° of abduction was limited to 30° for the first 6

FIGURE 6. The final suture engages the medial head of the coracohumeral ligament (MCHL) and the intra-articular subscapularis insertion (IASS), thus reconstructing the medial wall of the bicipital sheath. (MCHL, medial coracohumeral ligament; IASS, intraarticular subscapularis.)

26

W. F. BENNETT

FIGURE 7. A series of arthroscopic photographs showing a right shoulder from the posterior portal, illustrating the steps taken for tac repair of a subscapularis tear. This is a type-3 biceps subluxation/instability lesion (see Fig 9). (SS, subscapularis tendon; LT, lesser tuberosity; B, biceps tendon; Sghl/Mchl, superior glenohumeral/medial head coracohumeral ligament complex.) (A) Preparation of the bed of bleeding bone on the lesser tuberosity. (B) Insertion of first tac; notice how subscapularis tendon is elevated from bony insertion. The arrows represent the reduction that will take place with the tac. (C) The view can be achieved if the arthroscope is advanced to the anterior portion of the glenohumeral joint and the light source is positioned to look down over the edge of the subscapularis tendon. (D) The second tac is catching the reflection pulley, (E) a neuroprobe on the repaired IASS insertion, (F) same view as C after placement of both tacs.

weeks. No overhead active motion was allowed for the first 6 weeks. At 6 weeks, the patients began strengthening for internal and external rotation only, and only if passive range of motion in all planes was equal to the opposite side for all planes. They were instructed at 3 weeks to begin passive stretching exercises for all planes of motion except external rotation. At 6 weeks they were allowed to begin some assisted elevation of the arm. At 9 weeks they were allowed active elevation of the arm and encouraged to increase their strengthening incrementally. The criteria for the cessation of physical therapy was full passive range of motion in all planes, tensiometer measurements that were within 20% of the opposite side, and minimal to no pain. Home exercises were implemented at 3 months and patients were followed-up again incrementally until their strength was equal to the opposite side or, if the nondominant side was involved, they were within 10% of the opposite side. Scoring and Statistical Analysis The ASES index and the Constant score were used for scoring parameters. A 10-point VAS and 2 single questions were asked of the patients: “How do you rate the percent function of your affected extremity to your non-affected extremity,” and “Would you undergo this surgery and postoperative rehabilitation to achieve your functional result?” This last question was used to assess satisfaction. The ASES index and Con-

stant score were analyzed using a t test for preoperative to postoperative results and broken down according to sex. The Constant score (100 points) was broken down into the subjective component (35) and the objective component (65) in an effort to better look at the objective changes of the repair. The objective component consists of range of motion (40 points) and strength (25 points). The strength tests the supraspinatus. There is only one question worth 10 points of the total 40 points that is directly related to the subscapularis tendon, active internal rotation. The results of the Constant score were evaluated with respect to the sex and fixation device. The VAS and percent function test were also analyzed preoperatively to postoperatively and between sexes and analyzed with a t test. The final question was asked only at the latest follow-up. Routine radiographs were obtained and all patients had either an MRA or an MRI. The MRI examinations initially were without contrast. Later, routine MRA was performed13 because the use of contrast helps in visualizing subscapularis tears. RESULTS All 35 patients had follow-up data available to calculate the postoperative scores. However, between 1995 and 1997, there were 16 patients who did not have complete data to calculate the preoperative Constant score and ASES Index. However, all 35 patients

CLASSIFICATION OF BICEPS SUBLUXATION TABLE 1. ASES Index

Tac Tie Total

27

TABLE 3. Subjective Constant Score

Preoperative Score

Follow-up Score

P Value

33 ⫾ 15 31 ⫾ 23 31 ⫾ 19

88 ⫾ 12 72 ⫾ 11 80 ⫾ 14

.001 .002 .001

Total possible points, 100.

Tac Tie Total

Preoperative Score

Follow-up Score

P Value

10 ⫾ 5 13 ⫾ 9 12 ⫾ 8

30 ⫾ 5 31 ⫾ 3 30 ⫾ 4

.001 .001 .001

Total subjective points, 35.

had preoperative VAS, percent function, and answered the satisfaction postoperative question, “Would you go through the surgery and postoperative rehabilitation to achieve the result that you have today?” Because 16 patients did not have complete data for the preoperative Constant scores and ASES Index, and while they had postoperative scores available, these patients were not included in the statistical analysis to compare preoperative with postoperative scores. All of the affected extremities were the dominant side for both men and women. All patients were evaluated for postoperative clinical findings and radiographic findings. All patients except 2 had full passive range of motion, near complete alleviation of their pain and, clinically, 5/5 strength of the affected extremity by 4 months postoperatively. However, actual tensiometer increases were noted up to 1 year with minimal increases noted in the subsequent year. There were 19 patients who had consecutive data available for the complete analysis for the Constant score and ASES index, in addition to the 3 other parameters. The 4 shoulder rating systems used for evaluation showed a statistically significant improvement from preoperative to postoperative, except for the objective Constant score for the tie group; it was off by 8/1,000th of a P value. In other words instead of the results happening by chance 5% of the time, this suggests that it may occur 5.8% of the time. For the combined group, there were 24 men, average age 60 years. There were 11 women, average age 63 years. For the combined group, the total Constant score, average 53 and average 77, the objective Constant score, average 41 and average 47, the subjective

Constant score, average 12 and average 30, the ASES index, average 32 and average 79, the VAS, average 7 and average 2, and the percent function, average 42 and average 84, preoperative to postoperative, were statistically significant (Tables 1-6). There were no reoperations for retearing or nonhealing. There was no difference in outcomes based on gender. Finally, the study group was broken down into 2 subsets to determine the effect of the fixation device used for the repair of the subscapularis tendon. The supraspinatus tendon was repaired in the same fashion for all patients using a 5-mm Arthrex Corkscrew with two #2 Tevdeks per eyelet. The subscapularis was repaired with 8-mm Accufex Suretacs for patients with a robust reflection pulley and with two 5-mm Corkscrews with tie technique for patients who had a flimsy, torn, or frayed reflection pulley. There was no gender difference in the outcome. For the tac subgroup, there were a total of 9 patients, 5 men, average age 57 years, and 4 women, average age 60 years. For the tac subgroup, the total Constant score, average 50 and average 77, the objective Constant score, average 40 and average 47, the subjective Constant score, average 10 and average 30, the ASES index, average 33 and average 88, the VAS, average 7 and average 1, and the percent function, average 36 and average 86, preoperative to postoperative, were statistically significant (Tables 1-6). For the tie subgroup, there were a total of 10 patients, 7 men, average age 66 years, and 3 women, average age 59 years. For the tie subgroup, the total Constant score, average 55 and average 77, the objective Constant score, average 42 and average 46, the TABLE 4. Objective Constant Score

TABLE 2. Total Constant Score

Tac Tie Total

Preoperative Score

Follow-up Score

P Value

50 ⫾ 10 55 ⫾ 16 53 ⫾ 13

77 ⫾ 12 77 ⫾ 8 77 ⫾ 10

.001 .001 .001

Total Constant points, 100.

Tac Tie Total

Preoperative Score

Follow-up Score

P Value

40 ⫾ 8 42 ⫾ 8 41 ⫾ 8

47 ⫾ 9 46 ⫾ 7 47 ⫾ 7

.050 .058 .050

Total points, 65: 40 points range of motion, 25 points tensiometer strength testing.

28

W. F. BENNETT TABLE 5. Pain Scale

Tac Tie Total

Preoperative Score

Follow-up Score

P Value

7⫾2 7⫾3 7⫾3

1⫾1 2⫾2 2⫾2

.001 .002 .001

Total points, 10.

subjective Constant score, average 13 and average 30, the ASES index, average 31 and average 72, the VAS, average 8 and average 2, and the percent function, average 47 and average 83, preoperative to postoperative, were statistically significant (Tables 1-6), with the exception of the objective Constant score, P ⫽ .058. Specific intraoperative findings for the status of the biceps tendon included 8 patients who had tenosynovitis of the biceps tendon, 6 who had tendonitis/ fraying of the biceps tendon, 3 who had a tendon in a subluxated position, and 1 whose tendon was avulsed. There were 5 patients who had tears of the SGHL/ MCHL complex and 1 who had a tear of the MCHL without tearing of the subscapularis insertion. There were 3 who had plastic deformation of the SGHL/ MCHL complex. There was 1 patient who had involvement of the lateral head of the coracohumeral ligament (Fig 8). Of the 19 patients, 18 had some type of biceps tendon pathology (Table 7). The subscapularis tears were classified as follows: complete IASS (retracted versus nonretracted) and partial disruption of the IASS. There were 6 noncomplete disruptions of the IASS and 13 complete disruptions of the IASS with various degrees of retraction. The operative technique was similar for all types of tears. The only difference was in the number of softtissue fasteners. All repairs were checked for gap formation by bringing the arm from full internal rotation to full external rotation. None exhibited evidence of gap formation or loosening of the fasteners. This was assessed by arthroscopically visualizing the repair while bringing the arm into full external rotation with the arm at the side and with the arm in 90° of abducTABLE 6. Percent Function

Tac Tie Total

Preoperative Score

Follow-up Score

P Value

36 ⫾ 16 47 ⫾ 16 42 ⫾ 17

86 ⫾ 17 83 ⫾ 12 84 ⫾ 14

.001 .002 .001

Total percent, 100.

FIGURE 8. View of right shoulder, posterior portal, subacromial space. The supraspinatus tear on occasion can involve the lateral head of the coracohumeral ligament (LCHL) or the lateral wall of the bicipital sheath. This is a type-4 biceps subluxation/instability lesion (see Fig 9). (LCHL, lateral head coracohumeral ligament.)

tion. Patients who had a successful result were those who at follow-up had minimal to no pain, near full passive range of motion for all planes, and did not exhibit evidence of increased external rotation compared with the opposite extremity. Additionally, the lift-off, belly press, and Yergason’s and new biceps subluxation tests had to be negative. There were 2 patients who had positive findings (i.e., positive belly press), but did not require repeat surgery because pain was alleviated. In terms of morbidity, 1 biceps tendon disruption was noted. Interestingly, this was in a patient who had a tenodesis. There were 2 patients who had residual inflammation of the biceps tendon.

TABLE 7. Subscapularis Tear Qualification/Associated Biceps Pathology/Coracohumeral Ligament Pathology Lesions

Occurence

IASS partial IASS complete Biceps tenosynovitis Biceps tendonitis/fraying Biceps subluxation Biceps avulsion Pulley tear Pulley stretch

6 13 8 6 3 1 5 3

NOTE. Patients can have more than one lesion.

CLASSIFICATION OF BICEPS SUBLUXATION DISCUSSION No article was found in the literature that reported on the outcomes of arthroscopic repair for anterosuperior rotator cuff tears. The incidence of subscapularis involvement in the gamut of arthroscopically treated shoulders for various types of pathology was 27%, and specifically for rotator cuff pathology was 35%.11 This may be higher than previously thought.1 Routine arthroscopic visualization of the rotator interval10 and a better understanding of MRI, particularly with gadolinium,13 have enhanced our ability to make these diagnoses. In times past, series of the open repair of supraspinatus tears may have missed subscapularis tears, particularly the noncomplete tears. As such, studies that have reported on supraspinatus repair results may not have included the anterosuperior rotator cuff lesion. As noted previously, the term anterosuperior rotator cuff tear has at times2 been used to differentiate massive posterosuperior rotator cuff tears from those that included the subscapularis tendon. This study has addressed shoulders that had pathology of the subscapularis and supraspinatus tendons without infraspinatus tendon involvement, and with/without hidden rotator interval lesions. It seems that, when classifying tendon tears by tendon involvement, it would be easier to report anterosuperior tears as those that involve the subscapularis and supraspinatus, posterosuperior tears as those that involve the supraspinatus and infraspinatus, and massive tears as those that involve 3 or more tendon tears, i.e., subscapularis, supraspinatus, infraspinatus, and/or teres minor. “Hidden” rotator interval lesions6,7 have been implicated in and found in association with biceps tendon inflammation, widening, subluxation, and fraying.6,7,14 In a previous article, this author has presented a classification system for lesions that allow for biceps subluxation/instability.11 As noted in this article, biceps subluxation can occur as follows: type 1, with tears of the subscapularis (IASS) without involvement of the medial head of the coracohumeral ligament; type 2, without tears of the subscapularis (IASS) with involvement of the medial head of the coracohumeral ligament; type 3, with tears of the subscapularis (IASS) and with involvement of the medial head of the coracohumeral ligament; type 4, with tears of the supraspinatus and the lateral head of the coracohumeral ligament (Fig 8); and type 5, if there is a tear of the subscapularis (IASS), medial and lateral head of the coracohumeral ligament involving the leading edge of the supraspinatus tendon. Type 5 represents a

29

complete loss of any integrity to the medial and lateral walls of the bicipital sheath. Figure 9 is a simplified artistic rendition of the classification system for biceps subluxation/instability. There is a subtle difference between types 1 and 2. If the SGHL/MCHL complex is intact and the subscapularis is torn, the biceps displays intrasheath subluxation. The internal reflection or biceps pulley undergoes plastic deformation and without the robust insertion of the subscapularis insertion, the sheath becomes enlarged and the biceps can display increased medial/lateral motion. If the internal reflection, biceps pulley, SGHL/MCHL, medial wall is dis-

FIGURE 9. Biceps subluxation/instability classification. (s, subscapularis; m, medial head coracohumeral ligament; b, biceps; L, lateral head coracohumeral ligament.) Arrows represent biceps subluxation/instability direction. Type 1, with tears of the subscapularis (IASS) without involvement of the medial head of the coracohumeral ligament. Type 2, without tears of the subscapularis (IASS) with involvement of the medial head of the coracohumeral ligament. Type 3, with tears of the subscapularis (IASS) and with involvement of the medial head of the coracohumeral ligament. Type 4, with tears of the supraspinatus and the lateral head of the coracohumeral ligament. Type 5, with tears of the subscapularis (IASS), with medial and lateral head of the coracohumeral ligament including the leading edge of the supraspinatus tendon.

30

W. F. BENNETT

rupted (a capsular lesion), without a disruption of the subscapularis tendon, the bicipital sheath is also plastically deformed and enlarged and the biceps will exhibit increased medial/lateral motion. The main difference in area of untoward motion is that with the first situation the abnormal motion is toward the joint on the inner surface of the remaining subscapularis tendon and in the third situation, the motion is medial but the biceps subluxates on the outer surface of the subscapularis tendon but is still held within a relative sheath because the coracohumeral ligament fibers that cross to the lateral sheath hold the tendon from subluxing completely into the subcoracoid space. When both the internal reflection and subscapularis insertion are disrupted, the biceps can dislocate into the joint. In a rare condition when the supraspinatus tendon is torn with the lateral head of the coracohumeral ligament, the biceps tendon can sublux into the subcoracoid space. Whether one is using a tac or tie technique when repairing the subscapularis tendon, it is important to include the medial head of the coracohumeral ligament (internal reflection) in the repair to further aid in the reconstruction of a medial buttress to the bicipital sheath. If the subscapularis is repaired alone, the biceps tendon, can potentially sublux external to the subscapularis tendon and internal to the body of the coracohumeral ligament. If a surgeon were to repair just the subscapularis insertion and not the internal reflection, one might actually create a type-2 lesion. The only preoperative finding that may correlate with lesions of the SGHL/MCHL complex is an inferior drawer that does not diminish with external rotation.1 MRA is not helpful in discerning pulley lesions and it is only with direct arthroscopic visualization or opening of the rotator interval by open surgery that these lesions can be detected. The biceps is often involved in anterosuperior rotator cuff lesions and, as stated by Nove-Josserand et al.,1 “. . .although we do not know whether a large tendon causes a pulley lesion or a pulley lesion causes the biceps to become thick and flattened, the second hypothesis is more compatible with our experience.”1 This author agrees with this opinion. Additionally, fraying of the biceps tendon is usually localized to the area where the biceps tendon slips over the small tuberosity just medial to the IASS insertion. With lateral head coracohumeral ligament lesions, the fraying may be lateral but can also be medial. Biceps inflammation (tenosynovitis), subluxation and/or tendonitis/fraying have been traditionally addressed by tenodesis, debridement, and tenolysis. As such, this

approach does not treat the probable inciting factor, which is the tear of the subscapularis tendon and/or reflection pulley/medial head coracohumeral ligament.6,7,14 Arthroscopic observation of patients who have various hidden rotator interval lesions usually have some type of damage to the biceps tendon, whether it be inflammation (tenosynovitis), tendonitis/fraying, and/or subluxation. Eighteen of 19 patients in this study had biceps pathology. Additionally, arthroscopic observations of patients who have significant partial-thickness tears of the subscapularis almost routinely have some plastic deformation of the medial sheath of the bicipital wall. As was noted in a previous study,11 one type of biceps subluxation is intra-sheath, secondary to the plastic deformation of the sheath. While some1 reported that the open reconstruction of the sheath has not been successful and that the biceps tendon ruptures within 1 year in 33% of the cases, this study does not support that conclusion. The only bicep disruption that was noted in the follow-up period was the patient who had a tenodesis. This patient had minimal residual achiness. Please note that the literature on the reconstruction of the bicipital sheath that has shown a high incidence of biceps rupture, and has not been advocated, has dealt with a various number of open techniques. Some of them include deepening of the bony groove and transfers of tissue normally not found in that area. The arthroscopic reconstruction reported herein is in actuality a primary repair. First, an identification of the torn structures and then a reduction of these anatomic structures to their relative areas of insertion is performed. This approach is vastly different than deepening the groove. No patient after a successful repair had signs of biceps subluxation. Two patients had continued biceps inflammation that eventually resolved with time, less than 6 months postoperatively. This residual inflammation was found in the tac group. One explanation for the continued irritation of the biceps tendon in these patients may be secondary to irritation by the tac head or by irritation secondary to tac particles during resorption. Debridement is an attempt to save the tendon but does not allow for addressing the inciting factor. Seemingly, reconstruction that salvages the biceps tendon would be better suited for the patient than tenolysis and tenodesis. It has been stated in the literature that patients with a subscapularis tear will have increased passive external rotation.8 Increases in external rotation may not be detectable with acute subscapularis tears as there may

CLASSIFICATION OF BICEPS SUBLUXATION be pain inhibition. Additionally, subscapularis tendon tears will exhibit increased external rotation in complete tears but subtle differences may be impossible to detect in partial-thickness tears. These patients will have anterior pain, may have traditional impingement signs, and additionally may experience pain at shoulder elevation levels lower than those found in patients with traditional impingement symptoms. Subtleties in the Gerber lift-off, belly press sign, and lag signs should be notable in partial-thickness tears. Additionally, the patient may have weakness of internal rotation and the inability to bring the arm to the posterior aspect of the lumbar area. My clinical observations have determined that the belly press test and lift-off test can detect partial- and full-thickness tears of the subscapularis tendon only if there is normal internal rotation of the shoulder when it is at 90° of abduction as compared with the normal side. If there is any compromise of internal rotation at this position compared with internal rotation on the contralateral side, this test provides for false positive findings. In a previous study,11 subscapularis tears were classified with respect to length and thickness. The prior classification of subscapularis tears may be somewhat confusing and in this study it has been simplified to complete IASS and partial (noncomplete) disruption of the IASS. It has been noted that, with partial tears of the subscapularis tendon, the arthroscope may not reveal the full extent of the tear.1 The partial-thickness tearing can be extra-articular. This is where preoperative MRA and a thorough understanding of the findings will help the surgeon look for this extended lesion, usually a spot of bare bone or mixed signal intensity (representing scar tissue) on the MRA. Arthroscopically, this is where the take-down of a small rim of articular cartilage immediately superior to the insertion site of the IASS renders itself helpful. With debridement of the fluff of tissue found at the insertion site, the lesion can be better visualized to its inferior extent. The partial-thickness IASS tear may be more accurately termed the noncomplete IASS tear because these tears are often quite extensive1 but hidden. This study does not define a complete subscapularis tear as one that is completely detached and retracted to the glenoid neck. It is defined as a complete detachment of the IASS. The IASS, when detached from the trough where it belongs and if the lateral head of the coracohumeral ligament and lower fibers of the subscapularis are still attached to the lateral side of the bicipital groove, the entire complex may appear normal. However, when torn completely, the tendinous portion migrates inferiorly and medially and there is

31

subtle lengthening of this complex. With time the only way to tell is to locate the trough, usually behind a fluff of tissue. Additionally, the actual insertion site of the tendinous portion is usually thicker than other parts. During the arthroscopic visualization of this area, one may pull the biceps tendon into the joint. With lesions of the subscapularis, the biceps tendon is seen to “fall” medially and inferiorly. This is an arthroscopic indicator of biceps subluxation.5 However, the author has not correlated this with cadaveric cutting studies. Following the repair of the IASS and the medial head of the coracohumeral ligament, the biceps should be minimally medially displaced with a neuroprobe and the bicipital sheath should be seen to encircle the biceps tendon without much space (Fig 10). The weaknesses of this study include the lack of randomization and the subjective selection of patients that received tac versus tie fixation of the subscapularis tendon. Additionally, the study was not blinded. Finally, definitive tests for rerupture or nonhealing were not used and conclusions were based on clinical examination findings and the acromiohumeral interval. There is a built-in weakness in the Constant score when it used to evaluate the effect of repair of the subscapularis tendon. As mentioned in the Methods section, there are only 10 points of the entire 100 points (10%) that are related to the subscapularis tendon directly and in an objective manner. One other

FIGURE 10. Following the repair of the IASS and the medial head of the coracohumeral ligament, the biceps should be held in a stable position without falling off the front of the humeral head. The bicipital sheath is seen to encircle the biceps tendon without much of a patulous area. The 2 arrows represent the knots of the repaired subscapularis tendon and reflection pulley. The biceps tendon no longer drops inferiorly with neuroprobe pressure.

32

W. F. BENNETT

question may be indirectly related to subscapularis findings: “At what level can you use the shoulder without pain?” Lower levels of activity producing pain (that is below shoulder height), may be indicative of subscapularis involvement. The previous question is weighted for only 10 points of the 65-point objective component and is related to active internal rotation levels. Twenty-five points of the 65 points are related to strength testing of the supraspinatus. As such, only 15% of the objective component of the Constant score is directly affected by abnormalities of the subscapularis tendon. Every parameter evaluated preoperatively to postoperatively in this study was statistically significantly different except one, the objective component of the Constant score for the tie group. The P value was .058, 8/1,000th more than is allowed by statisticians. Clinically this outcome difference is significant and the numbers may be affected by the above fact; there is only 15% of the objective component of the Constant score directly related to the subscapularis tendon. Finally, because of the decision to treat patients who had less than a robust tendon with the tie technique, this may have skewed the results slightly, i.e., the lesion may have been worse to begin with. So when evaluating these lesions, the negation of preoperative clinical findings related to subscapularis insufficiency and biceps subluxation and/or inflammation is as important as the number score. In this study, there were no patients with residual biceps subluxation findings, none with lag signs, and none with a positive belly press test. There were 2 patients with improved lift-off tests but they were not felt to be completely normal. It is unclear whether these 2 patients had residual tears of the subscapularis tendon or if the clinical examination finding was affected by the lack of internal rotation with the arm at 90° of abduction, i.e., a tight posterosuperior capsule. The repair of these anterosuperior lesions should afford for healing of both tendons and decrease untoward medial/lateral motion of the biceps tendon. The possible omission of detailed attention to the lesions immediately proximate to the bicipital sheath may be a possible reason for the previously noted poorer results found in repair of the anterosuperior rotator cuff tear compared with the posterosuperior tear.3 The total preoperative Constant score in this study was 53 ⫾ 13 and the postoperative Constant score was 77 ⫾ 10. The only other study that this value could be compared with was Warner’s article2 on open repair. The values can only be evaluated by percent increase from preoperative to postoperative. The 31% increase

is nearly identical to the 31.1% increase found in this study. Fourteen of 19 patients had involvement of the infraspinatus tendon, whereas in this study, infraspinatus involvement was excluded. The arthroscopic repair of these anterosuperior lesions affords the patient an increased ability to use the arm in the internally rotated position for maneuvering the arm behind the back, provides alleviation of subscapularis findings in patients without a tight posterosuperior capsule, and removes clinical findings of biceps subluxation and biceps inflammation at a minimum of 2 years of follow-up. Additionally, the improvement in pain experienced by these patients was found to be statistically significant. To determine whether this was a result of the subscapularis repair or the supraspinatus repair is not elucidated in this study. However, with this operative approach, 100% of the patients (35) would undergo this procedure and go through the postoperative rehabilitation to achieve their respective functional result. This approach returns patients to a functional level quickly and with minimal morbidity. Acknowledgment: Special thanks for the statistical analysis to John C. Pezullo, Ph.D., Associate Professor, Clinical Pharmacology and Biostatistics, Georgetown University, Washington, D.C. The author thanks Lawrence May, Bradenton, Florida, for his artist renditions.

REFERENCES 1. Nove-Josserand L, Gerber C, Walch G. Lesions of the anterosuperior rotator cuff. In: Warner JJP, Ianotti JP, Gerber C, eds. Complex and revision problems in shoulder surgery. Philadelphia, Lippincott-Raven, 1997;165-176. 2. Warner JJP, Higgins L, Parsons IM, Dowdy P. Diagnosis and treatment of anterosuperior rotator cuff tears. J Shoulder Elbow Surg 2001;10:37-46. 3. Frankle MA, Cofield RH. Rotator cuff tears including the subscapularis. Presented at the 5th International Conference of Surgery of the Shoulder, Paris, France, 1992, p 52. 4. Ticker JB, Warner JJP. Single-tendon tears of the rotator cuff. Evaluation and treatment of subscapularis tears and principles of treatment for supraspinatus tears. Orthop Clin North Am 1997;28:99-116. 5. Sakurai G, Ozaki J, Tomita Y, Kondo T, Tamai S. Incomplete tears of the subscapularis tendon associated with tears of the supraspinatus tendon: Cadaveric and clinical studies. J Shoulder Elbow Surg 1998;7:510-515. 6. Walch G, Nove-Josserand L, Levigne C, Renaud E. Complete ruptures of the supraspinatus tendon associated with “hidden lesions” of the rotator interval. J Shoulder Elbow Surg 1994; 3:353-360. 7. Nove-Josserand L, Levigne C, Noel E, Walch G. Isolated lesions of the subscapularis muscle: Apropos of 21 cases [in French]. Rev Chir Orthop Reparatrice Appar Mot 1994;80: 595-601.

CLASSIFICATION OF BICEPS SUBLUXATION 8. Gerber C, Hersche O, Farron A. Isolated rupture of the subscapularis. Results of operative repair. J Bone Joint Surg Am 1996;78:1015-1023. 9. Gerber C, Krushell RJ. Isolated rupture of the tendon of the subscapularis muscle. Clinical features in 16 cases. J Bone Joint Surg Br 1991;73:389-394. 10. Bennett WF. Visualization of the anatomy of the rotator interval. Arthroscopy 2001;17:107-111. 11. Bennett WF. Subscapularis, medial and lateral head coracohumeral ligament insertion anatomy: Arthroscopic appearance

33

and incidence of “hidden” rotator interval lesions. Arthroscopy 2001;17:173-180. 12. Klapper RC, Jobe FW, Matsura P. The subscapularis muscle and its glenohumeral like bands. A histomorphologic study. Am J Sports Med 1992;20:307-310. 13. Pfirrmann CWA, Zanetti M, Gerber C, Hodler J. Subscapularis tendon tears: Detection and grading at MR arthrography. Radiology 1999;213:709-714. 14. Slatis P, Aalto K. Medial dislocation of the tendon of the long head of the biceps brachii. Acta Orthop Scand 1979;50:73-77.