Complications of Rotator Cuff Repair

Joseph Iannotti

Michael Codsi

Pain or disability following a rotator cuff repair is not always due to a recurrent tear. In fact, subjective outcomes after recurrent tears can be quite good. Common complications of rotator cuff repair include failed repairs, stiffness, deltoid injury, persistent subacromial impingement, and infection. Other issues include heterotopic ossification, acromial fractures, and anterosuperior humeral head subluxation. A precise diagnosis of the etiology of the patient’s complaints is the key to the treatment of complications of rotator cuff repair.

FAILED REPAIR

Successful open repair of the rotator cuff depends on many patient factors including cuff tear size, patient age, tendon quality, tendon retraction, and degree of muscle atrophy. The clinical success rate varies between 85% and 90% when measured in terms of patient satisfaction or patient outcome scores. These high success rates, however, do not directly correlate well with the rate at which tendon repairs heal when measured by postoperative imaging. Harryman et al. (1) described their findings in a series of 105 rotator cuff repairs, which were followed by an ultrasound exam to document the integrity of the repair. In all, 80% of the supraspinatus repairs were intact, 57% of the supraspinatus/infraspinatus repairs were intact, and 32% of three tendon tears were intact at an average of 5-years follow-up. Older patients with larger tears were more likely to have a recurrent defect. Nineteen patients in this series had revision cuff repair and they were more likely to have a recurrent defect than the primary repairs. Patients who had an intact repair had better function, motion, and pain relief, and the patients with recurrent defects had worsening function, motion and strength with increasing cuff defects. Overall, 94% of patients were satisfied with the repairs; this did not correlate with cuff integrity.

The lack of correlation between rotator cuff integrity and clinical outcome is commonly seen. Jost et al. (2) prospectively evaluated a series of 65 rotator cuff repairs and found 20 reruptures by magnetic resonance imaging (MRI) scan. Sixteen were smaller than the initial repair. Atrophy, fatty degeneration, and glenohumeral osteoarthritis progressed in all 20 patients, yet only three patients were not satisfied with the procedure and one patient underwent revision surgery. In the 20 reruptures, the average Constant score, adjusted for age- and sex-related normal values, increased from 49% preoperatively to 83% postoperatively; the scores did correlate with the number of tendons retorn.

An even larger disparity between patient satisfaction and cuff integrity was reported in the series by Galatz et al. (3). Eighteen patients underwent arthroscopic repair of large and massive tears and ultrasound evaluation of the cuff 12 months after surgery. Recurrent tears were seen in 17 patients. Despite the high rate of failed repair, 16 patients had an Adult Self-Expression Scale (ASES) score lower than 90, 12 patients had no pain, and all patients regained above shoulder motion to an average of 152 degrees. All patients were satisfied with the procedure. At the 2-year follow-up exam, the ASES scores decreased by 5 points and their forward elevation decreased 10 degrees. Outcome scores were not correlated with number of tendons retorn.

Etiology of Failed Repairs

The causes of a failed repair have been correlated to cuff tear size (number of tendons and amount retraction), degree of muscle atrophy (which in turn is related to size of the tear),
and chronicity of the tear. All of these factors influence the quality of the tendon tissue. In addition, other factors associated with failure of the tendon repair to heal can be attributed to inadequate subacromial decompression, poor tendon mobilization, excessive tendon tension, inappropriate suture technique, patient age, improper physical therapy, and patient compliance.

There are several main risk factors for rerupture following rotator cuff repair. Jost et al. (2) reported that in their series of 80 revision repairs, the initial tear was massive in 28% and large in 35% of the revisions. In all, 49 patients had an inadequate subacromial decompression and four patients underwent complete lateral acromionectomy. Twenty-four patients had inadequate tissue for a secure repair of the cuff at the initial repair, resulting in 17 repairs requiring arm abduction in order to bring the tendon to its insertion site. Also, nine patients had inappropriate rehabilitation because they either initiated strengthening too early or neglected passive motion exercises too late. The active motion before revision was 105 degrees of forward flexion, 39 degrees of external rotation, and internal rotation to the 11th thoracic vertebrae.

Similarly, Bigliani et al. (4) reviewed 31 patients who underwent revision surgery for their failed rotator cuff repairs. In all, 97% of the patients had large or massive tears at the initial procedure and 90% of the patients had persistent subacromial impingement before their revision procedure. Fifteen patients did not have adequate mobilization of the cuff tendon for attachment to the bone at the initial repair.

DeOrio et al. (5) attributed the failure of rotator cuff repairs to the size of the tears, inadequate subacromial decompression, and possibly inadequate postoperative external support. The average patient age was 52 years and 66% of the patients had a large or massive tear. Seven patients had tenderness over the supraspinatus tendon, a painful arc of motion, and a positive impingement sign, suggesting that the subacromial pathology was still present after the initial repair.

These three studies all had a retrospective study design and it is not known the frequency in which these presumed negative factors were present in successful surgery (2,4,5). The data and conclusion presented are suggestive but not definitive for a correlation between initial tear size, subacromial pathology, and clinical outcome.

In addition to these factors, Davidson et al. (6) studied the effects of rotator cuff tension on the subjective and objective outcomes after 67 rotator cuff repairs. A tissue tensiometer was used to measure the tension on the tendon after it was stretched to its insertion site on the humerus. ASES scores, isokinetic strength measurements, and visual analog pain scores were all worse if the tension of the repair was greater than 8 pounds.

Effect of Fixation Technique

Many studies have evaluated the effects of suture techniques on the strength of rotator cuff repairs. Gerber et al. (7) used the modified Mason-Allen stitch and a simple stitch to repair the cuff tendons in sheep. He found that the simple stitch pulled out of the tendon in all cases, whereas the Mason-Allen stitch failed in only 2 of 17 repairs. The fixation through the bone was also varied between augmentation with cortical-bone polylactide plate and no augmentation. The augmented group never failed, whereas the nonaugmented group failed in 8 of 16 cases.

The mode of failure for rotator cuff repairs was prospectively evaluated by Cummins et al. (8) in a series of 342 patients who underwent repair with suture anchors. Repairs were done with either one or two rows of anchors, depending on the size of the tears. Horizontal mattress sutures were used in all repairs. In all, 21 repairs failed and underwent revision surgery, and 19 of those failed at the suture-tendon junction. One suture anchor pulled out of the bone and two retears occurred at a different location.

The results of tendon repair found in the animal model were tested in a clinical trial by Gerber et al. (9). Repair of 29 massive rotator cuff tears involving complete detachment of at least two tendons was performed using no. 3 braided polyester sutures placed in a modified Mason-Allen technique. Both ends of the sutures were brought through the greater tuberosity and tied over a titanium plate used for cortical bone augmentation. After a minimum 2-year follow-up, the Constant score improved from 49% to 85%. The overall rate of retears according to MRI was 34%. The Constant score was 91% in group with an intact repair compared to 61% in the group with a retear (p = 0.016).

Workup

Clinical Evaluation

Patients with recurrent rotator cuff tears may complain of pain and weakness in the shoulder. It is imperative that the clinical assessment of a patient with persistent pain after rotator cuff surgery include an evaluation of all possible causes of shoulder pain beyond that of an anatomic finding of a persistent cuff defect. This includes symptomatic biceps tendon, acromioclavicular arthritis, cervical spine disease, or frozen shoulder. The importance of this fact is underscored by the extensive literature that clearly demonstrates that many patients with persistent cuff defects after rotator cuff repair have subjective outcome (using patient directed outcome assessment tools) equal to that of patients without persistent defects.

Many patients require narcotics before their revision surgery (4). The pattern of pain may help differentiate between a recurrent tear and persistent subacromial pathology, but this is often unreliable. Weakness, especially when it is not caused by pain, can predict which cuff tendon remains functional. The external rotation lag sign tests the integrity of the external rotators (10). If the arm cannot be actively held at maximum passive external rotation, then the infraspinatus tendon is substantially involved in the tear.

Testing the subscapularis can be done with the lift-off test or the belly-press test (Napoleon test). The lift-off test is the most reliable, but if the patient has pain or lacks sufficient internal rotation to get the hand to the lumbar spine, then the test is not reliable. The belly-press test is positive if the patient is unable to push their hand against their belly while keeping their elbow at or anterior to the coronal plane of the body. If the wrist flexes or the elbow extends behind the patient, then the test is positive. This test is sometimes difficult to perform for the patient and it can give a false-positive result if there is loss of passive internal rotation. A study of 25 patients with subscapularis tendon tears by Burkhart et al. (11) described the abdominal compression or Napoleon test and its relationship to the operative findings. Eight of nine patients with a positive Napoleon test had complete tears of the subscapularis, whereas seven patients with a negative Napoleon test had a tear of the upper half only. The authors warn that the reliability of the Napoleon test is altered if the patient lacks passive internal rotation.

Imaging

Radiographic evaluation of a patient with a suspected rotator cuff retear should start with plain films, including an outlet view to look for subacromial spurs, a Greise view or true anteroposterior view of the shoulder joint to evaluate the joint space, and an axillary lateral view to inspect the acromion for possible fracture or os acromiale. Findings particular to rotator cuff pathology include displaced suture anchors. Pearsall et al. (12) showed that greater tuberosity sclerosis, osteophytes, subchondral cysts and osteolysis are associated with rotator cuff tears, but these changes are not reliable in the postoperative setting. Significant superior migration of the humeral head is associated with large and massive chronic rotator cuff tears. When the acromial humeral space is less than 7 mm, the tear is often irreparable or, if reparable, has a low chance to heal.

Ultrasound can effectively evaluate the rotator cuff in patients who have sufficient pain-free motion to expose the cuff tendons from underneath the acromion. Pricket et al. (13) reported the results of 44 rotator cuff repairs that underwent both ultrasound and MRI exams. The ultrasound sensitivity was 91% and the specificity was 86%. The accuracy was 89%. Another study by Teefey et al. (14) compared the ultrasound findings and arthroscopic findings of 100 patients with shoulder pain who did not undergo surgery. All full thickness tears were correctly diagnosed with ultrasound, resulting in a specificity of 100%. There were 17 true-negative and 3 false-negative ultrasounds, resulting in a specificity of 85%.

Gaenslen et al. (15) compared the findings of MRI exams to the operative findings of 30 shoulders that had failed rotator cuff repairs. The sensitivity for full-thickness tears was 84% and the specificity was 91%. The sensitivity for partial-thickness tears was 83% and the specificity was 83%. The positive-predictive value for full-thickness tears was 94% and the negative predicted value was 77%. The MRI scans in this study also diagnosed articular cartilage damage and a ganglion cyst in the supraspinatus muscle, both of which could not be diagnosed with ultrasound.

Results of Revision

The results of revision rotator cuff repairs in patients without concomitant pathology, such as deltoid injury or acromioclavicular arthritis, are not easily found in the literature. The first published report of revision cuff repairs by DeOrio et al. (5) included 27 patients with a minimum 2-year follow-up. In all, 37 percent of patients had no, slight, or moderate pain, and 76% of patients had a reduction in pain. Seven patients had a third attempt to repair the rotator cuff and three of those patients went on to have a shoulder fusion. Overall, only four patients were considered to have a good result.

Bigliani et al. (4) reported slightly better results in their series of revision cuff repairs. Thirteen patients had attempts at mobilization and repair of the detached deltoid origin in addition to revision repair of the rotator cuff. At the time of revision, a massive tear was found in 12 patients, a large tear in 10 patients, a medium tear in seven patients, and a small tear in two patients. The results were excellent in 19%, good in 32%, fair in 23%, and poor in 26% of patients. An inferior result was associated with previous lateral acromionectomy, detached deltoid origin, and poor tendon quality at the time of revision.

Neviaser et al. (16) reported their results after reoperation for failed rotator cuff repairs in 50 patients. The patients included 39 men and 11 women with an average age of 54.5 years. The average number of previous operations was 1.6. Six patients had deltoid dysfunction. In all, 28 of 50 patients started resistance exercises within the first 3 months of their first rotator cuff repair. There were 6 small tears, 23 large tears, and 21 massive tears in the group. After operative fixation of the tear through bone troughs, the patients were allowed to do passive shoulder motion for 6 weeks, followed by active motion until 3 months after the repair. Resistive exercises were then started. After a mean follow-up of 30 months, 19 patients had no pain and 27 patients had only slight shoulder pain. Four patients had no change in pain. Improved motion was found in 26 patients, whereas 22 patients had the same shoulder motion.

Revision rotator cuff repair in 80 patients was more recently reported by Djurasovic et al. (17). The average initial tear size was massive in 22, large in 28, medium in 10, and small in 13 patients. In all, 45% of repairs required an anterior interval slide to aid in cuff mobilization and 24 patients had additional procedures to repair the deltoid to the acromion. The overall results were excellent in 33%, good in 25%, fair in 9%, and poor in 31% of patients. The average pain rating improved from 3 to 7.4 points out of a maximum pain score of 10. A satisfactory outcome was
reported in 78% of patients with an intact deltoid origin, compared to 57% of patients who did not have a deltoid injury. Other factors related to outcome were cuff tendon tissue quality, preoperative active motion, and number of prior procedures.

The most recent report of revision rotator cuff repairs was done by Lo and Burkhart (18). They arthroscopically repaired 14 failed repairs that consisted of 11 massive, one large, and two medium tears. Seven patients were involved in worker’s compensation claims and five patients had superior migration of the humeral head. Concomitant pathology found at the revision surgery included five patients with persistent acromial spurs, two superior labrum anterior posterior (SLAP) lesions requiring repair, and three biceps tendonopathies. Four patients were found to have a U-shaped tear repaired directly to bone instead of using a margin convergence technique. The University of California Los Angeles (UCLA) scores improved from 13.1 to 28.6, which translated into four excellent, five good, four fair, and one poor result. Five patients returned to previous level of work activity, two patients returned to less strenuous work activity, and one patient did not return to work.

Pearls

Postoperative pain needs a systematic evaluation that includes a combination of clinical examination and imaging studies. The first step is to determine if the pain is from the shoulder or from some other source outside of the shoulder such as the brachial plexus, cervical spine, or a nonorganic cause. If the pain is thought to come from the shoulder, then selective injections can on occasion help localize the pain to the glenohumeral joint, subacromial space, or acromioclavicular joint. It also helps in some cases validate that the pain is organic and localized to the shoulder. Addressing all of the causes of shoulder pain both within the shoulder, outside of the shoulder, organic, and nonorganic is critical to both determining operative versus nonoperative treatment and, if surgery is indicated, the nature of the surgery to be performed.

When a persistent rotator cuff tear is present, the questions that need to be addressed for treatment include:

Is it the source of pain?
What are the patient’s primary objectives: pain relief or return of strength?
What are the patient’s functional goals?
Can these goals be achieved with a reasonable degree of success (75%)?
Is the cuff repairable?
What is the likelihood that the cuff will heal if reparable?
What are the consequences if a second repair is attempted and it does not heal (i.e., will the results of surgery be a failure to achieve some of the goals of surgery)?

If the MRI arthrogram demonstrates a persistent tear that is large or massive with retraction medial to the midhumeral head and there is associated moderate to severe atrophy, then it is likely that the tear is irreparable or, if reparable, the repair will not heal. If strength is a primary objective of surgery, then a muscle transfer (latissimus and or teres major for a posterior superior defect or a sternocostal portion of the pectoralis major for an anterior superior cuff tear) should be planned as part of the revision surgery. The best results of muscle transfer surgery is for a patient with tolerable pain, good passive range of motion, good deltoid function, an intact cuff on the side opposite to the transfer, a nonarthritic joint, active elevation to shoulder height, ability to control the arm in decent when placed in a full elevated position, an intact coracoacromial arch, and physiologic youthfulness both in mind and activity level. Multiple negative attributes in these categories are, in the senior author’s experience, associated with poor clinical results.

Candidates for arthroscopic debridement are patients with a large or irreparable tear, pain as a primary indication for surgery, the ability to actively elevate the arm to shoulder height, the ability to control the decent of the arm from a passively fully elevated arm, and an intact coracoacromial arch. Arthroscopic debridement includes smoothing of the arch (do not decrease the anterior-posterior dimension of the acromion or remove the coracoacromial ligament) and removal of all greater tuberosity osteophyte (tuberoplasty) and foreign material (sutures, loose anchors, etc.). The response (pain and function) to a preoperative injection of lidocaine into the glenohumeral joint is often a good indicator of the response to a limited-goal arthroscopic debridement.

When a persistent tear is repairable and the patient is symptomatic, then the tear should be repaired. All causes of persistent pain should be addressed at the time of surgery and/or with the pre- and postoperative rehabilitation program.

For a massive cuff tear with coracoacromial arch deficiency in a patient of advanced age (greater than 70 years), with anticipated postoperative activity level to include only activities of daily living, the only surgery that the senior author has found to be reliable in providing pain relief and improvement in functions is a reverse total shoulder replacement. This surgery should be reserved for this older and lower-activity level patient.

For revision rotator cuff repair or muscle transfer, the authors always use a postoperative abduction brace for 4 to 6 weeks after surgery. During this time, the brace is removed for dressing and washing and eating activities, as well for the exercise program, which includes pendulum, passive range-of-motion in supine forward flexion. Other than for these activities, the brace is worn during the day and through the night. No. 2 fiberwire suture or its equivalent is used through bone tunnels as in a primary open rotator cuff repair; bone augmentation with a plastic button is used if the bone quality is poor. The principles of scar
excision and cuff mobilization are achieved to the maximum extent possible.

DELTOID INJURY

The deltoid can be injured during open rotator cuff repair by an overaggressive subacromial decompression, inadequate deltoid origin repair, injury to the axillary nerve, or compression from retractors. Early reports of complete acromionectomy or lateral acromionectomy for the treatment of impingement syndrome detailed the disadvantages of trading a treatable pathology for an untreatable condition. Patients no longer complained of anterior shoulder pain, but some patients lost the ability to elevate the arm. Neer and Marberry (19) treated 30 patients who previously underwent a radical acromionectomy and all the patients had poor results. The authors concluded that the acromion allowed for a mechanical advantage necessary for forward elevation of the arm, and that loss of the deltoid muscle integrity resulted in significant disability.

Bosley (20) reported his results after treating 38 shoulders with total acromionectomy between 1969 and 1989. Six patients were laborers and 19 shoulders also had a concomitant rotator cuff repair. The age range was from 31 to 70 years, 10 patients were women, and 21 patients were men. Postoperatively, the shoulders were immobilized in abduction for 5 to 9 weeks, which depended on status of the rotator cuff repair. After a minimum 2-year follow-up, 25 shoulders had excellent results, four shoulders had good results, three shoulders had fair results, and one had a poor result. Poor results of complete acromionectomy are often associated with persistent cuff tears and deltoid problems. That being said, a complete or even subtotal acromionectomy is not advocated by the authors, but this article is cited to give a balanced review of the literature.

Only gold members can continue reading. Log In or Register to continue