Patient Selection
Inclusion criteria were (1) arthroscopic repair of a full-thickness tear of
the subscapularis in conjunction with either a tear of the supraspinatus or
tears of the supraspinatus and infraspinatus, (2) availability of preoperative
and postoperative magnetic resonance imaging scans of the involved shoulder,
and (3) a minimum duration of follow-up of two years. Excluded were patients
with (1) an irreparable rotator cuff tear, (2) a partial rotator cuff repair,
(3) stage-3 fatty infiltration (as much fat as muscle) or stage-4 fatty
infiltration (more fat than muscle) of the multiple rotator cuff muscles seen
on T1-weighted magnetic resonance
imaging1,18,
(4) cuff tear arthropathy demonstrated on preoperative plain radiographs and
at arthroscopy19,
(5) a failed prior rotator cuff repair, or (6) a Workers' Compensation claim.
A tear was considered to be reparable if the tendon could be restored to, or
within 10 mm of, its anatomical insertion when traction was applied to it. If
the tendon could not be positioned appropriately, sites at which its mobility
was limited were inspected and released as necessary. If the torn tendon could
not be mobilized adequately or if the tendon was absent, the lesion was
considered to be irreparable.
Between April 2001 and October 2004, 158 arthroscopic rotator cuff repairs
were performed at our institution. Of these, twenty-eight (17.7%) were done in
patients with a subscapularis tear combined with another rotator cuff tendon
tear (or tears). One patient with a partial repair of a torn rotator cuff, two
patients with a failed prior rotator cuff repair, four patients who were
unable to undergo postoperative magnetic resonance imaging, and one patient
who had filed a Workers' Compensation claim were excluded, leaving twenty
consecutive patients (twenty shoulders) for evaluation. There were seventeen
men and three women; their mean age was 61.7 years (range, forty-five to
seventy-nine years), and they were followed up for a mean of 36.1 months
(range, twenty-four to sixty months). Fifteen involved shoulders were on the
right, dominant side. All twenty patients had full-thickness tears of the
subscapularis and supraspinatus, and seven (35%) had a concomitant
infraspinatus tear.
All twenty patients had sustained an injury that was associated with the
acute onset of shoulder pain and followed by functional impairment of the
involved arm. Twelve patients were injured in a fall; in seven patients, the
mechanism of injury was resistance to an external rotation force with the
shoulder in a position of abduction and external rotation; and one patient was
injured in a motor-vehicle accident. The mean time from the injury to the
surgery was 2.7 months (range, one to six months). All patients reported pain
and weakness when performing tasks that required internal rotation of the
shoulder such as reaching behind the body or placing the hand in the back
pocket, seventeen patients had difficulty when reaching for an overhead shelf,
and sixteen had shoulder pain at night.
All patients provided written informed consent stating that they understood
the purpose of the study as well as the potential risks and benefits of the
operation. Institutional review board approval was not considered necessary
according to the standards of our institution.
Clinical Assessment
Preoperative and postoperative clinical assessment consisted of a
structured interview, a detailed physical examination, and evaluations with
use of the shoulder rating scale of the University of California at Los
Angeles (UCLA)20
and the Japanese Orthopaedic Association (JOA)
assessment17,21,22.
Shoulder Rating Scale
The JOA shoulder score includes ratings of the level of shoulder pain,
function including abduction strength, and the range of motion. The maximum
possible total score is 100 points. The outcome was classified as excellent
when the total JOA shoulder score was >90 points), good when it was 81 to
90 points, fair when it was 71 to 80 points, and poor when it was <71
points. When a patient required a reoperation, the outcome was classified as
poor.
Physical Examination
Physical examination consisted of measurements of the range of motion with
a goniometer; a manual muscle-strength test; the
Neer23 and
Hawkins24
impingement tests; and the
supraspinatus25,
infraspinatus26,
lift-off11,
belly-press12, and
Speed tests27. The
range-of-motion assessment included measurement of elevation in the sagittal
plane, external rotation in the coronal plane with the arm at the patient's
side, and internal rotation according to the highest vertebral level that the
patient could reach. The lift-off test was performed by placing the hand of
the affected arm on the back and asking the patient to internally rotate the
arm to lift the hand posteriorly off of the back. The test was considered
positive if the patient was unable to lift the hand posteriorly off of the
back or if the patient performed the lifting maneuver by extending the elbow
or the shoulder. The belly-press test was performed, with the arm at the side,
the elbow flexed to 90°, and the wrist straight, by having the patient
press the palm into his or her abdomen by internally rotating the shoulder.
The test was considered positive if the patient pushed the hand against the
abdomen by means of shoulder extension or wrist flexion. All patients were
evaluated at the latest follow-up by an orthopaedic shoulder specialist other
than the primary surgeon.
Radiographic Evaluation
Anteroposterior (adduction and abduction), axillary (West Point), and
scapulolateral radiographs and magnetic resonance imaging scans were made of
all shoulders before and after surgery. The coracohumeral
interval28 was
measured as the shortest distance from the tip of the coracoid process to the
anterior aspect of the humerus on the preoperative magnetic resonance image
made with magnification control with the arm in neutral rotation. All patients
had a magnetic resonance imaging study at the latest follow-up visit, at a
mean of 36.1 months (range, twenty-four to sixty months) postoperatively. The
structural integrity of the repaired rotator cuff, including the subscapularis
tendon, was evaluated on the postoperative magnetic resonance image. When
fluid-equivalent signal was found in the way of a tendon or when there was
complete nonvisualization of a tendon in at least one section of a
fluid-sensitive sequence, a diagnosis of a full-thickness rotator cuff tear
was made. The magnetic resonance imaging scans were independently reviewed by
an orthopaedic shoulder specialist and a radiologist trained in the
interpretation of musculoskeletal images. Interobserver reliability was almost
perfect (kappa statistic = 0.816). In our previous diagnostic study of
full-thickness rotator cuff tears, the sensitivity, specificity, and accuracy
of preoperative magnetic resonance imaging were 100%, 76.9%, and 89.2%,
respectively29.
Surgical Technique
The passive range of motion and anterior, posterior, and inferior
translation were assessed with the patient under general anesthesia and lying
in the lateral decubitus
position30,31.
A shoulder distraction device (3-Point Shoulder Distraction System; Arthrex,
Naples, Florida) and an arthroscopy pump (Continuous Wave-2 Arthroscopy Pump;
Arthrex) were employed. The arm was maintained in 10° to 50° of
abduction with longitudinal traction applied with use of a 3 to 4-kg weight.
Diagnostic arthroscopy was performed through posterior, anterior,
anterosuperior, and lateral subacromial
portals17,30.
Arthroscopic Evaluation and Classification of Subscapularis Tear
The size of the tear was measured with a calibrated probe. The
subscapularis tears involved 15 to 30 mm (mean, 25.7 mm) of the subscapularis
footprint from the lesser tuberosity in the superior-to-inferior direction.
The full-thickness tears of the subscapularis were classified, according to
their size and amount of tendon retraction, as grade 1 when the superior
one-third of the attachment (the superior one-half of the tendon portion) was
torn (five tears), as grade 2 when the superior two-thirds of the attachment
(the entire tendon portion) was involved (fifteen), and as grade 3 when all of
the attachment (the tendon and muscular portions) was torn (no tears received
this grade in our series). Tendon retraction was classified as minimal when
the edge of the tear was over the lateral articular cartilage margin in
neutral rotation (two tendons), as moderate when it was in an area lateral to
the glenoid surface (eight), and as severe when it was on the glenoid surface
level or in an area medial to the glenoid surface (ten).
The tears of the supraspinatus and infraspinatus involved 10 to 30 mm
(mean, 19.3 mm) of the footprint on the greater tuberosity in the
anterior-to-posterior direction. The tear pattern was classified as involving
either the subscapularis and supraspinatus only (thirteen patients) or the
subscapularis, supraspinatus, and infraspinatus (seven). The long head of the
biceps tendon was completely torn in five patients, was dislocated or
subluxated in six patients, and was partially torn in three. Three patients
presented with a combination of a subluxation and a partial tear of the long
head of the biceps tendon. Only three patients had a normal long head of the
biceps tendon.
Arthroscopic Procedure
For the subscapularis repair, the anterosuperior portal was mainly used as
the viewing portal; the other portals were the working portals. When
necessary, an additional accessory anterior portal was created with use of a
30°-angle arthroscope under spinal needle guidance. Seven lesions of the
long head of the biceps tendon were addressed with arthroscopic tenodesis and
use of suture anchors, and five were treated with a tenotomy.
As viewed from the anterosuperior portal, the attachment of the inferior
muscular portion of the subscapularis to the humerus was detected as a guide
with which to find the torn subscapularis tendon. The inferior muscular
portion of the subscapularis was retained in all patients. The torn
subscapularis tendon required mobilization in fifteen patients with a grade-2
tear and moderate or severe tendon retraction. Through the anterior portal,
the medial and lateral aspects of the inferior portion of the torn
subscapularis were released from adhesions to the glenohumeral joint capsule,
the deltoid fascia, or the conjoined tendon with use of a radiofrequency
device (VAPR; Mitek, Westwood, Massachusetts). To avoid neurovascular injury,
release of the medial-anterior aspect of the inferior portion of the torn
subscapularis was halted when the axillary and/or subscapularis nerve was
identified. The release was extended to the superior portion of the torn
subscapularis tendon. With a radiofrequency device, the tip of the coracoid
process was prepared to release the anterior aspect of the superior portion of
the torn subscapularis. Also with use of this device, the coracohumeral
ligament was released at the base of the coracoid process.
In six patients, we performed a coracoplasty—i.e., resection of the
lateral tip and posterior aspect of the coracoid process—with use of a
high-speed burr to create adequate space for the subscapularis tendon. The
edge of the torn subscapularis and the soft tissue on the lesser tuberosity
were débrided through the anterior portal with a shaver. The mobility
of the torn tendon and the extent of lateral translation were tested with a
grasper introduced through the anterior portal.
The bone on the lesser tuberosity was lightly decorticated with a
high-speed burr introduced through the anterior portal. A titanium-screw type
of suture anchor (Corkscrew, Arthrex; Fastin RC, Mitek; or Super Revo,
Linvatec, Largo, Florida) with two number-2 nonabsorbable sutures per anchor
was inserted into the lesser tuberosity through the anterior or accessory
anterior portal. Two suture anchors were used in the fifteen patients with a
grade-2 tear; one was placed superiorly and the other was placed inferiorly on
the subscapularis footprint of the lesser tuberosity. One suture anchor was
used in the five patients with a grade-1 tear.
Three techniques were used to pass the suture of the anchor through the
torn tendon. With the first technique, a suture hook loaded on a shuttle-relay
(Linvatec) was passed into the torn tendon through the anterior portal, and
the shuttle-relay was then retrieved through the posterior portal. The suture
of the anchor was likewise retrieved through the posterior portal. The suture
was fed in the shuttle-relay eyelet and moved back through the torn tendon
from the anterior portal (Figs. 1-A, 1-B,
1-C, 1-D, 1-E,
1-F). With the second
technique, a bird-beak penetrator (Arthrex) or clever hook (T.A.G. Medical
Products; Kibbutz Gaaton, Israel) was inserted into the torn subscapularis
from the anterior portal, and the suture was grasped and retrieved directly
from the anterior portal. With the third technique, sutures of the anchor were
passed directly with a flexible suture passer (Mitek). After all sutures were
passed through the torn subscapularis with use of these techniques, an
arthroscopic knot was created on a sliding or nonsliding knot followed by
three alternating post half-hitches to produce a simple stitch. Looking
through the anterosuperior portal, the surgeon created the knots of the
inferior anchor and then, looking through the lateral subacromial portal,
created the knots of the superior anchor.
After completion of the subscapularis repair, subacromial smoothing was
performed, with preservation of the coracoacromial ligament. This was followed
by arthroscopic repair of the supraspinatus tear (all twenty patients) and,
when appropriate, the infraspinatus tear (seven patients) with single-row
fixation as previously described by
us21. The final
evaluation of the construct was performed both subacromially and
intra-articularly.
Postoperative Management
The arm was supported in a sling with an abduction pillow for six weeks.
Passive-motion exercises performed by a physiotherapist and exercises
performed by a continuous-passive-motion
machine17,21
were started on the first postoperative day. Passive external rotation with
the arm at the side was limited to 0° and passive flexion or abduction was
limited to 90° for the first four weeks. Active-assisted range-of-motion
exercises with use of a stick were started four to six weeks after the
surgery, and overhead active motion was initiated at six weeks. Shoulder
strengthening
exercises32 were
begun nine to twelve weeks postoperatively. Rehabilitation continued for three
to six months. Strenuous manual work and overhead activities were allowed
after good shoulder strength had been restored.
Statistical Analysis
The Wilcoxon signed-rank test was used for matched pairs. Differences in
values between two unpaired groups were analyzed with the Mann-Whitney U test.
Multivariate analysis was done with the chi-square test. A p value of <0.05
was considered significant. With the level of alpha set at 0.05, post hoc
sample-size calculations showed that power ranged from 63% to 99%.
Shoulder Rating Scales
Table I shows the
preoperative and postoperative shoulder scores of the twenty patients. After
the procedure, the mean UCLA and JOA scores improved significantly (p <
0.0001). The patients had significantly higher scores for the pain, function,
range-of-motion, abduction strength, elevation, external rotation, and
internal rotation components of the JOA rating scale. The outcome was
excellent for thirteen patients (65%), good for five (25%), fair for one (5%),
and poor for one (5%).
Range of Motion
Postoperatively, the mean active elevation was 162° (range, 100° to
175°) compared with 97° (range, 30° to 170°) preoperatively (p
< 0.0001), the mean active external rotation was 43° (range, 20° to
60°) compared with 30° (range, —15° to 60°) (p =
0.0115), and the mean active internal rotation was to T10 (range, T9 to T12)
compared with L2 (range, T10 to the gluteal level) (p = 0.0007).
Lift-Off and Belly-Press Tests
Preoperatively, thirteen patients had a positive result and one had a
negative result of the lift-off test; six patients could not be tested because
they had a restricted range of internal rotation. The belly-press test was
positive for nineteen patients and negative for one. Postoperatively, six
patients had a positive lift-off test and fourteen had a negative lift-off
test; six had a positive belly-press test and fourteen had a negative
belly-press test.
Radiographic Evaluation
No suture anchor failures were detected on radiographs made at the latest
follow-up assessment. The mean minimal coracohumeral interval on the
preoperative magnetic resonance imaging scans was 9.0 mm (range, 3 to 15 mm);
the interval was <6 mm in six patients who had undergone a coracoplasty
(Figs. 2-A and 2-B).
Preoperative magnetic resonance imaging showed subscapularis
(Fig. 3-A) and supraspinatus
tears in all twenty patients and an infraspinatus tear in seven patients.
The repair was intact in thirteen patients and had failed in seven patients
as seen on postoperative magnetic resonance imaging. Fourteen of the twenty
patients were observed to have structural integrity of the subscapularis
(Fig. 3-B); thirteen of the
twenty, to have structural integrity of the supraspinatus
(Fig. 3-C); and three of seven,
to have structural integrity of the infraspinatus.
Analysis of Repair Integrity and Clinical Results
Seven patients, six men and one woman, had a failed repair. Preoperatively,
four of these seven patients had had a three-tendon tear; three of the four
had postoperative recurrence of the tears of all three tendons, and one
patient had a healed subscapularis and recurrent tears of the supraspinatus
and infraspinatus. The other three patients had had a two-tendon tear before
the procedure and had recurrent tears of both tendons postoperatively. The
patients who had an intact repair are compared with those who had a failed
repair in Table II. The mean
age of the patients with a failed repair (68.4 years) was significantly higher
than that of the patients with an intact repair (58.1 years) (p = 0.014).
Preoperatively, the mean JOA shoulder scores did not differ significantly
between the two groups. Postoperatively, they improved significantly in both
groups (p < 0.0001); however, they were significantly lower in the patients
with a failed repair (p = 0.0034). The score for the function component of the
JOA score was significantly higher for the patients with an intact repair (p =
0.0324), but the scores for pain and the range of motion did not differ
significantly with the small numbers reviewed for this study (p = 0.0961 and
0.1322, respectively). Of the thirteen patients with an intact repair, twelve
had an excellent result and one had a poor result. Of the seven patients with
a failed repair, one with a healed subscapularis had an excellent result, five
had a good result, and one had a fair result. All six patients with a
recurrent tear of the subscapularis had positive postoperative lift-off and
belly-press tests.
Analysis of Repair Integrity and Tear Configuration
Table III shows the number
of patients and failed repairs categorized according to the tear size, tendon
retraction, and tear pattern on arthroscopic evaluation. No patient had a
grade-3 tear. The prevalence of recurrent tears was significantly higher in
patients with severe preoperative tendon retraction (six of ten) than it was
in those with minimal or moderate preoperative tendon retraction (one of ten)
(p = 0.0191). With the numbers in this study, the tear pattern and size were
not found to be significantly associated with tear recurrence (p = 0.1276 and
0.4168, respectively).
Unsatisfactory Results
Two patients had a fair or poor result. The poor result was in a
fifty-year-old woman with a three-tendon tear on the dominant side
preoperatively. At the index operation, she had a grade-1 tear and moderate
tendon retraction; the long head of the biceps tendon was normal. She
experienced tenderness over the bicipital groove and had positive impingement
and Speed tests and negative lift-off and belly-press tests ten months after
the surgery. The pain was transiently reduced after a local anesthetic
injection into the bicipital groove. Postoperatively, elevation was 150°,
external rotation was 40°, and internal rotation was to T10. A magnetic
resonance imaging scan made twelve months after the procedure showed healing
of the three repaired tendons, and the patient elected to undergo a second
operation fourteen months after the first. Second-look arthroscopy
demonstrated healing of the three repaired tendons and a partial tear of the
long head of the biceps tendon, and the patient underwent arthroscopic
tenotomy of the long head of the biceps tendon. She had pain relief after the
second operation.
The result was fair in a seventy-six-year-old man. Preoperatively, he had
had a three-tendon tear on the nondominant side. At the time of surgery, he
had a grade-2 tear with severe tendon retraction. Postoperatively, he
experienced shoulder pain, and magnetic resonance imaging scans made at ten
months after the surgery showed recurrent tears of the three tendons. He did
not elect to undergo a second operation. He had positive lift-off and
belly-press tests, elevation to 100°, external rotation to 20°,
internal rotation to L2, and grade-3 (fair) abduction strength as assessed
with the manual muscle test fifty months after the surgery.
Complications
At the time of the latest follow-up, none of the patients manifested
permanent neurological compromise. However, one patient experienced transient
dysfunction of the anterior interosseous nerve that resolved two months after
the surgery without specific treatment. We postulated that this complication
may have been due to arm traction in the lateral position during surgery.
There were no infections related to the arthroscopic procedure. No hardware
complications were identified, and no patient required manipulation because of
postoperative stiffness.
To our knowledge, this study represents the first prospective evaluation,
with use of postoperative magnetic resonance imaging, of arthroscopic repairs
of anterosuperior rotator cuff tears. In our previous study of arthroscopic
repair of rotator cuff tears, from which subscapularis tears were excluded,
the postoperative mean shoulder scores in patients with a large-to-massive
tear were significantly lower than those in patients with a small-to-medium
tear17. The
outcomes of the arthroscopic repairs of anterosuperior rotator cuff tears in
the present study were comparable with those of repairs of large-to-massive
posterosuperior rotator cuff tears in the previous study. However, direct
comparison of our results with those of previous studies on open or
arthroscopic repair of combined rotator cuff tears that included the
subscapularis is difficult because various rating systems were used. Bigliani
et al.2 reported
excellent results in eleven patients with combined subscapularis and
supraspinatus tears, all of which were repaired through an anterosuperior
approach. Warner et
al.1, who used open
repair, reported excellent or good results in only eight of nineteen patients
with combined subscapularis, supraspinatus, and infraspinatus tears. Their
series included patients with stage-3 or 4 fatty infiltration of the torn
rotator cuff muscles as well as those with failed prior surgery. Warner et
al.1 stated that
repair earlier than six months after the injury was associated with a better
functional outcome because of less involution of muscle and tendon tissue. As
all patients in our study underwent surgery six months or less after the
injury and no patient had stage-3 or 4 fatty infiltration or a failed prior
rotator cuff repair, it is possible that the quality of the torn rotator cuff,
including the subscapularis, was better in our patients. This may explain why
we demonstrated better results. Kreuz et
al.9 reported that
fifteen of eighteen patients with open repair of combined subscapularis and
supraspinatus tears subjectively rated the outcome as excellent or good. In
their preliminary report, Burkhartand
Tehrany6 reported
excellent or good results in twenty-three of twenty-five patients who had
undergone arthroscopic repair of an isolated subscapularis tear (eight) or a
subscapularis tear accompanied by supraspinatus and infraspinatus tears
(seventeen).
Bennett8 found an
improvement in shoulder scores after arthroscopic repair of combined
subscapularis and supraspinatus tears in nineteen patients. While the clinical
results in those earlier
reports6,8,9
are similar to ours, the authors did not assess repair integrity at the time
of follow-up.
In our study, the repair was intact in thirteen of the twenty patients at
the time of follow-up, and repair integrity correlated significantly with the
clinical results. The arthroscopic repair was intact in ten of the thirteen
patients with a two-tendon tear and in three of the seven with a three-tendon
repair. Although all seven patients with a failed repair had improved shoulder
scores postoperatively, their scores were significantly lower than those for
the patients with an intact repair. The mean total shoulder score and score
for the function component of the JOA rating scale were significantly higher
for the patients with an intact repair than they were for the patients with
recurrent tears, although the scores for pain and range of motion did not
differ significantly between the groups. Flury et
al.10 reported that
an intact repair was found with ultrasonography in twenty-eight (88%) of
thirty-two patients who had undergone an open repair of combined subscapularis
and supraspinatus tears through a deltopectoral approach. Harryman et
al.33 examined 105
open rotator cuff repairs ultrasonographically to correlate functional results
with the integrity of the cuff. Their overall rate of cuff integrity was 65%,
with 80% of one-tendon repairs, 57% of two-tendon repairs, and 32% of
three-tendon repairs intact at the time of follow-up. The integrity of the
cuff correlated with the active range of motion, strength, and function as
well as with patient comfort and satisfaction. Gerber et
al.34 inspected
follow-up magnetic resonance imaging scans of twenty-nine patients who had
undergone open repair of a massive rotator cuff tear (two tendons or more).
They found 66% of the repairs to be intact and reported that patients with an
intact repair had better results than did patients with a failed repair. The
repair integrity in our series was comparable with that reported in those
previous
studies10,33,34.
Others35-38
have reported findings similar to ours—i.e., that despite structural
failure of a rotator cuff repair, patients can manifest significant clinical
improvement.
Our study clarified that older age and more severe tendon retraction are
negatively correlated with an intact repair as documented by magnetic
resonance imaging. The mean age of the patients with a failed repair was ten
years older than that of the patients with an intact repair (68.4 years and
58.1 years, respectively). Harryman et
al.33 reported that
older patients and patients in whom a larger tear had been repaired had a
greater prevalence of recurrent tears. In their study of structural integrity
after arthroscopic repair of the supraspinatus tendon, Boileau et
al.39 observed that
factors that were negatively associated with tendon healing were increased age
and associated delamination of the subscapularis or infraspinatus tendon.
Osteoporotic bone found in elderly patients may be negatively associated with
bone-to-tendon
healing40. In our
study, the prevalence of recurrent tears in patients with severe tendon
retraction was significantly higher than that in patients with minimal or
moderate tendon retraction. In contrast, Boileau et al. reported that the rate
of tendon healing was not significantly different between shoulders with a
minimally retracted tear and those with a more retracted
tear39. We agree
with Gerber et
al.41, who stated
that rotator cuff tears lead to substantial and progressive muscular changes
with a severity that is proportional to the amount of musculotendinous
retraction. If muscular function is to be preserved, it may be necessary to
perform a repair before marked retraction has occurred, or new or different
repair methods need to be developed to improve the biological healing of the
rotator cuff41.
Arthroscopic repair of subscapularis tears combined with another tear or
other tears of the rotator cuff has several advantages over open repair. Open
repair requires not only a deltopectoral approach but also an extended
superior approach to mobilize and repair the supraspinatus and infraspinatus
tendons1. Like open
repair, arthroscopic repair requires caution to avoid neurovascular injury.
With the arthroscopic procedure, it is possible to identify the axillary or
subscapular nerve on release of the posterior and medial aspects of the
inferior portion of the subscapularis. In contrast, a drawback of arthroscopic
repair may be its relative technical difficulty because the anterior deltoid
tends to drape itself tightly across the footprint of the subscapularis tendon
on the lesser tuberosity. This renders the working space relatively tight
during the procedure.
In our experience, the anterosuperior portal is useful as a viewing portal
in the arthroscopic repair of a subscapularis tear. The important first step
is to find the torn subscapularis tendon, which generally adheres to
surrounding tissue. As viewed from the anterosuperior portal, the attachment
of the inferior muscular portion of the torn subscapularis to the humerus can
usually be identified as a residual muscle even in patients with a retracted
tendon. At the insertion of the subscapularis to the humerus, the superior
two-thirds is the tendinous portion and the inferior third is the muscular
portion42-45.
Of our twenty patients, only six manifested a narrow coracohumeral interval
even after the preparation of the coracoid process and required coracoplasty
to create an adequate space for the subscapularis tendon. Our indication for
coracoplasty is a narrow coracohumeral interval (<6 mm). Preoperative
magnetic resonance imaging can provide useful information on the coracohumeral
interval (Figs. 2-A and
2-B).
The present study has several possible weaknesses. There was no control
group of patients who had undergone an open repair, and the number of treated
patients was relatively small. Also, bias in the selection of patients may
have influenced our results. We excluded patients who had filed a Workers'
Compensation claim, as various secondary issues may have affected the
treatment
outcome46. On the
other hand, our investigation was prospective, the arthroscopic procedure was
performed by one surgeon at one institution, the evaluator was not the surgeon
who performed the operations, repair integrity was evaluated with
postoperative magnetic resonance imaging scans, and the follow-up period
exceeded two years.
In conclusion, arthroscopic repair of combined rotator cuff tears that
include the subscapularis is a safe and effective procedure with respect to
the amelioration of shoulder pain and the improvement of function and the
range of motion in selected patients. Our data demonstrate that repair
integrity correlates significantly with clinical results and that patient age
and the degree of tendon retraction are significantly associated with repair
integrity. ?