Operative techniques have advanced sufficiently so that surgeons
can repair anterior-inferior glenohumeral instability arthroscopically.
The proposed advantages of arthroscopic stabilization include smaller
skin incisions, more complete inspection of the glenohumeral joint,
the ability to treat intra-articular lesions, access to all areas
of the glenohumeral joint for repair, less soft-tissue dissection, and
maximum preservation of external rotation33,34,49.
Early arthroscopic repairs were performed with use of a staple to
advance Bankart lesions superiorly and medially and were associated
with failure rates of up to 50 percent21,29.
Because of complications related to the placement of staples within
the glenohumeral joint60, later
investigators reported on suture repair of Bankart lesions49,58. The essential element of these
suture techniques was the passage of sutures through the avulsed
labrum and then through drill-holes in the scapular neck, where
the sutures were tied over soft tissue or bone36.
The initial success rate in twenty-five shoulders was 100 percent,
but these results deteriorated with a longer duration of follow-up36.
Later studies documented two flaws with this approach: the location
of the repaired labrum and the failure to address capsular laxity.
Neviaser41 was the first, as far
as we know, to report on the anterior labroligamentous periosteal
sleeve avulsion (ALPSA) lesion. He identified this lesion in shoulders
with anterior-inferior glenohumeral instability. The labrum-ligament
complex had detached and healed medially on the scapular neck and
allowed excessive humeral translation. It was apparent that the
staple and suture techniques, described above, repaired the labrum
medially and created an anterior labroligamentous periosteal sleeve
avulsion lesion. Caspari and Savoie10 arthroscopically
examined shoulders that had had a failed arthroscopic stabilization
(with the labrum repaired five millimeters medial to the glenoid rim),
and they were the first, to our knowledge, to point out that the
attachment of the repaired ligaments was critical. Their technique
was modified to move the entry position from the medial aspect of
the scapular neck to the glenoid articular surface. Savoie et al.49 reported improved results with the
newer technique. Wolf et al.58 pioneered
the use of bone-suture anchors in the arthroscopic treatment of
instability. Those investigators were able to repair the detached
labrum directly to the glenoid rim. Further investigation, however,
called into question the concept of whether the Bankart lesion was
solely responsible for anterior-inferior instability. Rodosky et
al.46 investigated the role of
the biceps-labrum complex and found that detachment of the superior
aspect of the labrum (a tear of the superior part of the labrum
from anterior to posterior [a SLAP lesion]) allowed increased anterior
translation of the humeral head. Speer et al.51,
in a cadaveric study, found that while a Bankart lesion allowed
increased translation of the humeral head it did not allow the humeral
head to dislocate. Those authors suggested that capsular stretch
or elongation may also be necessary for dislocation. Bigliani et
al.6 studied the tensile properties
of the shoulder capsule in patients with an acute dislocation and
noted that capsular damage was almost always present to some extent,
even in the face of a Bankart lesion. Baker et al.5 arthroscopically inspected the shoulders
of forty-five patients within ten days after an acute dislocation
and found that the capsule had been stretched or torn in all patients
with or without an associated Bankart lesion. This information was
elegantly summarized by Altchek et al.3.
Previous arthroscopic techniques have not included treatment
of the rotator interval. The role of the rotator interval in shoulder
instability was described by Neer and Foster39 and
by Rowe and Zarins48. The latter
investigators reported that a large opening between the supraspinatus
and the subscapularis was found in twenty of thirty-seven shoulders
in which the superior aspect of the musculotendinous cuff was explored.
We believe that the high failure rates in prior studies of arthroscopic
repair were due to technical factors (medial repair of the anterior
portion of the labrum) and also to the failure to treat other lesions that
are responsible for glenohumeral instability. The purpose of the
present study was to determine if the results of arthroscopic treatment
of instability could be improved with use of a technique in which
all of the components of anterior-inferior glenohumeral instability
are evaluated and repaired. We repaired lesions of the anterior,
inferior, and superior portions of the labrum; restored capsular
tension with a combination of sutures and thermal capsulorrhaphy;
and repaired the rotator interval, if necessary.
We report the results of a prospective study of fifty-three patients
who were managed with arthroscopic treatment of glenohumeral instability between
January 1994 and February 1997.
Study Hypotheses
As the purpose of this study was to evaluate the final outcome
of arthroscopic treatment of glenohumeral instability, we decided,
on the basis of a review of the literature2,4,7,11,25,33,47,48,
that a final mean score of more than 85 points according to the
system of Rowe et al.47 was a
satisfactory result. We made this decision because either functional
limitation (in work or sports) or discomfort when the arm is examined for
apprehension leads to the deduction of 15 points, resulting in a
maximum possible score of 85 points. We believed that the presence
of either functional limitation or discomfort on apprehension-testing
was unacceptable. Therefore, the null (primary) hypothesis (H0)
of this study was that arthroscopic repair of anterior-inferior
glenohumeral instability results in a final mean Rowe score of 85
points or less (H0: £ 85). The alternative hypothesis (Ha) was
that the final mean Rowe score would be greater than 85 points (Ha:
> 85). The alpha level was set at 0.05. The effect size was selected
on the basis of a literature review33,34,49.
We estimated that a mean score that was greater than 10 points from
= 85 would result in an unfavorable outcome. On the basis of this,
an effect size of 0.60 was estimated. A power analysis was then
performed to quantify the number of patients required to reject
our null hypothesis if, in fact, it were to be rejected. With use
of an effect size of 0.60, with a one-tailed or unidirectional analysis
at a desired power of 0.95, it was determined that approximately
thirty-five patients were required12.
Inclusion and Exclusion Criteria
The inclusion criterion was a preoperative diagnosis of anterior-inferior
glenohumeral instability that was confirmed at the time of the arthroscopic operation.
The diagnosis was made on the basis of a combination of signs and
symptoms: (1) the patient's description of shoulder dislocation
or a sensation of looseness and slipping, (2) pain or apprehension,
or both, on anterior-inferior instability tests, (3) radiographic
evidence of glenohumeral instability, and (4) findings during the arthroscopic
operation that documented anterior-inferior glenohumeral instability.
We operated on patients who had a clinical diagnosis of anterior-inferior
glenohumeral instability and made no attempt to exclude patients
who had particular lesions of the shoulder. Each shoulder was evaluated for
lesions at the time of the operation, and specific approaches were
selected to treat the different lesions that were encountered.
The exclusion criteria included multidirectional instability
(thirty-eight patients), posterior instability (twelve patients),
a prior operation for instability (twenty-six patients), and repair
of a full-thickness rotator-cuff tear (two patients). Patients with
multidirectional instability demonstrated painful, excessive translation
of the humeral head anteriorly, inferiorly, and posteriorly on physical examination.
Patients with posterior instability demonstrated painful, excessive
translation posteriorly and often inferiorly but not anteriorly.
Thirty-one patients who were receiving Workers' Compensation also
were excluded because of various issues that adversely affected
the outcome. Misamore et al.35 and
other investigators18,26 have
documented inferior results following shoulder operations in this
population.
Sixty-four patients met the criteria for inclusion in the study.
Four patients declined to participate because they did not think
that they could attend the number of clinic visits required for
the minimum duration of follow-up. Seven patients did not return
for the final follow-up evaluation. However, the demographic data
and operative findings demonstrated that they were representative
of the entire study-group population. Four of them did not return
for the one-year follow-up evaluation. The three other patients
were examined one year after the operation, but they did not return
for the minimum two-year follow-up evaluation. Therefore, the results
of fifty-three shoulder operations were included in our analysis.
The only exceptions to the requirement for a two-year-minimum follow-up
interval were patients who were considered to have had a failure
of the index operation but who did not return for the two-year evaluation.
These patients demonstrated postoperative subluxation or dislocation
and were included irrespective of the time to failure. We believed
that exclusion of these patients would have resulted in selection
bias and would have distorted our results.
Study Group
The study group consisted of forty-four male and nine female
patients. The mean age at the time of the operation was thirty-two
years (range, fifteen to fifty-eight years). The dominant shoulder
was involved in twenty-five patients and the nondominant shoulder,
in twenty-eight.
The preponderance of our patients participated in sports at a
recreational level; specifically, thirty-five patients described
their level of sports activity as recreational, seven participated
in high-school sports (four played football, two played basketball,
and one played softball), one participated in college sports (as
a football defensive lineman), and ten did not participate in sports.
No professional or semiprofessional athletes were included in our
study, and none of us were the designated physician for any high-school,
college, or professional sports team.
The patients were recruited from the private practice of a single
orthopaedic surgeon who performed all of the operations at a private
hospital. We gathered no information about annual income or ethnicity,
but all of the patients had private health insurance.
Preoperative Assessment
We collected sufficient data to rate the shoulders according
to the American Shoulder and Elbow Surgeons (ASES) Shoulder Index44, the scoring system of Constant
and Murley13, the scoring system
of Rowe et al.47, and the University
of California at Los Angeles (UCLA) Shoulder Score15. The system of Constant and Murley
typically is used to evaluate patients with rotator cuff lesions or
osteoarthritis, but it was included in this study in order to allow
our results to be compared with those of future studies as well
as with those for patients who have other shoulder conditions.
Prior to the operation, all patients completed a self-assessment
questionnaire to document their levels of satisfaction and function.
The patients recorded their pain level with use of a visual analog scale
as well as with use of a separate ordinal scale in order to allow
us to rate the shoulder according to the ASES Shoulder Index and
the UCLA Shoulder Score. Patients who had sustained the injury as a
result of a specific traumatic event were questioned about the mechanism
of injury and were asked to recall, if possible, the position of
the arm at the time of the traumatic event. The amount of pain at
the time of the initial injury and during the recovery period was
noted. All patients were questioned about the position of the arm
or the activity that reproduced the symptoms. In order to define the
sports activity accurately, we classified sports according to the
method described by Allain et al.1.
Type 1 indicated nonimpact sports, which consisted of swimming the
breaststroke, rowing, running, or sailing; type 2, high-impact sports, including
riding a bicycle, downhill skiing, soccer, or waterskiing; type
3, sports that required overhead use of the arm with hitting movements,
such as swimming the crawl or butterfly stroke, golf, tennis, throwing,
and weight lifting; and type 4, sports that involve overhead hitting
movements and sudden stops, such as basketball, handball, ice hockey,
judo, karate, kayaking, lacrosse, polo, rodeo, volleyball, windsurfing,
and wrestling. The level of sports participation was categorized
as high-school team sport (level 1), college team sport (level 2),
or recreational (level 3).
We measured active range of motion, which included forward elevation,
abduction, external rotation in abduction, and behind-the-back internal rotation,
according to the Constant and Murley rating system13. Passive elevation and external
rotation (with the arm adducted), as well as external rotation and
internal rotation with the arm abducted 90 degrees, were measured
with use of a handheld goniometer and recorded by the examiner to
the nearest 5 degrees. The operating surgeon recorded all measurements
during the initial physical examination and at subsequent clinic
visits. We made no attempt to increase the precision of the measurements
by employing techniques such as having the evaluation performed
by a blinded examiner, evaluating test-retest validity, measuring
interobserver and intraobserver reliability, instructing the patient
before the measurement, or asking the patient to perform warm-up
exercises before the evaluation. The strength of elevation was measured
with use of a dynamometer with the arm elevated 90 degrees in the
scapular plane and internally rotated; the result was recorded in
pounds. We did not inject an anesthetic into the shoulder before strength-testing,
so we were unable to quantify how much of the loss of strength was
due to pain.
Both shoulders were examined for stability. We compressed (loaded)
the humeral head into the glenoid during all maneuvers. Anterior
translation was assessed by applying an anterior force to the shoulder
with the arm in 90 degrees of abduction. Inferior-anterior translation
was evaluated with use of the Rowe test16.
For this examination, the patient stands and flexes the trunk from
the hips approximately 30 degrees. In this relaxed position, the
shoulders are effectively elevated 30 degrees. The examiner then
applies a distraction force to the shoulder. Inferior translation
was assessed by applying an inferior force to the shoulder with
the arm in 0 degrees of abduction (the sulcus test). Posterior translation
was examined with the arm elevated 90 degrees, adducted slightly,
and internally rotated approximately 30 degrees. A posterior force
was applied, and then the shoulder was extended. We recorded the
presence or absence of pain and apprehension during each instability
maneuver. We graded the amount of humeral head translation on the
glenoid surface as 0 (stable or trace laxity), 1 (up to 50 percent
translation), 2 (more than 50 percent translation but not dislocatable),
and 3 (dislocatable)14. The grading
of instability was subjective, as we made no attempt to measure
the degree of translation with fluoroscopic observation. We recorded the
presence of laxity in the contralateral shoulder and the elbows
and assessed the ability of the patient to bring the thumb to the
forearm. We did not utilize any formal system to grade the degree
of generalized ligamentous laxity; laxity was categorized simply
as present or absent on the basis of this examination. We excluded
other sources of shoulder pain (rotator cuff lesions, acromioclavicular
joint arthritis, thoracic outlet syndrome, brachial plexus lesions,
and glenohumeral arthritis) on the basis of the patient's history,
physical examination, and radiographic analysis.
Anteroposterior glenoid, axillary, and supraspinatus outlet40 radiographs were made routinely.
Imaging studies such as magnetic resonance imaging, computer-assisted
tomography, and arthrography were not routinely performed. Direct
radiographic evidence of glenohumeral instability consisted of a
finding of a dislocated humeral head. Indirect radiographic evidence
of instability included calcification adjacent to the anterior portion
of the glenoid, an osseous Bankart lesion, or a Hill-Sachs lesion.
Evidence of instability on magnetic resonance imaging and computer-assisted
tomography studies (ordered by the referring physician) included
not only the above findings but also detachment of the glenoid labrum
from the glenoid bone, capsular stripping from the glenoid, or ligamentous
insufficiency.
The primary indication for the operation was persistent shoulder
pain due to anterior-inferior glenohumeral instability that had
not responded to a program of at least six months of nonoperative treatment,
consisting of avoidance of painful activities, use of nonsteroidal
anti-inflammatory medication, and participation in a home physical therapy
program designed to maintain or improve the strength in the shoulder
girdle. Our goal was to improve the strength of the muscles responsible for
glenohumeral stability. Therefore, patients performed resistive
exercises for the deltoid, internal rotators, external rotators,
biceps, triceps, and scapular muscles with use of surgical tubing
and light weights (maximum, five pounds [2.3 kilograms])9. The only exceptions were the three
patients who desired operative repair acutely (within six weeks after
the onset of symptoms). The patients were evaluated at five postoperative
intervals within the first year (at two weeks, six weeks, three
months, six months, and one year) and then yearly thereafter.
Classification of Instability
In order to increase diagnostic precision, we classified each
shoulder with regard to the chronicity, degree, and etiology of
the instability. The chronicity of the instability was classified,
according to the patient's description, as chronic (present for six
weeks or more) or acute (present for less than six weeks). The degree
of instability was classified as recurrent dislocation, recurrent
subluxation after a single dislocation, or recurrent subluxation without
prior dislocation. The etiology of the instability was classified
as either traumatic (if the instability had developed after a traumatic
event of sufficient magnitude to damage the glenohumeral ligaments)
or atraumatic. The guidelines that were used to determine the etiology
were similar to those described by Wirth et al.57.
A traumatic etiology was suggested when the injury had occurred
with the arm forcefully abducted and externally rotated and extended
and there had been sudden, sharp pain, the need for manipulative reduction,
and residual aching in the shoulder for several weeks. Atraumatic
instability was characterized by either an insidious onset or the
development of symptoms after minor trauma and was associated with
mild pain and spontaneous reduction.
We also evaluated the shoulder with regard to the direction and
degree of instability. During the preoperative physical examination,
we noted the arm position that produced pain or apprehension and recorded
the amount of humeral head translation that occurred in association
with each provocative maneuver. These findings, combined with radiographic
findings and the patient's description of the position and the activity
that produced pain or apprehension, allowed us to categorize the
direction and degree of the instability. We observed movement of
the humeral head under direct arthroscopic visualization and also
found the location of intra-articular lesions to be helpful in the determination
of the predominant direction of instability. The lesions were located
in the humeral head and the glenoid (chondral or osteochondral defects),
the labrum (fraying or separation from the glenoid), and the capsular
ligaments (tearing or laxity).
Patient Compliance
We constructed a compliance scale to measure the completion of
the postoperative protocol. The components of the scale, which had
a possible maximum score of 100 points, were the appropriate use
of the sling (10 points), attendance at the five scheduled appointments
during the first postoperative year (10 points per visit, for a
total of 50 points), and performance of the prescribed rehabilitation
exercises (10 points per four phases of rehabilitation exercises,
as reported by the patient during office visits).
Operative Approach
Since our arthroscopic technique differs from those reported
in the literature, we describe our operative rationale, operative
technique, and intraoperative decision-making process in detail.
Operative Rationale
The underlying principle of our arthroscopic technique was to
identify and repair all lesions that contributed to glenohumeral
instability. Our technique included d衲idement, repair of ligamentous and
labral tears, capsular tensioning, and, if needed, repair of the
rotator interval.
The goals of d衲idement were to remove the sources of mechanical
irritation or functional instability43.
Only minor labral flap tears (those involving less than 50 percent
of the labral thickness) were removed, and every attempt was made
to repair the lesions.
The purpose of reattachment of the ligaments and the labrum to
bone was twofold. First, adequate capsular tension was impossible
to achieve unless the labrum and the ligaments were securely attached
to the glenoid. Therefore, we repaired all traumatic tears of the
superior, anterior, and inferior aspects of the labrum. We identified
these lesions at the time of the operation and believed that they
all contributed to glenohumeral instability. Second, anatomical
repair of the ligaments and the labrum restores concavity-compression
to the glenohumeral joint. Lippitt et al.31 demonstrated
that compression of the humeral head into the glenoid by muscular
force is an effective stabilizer to humeral translation and that resection
of the labrum decreases stability by 20 percent. However, reattachment
of the anterior-inferior aspect of the ligament-labrum complex to the
glenoid may not restore sufficient stability to the glenohumeral
joint. Speer et al.51 observed
only a small increase in humeral translation in association with
a simulated Bankart lesion and postulated that capsular stretching
or elongation was necessary to produce glenohumeral instability.
Therefore, the final portion of the operation was performed to restore
capsular tension.
We classified capsular elongation as primary or secondary. Primary
elongation refers to permanent deformation of the capsular fibers
due to a single traumatic event or to multiple episodes of instability.
The rate of speed of the injury may determine where the capsular
ligament is damaged. In a laboratory study, Bigliani et al.6 showed that a faster strain rate
was associated predominantly with ligamentous injuries, whereas testing
at a slower strain rate was associated with a higher percentage
of failures at the ligament insertion site. Secondary elongation
develops when there is a tear at the insertion site, thereby decreasing
capsular tension. This may occur within the anterior-inferior aspect
of the capsule after development of a Bankart lesion, but it may
also occur because of a tear of the superior part of the labrum. The
biceps-labrum complex contributes to anterior-inferior stability,
and its detachment results in increased humeral translation46. On the basis of this data, we believed
that we should repair all traumatic detachments of the superior
part of the labrum. Tears of the rotator interval and the superior
glenohumeral ligament also affect glenohumeral stability. In a laboratory study,
Harryman et al.24 found that tightening
of the rotator interval decreases inferior and posterior translation
of the humeral head. We observed, at the time of the operation,
that repair of the rotator interval decreased inferior translation
of the humeral head. If the repair also incorporated the superior
portion of the middle glenohumeral ligament, tension in the anterior
part of the capsule was increased. Thus, we restored capsular tension
with two methods: primary capsular elongation required an operation directly
on the capsule, and secondary elongation responded to repair of
tears at the insertion site.
We corrected primary capsular elongation with three techniques,
which were used singly or in combination: (1) advancement of the
capsule to the labrum, (2) advancement of the capsule to the glenoid
with suture anchors, and (3) thermal capsulorrhaphy. The goal of
this portion of the procedure was to restore ligamentous and capsular tension
and to eliminate excessive translation of the humeral head (defined
as translation involving more than 25 percent of the glenoid surface).The middle
glenohumeral ligament, the anteroinferior glenohumeral ligament,
the inferior aspect of the capsule, the posteroinferior glenohumeral
ligament, and the posterior aspect of the capsule were tightened
as necessary.
Our preference was to advance the capsule to the intact or repaired
labrum with use of monofilament sutures; only if the labrum was
small or absent was the capsule repaired to the glenoid rim with
bone-suture anchors. Drill-holes for the suture anchors were placed
through the glenoid articular surface, approximately one to two
millimeters from the lateral margin of the glenoid. The detached
labrum was sutured so that it was in contact with the scapular neck
and extended onto the glenoid articular surface in order to establish
the labrum as a bumper and to recreate optimal conditions for concavity-compression.
We estimated the appropriate amount of tightening on the basis of
both the degree and the direction of translation, according to guidelines
similar to those described by Warner et al.55 for
open repairs. A soft-tissue grasper was used to apply traction to
the various portions of the capsule while translation forces were
applied with the arm in varying degrees of abduction and external rotation.
We were technically unable to perform the repair with the arm in
complete abduction or external rotation, so we estimated the appropriate amount
of tension, returned the arm to 20 degrees of abduction and 30 degrees
of external rotation, and then completed the arthroscopic repair.
If areas of the capsule did not advance adequately, after suture
repair we used laser thermal application (Versalink Laser holmium:
YAG; Coherent Medical Group, Santa Clara, California) to contract
the capsule and the ligaments. We believe that it is not appropriate
to perform thermal capsulorrhaphy before suture repair for two reasons. First,
as the ligaments shorten with thermal capsulorrhaphy, they are more
difficult to repair to the glenoid rim. Second, we believe that
it is more difficult to determine the appropriate amount of ligamentous
advancement after thermal application. The use of thermal shrinkage
was limited to shoulders in which capsular suture tightening failed
to restore adequate soft-tissue tension to all areas of the capsule.
This was a common finding, and although thermal application was
not the primary operative technique during this study it was used for
the majority of patients.
We believed that one reason for the high rates of failure reported
by other investigators20,54 was
that they failed to address the quality of the ligaments. Although
suture tightening can increase ligamentous tension, it may, in certain
individuals, serve only to tighten attenuated soft tissue and does
not address plastic deformation at the cellular level. The findings
of laboratory studies have supported the concept that thermal contraction
accomplishes this goal27,32,42.
In our experience, the effectiveness of laser application has been
variable in that some ligaments have responded well while others
have shown minimal response. Although the long-term effectiveness
of thermal application has not been described in the literature
as far as we know, we thought that there was enough evidence in
laboratory studies17,23,53 to
justify its use. We also recognized that thermal capsulorrhaphy
added another variable to this study. Our goal, however, was not
to compare the efficacy of capsular suture repair with that of thermal
capsulorrhaphy but rather to determine the success of an arthroscopic
operation (comprising multiple techniques) to restore glenohumeral
instability.
Operative Technique
Prior to being placed under general anesthesia, all patients
received an interscalene block to diminish postoperative pain. The
anesthesiologist administered one gram of cephalosporin intravenously. We
placed the patient in the sitting position and examined both shoulders
as described above.
The shoulder joint was entered with a cannula and a blunt trocar
through a posterior skin incision placed 1.5 centimeters inferior
and 1.5 centimeters medial to the posterolateral border of the acromion.
An anterior portal was identified with a spinal needle so that the
cannula entered the shoulder joint immediately superior to the subscapularis tendon
and one centimeter lateral to the glenoid. The glenohumeral joint
was inspected. We reexamined the shoulder for translation while
viewing the shoulder through the arthroscope. The arthroscope was
removed and inserted through the anterior cannula to allow more
complete inspection of the posterior part of the glenohumeral joint.
The arthroscope was then returned to the posterior cannula. An arthroscopic
probe was used to assess labral attachment and ligamentous tension accurately.
All structures within the glenohumeral joint were examined systematically,
and any signs of instability were recorded. Such signs were variable
and included partial and complete tears of the rotator cuff, the
rotator interval, and the biceps tendon. We noted, as have others22,59, that the glenohumeral ligaments
may tear at their insertion on either the glenoid or the humeral
head. In order to evaluate the glenohumeral ligaments for mid-substance
tears or plastic deformation, we also assessed them for laxity by
directly observing and palpating them (with an arthroscopic probe) and
applying translational stresses as we rotated the shoulder. We documented
the location on the glenoid and the extent (superior to inferior
and medial to lateral) of labral detachment. Labra that were frayed
or had mid-substance tears were noted. The cartilage was inspected
for damage to the glenoid and the humeral head (a Hill-Sachs lesion).
The presence or absence of loose bodies was recorded. Once the diagnosis
of glenohumeral instability had been confirmed, an anterosuperior portal
was created and a second cannula was placed through the rotator
interval one centimeter lateral to the glenoid.
Intraoperative Decision-Making and Indications
D衲idement: We removed only minor labral flap
tears. Flap tears involving at least 50 percent of the labral thickness
were repaired with absorbable monofilament sutures. We found that
palpation of the labrum with a probe was necessary to adequately determine
the presence of minor flap tears, cleavage tears that existed within
the labral substance, and minor separations of the labrum from the
glenoid. Loose bodies were removed with surgical forceps.
Labral repair: The labrum normally is attached
securely to the glenoid bone anteriorly, inferiorly, and posteriorly,
distal to the glenoid equator, and we considered separations in
these areas as lesions. The anterosuperior aspect of the labrum
is usually not well attached to the glenoid, and separation in this
area was considered to represent a normal anatomical variant known
as a sublabral foramen56. The
attachment of the superior part of the labrum is variable30,52, and when the superior part of
the labrum was mobile without evidence of trauma it was not classified
as a tear of the superior part of the labrum from anterior to posterior
(a SLAP lesion)50. When the separation
of the superior part of the labrum is a normal variant, the superior
part of the glenoid is covered with smooth cartilage and the labrum
shows no evidence of trauma. Signs of traumatic separation include
tears within the substance of the superior part of the labrum, cartilage loss
with exposed bone at the site of labral attachment, and an increase
in separation of the superior part of the labrum with abduction
and external rotation of the arm8,37.
We repaired the superior part of the labrum anatomically. We made
no attempt to shift the anterior part of the labrum superiorly,
but we did, if necessary, shift the anterior part of the labrum
laterally so that it projected onto the glenoid surface and reestablished
the labrum as a bumper and an aid in concavity-compression. The
inferior part of the labrum was shifted superiorly onto the glenoid
surface for the same reasons.
Capsular tensioning: The ligament repair site
(and therefore the ligamentous tension) was estimated by grasping
the ligament and placing it at different locations on the glenoid.
Translation of the humeral head was performed with the torn ligament
positioned at each possible repair site until humeral head translation involved
less than 25 percent of the glenoid surface. Typically, five to
fifteen millimeters of lateral and superior ligamentous advancement
was required. The position of the arm affects tension within the
ligaments and the capsule, so we routinely maintained the shoulder
in 20 degrees of abduction and 30 degrees of external rotation during this
portion of the operation. We altered the arm position when we performed
the operation on the dominant arm of a patient who was a competitive throwing
athlete. In these patients, we determined the ligament repair site
after positioning the arm in 60 degrees of external rotation.
Rotator interval: If the shoulder demonstrated
persistent, excessive translation after d衲idement, labral repair,
and capsular tensioning, we turned our attention to the rotator
interval. If the direction of translation was inferior or inferior-posterior,
we passed a monofilament suture through the soft tissue immediately adjacent
to the anterior border of the supraspinatus and then through the
soft tissue superior to the subscapularis tendon. We placed the
suture as far laterally as possible so as not to interfere with postoperative
external rotation. Traction was applied on the suture, and we again
assessed the translation of the humeral head. If the correction was
adequate, the suture was tied. If the correction was inadequate,
the suture was removed and placed in a more medial position until
excessive translation was corrected. If the direction of persistent
translation was inferior-anterior, the inferior limb of the suture
was passed through the superior portion of the middle glenohumeral
ligament to increase tension in that portion of the capsule.
Thermal capsulorrhaphy: If persistent translation
was not corrected after labral and capsular suture repair and application
of tension on the rotator-interval suture, we performed a thermal
capsulorrhaphy. We used the laser to contract areas of the capsule
that corresponded to the direction of excessive movement of the
humeral head.
The repair sequence varied and depended on the specific combination
of lesions identified. In general, we followed a pattern of d衲idement,
ligamentous or labral reattachment, and capsular tensioning.
D衲idement was performed to smooth frayed labral fragments or
to remove torn labral fragments and, if necessary, to identify the
depth of partial-thickness rotator-cuff tears. Loose bodies were
removed.
We then treated labral or ligamentous insertion-site tears. The
labral tear or tears were then repaired, beginning with the inferior
part of the labrum and proceeding as necessary to the anterior and
superior aspects of the labrum. Technical considerations dictated
the order of labral repair. As the labrum (and its attached ligaments)
was repaired, our ability to displace the humeral head and to insert
bone-suture anchors and soft-tissue sutures was compromised. We
repaired the inferior part of the labrum first, as access to this
lesion became difficult after repair of the superior or anterior
aspect. We identified three types of anterior labral detachment.
Type A indicated that the labrum was separated from the glenoid
bone but remained at the level of the glenoid articular surface; type
B, that the labrum was separated and retracted medially; and type
C, that the labrum was retracted and had healed medially on the
glenoid (an ALPSA lesion41). Type-B
and type-C lesions required us to dissect the labrum from the glenoid
so that we could move the labrum laterally and place it on the glenoid
articular surface. We performed this procedure with a combination
of a thermal probe, a power burr, and blunt dissection. We repaired
forty-four tears of the anterior part of the labrum (twenty-five type-A,
fifteen type-B, and four type-C lesions) and two tears of the inferior
part of the labrum.
If the anterior or middle glenohumeral ligament was retracted
and adherent to the subscapularis, we released the ligament prior
to repair of the insertion site. We made an incision along the superior
border of the middle glenohumeral ligament. We then inserted a blunt
instrument (posterior to the capsule and anterior to the subscapularis
tendon) in order to separate the two structures. After we completed
labrum-and-ligament mobilization, the scapular neck was abraded
to a depth of one to two millimeters. The abraded area began at
the level of the glenoid cartilage and extended two centimeters
medially on the scapula. Drill-holes for the suture anchors were
placed through the glenoid articular surface approximately one to
two millimeters from the lateral glenoid margin. Drill-holes were
created in the anterior and inferior parts of the glenoid with a
power drill that was inserted through the anterior-inferior cannula.
The first suture anchor was inserted into the most inferior drill-hole.
The detached labrum was sutured so that it was in contact with the
scapular neck and extended onto the glenoid articular surface. The suture
strands were then tied. The number of suture anchors varied and
was dependent upon the size of the labral detachment.
After the inferior or anterior aspect of the labrum, or both,
had been repaired, the site of the labral tear from the superior
part of the glenoid bone was identified and abraded with a power
burr and two bone suture-anchors were inserted. The location of the
suture anchors varied and was dependent on the anatomy of the lesion,
but in general we placed one suture anchor one-third of the tear
length from the posterior margin and the second anchor one-third
of the tear length from the anterior margin of the tear. Nonabsorbable
number-2 braided suture was utilized5,9,37.
During the period of this study, a wide variety of suture anchors,
including metallic screw-in anchors, metallic expandable anchors,
and polyethylene expandable anchors, were used. We currently use
a metallic screw-in anchor (ROC 5; Orthopedic Biosystems, Scottsdale,
Arizona) exclusively.
We modified the repair technique when the labrum was intact but
the glenohumeral ligament had torn from the labrum. If the labrum
was of sufficient size to allow suture placement within its substance,
the ligament was repaired directly to the labrum with monofilament
suture. If the labrum was absent, the capsule was advanced onto
the glenoid articular cartilage surface and repaired with suture anchors
(as described above) in order to create a labral bumper. A mechanical
arm-holder greatly facilitated this step.
If the labrum-ligament complex was well attached to the glenoid
but the ligament lacked sufficient tension to contain the humeral
head, we operated directly upon the capsule with use of the methods described
above. The goal of this portion of the procedure was to restore
ligamentous and capsular tension and to eliminate excessive translation
of the humeral head.
Repair of the rotator interval was the last step performed within
the glenohumeral joint, as a cannula could not be inserted anteriorly
once this repair was completed. A suture-passer was used to place a
monofilament suture through the capsular tissue immediately anterior
to the supraspinatus tendon and then through the capsule superior
to the subscapularis tendon. If a greater degree of tightening was
required, then the superior capsular tissue was sutured to the middle
glenohumeral ligament. A sliding knot was used to tie the suture
extra-articularly. The details of this technique have been described
previously18.
If excessive translation remained after we had applied traction
to the rotator interval suture, the suture was not tied and we inserted
the laser. We identified the areas of the capsule that corresponded
to the direction of excessive translation (usually inferior or inferior-anterior).
Heat was applied over a very limited area of the capsule until the capsule
contracted. We always left normal capsule between any two areas
of laser application in order to avoid thermal necrosis. If further
capsular tension was necessary, we then tied the rotator interval
suture.
Postoperative Management
Postoperative management was similar for all patients. A soft
pillow-sling supported the arm in 15 degrees of abduction. An ice-pack
wrap decreased postoperative shoulder swelling and pain. One gram
of a cephalosporin was administered eight hours postoperatively.
Patients went home the morning after the operation. Active range-of-motion
exercises for the fingers, wrist, and elbow as well as isometric
exercises for the deltoid muscle were started the morning after
the operation and were continued at home for two weeks. At two weeks,
an anteroposterior radiograph was made to document the position
of any metallic suture anchors. The patients were allowed to remove
the sling for active elevation and external rotation exercises twice
daily but wore the sling at all other times. Active elevation was
limited to 120 degrees and external rotation, to 40 degrees. The
patients continued to wear the sling for six weeks, at which point
it was removed and unrestricted active range-of-motion and strengthening
exercises were begun. The range-of-motion and strengthening exercises
were continued for one year.
At each follow-up visit, after arrival at the clinic but before
the examination, each patient completed self-assessment forms to
document shoulder pain, function, satisfaction, and level of sports
activity. Active and passive ranges of motion as well as strength
were documented. No postoperative radiographic imaging studies such
as ultrasound, magnetic resonance imaging, or arthrography were
performed routinely. The level of patient compliance was recorded
at each clinic visit.
Data Analysis
Initial data screening was accomplished with use of scatterplots,
histograms, and frequency tables for all variables. Additional diagnostics
were completed on any potential outliers with use of regression
diagnostics and studentized residuals. Violations of linearity,
homoscedasticity, and independence were assessed on the scatterplots28. A chi-square goodness-of-fit test
was utilized to evaluate our null hypothesis that the mean postoperative
Rowe score would be 85 points or less. The alpha level was set at
0.05. Paired t tests were utilized to determine if there were any
differences between preoperative and postoperative scores. Within-subject
analyses of variance were used to evaluate whether there were any
differences among items (such as degree of instability, patient compliance,
and frequency of preoperative dislocations) with three variables
or more. Tukey post hoc testing was completed for
all possible pairwise comparisons, with the overall experimental
alpha level maintained at 0.05. Standard statistical software (SPSS,
Chicago, Illinois) was used to analyze the data.
The mean duration of symptoms prior to surgery was fourteen months
(range, one week to 120 months). The mean duration from the operation
to the final follow-up evaluation was thirty-three months (range,
twenty-six to sixty-three months).
Fifteen of the patients had had recurrent subluxation, thirty-three
had had recurrent dislocation, and five had had recurrent subluxation
after a single dislocation. Shoulder instability had developed after
a single traumatic event in forty-eight patients, and it had developed
without a traumatic event in five patients. We classified the instability as
chronic in fifty patients and as acute in three.
Radiographic Analysis
All patients were evaluated preoperatively with use of standard
radiographs. The radiographic findings were normal in thirty-six
patients and abnormal in seventeen. The abnormalities included anterior-inferior
dislocation in six patients, anterior-inferior dislocation and a
nondisplaced fracture of the greater tuberosity in one, an osseous
Bankart lesion in one, and a Hill-Sachs lesion in eleven; multiple
abnormalities often were observed in the same patient. Two patients
underwent computerized axial tomography prior to evaluation in our clinic;
a tear of the anterior part of the labrum was identified in one
of them and a Hill-Sachs lesion, in the other. Twenty-nine patients
underwent preoperative magnetic resonance imaging. The results were
normal in five patients and abnormal in twenty-four. The abnormal
findings included a tear of the anterior part of the labrum in eight
patients, a lesion of the superior part of the labrum in two, a
Hill-Sachs lesion in eight, a large tear of the anterior aspect
of the capsule in three, and a partial-thickness tear of the rotator
cuff in three. The operative findings in the five patients with
normal magnetic resonance imaging studies (all of which were done
without contrast material) included a tear of the superior part
of the labrum in one patient, a Bankart lesion in one patient, and
a tear of the superior part of the labrum and a Bankart lesion in
three patients. The mean final Rowe score for these five patients
was 96.4 points (range, 83 to 100 points).
Findings on Physical Examination
Abduction and external rotation produced pain in forty-eight
of the fifty-three patients and apprehension in thirty-eight. Five
patients had negative findings on examination, but four of them
had a Bankart lesion that was noted at the time of the operation.
The Rowe test16 produced pain
in forty-four patients and apprehension in thirty-seven. The sulcus
test was negative in thirty-four patients, and it produced pain
in fifteen patients and apprehension in four.
Operative Findings
The operative findings varied, and most patients had more than
a single lesion (Table I).
Correlation of Physical Examination and Operative
Findings
When patients who had pain with abduction and external rotation
were compared with those who had apprehension, we could find no
difference in the operative findings. When patients who had negative
findings on physical examination were compared with those who had
apprehension, we noted a difference in the operative findings related to
the anterosuperior aspect of the glenohumeral joint. Specifically,
none of the patients who had negative findings on physical examination
had a tear of the superior glenohumeral ligament or the rotator
interval. Only one of the five patients who had negative findings
on physical examination had a tear of the middle glenohumeral ligament,
but all had a lesion of the anteroinferior aspect of the ligament-labrum
complex. None of the patients who had negative findings on physical
examination had loose bodies or synovitis, whereas eight patients who
had positive findings on physical examination had such abnormalities.
Patients who had a negative result on the Rowe or sulcus test also
did not have any abnormalities of the superior glenohumeral ligament,
rotator interval, or superior part of the labrum, while eighteen
of the thirty patients who had a positive result on either of these
tests had lesions in this area.
Operative Repair
A variety of lesions were repaired at the time of the operation,
and most patients had more than one lesion (Table II). We inserted
suture anchors in fifty-two patients. A mean of 2.4 anchors (range,
zero to five anchors) were used in each patient. Thermal capsulorrhaphy
was used to increase ligamentous or capsular tension after suture
repair in forty-eight of the fifty-three shoulders. There were no
cases in which we thought that thermal capsulorrhaphy without suture
repair would have been adequate to restore soft-tissue tension.
Postoperative Shoulder Scores
All four rating systems revealed significant improvement in the
status of the shoulder when the preoperative scores were compared
with the scores that were recorded at the time of the most recent follow-up.
Specifically, the mean score (and standard deviation) increased
from 45.5 18.6 points to 91.7 13.7 points with the system of the
American Shoulder and Elbow Surgeons (ASES)44,
from 56.4 13.3 points to 91.8 11.3 points with the system of Constant
and Murley13, from 11.3 5.7 points
to 91.9 20.8 points with the system of Rowe et al.47, and from 17.6 4.8 points to 32.0
4.7 points with the system of the University of California at Los
Angeles (UCLA)15 (p = 0.001 for
all comparisons; paired t test). Neither the Constant scoring system
nor the ASES scoring system includes a specific description of the
scores that are considered to represent an excellent or poor rating.
Ellman et al.15 categorized a
UCLA Shoulder Score of 29 to 35 points as good or excellent and
a score of less than 29 points as fair or poor. Rowe et al.47 rated a score of 90 to 100 points
as excellent and a score of 75 to 89 points as good. In the present study,
forty-nine (92 percent) of the fifty-three shoulders had an overall
rating of good or excellent according to both the Rowe score and
the UCLA Shoulder Score.
Primary Hypothesis Testing
The final mean Rowe score for all patients was 91.9 points. To
evaluate the results of our null hypothesis (that arthroscopic treatment
of instability results in a postoperative mean Rowe score of 85 points
or less), a chi-square goodness-of-fit statistic was calculated.
The chi-square statistic was 862.38 with 15 degrees of freedom (p < 0.0005). Therefore,
we rejected the null hypothesis and accepted the alternative, namely,
that an arthroscopic operation to treat instability produces a mean Rowe
score of greater than 85 points.
Satisfaction
Patients rated their level of satisfaction with use of the UCLA
Shoulder Score. Preoperatively, none of the patients rated their
satisfaction as good or excellent (4 or 5 of a possible 5 points).
Postoperatively, forty-eight (90 percent) of the fifty-three patients
rated their satisfaction as good or excellent (4 or 5 points) and
five (10 percent) rated it as fair or poor (0 to 3 points).
Pain
Preoperatively, ten patients rated the pain as minimal (a score
of 0, 1, or 2 points) on the 10-point visual analog scale. All ten
patients described the typical level of shoulder pain as minimal
but rated the pain as severe (10 points) when the shoulder was dislocated.
Postoperatively, pain was minimal (0, 1, or 2 points) in forty-eight
patients (91 percent), mild (3 or 4 points) in four, and moderate
to severe (5 to 10 points) in one.
Function
The patients completed a self-assessment questionnaire on shoulder
function with use of the four shoulder-scoring systems. Preoperatively,
no patient rated function as good to excellent (a score of 35 to
50 points) on the Rowe function subscale (maximum possible score,
50 points). Postoperatively, forty-eight patients (91 percent) rated
function as good to excellent and five rated it as fair to poor.
On the ASES Shoulder Index, significant improvement was demonstrated
in all ten activities of daily living (function items) (p = 0.0001) and
the score for total function improved from a mean of 24.4 points
to a mean of 47.1 points (p = 0.0001).
Range of Motion
No patient lost more than 5 degrees of elevation. The mean external
rotation with the shoulder in 90 degrees of abduction measured 88.2
degrees. External rotation measured 85 degrees in three patients
and 70 and 60 degrees in one patient each. These five patients rated
their satisfaction (according to the UCLA Shoulder Score) as 5 of
a possible 5 points. The patient who had the operation on the dominant
side was nonathletic. The four other patients had the operation
on the nondominant side.
Strength
The strength of elevation improved 60 percent, from a preoperative
mean of 12.9 pounds (5.9 kilograms) to a mean of 20.6 pounds (9.3
kilograms) at the final follow-up evaluation; this finding was significant
(p = 0.0001). Since an anesthetic was not injected prior to strength
measurement, this increase in strength was probably due to the postoperative
program of strengthening exercises and to a decrease in the level
of pain.
Return to Sports Participation
Forty-three patients had participated actively in sports prior
to the onset of the shoulder symptoms. According to the sports category,
none of the patients had participated in type-1 sports; five, in type-2
sports; thirty, in type-3 sports; and eight, in type-4 sports. Seven
patients had participated at level 1 (high-school team sports);
one, at level 2 (college team sports); and thirty-five, at level
3 (recreational sports). At the final follow-up evaluation, five
patients reported that they did not participate in sports because
of issues that were unrelated to the shoulder. The reasons most
commonly cited were work or family commitments, graduation from
high school or college (and the associated lack of team sports),
and injuries to the knee or lumbar spine. At the final follow-up
examination, thirty-eight patients who did not have such issues
participated in sports. One patient participated in type-1 sports;
six, in type-2 sports; twenty-six, in type-3 sports; and five, in
type-4 sports. Three were involved in level-1 sports; none, in level-2
sports; and thirty-five, in level-3 sports. Four patients with persistent
shoulder instability had decreased their level of participation.
Degree of Instability
The final mean Rowe score was found to be 90 points for the thirty-three
patients with recurrent dislocation, 92 points for the five patients
with recurrent subluxation after an initial dislocation, and 97
points for the fifteen patients with recurrent subluxation. The
differences among these three groups (analyzed with a between-subjects
analysis of variance) were not found to be significant (p = 0.519).
Traumatic Compared with Atraumatic Instability
The final mean Rowe score was 91.4 points for the forty-eight
patients with shoulder instability that followed a traumatic event
and 99.6 points for the five patients with an atraumatic etiology.
The difference between these groups was found to be significant
(p = 0.01). Four of the five patients who did not have a traumatic
event were found, at the operation, to have a Bankart lesion.
Number of Preoperative Dislocations
The final mean Rowe score was tabulated according to the number
of preoperative dislocations. We grouped patients into three categories
on the basis of the number of reported dislocations; ten patients had
had one or two dislocations, fifteen had had three to nine dislocations,
and thirteen had had ten dislocations or more. No significant difference among
the groups was identified, with the numbers available (p = 0.8).
We noted a trend toward an increase in the severity of the lesion
in the anteroinferior part of the labrum in association with an
increase in the number of preoperative dislocations. Similar findings
were noted by Habermeyer et al.22.
Type-C labral tears were found only in shoulders with three dislocations
or more. Operative correction of the inferior aspect of the capsule
or labrum was required in only eight of the twenty-eight patients
with three dislocations or more.
Age at Time of Operation
The final mean Rowe score was 84 points for the nine patients
who were less than twenty years old at the time of the operation
and 93 points for the forty-four patients who were twenty years
or older. Both Grana et al.20 and
Savoie et al.49 noted poorer results
in younger patients. Although there was a tendency toward poorer
results in younger patients in the present series, the difference
was not found to be significant, with the numbers available (p =
0.240).
Gender
The final mean Rowe score was 92 points for the male patients
and 91 points for the female patients. This difference was not found
to be significant, with the numbers available (p = 0.850).
Arm Dominance
The final mean Rowe score was 91 points for the patients in whom
the dominant side was involved and 93 points for the patients in
whom the nondominant side was involved. This difference was not
found to be significant, with the numbers available (p = 0.740).
Duration of Instability
The final mean Rowe score was 91.4 points for the fifty patients
in whom the instability was classified as chronic (present for more
than six weeks before the operation) and 100 points for the three
patients in whom it was classified as acute. The difference was
found to be significant (p = 0.007).
Ligamentous Laxity
The final mean Rowe score was 94 points for the forty-seven patients
without evidence of ligamentous laxity and 74 points for the six
patients with ligamentous laxity. This difference was found to be
significant (p = 0.02). The inferior results in the patients with
ligamentous laxity may have been the result of a technically inadequate
repair or may suggest that patients with anterior-inferior instability
and generalized ligamentous laxity require an open capsular reconstruction
to achieve adequate soft-tissue tension.
Patient Compliance
The mean compliance score was 7.8 points (range, 2 to 10 points).
Four patients had a low compliance score (0 to 3 points), twenty-five
had a medium compliance score (4 to 7 points), and twenty-four had
a high compliance score (8, 9, or 10 points). The final mean Rowe
score was 88 points for patients with low compliance, 87 points
for those with medium compliance, and 97 points for those with high
compliance. The differences among the three categories (analyzed
with a one-way analysis of variance) were not found to be significant,
with the numbers available (p = 0.263).
Analysis of Unsatisfactory Results
The result in four patients was rated as fair or poor (a score
of less than 29 points according to the UCLA Shoulder Score or a
score of less than 75 points according to the system of Rowe et
al.) at the final follow-up evaluation. Two patients (both of whom
had a preoperative diagnosis of recurrent dislocation) continued
to have dislocation of the shoulder after the operation. Two patients
(one with a preoperative diagnosis of recurrent dislocation and
one with a preoperative diagnosis of recurrent subluxation after
a single dislocation) were noted to have recurrent subluxation after
arthroscopic repair. We were unable to distinguish this subset of
four patients on the basis of the operative findings or the repair
technique. All four patients were male, and their mean age (twenty-one
years; range, fifteen to thirty-one years) was lower than that for
the rest of the study group. The mean compliance score of the four
patients was 4.5 points, and the final mean Rowe score was 31 points (range,
10 to 53 points). The time to failure (defined as postoperative
dislocation or subluxation) averaged thirteen months (range, nine
to eighteen months). No patient who had a stable shoulder at eighteen
months postoperatively had development of late instability after
as many as five years of follow-up.
Complications
No major intraoperative or perioperative complications (permanent
nerve injuries or wound infections) occurred. Two patients had paresthesias
in the musculocutaneous nerve distribution that resolved by the
six-week postoperative follow-up examination. One patient noted
minor wound drainage that resolved within one week without the use
of antibiotics. We did not observe any complications related to
the use of suture anchors.
Revision Operations
One patient (with an original diagnosis of recurrent anterior-inferior
subluxation) noted a return of the subluxation fourteen months after
the index procedure. At the time of the revision operation, the
original repairs (for a Bankart lesion, a SLAP lesion, and an abnormality
of the rotator interval) were intact but we noted inferior and posterior capsular
laxity. Since the insertion-site lesions that had been repaired
at the time of the index operation were intact, the inferior portion
of the capsule was tightened with monofilament sutures. We performed
laser thermal capsulorrhaphy to further tighten the posteroinferior
aspect of the glenohumeral ligament. The patient had a successful result
with a return to full activities and normal stability on physical
examination.
The wide variety of lesions, the patient population, the operative
techniques, the duration of follow-up, and the use of multiple scoring
systems complicate comparison of the results of this report with the
results of open repairs. However, on the basis of the level of improvement
in the various parameters described in the present investigation,
we concluded that the outcome of arthroscopic repair of anterior-inferior
glenohumeral instability is better than that of previous arthroscopic
treatments and is equivalent to that of open repair. Wirth et al.57 described the results of open repair
in a study of 138 patients (142 shoulders). They did not report mean
scores, but 93 percent (132) of the 142 shoulders had a good or
excellent result according to the Rowe scoring system.
The spectrum of operative findings in the present study did not
support the concept of any "essential lesion." On the contrary,
it appeared that the etiology of anterior-inferior glenohumeral
instability was multifactorial and that successful treatment required
that any operative approach be sufficiently flexible to deal with
the variety of lesions found. Our arthroscopic approach allows the
surgeon to identify and treat all of the lesions of shoulder instability.
We believe that the success of our arthroscopic approach was due
to our ability to perform an anatomical repair of tears of the anterior,
superior, and inferior parts of the labrum; to correct capsular
elongation; and, if necessary, to repair the rotator interval.
The present study has a number of weaknesses. Although the investigation
was prospective, the patients were not randomized and the investigator was
not blinded. In addition, the follow-up period was relatively short;
we are continuing to follow these patients so that we can evaluate
the results over a longer interval. We also believe that these results
may deteriorate with time and may parallel the experience after
open repair38.
Currently, we use the arthroscopic technique whenever operative
treatment of glenohumeral instability is indicated; no open repairs
are performed. The arthroscopic technique allows us to inspect the
entire glenohumeral joint and to avoid soft-tissue dissection. No
division of the subscapularis is required. Although we are unable
to document our impressions statistically, we believe that arthroscopic
repair provides an improved cosmetic appearance, decreased postoperative
pain, and more rapid gains in motion when compared with open operative
treatment of similar lesions.
Our technique can be recommended only to experienced orthopaedic
surgeons who are familiar with the normal and abnormal anatomy seen
during both open and arthroscopic shoulder operations. A thorough
understanding of the various conditions that produce pain in the
shoulder is needed. Orthopaedic surgeons who infrequently perform
open repair for glenohumeral instability should not undertake the
arthroscopic procedure. The arthroscopic operation requires advanced
arthroscopic techniques and is still in the developmental stage.
Arthroscopic repair for anterior-inferior glenohumeral instability
produced good or excellent results in forty-nine of fifty-three
patients. The mean external rotation with the shoulder in 90 degrees
of abduction measured 88.2 degrees. Thirty-four of thirty-eight
patients were able to return to their desired level of sports activity.
Four patients who had persistent instability were considered to
have had a failure of the index operation. We believe that the improved
rate of success demonstrated in our study is the result of repair
not only of anterior-inferior (Bankart) lesions but also (when necessary)
of inferior and superior labral tears. Additionally, soft-tissue
tension within the capsule and ligaments was corrected with use
of a suture technique but was supplemented with laser thermal capsulorrhaphy,
if necessary. Repair of the rotator interval was considered a critical
factor in selected patients.
Allain, J.; Goutallier, D.; and Glorion, C.: Long-term results of the Latarjet procedure for the treatment
of anterior instability of the shoulder. J. Bone and Joint Surg.,80-A: 841-862, June 1998.80-A841
1998
Altchek, D. W.; Warren, R. F.; Skyhar, M. J.; and Ortiz, G.: T-plasty modification of the Bankart procedure for multidirectional
instability of the anterior and inferior types. J. Bone and Joint Surg.,73-A: 105-112, Jan 1991.73-A105
1991
Altchek, D. W.; Arciero, R. M.; Gross,
R. M.; Hawkins, R. J.; and Warren, R. F. [moderator]: Open and arthroscopic
glenohumeral instability repairs. Instructional course at the Annual
Meeting of the American Academy of Orthopaedic Surgeons, New Orleans,
Louisiana, March 21, 1998.
Arciero, R. A.; Taylor, D. C.; Snyder, R. J.; and Uhorchak, J. M.: Athroscopic bioabsorbable tack stabilization of initial
anterior shoulder dislocations: a preliminary report. Arthroscopy,11: 410-417, 1995.11410
1995
[PubMed]
Baker, C. L.; Uribe, J. W.; and Whitman, C.: Arthroscopic evaluation of acute initial anterior shoulder
dislocations. Am. J. Sports Med.,18: 25-28, 1990.1825
1990
[PubMed]
Bigliani, L. U.; Pollock, R. G.; Soslowsky, L. J.; Flatow, E. L.; Pawluk, R. J.; and Mow, V. C.: Tensile properties of the inferior glenohumeral ligament. J. Orthop. Res.,10: 187-197, 1992.10187
1992
[PubMed]
Bigliani, L. U.; Kurzweil, P. R.; Schwartzbach, C. C.; Wolfe, I. N.; and Flatow, E. L.: Inferior capsular shift procedure for anterior-inferior
shoulder instability in athletes. Am. J. Sports Med.,,2: 578-584, 1994.2578
1994
Burkhart, S. S., and Morgan, C. D.: The peel-back mechanism: its role in producing and extending
posterior type II SLAP lesions and its effect on SLAP repair rehabilitation. Arthroscopy,14: 637-640, 1998.14637
1998
[PubMed]
Burkhead, W. Z., Jr., and Rockwood, C. A., Jr.: Treatment of instability of the shoulder with an exercise
program. J. Bone and Joint Surg.,74-A: 890-896, July 1992.74-A890
1992
Caspari, R. B., and Savoie, F. H.:
Arthroscopic reconstruction of the shoulders. The Bankart repair.
In Operative Arthroscopy, pp. 507-515. Edited by J. B. McGinty.
New York, Raven Press, 1991.
Cofield, R. H.; Kavanagh, B. F.; and
Frassica, F. J.: Anterior shoulder instability. In Instructional Course
Lectures, American Academy of Orthopaedic Surgeons. Vol. 34, pp.
210-227. St. Louis, C. V. Mosby, 1985.
Cohen, J.: Statistical Power Analysis
for the Behavioral Sciences. Ed. 2, pp. 25-41. Hillsdale, New Jersey,
Lawrence Erlbaum Associates, 1988.
Constant, C. R., and Murley, A. H.: A clinical method of functional assessment of the shoulder. Clin. Orthop.,214: 160-164, 1987.214160
1987
[PubMed]
Cooper, N. A., and Brems, J. J.: The inferior capsular-shift procedure for multidirectional
instability of the shoulder. J. Bone and Joint Surg.,74-A: 1516-1521, Dec 1992.74-A1516
1992
Ellman, H.; Hanker, G.; and Bayer, M.: Repair of the rotator cuff. End-result study of factors
influencing reconstruction. J. Bone and Joint Surg.,68-A: 1136-1144, Oct 1986.68-A1136
1986
Ellman, H., and Gartsman, G. M.: Arthroscopic Shoulder
Surgery and Related Procedures, pp. 268-269. Philadelphia, Lea and
Febiger, 1993.
Fanton, G. S.:: Schulterarthroskopie unter Verwendung des Holmium:YAG-Lasers.
Stand 1994. Orthopade, ,25: 79-83, 1996.2579
1996
Gartsman, G. M.: Arthroscopic acromioplasty for lesions of the rotator
cuff. J. Bone and Joint Surg.,72-A: 169-180, Feb 1990.72-A169
1990
Gartsman, G. M.; Taverna, E.; and Hammerman, S. M.: Arthroscopic rotator interval repair in glenohumeral instability:
description of an operative technique. Arthroscopy,15: 330-332, 1999.15330
1999
[PubMed]
Grana, W. A.; Buckley, P. D.; and Yates, C. K.: Arthroscopic Bankart suture repair. Am. J. Sports Med.,21: 348-353, 1993.21348
1993
[PubMed]
Gross, R. M.: Arthroscopic shoulder capsulorrhaphy: does it work?. Am. J. Sports Med.,17: 495-500, 1989.17495
1989
[PubMed]
Habermeyer, P.; Gleyze, P.; and Rickert, M.: Evolution of lesions of the labrum-ligament complex in
posttraumatic anterior shoulder instability: a prospective study. J. Shoulder and Elbow Surg.,8: 66-74, 1999.866
1999
Hardy, P.; Thabit, G., III; Fanton, G. S.; Blin, J. L.; Lortat-Jacob, A.; and Benoit, J.: Arthroskopische Behandlung der rezidivierenden vorderen
Schulterluxation durch Kombination der Labrumnaht mit einer anteroinferioren Kapselschrumpfung
mit dem Holmium:YAG-Laser. Orthopade,25: 91-93, 1996.2591
1996
[PubMed]
Harryman, D. T.; Sidles, J. A.; Harris, S. L.; and Matsen, F. A., III: The role of the rotator interval capsule in passive motion
and stability of the shoulder. J. Bone and Joint Surg.,74-A:: 53-66, Jan 1992.74-A:53
1992
Hawkins, R. B.: Arthroscopic stapling repair for shoulder instability:
a retrospective study of 50 cases. Arthroscopy,5: 122-128, 1989.5122
1989
[PubMed]
Hawkins, R. J.; Chris, T.; Bokor, D.; and Kiefer, G.: Failed anterior acromioplasty. A review of 51 cases. Clin. Orthop.,243: 106-111, 1989.243106
1989
[PubMed]
Hayashi, K.; Thabit, G., III; Bogdanske, J. J.; Mascio, L. N.; and Markel, M. D.: The effect of nonablative laser energy on the ultrastructure
of joint capsular collagen. Arthroscopy,12: 474-481, 1996.12474
1996
[PubMed]
Huck, S. W., and Cormier, W. H.: Reading
Statistics and Research. Ed. 2, pp. 128-145. New York, Harper Collins,
1996.
Johnson, L. L.: Diagnostic and Surgical
Arthroscopy of the Shoulder, pp. 318-321. St. Louis, C. V. Mosby,
1993.
Kohn, D.: The clinical relevance of glenoid labrum lesions. Arthroscopy,3: 223-230, 1987.3223
1987
[PubMed]
Lippitt, S. B.; Vanderhooft, J. E.; Harris, S. L.; Sidles, J. A.; Harryman, D. T., II; and Matsen, F. A., III: Glenohumeral stability from concavity-compression: a quantitative
analysis. J. Shoulder and Elbow Surg.,2: 27-35, 1993.227
1993
Lopez, M. J.; Hayashi, K.; Fanton, G. S.; Thabit, G., III;; and Markel, M. D.: The effect of radiofrequency energy on the ultrastructure
of joint capsular collagen. Arthroscopy,14: 495-501, 1998.14495
1998
[PubMed]
McIntyre, L. F.; Caspari, R. B.; and Savoie, F. H., III: The arthroscopic treatment of multidirectional shoulder
instability: two-year results of a multiple suture technique. Arthroscopy,13: 418-425, 1997.13418
1997
[PubMed]
McIntyre, L. F.; Caspari, R. B.; and Savoie, F. H., III: The arthroscopic treatment of posterior shoulder instability:
two-year results of multiple suture technique.. Arthroscopy,13: 426-432, 1997.13426
1997
[PubMed]
Misamore, G. W.; Ziegler, D. W.; and Rushton, J. L., III: Repair of the rotator cuff. A comparison of results in
two populations of patients. J. Bone and Joint Surg.,77-A: 1335-1339, Sept 1995.77-A1335
1995
Morgan, C. D., and Bodenstab, A. B.: Arthroscopic Bankart suture repair: technique and early
results. Arthroscopy,3: 111-122, 1987.3111
1987
[PubMed]
Morgan, C. D.; Burkhart, S. S.; Palmeri, M.; and Gillespie, M.: Type II SLAP lesions: three subtypes and their relationships
to superior instability and rotator cuff tears. Arthroscopy,14: 553-565, 1998.14553
1998
[PubMed]
Morrey, B. F., and Janes, J. M.: Recurrent anterior dislocation of the shoulder. Long-term
follow-up of the Putti-Platt and Bankart procedures. J. Bone and Joint Surg.,58-A: 252-256, March 1976.58-A252
1976
Neer, C. S., and Foster, C. R.: Inferior capsular shift for involuntary and multidirectional
instability of the shoulder. A preliminary report. J. Bone and Joint Surg.,62-A: 897-908, Sept 1980.62-A897
1980
Neer, C. S., II, and Poppen, N. K.: Supraspinatus outlet. Orthop. Trans.,11: 234, 1987.11234
1987
Neviaser, T. J.: The anterior labroligamentous periosteal sleeve avulsion
lesion: a cause of anterior instability of the shoulder. Arthroscopy,9: 17-21, 1993.917
1993
[PubMed]
Nottage, W. M.: Laser-assisted shoulder surgery. Arthroscopy,13: 635-638, 1997.13635
1997
[PubMed]
Pappas, A. M.; Goss, T. P.; and Kleinman, P. K.: Symptomatic shoulder instability due to lesions of the
glenoid labrum. Am. J. Sports Med.,11: 279-288, 1983.11279
1983
[PubMed]
Richards, R. R.; An, K.-N.; Bigliani, L. U.; Friedman, R. J.; Gartsman, G. M.; Gristina, A. G.; Iannotti, J. P.; Mow, V. C.; Sidles, J. A.; and Zuckerman, J. D.: A standardized method for the assessment of shoulder function. J. Shoulder and Elbow Surg.,3: 347-352, 1994.3347
1994
Rodosky, M. W.; Rudert, M. J.; Harner, C. H.; Luo, L.; and Fu, F. H.:: Significance of a superior labral lesion of the shoulder:
a biomechanical study. Trans. Orthop. Res. Soc.,15: 276, 1990.15276
1990
Rodosky, M. W.; Harner, C. D.; and Fu, F. H.: The role of the long head of the biceps muscle and superior
glenoid labrum in anterior stability of the shoulder. Am. J. Sports Med.,22: 121-130, 1994.22121
1994
[PubMed]
Rowe, C. R.; Patel, D.; and Southmayd, W. W.: The Bankart procedure. A long-term end-result study. J. Bone and Joint Surg.,60-A: 1-16, Jan 1978.60-A1
1978
Rowe, C. R., and Zarins, B.: Recurrent transient subluxation of the shoulder. J. Bone and Joint Surg.,63-A: 863-872, July 1981.63-A863
1981
Savoie, F. H., III; Miller, C. D.; and Field, L. D.: Arthroscopic reconstruction of traumatic anterior instability
of the shoulder: the Caspari technique. Arthroscopy,13: 201-209, 1997.13201
1997
[PubMed]
Snyder, S. J.; Karzel, R. P.; Del Pizzo, W.; Ferkel, R. D.; and Friedman, M. J.: SLAP lesions of the shoulder. Arthroscopy,6: 274-279, 1990.6274
1990
[PubMed]
Speer, K. P.; Deng, X.; Borrero, S.; Torzilli, P. A.; Altchek, D. A.; and Warren, R. F.: Biomechanical evaluation of a simulated Bankart lesion. J. Bone and Joint Surg.,76-A: 1819-1825, Dec 1994.76-A1819
1994
Vangsness, C. T., Jr; Jorgenson, S. S.; Watson, T.; and Johnson, D. L.: The origin of the long head of the biceps from the scapula
and glenoid labrum. An anatomical study of 100 shoulders. J. Bone and Joint Surg.,76-B(6): 951-954, 1994.76-B(6)951
1994
Vangsness, C. T., Jr., and Smith, C. F.:: Arthroscopic shoulder surgery with three different laser
systems: an evaluation of laser applications. Arthroscopy,11: 696-700, 1995.11696
1995
[PubMed]
Walch, G.; Boileau, P.; Levigne, C.; Mandrino, A.; Neyret, P.; and Donell, S.: Arthroscopic stabilization for recurrent anterior shoulder
dislocation: results of 59 cases. Arthroscopy,,11: 173-179, 1995.11173
1995
Warner, J. J.; Johnson, D.; Miller, M.; and Caborn, D. N.: Technique for selecting capsular tightness in repair of
anterior-inferior shoulder instability. J. Shoulder and Elbow Surg.,4: 352-364, 1995.4352
1995
Williams, M. M.; Snyder, S. J.; and Buford, D., Jr.: The Buford complex - the "cord-like" middle glenohumeral
ligament and absent anterosuperior labrum complex: a normal anatomic
capsulolabral variant. Arthroscopy,,10: 241-247, 1994.10241
1994
Wirth, M. A.; Blatter, G.; and Rockwood, C. A., Jr.: The capsular imbrication procedure for recurrent anterior
instability of the shoulder. J. Bone and Joint Surg.,78-A: 246-259, Feb 1996.78-A246
1996
Wolf, E. M.; Wilk, R. M.; and Richmond, J. C.: Arthroscopic Bankart repair using suture anchors. Op. Tech. Orthop.,1: 184-191, 1991.1184
1991
Wolf, E. M.; Cheng, J. C.; and Dickson, K.: Humeral avulsion of glenohumeral ligaments as a cause
of anterior shoulder instability. Arthroscopy,11: 600-607, 1995.11600
1995
[PubMed]
Zuckerman, J. D., and Matsen, F. A., III: Complications about the glenohumeral joint related to
the use of screws and staples. J. Bone and Joint Surg.,66-A: 175-180, Feb 1984.66-A175
1984