The Journal of

Bone and Joint Surgery

American Volume
VoLUME 60-A, No. I JANUARY 1978

(‘ssp right i)7h h The Jssur,tal si ROt teat ul Joi’tt Surgers . Incorporated

The Bankart Procedure

A LONG-TERM END-RESULT STUDY

BY CARTER R. ROWE, M.D.*, DINESH PATEL, M.D.*,

AND WILLIAM W. SOUTHMAYD, M.D.*, BOSTON, MASSACHUSETTS

Front f/it’ Deparitneni of Orthopaedics, Massachusetts General Hospital, Boston

ABSTRACT: Of 161 patients with 162 shoulders op- anterior dislocation of the shoulder and his method of re-
erated on during a thirty-year period (1946 to 1976), pair L’ His comments prompted lively , as
124 were re-examined and twenty-one answered a was evident in an excellent review of shoulder dislocations
questionnaire. The lesions found at surgery were sep- published in 1948
aration of the capsule from the anterior glenoid rim in It is the purpose of this report to document the
85 per cent, a Hill-Sachs lesion of the humeral head in findings after long-term follow-up of shoulders repaired by
77 per cent, and damage to the anterior glenoid rim one specific technique which had not varied, except for
(including fracture) in 73 per cent. There were five re- minor changes, since 1946. The procedure to be described
currences (3.5 per cent) after repair by the method de- closely parallels Bankan’s original method. Only patients
scribed in the 145 shoulders that were followed. Only whose surgery was performed either by the senior author
one of the forty-six patients with dislocation on the (C. R. R.) or while he was present were included. There-
dominant side and one of the thirty-one with disloca- fore, the preoperative evaluation, operative technique, and
tion on the non-dominant side failed to return to the postoperative care were uniform.
competitive athletic activities in which they had par-
ticipated prior to injury. The results at follow-up were Clinical Material
rated excellent in 74 per cent, good in 23 per cent, and One hundred and sixty-one patients ( 138 male and
poor in 3 per cent. Ninety-eight per cent of the patients twenty-three female patients) had 162 shoulders operated
rated their result as excellent or good. Sixty-nine per on between 1946 and 1976. Patients with voluntary dislo-
cent of the shoulders had a full range of motion, and cations of the shoulder were excluded. Ofthe 161 patients
only 2 per cent of these shoulders redislocated. A frac- included, sixteen could not be located for follow-up,
ture of the rim of the glenoid did not increase the risk although preoperative and operative findings were avail-
of recurrence, while a moderate to severe Hill-Sacks able; 124 were examined personally at follow-up and their
lesion increased the risk only slightly. results were graded according to a standard rating scale
We concluded that with the meticulous technique (Table I); and twenty-one, unable to come in for examina-
of the Bankart repair as described, postoperative im- tion, answered a detailed questionnaire that included nine
mobilization is not necessary, early return of motion questions relative to recurrence, stability, the percentage
and function can be expected, and resumption of ath- of motion compared with the opposite shoulder, any limi-
letic activities with no limitation of shoulder motion is tations in sports or work, and current work and sports ac-
possible for most patients. tivities.
Pathological lesions were recorded for 1 58 of the 162
It has been fifty-five years since Bankart presented his shoulders. The result was not evaluated in any patient
concept of the pathological lesion responsible for recurrent whose surgery was performed less than one year prior to
follow-up examination.
* Massachusetts General Hospital. Warren Building. 275 Charles Ofthe 145 patients (146 shoulders), ninety-seven (67
Street. Boston. Massachusetts 021 14. per cent) were followed for one to five years and forty-
2 C. R. ROWE, DINESH PATEL, AND W. W. SOUTHMAYD

TABLE I
RATING SHEET FOR BANKART REPAIR

Excellent Good Fair Poor

Scoring System Units (100-90) (89-75) (74-51) (50 or Less)

Stability
No recurrence, subluxa- 50 No recurrences No recurrences No recurrences Recurrence of dis-
tion, or apprehension location
or
Apprehension when 30 No apprehension Mild apprehension Moderate apprehen- Marked apprehension
placing arm in certain when placing arm when placing arm sion during eleva- during elevation or
positions in complete dc- ifl elevation and tion and external extension
vation and external external rotation notation
rotation
Subluxation (not requiring 10 No subluxations No subluxations No subluxations
reduction)

Recurrent dislocation 0

Motion
100% of normal external 20 100% of normal 75% of normal 50% of normal No external rotation;
rotation , internal rota- external rotation; external rotation; external rotation; 50% of elevation
tion , and elevation complete elevation complete elevation 75% of elevation (can get hand only
75% of normal external 15 and internal and internal and internal to face) and 50% of
rotation, and normal rotation notation rotation internal rotation
elevation and internal
rotation
50% of normal external 5
rotation and 75% of nor-
mal elevation and in-
ternal rotation
50% of normal elevation 0
and internal rotation; no
external rotation

Function
No limitation in work or 30 Performs all work and Mild limitation in Moderate limitation Marked limitation;
sports; little or no sports; no limitation work and sports; doing overhead unable to perform
discomfort in overhead activi- shoulder strong; work and heavy overhead work and
Mild limitation and 25 ties; shoulder minimum dis- lifting; unable to lifting; cannot
minimum discomfort strong in lifting, comfort throw, serve hard in throw, play tennis,
Moderate limitation 10 swimming, tennis, tennis, or swim; or swim; chronic
and discomfort throwing; no dis- moderate disabling discomfort
Marked limitation and pain 0 comfort pain

Total units possible 100

history. This incidence was low compared with that in

eight (33 per cent), for five to thirty years. The average
follow-up was six years. previous reports 15.34,36

Preoperative Findings Hand Dominance

Of the 162 dislocations, 142 (88 per cent) were corn- Of the 162 dislocated shoulders in 161 patients,
plete recurrent anterior dislocations and twenty (12 per eighty-three were on the right and seventy-nine, on the left
cent) were transient, in which the shoulder had always re- side. Nineteen (12 per cent) of the patients had bilateral
duced spontaneously before roentgenograrns could be dislocation (an incidence comparable to the 1 0 per cent
made. Eight of the shoulders had been operated on previ- found by Moseley and Overgaard38), although only one
ously but the dislocation had recurred. One hundred and had both shoulders repaired.
forty (86 per cent) of the initial dislocations had been pro- The hand dominance was known for 1 24 patients , of

duced by a definite injury (the traumatic group) and whom 106 (85 per cent) were right-handed; fifteen (13 per
twenty-two (14 per cent), by a natural movement of the cent), left-handed; and three (2 per cent), ambidextrous.
arm (the atraumatic group). Gallie and LeMesurier, in Surgery was performed on fifty-two (49 per cent)
their series, reported a 17 per cent incidence of atraumatic right and fifty-four (5 1 per cent) left shoulders in the 106
initial dislocations 24 right-handed patients; on eleven (73 per cent) right and
four (27 per cent) left shoulders in the fifteen left-handed
Family History
patients; and on one right and two left shoulders in the
Information was available relative to a family history three ambidextrous patients. Thus, in the right-handed pa-
of recurrent dislocation for fifty-five of the 161 patients. tients, there was no appreciable difference in the fre-
Of these fifty-five, forty (73 per cent) denied any familial quency of dislocation on the dominant and non-dominant
incidence and fifteen (27 per cent) gave a positive family sides, while in the left-handed patients, there was a sig-

THE JOURNAL OF BONE AND JOINT SURGERY

 THE BANKART PROCEDURE 3

nificantly increased incidence on the non-dominant side. the deltoid muscle from the clavicle since 1960.
9 reported an over-all increased incidence of disloca- 7. The coracoid process is routinely osteotomized,
tions on the weaker side. allowing the coracobrachialis and the short head of the
biceps to retract mesially (Fig. I-A, C).
Age at Surgery
8. The arm, still at the side of the body, is then exter-
Of the 161 patients, eighty-one were less than nally rotated, exposing the subscapularis muscle. The
twenty-one years old and eighty were twenty-one years old circumflex vessels along the inferior border of the muscle
or older. The youngest patient was fifteen years old and the are ligated (Fig. 1-A, C).
oldest, forty-seven. 9. Starting distally, the subscapularis muscle is care-
fully separated from the capsule in toto (Fig. 1-B, D).
Causes of Initial Dislocations
This is a very important step in the operation and can be
Information relative to the specific mechanism of the effectively accomplished by holding the knife blade in the
initial dislocation was available on eighty-six patients. horizontal plane and separating the tendon from the cap-
Forceful extension or abduction of the arm was responsible sule by sharp dissection. To avoid entering the joint, a
for the first dislocation in twenty-six shoulders (30 per small amount of tendon is left on the capsule. Once the
cent); forceful elevation and external rotation, in twenty- tendon has been separated from the capsule (the attach-
one (24 per cent); a direct blow to the shoulder, in ment of the tendon usually extends over a distance of 2.5
twenty-five (29 per cent); and a fall on the outstretched centimeters in the medial-to-lateral direction), the muscle
hand, in fourteen ( 16 per cent). can be separated from the capsule by blunt dissection,
Our findings do not substantiate Bankart’s using a wing-type periosteal elevator.
as to the mechanism of this injury, since 30 per cent of the 10. The arm is completely externally rotated before a
recurrent dislocations in our series were caused initially by vertical incision is made into thejointjust lateral to the rim
a forceful abduction or extension of the arm, a dislocation ofthe glenoid (Fig. 1-B, E). This gives an excellent expo-
that Bankart claimed never recurs. Also, only 29 per cent sure of the joint and the entire anterior rim and ensures that
of the recurrent dislocations were caused initially by a di- the lateral flap will be of proper length to permit adequate
rect blow to the shoulder and elbow, an injury that Bankart external rotation of the shoulder postoperatively (Fig. 1-B,
thought was the sole cause of the recurrent dislocation. In F).
our opinion, both the so-called ordinary (non-recurrent) 1 1 A humeral-head
. retractor (Fig. 2) is inserted into
and the recurrent dislocations of Bankart may produce the the joint and is used to displace the humeral head pos-
same lesion. terolaterally (Fig. 1-B, H). Ifthe capsule is separated from
the glenoid rim, a three-pronged retractor (Fig. 2) is in-
Surgical Technique serted into the glenoid neck and used to retract the medial
We employed the following technique in this series part of the capsule (Fig. 1-B, H).
(Figs. 1-A, 1-B, and 1-C): 12. The rim of the glenoid and the neck of the
1 . General anesthesia with endotracheal intubation is scapula can now be freshened with a small osteotome or
used. curette. Three holes are made through the rim of the
2. The stability of the shoulder is tested after the pa- glenoid (at one, three, and five o’clock in the right shoul-
tient is anesthetized. Two patients referred with a diag- der and at eleven, nine, and seven o’clock in the left
nosis of recurrent posterior dislocation were found to have shoulder) using a small glenoid punch to initiate the holes
anterior instability. and then a forceps with three-edged cutting points and a
3. The patient is placed supine with a folded blanket special awl to complete them (Fig. 1-B, G through 1).
under the arm rather than under the shoulder, so that the 13. A double No. 0 cotton suture is passed through
humeral head can be displaced posteriorly more easily dur- each hole, using a No. 5, one-half-taper Mayo needle.
ing the procedure. We do not use the semi-sitting position This tough little needle is perfectly curved for this pur-
so frequently described. The anesthetist is on the opposite pose. Each double suture is passed through the edge of the
side of the patient to allow more room at the head of the lateral capsular flap (Fig. 1-C, J) and tied so as to hold the
table. lateral flap securely against the freshened rim of the
4. Good exposure, adequate help, and proper instru- glenoid (Fig. 1-C, K). One limb ofeach of the top and bot-
ments (Fig. 2) are essential. tom sutures (A and D in Fig. 1-C, K) is then cut offand the
5. A straight incision is made from the coracoid pro- four remaining limbs of the sutures (A, B, C, and D in Fig.
cess to the axilla, a shorter incision being used in female 1-C, L) are passed through the medial flap and tied to one
patients (Fig. 1-A, A). another(Ato B and C to D in Fig. 1-C, N). This procedure
6. The deltopectoral interval is identified and de- reinforces the capsule at the rim of the glenoid and along
veloped down to the cephalic vein, which in the majority the neck of the scapula. The arm at this stage can be easily
of instances is ligated proximally and distally and removed externally rotated 25 to 30 degrees beyond neutral.
(Fig. 1-A, B). This eliminates oozing from the vein during 14. Closure of the wound is accomplished by return-
the procedure. We have not found it necessary to separate ing all tissues to their normal insertions. No staples, wires,

VOL. 60-A, NO. I, JANUARY 1978

4 C. R. ROWE, DINESH PATEL, AND W. W. SOUTHMAYD

screws, or bone grafts are used. The subscapularis tendon The average duration of surgery was two and a half
is replaced in its original position on the lesser tuberosity hours. No patient in this series required blood replacement
and secured with interrupted sutures composed of double during or following operation.
strands of No. 20 cotton or, rarely, one strand of No. 0
Postoperative Routine
cotton in a heavily muscled individual. In this way the
muscle is not shortened, overlapped, or transplanted. The Until the early 1960’s, we imniobilized the shoulder
coracoid process is reattached by making a single hole in for three to six weeks in a special shoulder sling, but dur-
the osteotomized fragment and in the base of the process, ing the last ten years most patients have used the sling for
using a small scaphoid gouge. and then passing a double only two to three days, after which the arm has been com-
strand of No. 0 suture through these holes (Fig. 1-C, P); pletely free. During this time the patient can take showers
the strands are tied as a single suture. with the fragments and dress normally. Pendulum exercises as well as light
held in proper position. A reinforcing suture passed activities are begun in the hospital. No formal physical
through the attachment of the common tendon on the therapy is used; instead, the patient is instructed to in-
coracoid process and the coraco-acromial ligament then crease gradually the motion and function of the extremity,
completes the fixation (Fig. 1-C, P). This method of fixa- and is usually back at office work or school in two weeks.
tion is simple and appears to be adequate, since no separa- In six weeks, swimming or rowing is begun. By three
tion of the fragments was seen on any of the follow-up months the patient should have regained 70 per cent of ex-
roentgenograms. ternal rotation and elevation of the shoulder. Tennis, golf,

end of
coracoid process

Retracted biceps (short head

and coracobrachialis

Circumflex vessels

FIG. I-A

Figs. I-A. I-B. and I-C: Technique of the Bankart procedure used in this series.

THE JOURNAL OF BONE AND JOINT SURGERY

 THE BANKART PROCEDURE 5

and resistive exercises are then begun. At six months, the

Muscles
patient should have from 75 to 100 per cent of normal mo-
tion and strength in the shoulder, and be in condition to Of the 161 patients, 130 (81 per cent) had normal
resume all activities including contact sports. muscle development, eighteen (11
cent) were thin per
and of slight build, and thirteen (8 per cent) were definitely
Operative Findings loose-jointed. In the 158 shoulders at surgery for which
The operative findings were documented adequately data were available, the subscapularis muscle appeared to
in 158 of the 162 shoulders. Since the subscapularis mus- be normal in 132 (83 per cent), “attenuated” or “in-
cle was not severed or divided while exposing the shoul- adequate” in fifteen (10 per cent), and definitely ruptured
der, but was carefully removed from its insertion and the (within the muscle belly) in eleven (7 per cent) (Fig. 3). Of
capsule, the condition of the muscle and capsule could be these ruptures, seven were in the lower half and four were
assessed. The size, shape, and inclination of the glenoid in the upper half involving the junction of the sub-
fossa were not recorded except for obvious abnormalities, scapularis and supraspinatus muscles.
nor did we determine the degree of retrotorsion of the
humeral head, as described by Saha47 and by Debevoise Capsule
and associates’4. The capsule was completely avulsed or separated

5of joint E Shoulder in complete

external rotation

3scapulari S
m.

horIzontally

Lateral edge of capsule

. Head of humerus

FIG. 1-B

VOL. 60-A, NO. I, JANUARY 1978

6 C. R. ROWE, DINESH PATEL, AND W. W. SOUTHMAYD

,ing lateral flap to

, . leave sutures

Suturing lateral flap of capsule

directly to glenoid rim A

.5 -

Then pass sutures through

medial capsule

Tying A to B and C to D Strongly enforcing

anterior glenoid rim

ichment of
ii d process
louble

Subscapularis
to its original
rn resutured
insertion
FIG. 1-C
I sutures

from the anterior glenoid rim in 135 shoulders (Fig. 4), an the rim in fifteen (13 per cent), and well developed but dis-
incidence of 85 per cent, and was normally attached to the placed into the joint across the glenoid (resembling a
rim of the glenoid, with the labrum intact, in twenty-three bucket-handle split of a meniscus in the knee) in seventeen
shoulders (15 per cent). The incidences of an intact cap- (14 per cent) of the 1 18 shoulders (Fig. 5).
sule ranged from 13 to 28 per cent in other series 1.9.13.17#{149}
Glenoid Rim
Considering all 158 shoulders, twenty-nine had a pouched,
stretched, or redundant capsule, which was normally at- Damage to the glenoid rim was present in 1 16 (73 per
tached to the glenoid rim in thirteen, completely separated cent) and absent in forty-two of the 158 shoulders. Sixty-
in twelve, and ruptured in four. five (56 per cent) of the damaged rims were eburnated or
eroded, three of them being in the atraumatic group, and
Labruin
fifty-one (44 per cent) were fractured. Ofthe fifty-one frac-
The condition ofthe labrum at operation was adequately tures, eighteen (35 per cent) involved one-sixth; twenty-
described in 1 1 8 shoulders It was absent or completely
. de- six (51 per cent), one-quarter; and seven (14 per cent),
stroyed in eighty-six (73 percent), intact but separated from one-third of the area of the joint surface.

THE JOURNAL OF BONE AND JOINT SURGERY

 THE BANKART PROCEDURE 7

A B C D E F
FIG. 2
Instruments used in the Bankart procedure. A, curved spike used to initiate holes in the glenoid rim; B, clamps with three-edged cutting points to
enlarge the hole; C, curved awls used to complete the hole; D, retractor for the medial capsular flap; and E and F, two types of humeral-head retractor
to fit different-sized heads and glenoid cavities.

Shoulders with No Bankart Lesion

tiple loose bodies, an incidence of 8 per cent, which was
In these twenty-three shoulders, the initial dislocation comparable to the 9 per cent reported by .

had been traumatic in nine and atraumatic in fourteen pa- A bone cyst was present in the neck of the scapula
tients . The capsule of the shoulder was ‘ ‘herniated’ ‘ or
‘ ‘redundant’ ‘ in thirteen (57 per cent) and normal in ten.
The subscapularis muscle was “deficient” in one of the
twenty-three shoulders, while the glenoid fossa was de-
scribed as deficient and shallow in two. A Hi11-Sachs le-
sion of the humeral head, present in seven (30 per cent) of
the twenty-three shoulders , was of moderate size in four;
severe, in one; and mild, in two. None of these twenty-
three shoulders showed separation of the capsule from the
rim of the glenoid or evidence of trauma to the rim.

Humeral Head

Of the 162 shoulders, 142 had satisfactory anteropos-

terior preoperative and postoperative roentgenograms
made with the arm in neutral position, in 60 degrees of in-
ternal rotation, in 60 degrees
and of external rotation. The
27 head defect (a compression fracture of the
humeral head due to impact against the anterior rim of the
glenoid) was present in 1 10 (77 per cent) and absent in
thirty-two (23 per cent). The defect was mild (Fig. 6-A) in
thirty (27 per cent), moderately severe in sixty-four (58
per cent), and severe (Fig. 6-B) in sixteen (15 per cent). Of
these 1 10 shoulders, 105 had had initial dislocations that
were traumatic.

Other Lesions
. . . Rupture (arrow) of the inferior half of the subscapularis muscle and
In our series, six shoulders had one and six had mul- the capsule of the left shoulder.

VOL. 60-A, NO. 1, JANUARY 1978

8 C. R. ROWE, DINESH PATEL, AND W. W. SOUTHMAYD

1
I

i/I

Avulseil

iLQ2!4!i.
ijnft’rior

FIG. 4
(iIS ‘

. . i
A typical Bankart lesion of the left shoulder with avulsion of the capsule from the anterior rim of the glenoid. The labrum s missing, apparently
completely worn away.

prior to surgery in two patients, both lesions being just

proximal to the articular surface. The cysts were not
explored at surgery and no histological diagnosis was es-
tablished. Clinically they appeared to be benign bone
cysts. After operation one had disappeared, and the other
was filling in at one year.
Muscle anomalies were seen in three shoulders. The
pectoralis major was absent in one and the pectoralis minor
inserted on the lesser tuberosity of the humerus in another,
the second such anomaly seen by one of us (C. R. R.). The
coracobrachialis and short head of the biceps muscle arose
from the rotator cuff in one patient. In another patient,
who had had a previous unsuccessful Bankart procedure,
the coracoid process was found to be ununited at the sec-
ond procedure.

Results

The results in the 145 shoulders evaluated were

graded excellent, good, and poor using both the rating
scale summarized in Table I and the patient’s own evalua-
tion of the shoulder. Based on the examining physician’s
evaluation, 108 (74 per cent) were graded excellent;
thirty-three (23 per cent), good; and four (3 per cent),
poor. The patients, however, rated their results as 120 (83
per cent) excellent, twenty-two (15 per cent) good, and
three (2 per cent) poor. Thus, by the surgeon’s rating there
FIG. 5 were 97 per cent excellent to good results and by the pa-
lntra-articular bucket-handle split of the labrum in the left shoulder. tients’ evaluations, 98 per cent excellent to good results.

THE JOURNAL OF BONE AND JOINT SURGERY

 THE BANKART PROCEDURE 9

FIG. 6-A
Figs. 6-A and 6-B: Typical Hill-Sachs lesion.
Fig. 6-A: A mild lesion of the left shoulder.

FIG. 6-B
A severe lesion, before (left) and after (right) reduction of a dislocation. Note the compression fracture of the superolateral aspect of the humeral
head.

Recurrences capsule was partially separated from the glenoid rim and there was a
moderate Hill-Sachs lesion. After the Bankart repair he had three recur-
There were five recurrences in our series of 145 rences within a year. He was then put on a program of specific resistive
shoulders, an incidence of 3.5 per cent. The recurrence exercises, and during the ensuing ten years he had no recurrences.
rates following surgical repair in most series reported since
CASE 2. An eighteen-year-oldman who was loose-jointed had had
1948 are summarized in Table II. The recurrences in our an atraumatic dislocationthe left shoulder
of initially. Operation re-
series are described in the following case reports. vealed a severe Hill-Sachs lesion and a redundant capsule that was sepa-
rated from the rim of the glenoid. Two years after surgical repair the pa-
CASE 1 . A twenty-year-old loose-jointed man dislocated the right tieni first noted a ‘ ‘popping-out’ ‘ of the shoulder on forceful hyperex-
shoulder initially without appreciable injury. At surgery in 1959, the tension ofthe arm; two nights later, while sleeping face down, the shoul-

VOL. 60-A, NO. 1, JANUARY 1978

10 C. R. ROWE, DINESH PATEL, AND W. W. SOUTHMAYD

der dislocated. This patient also was put on a program of resistive exer- operation. Of the other two patients who had had an un-
cises for the shoulder muscles, and the dislocation had not recurred one successful Magnuson procedure prior to their Bankart op-
year later.
eration, one had a good result one year later and the other
was lost to follow-up.
These two patients, with very lax ligaments, would
In the two patients who had had a Putti-Platt repair,
not be operated on at the present time unless they had
the capsule was found to be completely detached from the
failed to respond to a program of resistive exercises. We
glenoid rim in one, while in the other it was lax and the
have found that this type of shoulder instability responds
subscapularis muscle was “deficient’ ‘ . After routine Ban-
very well to exercises as the primary treatment, and that
kart repair these two patients had had no recurrences, one
the results of surgical repair are unpredictable.
and three years after operation.
In the other three patients, the postoperative recur-
In the two patients with failed Bankart repairs, the
rences were caused by severe trauma, sufficient to produce
capsule was detached from the rim in one; a DuToit staple
a primary dislocation.
had pulled out in the other, exposing a fracture of the
glenoid rim that involved approximately one-sixth of the
CASE 3 . A twenty-two-year-old man had had a traumatic dislocation
of the left shoulder initially, followed by many recurrences. At surgery, glenoid fossa, and there was also a severe Hill-Sachs le-
extensive damage to the capsule and glenoid rim was found but there was sion of the humeral head. After routine Bankart procedures
no Hill-Sachs lesion. Ten years after surgery, this man had one recur- both patients returned to full activities, one with an excel-
rence while roping a steer in a rodeo. However, during the next fifteen lent and the other with a good result.
years he had no dislocations. He graded his result as excellent even
The patient with the failed Nicola procedure had se-
though we had to grade it poor because the dislocation had recurred once
following surgery. vere separation of the capsule from the glenoid rim but had
an excellent result at follow-up, ten years after the Bankart
CASE 4 . A fifteen-year-old boy had had an initial traumatic disloca-
repair.
tion of the left shoulder followed by many recurrences. Operation re-
vealed a well developed Bankart lesion, a fracture of the glenoid rim, Of these eight patients whose previous surgery had
and a moderate-sized Hill-Sachs lesion. This boy was very belligerent not been successful, three had excellent and four had good
and had two recurrences of the dislocation during violent fights within a results after follow-ups ranging from one to ten years, and
few months after surgery. Several years later, he was killed in an au- one was lost to follow-up. In each instance, reoperation
tomobile accident.
disclosed adequate cause for failure.
CASE A thirty-four-year-old
5. male epileptic who had had many
dislocations of the right shoulder had a severe Hill-Shs lesion and se- Results in Shoulders without a Bankart Lesion
vere damage to the rim of the glenoid. After repair he sustained several
At operation in the twenty-three shoulders in which
dislocations during seizures within the first twelve months after surgery.
no Bankart lesion was found, holes were made through the
rim at the base where the medial capsule was attached. Su-
The results at follow-up in our series were correlated
tures were passed through these holes and into the lateral
with several factors, including previous unsuccessful re-
capsular flap and were used to attach the lateral flap se-
pair, absence of a Bankart lesion, fracture of the glenoid
curely to the glenoid rim. The medial flap of the capsule
rim, a Hill-Sachs defect, external rotation of the shoulder,
was then sutured over this as reinforcement. At follow-up
athletic activity , and epilepsy.
examination after an average of seven years (range, one to
Results after Failed Surgery twenty-five years), none had had a recurrence; thirteen
were graded excellent and eight, good. The other two pa-
Eight patients were referred to us because of recurrent
tients were lost to follow-up.
dislocations after surgical repair. Three had had a Magnu-
son repair; two, a Putti-Platt procedure; two, a Bankart re-
pair; and one, a Nicola repair. Seven had had a traumatic Results in Shoulders with Fracture of the Glenoid Rim

dislocation initially and one, an atraumatic dislocation. All There were fifty-one shoulders with a fracture of the
had normal musculature; none were loose-jointed. In the anterior glenoid rim. Of the eighteen with one-sixth of the
three patients who had had a Magnuson repair, at reopera- glenoid fossa involved, ten were graded excellent, five
tion the capsule was found to have been avulsed from the were good, and three were lost to follow-up. None had re-
glenoid rim. One of the three had been a very promising currences.
collegiate basketball center, but was unable to play after Of the twenty-six with one-fourth of the glenoid fossa
the Magnuson repair because of recurrent dislocation of avulsed, fifteen had an excellent and eight, a good result,
the non-dominant shoulder. At reoperation the sub- while one had a recurrence. The other two were lost to
scapularis muscle was released from its transplanted posi- follow-up. Of the other seven shoulders with one-third of
tion, revealing complete avulsion of the capsule from the the glenoid fossa fractured off, five were graded excellent
rim of the glenoid. After Bankart repair of the type de- and two were lost to follow-up.
scribed, this patient again played basketball, and was Of these fifty-one shoulders, forty-four were re-
named to the All-New England College Team during his examined one to twenty-five years after repair (an average
senior year (Fig. 8). At follow-up, his shoulder had a full follow-up of ten years); forty-three (98 per cent) were
range of motion, was strong, and had not dislocated since rated excellent to good (69 per cent excellent and 29 per

THE JOURNAL OF BONE AND JOINT SURGERY

 THE BANKART PROCEDURE 11

cent good) and one had had a recurrence. Therefore, the twenty-two and good in eight after follow-up of one to
rate of recurrence in this group was 2 per cent, which, twenty-six years (average, 6.3 years). In the sixty-four
surprisingly, was lower than the over-all recurrence rate of shoulders with a moderately severe Hill-Sachs lesion,
3.5 per cent. there were three recurrences (4.7 per cent) and forty excel-
In these shoulders, no bone grafts or muscle trans- lent, twenty-one good, and three poor results after an aver-
plants were used to reinforce the glenoid rim, and the frac- age follow-up of 5.3 years (range, one to twenty-three
ture fragment was not replaced. Rather, it was either ex- years). In the sixteen shoulders with severe Hill-Sachs le-
cised or left in the medial flap of the capsule, while the sions, there was one recurrence (6 per cent incidence).
lateral flap was sutured in the usual way to the margin of There were eleven excellent, four good, and one poor re-
the intact part of the glenoid fossa. A stable shoulder with suit after an average follow-up of three years (range, one
an excellent range of motion and strength was obtained in to eight years).

TABLE II
PATHOLOGY, 1948 TO 1976

Trauma to
Bankart Abnormalities of Hill-Sachs Anterior Incidence of
Series Lesion Subscapularis Lesion Glenoid Rim Recurrence
(Per cent) (Per cent) (Per cent) (Per cent) (Per cent)

Adamst (1948) 87 None 82 “Unusual” 5.5

Bateman (1972) 2.1
Boyd and Hunt8 (1965) 4.1
Bray9 (1955) 86 32 7.3
Connolly (1969) 98 10.0
D’Angelo’2 (1970) 100 “Infrequent”, 0cc. 100 “Extremely frequent”: 1.7
atrophy fract. 31% ,
De Anquin 3 (1965) 72 None 100 Erosion. frequent: 0.7
fract. , 2%
DePalma”t (1973) 45 ‘ ‘Lax” in 100% 75 Erosion, 46%; fract. , 8.7
11%
Dickson and (1955 and 64 4.0
1957) (Bankart’s own patients)
DuToit2 (1976) 98 26 24 7.0
DuToit and Roux2’ (1955) 99 33 31 5.0
Eyre-Brook22 (1948) 76 65 0
Gallie and LeMesurier24 (1948) 4.0
Helfet2 (1958) 7.0
Hermodsson2” (1963) 100
Lindholmtt (1974) 4.0
Lombardo and associates3t (1976) 2.0
May35 (1970) 0
Morrey and Janes#{176} ( 1976) 53 31 2.0 (1963)
11.0 (1975)
Moseley and Overgaard ( 1962) 84 ‘Lax in all cases” 100 1.0
Osmond-Clarke4t ( 1965) 90 89 1.4
Palmer and Widen42 (1948) 45 100 Fract., 20% 7.0
Quigley and Freedman44 (1973) 27 52 5.1
Rowe and associates (present series) 85 17 77 Erosion, 73%; fract. . 3.5
(1976) 44%
Saha4’ (1969) 25 2
Symeonides4’ (1972) 62 “Laxineverycase” 53 Fract.. 18% 3.0
Viek and BeIl2 (1959) 80 51 2.6
Watson-Jonest ( 1948) 70 2.0

each instance after the lateral flap was sutured securely to Thus, there were four recurrences among the eighty
the fractured margin of the glenoid and the medial flap was shoulders with moderate to severe Hill-Sachs lesions, an
sutured over it as reinforcement. incidence 5 per cent, of which would indicate that the
presence of a sizable Hiii-Sachs lesion is a more important
Results ii I JO Shoulders with Hill-Sachs Defects
factor causing instability after a Bankart procedure per-
In the thirty shoulders with a mild Hill-Sachs lesion, formed by the technique described than is a fracture of the
there were no recurrences and the results were excellent in anterior glenoid rim. However, a 5 per cent recurrence

VOL. 60-A, NO. I, JANUARY 1978

12 C. R. ROWE, DINESH PATEL, AND W. W. SOUTHMAYD

rate, compared with the over-all recurrence rate of 3.5 per shoulder repair, including two college pitchers, two col-
cent, is an acceptable incidence. lege catchers, one triple-letter man in college, one profes-
sional basketball player, one collegiate tennis champion,
Results in Shoulders with Complete
seven football players, and one backstroke swimming
External Rotation Restored champion at the U.S. Naval Academy. One patient be-
Of the 124 patients whose shoulders were evaluated came a Marine and was in combat in Korea (Figs. 7-A,
by personal interview and examination, eighty-six (69 per 7-B, and 7-C). Only one ofthe forty-six patients in whom
cent) had complete elevation and external rotation com- the dominant shoulder was repaired failed to return to his
pared with the opposite shoulder, and only two (2 per cent) original sports activities.
of these eighty-six shoulders had redislocated after surgi- Of the thirty-one patients in whom the non-dominant
cal repair. Another thirty (24 per cent) of the 124 shoulders shoulder was repaired, only one was not able to return to
had regained 75 per cent of normal external rotation and the sports activities in which he had participated before his
had had no recurrences. Thus, of the 1 16 shoulders with injury. The other thirty had no limitations and in many in-
return of 75 to 100 per cent of normal external rotation, stances were superior athletes, including ten college foot-
only two (1.7 per cent) had recurrent dislocation, while ball players, eight three-letter men, five competitive
two of the other eight that had regained 50 per cent of ex- swimmers, one hammer-thrower who placed third in the
ternal rotation or less had had a recurrence, an incidence of Olympic tryouts in the East, an All-New England basket-
25 per cent. The return of complete external rotation fol- ball center (Fig. 8), a member of the U.S. Olympic Ski
lowing surgery therefore was not associated with an in- Team, a college hockey goal-tender. and two college
creased incidence of instability or recurrent dislocation of weight-lifters.
the shoulder, but rather with a lower incidence. These results compare favorably with those reported
by other investigators. Of Lombardo and associates’
Return to Athletic Activity twenty-seven patients with the dominant shoulder in-
Seventy-seven of our 161 patients had been involved volved, none returned to their original level of perform-
in athletics prior to their shoulder injury, which was on the ance after a modified Bristow procedure . All of the pa-
dominant side in forty-six and on the non-dominant side in tients of Gallie and 24 and 88 per cent of those
thirty-one. Of the forty-six patients whose dominant side of Palmer and Widen ‘ returned to their normal activities,
was involved, thirty had engaged in throwing sports. Fol- while in the series of Morrey and Janes iii 5 per cent gave
lowing surgical repair, ten (33 per cent) of the thirty were up sports and 22 per cent were forced to limit their athletic
able to throw or pitch a baseball as hard and a football as activities.
far as they had before injury, and they could serve hard in
Results in Epileptics
tennis, swim hard with an overhead stroke, or “spike”
(forcefully hit the ball downward over the net) while play- There were four patients in our series whose initial
ing volleyball. The other twenty (67 per cent) could throw dislocation and subsequent recurrences were caused by
a football or softball hard and serve hard in tennis, but epileptic seizures. All had anterior dislocations (epileptics
could not throw a baseball as hard as formerly. Some of usually have posterior dislocations). Three had severe and
these forty-six patients had become superior athletes after one, a moderately severe Hill-Sachs lesion. All four had

FIG. 7-A FIG. 7-B FIG. 7-C

Figs. 7-A. 7-B. and 7-C: Fifteen years following Bankart repair of the right (dominant) shoulder. Six months after surgery this patientjoined the
U.S. Marines and fought in combat in Korea. At present he is a competitive swimmer and tennis player and has full strength and motion. There have
been no recurrences.

THE JOURNAL OF BONE AND JOINT SURGERY

 THE BANKART PROCEDURE 13

complete separation of the capsule from the glenoid rim, months after surgery. After removal of silk sutures the
and one had a severe fracture of the anterior glenoid rim. sinuses promptly closed up, and there had been no sign of
After repair, two patients had had no recurrences and had deep sepsis or other sinuses during the ensuing twelve
good shoulder function at two and five years, despite con- years while he continued full-time work as a carpenter.
tinuing seizures; one (Case 5) had had recurrences; and After this case we began to use cotton sutures routinely,
one was lost to follow-up. and have seen no further reaction to suture material. One
patient had a postoperative hematoma that necessitated
Lou’ Roeiitgenogi’aphic Changes
evacuation and closure. His wound then healed unevent-
of the Glenohuineral Joint
fully. This patient was found to have a qualitative platelet
In the 124 patients seen personally at follow-up, there defect due to aspirin that he had taken preoperatively. Four
was no evidence of late degenerative changes in the other patients had mild postoperative hematomas which
glenohumeral joint or of myositis ossificans. We believe absorbed. One patient had thrombophiebitis of the
cephalic vein which cleared up with warm compresses.
Another had a weak deltoid muscle postoperatively, which
gradually improved. Electromyograms confirmed that his
axillary nerve was intact. There were no non-unions of the
osteotomized coracoid process after repair by the method
described.

Discussion

There was no evidence in this series that there is a

single essential lesion responsible for recurrent disloca-
tions of the shoulder. However, the commonest findings at
surgery were separation of the capsule from the anterior
glenoid rim (85 per cent), different degrees of the Hill-
Sachs lesion of the humeral head (77 per cent), and dam-
age to the anterior glenoid rim (73 per cent). In twenty-
three (14.2 per cent) of the shoulders there was flO cvi-
dence of the so-called Bankart lesion. In these patients, the
capsule was redundant in nine, the subscapularis muscle
was deficient in one, the glenoid fossa was deficient in
two, and a Hill-Sachs lesion of the humeral head was pres-
ent in seven. No findings other than absence ofthe Bankart
lesion were noted in the remaining four shoulders.
The pathological lesions observed at operation for re-
current dislocation of the shoulder in twenty-eight series
are summarized in Table II. The high incidence of
pathological changes in the subscapularis muscle de-

F1;. 8 scribed by other authors 15,16,48 was not observed in this

One year after a Bankart procedure on the left (for a failed Magnuson
series. The towel-clip test for laxity of the subscapularis
procedure), this patient was elected All-New England center in college muscle employed by these investigators did not seem to be
basketball.
an accurate way to determine the functional state of the
subscapularis muscle. In our series, obvious thinning and
that in the Bankart procedure used in this study, the soft attenuation of the subscapularis was present in 10 per cent
tissues are not traumatized and stability does not depend and direct rupture, in 7 per cent of the shoulders. Since in
Ott scar tissue. Surgical trauma is lessened because the our patients this muscle was normal in most shoulders and
holes in the glenoid rim are small in diameter and because at the close of the procedure it was returned to its original
the instruments are specially designed and smaller than insertion without advancement or shortening, we doubt
those ordinarily used (Fig. 2). After Eden-Hybbinette re- that abnormality of the subscapularis is an essential lesion.
pairs, Lindholm found that 8 per cent of the shoulders Adams ‘ also found no abnormalities of the subscapularis,
showed osteoarthritic changes of the glenohumeral joint and noted that the capsule was stripped from the front of
and 4 per cent had myositis ossificans, but he did not mdi- the glenoid neck but was otherwise normal. Dc Anquin ,

cate how long the patients were followed in his review of 150 operations for recurrent dislocations
of the shoulder, also did not find sufficient lesions of the
Complications
subscapularis muscle to account for the shoulder instabil-
Iii one patient, a forty-seven-year-old carpenter who ity.
was operated Ofl in 1953, two sinuses developed ten The labrum appears to be a variable structure. Town-

VOL. 60-A, NO. 1. JANUARY 978

14 C. R. ROWE, DINESH PATEL, AND W. W. SOUTHMAYD

ley l, Moseley and Overgaard 37,38, and DePaima i5.16 lower than in the whole series (3.5 per cent). Eighty-four
concluded that it is an extension of the capsular ligament per cent of these fifty-one patients also had a Hill-Sachs
along the rim of the gienoid. In our series, it was totally lesion of the humeral head. The effectiveness of the Ban-
absent or destroyed in 73 per cent of the shoulders, well kart repair in patients with a fracture of the gienoid rim,
formed in 13 per cent, and displaced into the joint (re- which was a surprise to us, emphasizes the importance of
sembling a bucket-handle tear of a meniscus in the knee) in reconstructing a stable capsular barrier to the humeral head
14 per cent (Fig. 5). Palmer and Widen “ found a 3 per along the anterior rim of the glenoid.
cent incidence of bucket-handle tears of the labrum in their Opinions as to the effect of the Hill-Sachs lesion on
series, and DuToit and Roux 21 , an 1 1 per cent incidence. the stability of the shoulder after Bankart repair vary.
Watson-Jones 52, Bateman , and BraV mentioned Palmer and Widen “ stated that the Hiil-Sachs defect was
bucket-handle tears of the labrum, but did not give their the essential lesion of recurrent anterior dislocation, and
incidence. We agree with D’Angelo 12 that the labrumper that when it is present, dislocation may recur even after the
se plays a minor role in stability of the shoulder. capsule and labrum have been repaired unless external ro-
Separation of the capsule from the rim of the gienoid tation is restricted, preventing the defect from coming in
was the most significant and frequently found lesion in our contact with the glenoid rim. They recommended placing a
series (85 per cent) (Fig. 4). However, Magnuson con- bone graft at the glenoid rim (Hybbinette-Eden technique)
cluded that the capsule of the shoulder has nothing what- to prevent the head defect from slipping over the rim.
ever to do with holding the head of the humerus in the Connolly transplanted the tendon of the infraspinatus
glenoid, and others 16.48 agreed with him. Conversely, the muscle into the head defect, using the procedure described
first line of defense against recurrent anterior dislocation by McLaughlin for recurrent posterior dislocations . In
was considered to be strong reinforcement of the capsule our series, the size of the head defect did influence the in-
along the anterior rim of the glenoid by many other au- cidence of recurrence since the recurrence rate was 4.7 per
thors, whether by a bone block 13,22.28,41 fascial rein- cent in the presence of a moderately severe defect and 6
forcement 5.24, metal implants 37.38, bone-pedicie trans- per cent in the presence of a severe defect. These rates
plants 25.35 muscle and capsule reinforcement 1.2.9.39,40, or compare favorably with the recurrence rate of 7 per cent in
direct suture of the capsule to the rim of the glenoid or Palmer and Widen’s series ‘ and the 10 per cent rate in
neck of the scapula i4.8.iO.17.18.20.2i .23.42.4446.52W Connoliy’s series ‘.
The question is often asked, ‘ ‘ Does shortening of the Although it has been stated that return of complete ex-
capsule account for the effectiveness of the Bankart proce- ternal rotation of the shoulder following surgical repair
dure, rather than its reattachment to the glenoid rim?” is associated with an increased incidence of recur-
Shortening undoubtedly is a factor, especially in patients rence 6.i6.4i , only two (2 per cent) of our eighty-six pa-
in whom no Bankart lesion is found at surgery. In our se- tients with complete external rotation and a complete range
ries, we did not deliberately attempt to shorten the capsule of motion had recurrences, and none of the thirty patients
or restrict external rotation of the shoulder. However, with 75 per cent of normal external rotation had a recur-
whenever the capsule is opened and repaired it must be rence. Conversely, in our eight patients whose external ro-
shortened to some extent. To avoid restricting external ro- tation was limited to less than 50 per cent of normal, two
tation, as previously noted, the shoulder should be exter- had recurrences, an incidence of 25 per cent. Therefore, in
nally rotated completely before the vertical incision is our series the return of maximum external rotation was as-
made in the capsule just lateral to the glenoid rim (less than sociated with an increase rather than a decrease in stabil-
0.5 centimeter). By doing this, the shoulder with the arm ity. We found in this follow-up study that any restriction of
at the side could be rotated 25 to 30 degrees with ease at external rotation can be a handicap in athletes who need
the end of the operative procedure in this series. complete elevation and external rotation in such above-
Although fractures of the anterior rim of the gienoid the-shoulder activities as serving in tennis, pitching a
were noted by several investigators, their role in recurrent baseball, throwing a football, making a lay-up in basket-
dislocation has not been established. D’Angelo 12 found a ball, swimming, and gymnastics. Some types of work,
3 1 per cent incidence of fractures of the anterior glenoid such as plastering, painting, and paper-hanging, also re-
rim in his series, while the incidence found by Palmer and quire full shoulder motion.
Widen ‘ was 20 per cent; by Symeonides 48, 18 per cent; A frequent question is, ‘ ‘What does one do when no
and by DePalma 15.16 1 per cent. DePalma stated that un- Bankart lesion is found at surgery?’ In the small group of ‘

less the fractured glenoid fossa was built up by a bone twenty-three such patients in our series, the most consis-
graft, it ‘ ‘ may be virtually impossible to restore muscle tent operative findings were a “herniated’ ‘ or “redun-
balance’ ‘ but this statement
16 was not substantiated in our dant’ capsule
‘ in 5per cent and a Hill-Sachs lesion of the
fifty-one patients who had fractures of the glenoid rim in- humeral head in 30 per cent. There were evidently other
volving from one-sixth to one-third of the glenoid fossa. factors that we did not identify, such as neuromuscular
After direct suture of the capsule to the remaining glenoid imbalance (as emphasized by DePalma 15.16 and
rim with no bone grafts or transplants, only 2 per cent of Symeonides 48), and retrotorsion of the humerus (as de-
these dislocations recurred, an incidence 1 .5 per cent scribed by Saha 47) . The most reasonable procedure to

ThE JOURNAL OF BONE AND JOINT SURGERY

 THE BANKART PROCEDURE 15

carry out in this group, it seemed to us, was to reinforce the second half of the series (eighty patients with an aver-
the capsule along the anterior rim of the glenoid as already age follow-up of three years) and three to six weeks of
described. This evidently was effective, since after an av- immobilization was used in the first half (sixty-five pa-
erage follow-up of eight years eighteen patients had had no tients with an average follow-up of 9.5 years). Of the five
recurrence. One had had recurrences and four were lost to recurrences, two were in the former group and three, in the
follow-up. latter group.
Also asked is the question, ‘ ‘ What have been the Consequently , postoperative immobilization did
findings in shoulders in which no Hill-Sachs lesion was not seem to be a significant factor.
present, and how should they be treated?’ ‘ In our series Is the incidence of recurrence following surgical re-
there were twenty-nine shoulders in which no Hill-Sachs pair higher in patients with a family history of shoulder
lesion was found. Of these shoulders, twenty-three (72 per dislocations? Morrey and Janes 36 reported that 30 per cent
cent) did and nine (28 per cent) did not have a typical Ban- of the postoperative recurrences in their series were in
kart lesion. Therefore, a Bankart lesion was found more such patients. Information concerning family history was
frequently in shoulders with no Hill-Sachs lesion (72 per available in only one of our five patients with recurrence.
cent) than a Hi11-Sachs lesion was found in shoulders with In that instance, no one in the family had had a shoulder
no Bankart lesion (30 per cent). The recurrence rate in dislocation.
shoulders with no Hill-Sachs lesion was 4 per cent, almost In our five patients whose shoulder dislocations re-
as high as the rate in those with a Hill-Sachs lesion (5.4 per curred, there were several significant factors that appeared
cent) and much higher than the rate in shoulders with no to contribute to the recurrence. Two of the five were
Bankart lesion (zero per cent). loose-jointed and their initial dislocation and subsequent
Money and Janes 36 cautioned against a short-term recurrences had been atraumatic or produced by minimum
follow-up study, pointing out that such a report from the trauma. The experience of DuToit and Roux 20 was similar
Mayo Clinic in 1 949 gave a recurrence rate of only 1 .4 per to ours, three of their seven recurrences being in patients
cent, whereas a subsequent report from the same clinic, with excessive ligament laxity. Our present approach to
with long-term follow-up, showed a recurrence rate of 11 this type of shoulder instability is to start the patient on a
per cent. We agree that a short-term follow-up report can schedule of specific resistive exercises to the shoulder. In
be misleading; however, in our series, which included the majority of our patients the shoulder instability was
forty-eight patients followed for five to thirty years, there eliminated after muscle strength improved, and surgery
were five recurrences: three within one year of surgery, was not needed.
one after two years, and one after ten years. Of the eight The other three patients whose dislocations recurred
patients referred to us because of recurrence after repair, had major trauma to the shoulder after repair, similar to
the recurrence had been sustained during the first post- that which produced the initial dislocation (roping a steer
operative year in three, within two years of surgery in at a rodeo, an epileptic fit, and a violent fight). One of
three, and after five years in one (leaving one patient for these patients had a single recurrence and after resistive
whom no information was available). Therefore, of these exercises for the shoulder muscles he had no more recur-
eleven postoperative recurrences, nine (82 per cent) oc- rences during the ensuing fifteen years.
curred within two years of surgery. The lessons learned from this study, we believe, are
Morrey and Janes 36 suggested that a short period of that if the meticulous technique described is used, de-
postoperative immobilization may be a factor contributing generative changes in the joint can be avoided, as well as
to an increased recurrence rate. This, we think, would de- myositis ossificans. In addition, the patient can regain a
pend on the type of surgical repair employed. Procedures full range of shoulder motion and return to full participa-
such as muscle and tendon transplants (Bristow and Mag- tion in sports: ( 1) if the capsule is incised vertically just
nuson operations) or bone-block operations (Hybbinette- lateral to the glenoid rim while the shoulder is held in full
Eden and De Anquin repairs) would require a period of external rotation, thereby ensuring that the repaired an-
immobilization long enough to ensure healing of the terior capsule is not too tight; and (2) if the shoulder is not
transplanted tendon or bone. With the technique used in immobilized postoperatively, so that early resumption of
our series, no postoperative immobilization was used in motion and function is possible.

References
1. ADAMS, J. C.: Recurrent Dislocation of the Shoulder. J. Bone and Joint Surg. 30-B: 26-38, Feb. 1948. ,
2. ADAMS, J. C.: Personal communication, 1976.
3. BANKART, A. S. B.: Recurrent or Habitual Dislocation of the Shoulder-Joint. British Med. J., 2: 1132-1 133, 1923.
4. BANKART, A. S. B.: The Pathology and Treatment of Recurrent Dislocation of the Shoulder-Joint. British J. Surg.. 26: 23-29, 1938.
5. BATEMAN, J. E.: The Shoulder and Neck. Philadelphia, W. B. Saunders, 1972.
6. BONNIN. J. G.: Transplantation of the Tip of the Coracoid Process for Recurrent Anterior Dislocation of the Shoulder. In Proceedings of the
North-West Metropolitan Orthopaedic Club. J. Bone and Joint Surg. 51-B: 579, Aug. 1969. ,
7. BosT, F. C. and INMAN, . V. T.: The Pathological Changes in Recurrent Dislocation ofthe Shoulder. A Report of Bankart’s Operative Procedure.
J. Bone and Joint Surg. , 24: 595-613, July 1942.
8. BOYD, H. B., and HUNT, H. 0.: Recurrent Dislocation ofthe Shoulder. The Staple Capsulorrhaphy. J. Bone and JointSurg., 47-A: 15 14-1520,
Dec. 1965.
9. BRAy, E. A.: An Evaluation of the Putti-Platt Reconstruction Procedure for Recurrent Dislocation of the Shoulder. J. Bone and Joint Surg.,
37-A: 731-741, July 1955.

VOL. 60-A, NO. 1, JANUARY 1978

 V

16 C. R. ROWE, DINESH PATEL, AND W. W. SOUTHMAYD

10. CAVE, E. F., and RowE, C. R.: Capsular Repair for Recurrent Dislocation ofthe Shoulder: Pathological Findings and Operative Technic. Surg.
Clin. North America, 27:
1289-1294, l947.
I I . CoNN0II.Y, J. F.: Humeral Head E)efects Associated with Shoulder Dislocations. Thesis, Department of Orthopedic Surgery, Vanderbilt Uni-

12.
13.

14.
versity
D’ANGEIo.
DE ANQUIN,
thopaedic
DEKEVOISE.
School

C.
Surgeons.
of Medicine.
D.: Luxacao
E.: Recurrent
Nashville,
recidivante
Dislocation
J. Bone and Joint Surg.
N. T.; HYATT, G. W.: and TOWNSEND.
Tennessee.
anterior
1969.
de ombro.
of the Shoulder
Thesis.

47-A: 1085, July 1965.

G. B.: Humeral
.
Rio de Janeiro,
Roentgenographic

Torsion
-Brazil,

in Recurrent
1970.
Swdy. In Proceedings

Shoulder Dislocations.
of The

A
American

Technic
Academy

ofDetermination
of Or-

by
X-Ray. Clin. Orthop.. 76: 87-93. 1971.
IS. DFPAI MA. A. F.: Surgery of the Shoulder. Philadelphia. J. B. Lippincott. 1950.
l6. DEPALMA. A. F.: Surgery of the Shoulder. Ed. 2, p. 416. Philadelphia, J. B. Lippincott, 1973.
17. DICKSON. J. W.. and DEVAS, M. B.: Recurrent Dislocation ofthe Shoulder. In Proceedings ofthe British Orthopaedic Association. J. Bone and
Joint Surg.. 37-B: 721, Nov. 1955.
18. DICKSON. J. W., and DEvAS, M. B.: Bankart’s Operation for Recurrent Dislocation ofthe Shoulder. J. Bone and Joint Surg.. 39-B: 114-119,
Feb. l957.
19. DIscussIoN ON RECURRENT DISLOCATIONS OF THE SHoUlDER: J. Bone and Joint Surg., 30-B: 7-8, 46-48, Feb. 1948.
20. DUTOIT, G. T.: Personal communication, 1976.
21 . DUTOIT. 0. 1., and ROUX, D.: Recurrent Dislocation ofthe Shoulder. Twenty-four Year Study ofthe Johannesburg Stapling Operation. J. Bone
and Joint Surg. .
37-A: 633, June 1955.
22. EDEN, RuDolF: Zur Operation der habituellen Schulterluxation unter Mitteilung eines neuen Verfahrens bei Abriss am inneren Pfannenrande.
Deutsche Zeitschr. f. Chir. , 144:
268-280, 1918.
23. EYRE-BROOK, A. L.: Recurrent Dislocation of the Shoulder. Lesions Discovered in Seventeen Cases, Surgery Employed, and Intermediate Re-
port on Results. J. Bone and Joint Surg. 30-B: , 39-46. Feb. 1948.
24. GAI.LIE. W. E.. and LEMESURIER, A. B.: Recurring Dislocation of the Shoulder. J. Bone and Joint Surg., 30-B: 9-18, Feb. 1948.
25. HEI.FE1 ,
A. J.: Coracoid Transplantation for Recurring Dislocation of the Shoulder. J. Bone and Joint Surg. 40-B: ,
198-202, May 1958.
26. HERMODSSON, I.: Roentgenographic Studies of Traumatic and Recurrent Dislocation of the Shoulder. Edited by H. F. Mosely and B. Overgaard.
Montreal, McGill University Press, 1963.
27. HILl., H. A., and SACHS, M. 0.: The Grooved Defect of the Humeral Head. A Frequently Unrecognized Complication of Dislocations of the
Shoulder Joint. Radiology, 35: 69()-700, 1940.
28. HYBBINETTE. S.: De Ia transplantation dun fragment osseux pour rem#{233}dier aux luxations r#{233}cidivantes de l’#{233}paule; constatations et r#{233}sultats
operatoires. Acta Chir. Scandinavica. 71: 41 1-445, 1932.
29. JENS. J.: The Role of the Subscapularis Muscle in Recurrent Dislocation of the Shoulder. In Proceedings of the Joint Meeting of Orthopaedic
Associations. J. Bone and Joint Surg., 46-B: 780-781 Nov. ,1964.
3(). LINDHOLM, T. S.: Results of Treatment for Anterior Recurrent Dislocation of the Shoulder Joint with Eden-Hybbinette Operation. Acta Orthop.
Scandinavica, 45: 508-517. 1974.
31 . LOMIARDO. S. J.; KERLAN, R. K., JOKE, F. W.; CARTER, V. S.; BLAZINA, M. E.; and SHIELDS, C. L., JR.: The Modified Bristow Procedure for
Recurrent Dislocation of the Shoulder. J. Bone and Joint Surg., s-A: 256-261 ,
March 1976.
32. MAGNUSON, P. B.: Treatment of Recurrent Dislocation of the Shoulder. Surg. Clin. North America, 25: 14-20, Feb. 1945.
33. MAGNUSON, P. B.: Discussion’.
34. MAGNUSON. P. B.. and STACK. J. K.: Bilateral Habitual Dislocation ofthe Shoulders in Twins. A Familial Tendency. J. Am. Med. Assn., 114:
2103. 1940.
35. MAY. V. R.. JR.: A Modified Bristow Operation for Anterior Recurrent Dislocation ofthe Shoulder. J. Bone and Joint Surg., 52-A: lOlO-l0l6,
July 1970.
36. MORREY. B. F.. and JANES. J. M.: Recurrent Anterior Dislocation ofthe Shoulder. Long-Term Follow-up ofthe Putti-Platt and Bankart Proce-
dures. J. Bone and Joint Surg. .
58.A: 252-256, March 1976.
37. MOSEI.EY. H. F.: Shoulder Lesions. Ed. 3, pp. 140-205. Edinburgh, E. and S. Livingstone. 1969.
38. MOSEI.EY. H. F.. and OVERGAARD, B.: The AnteriorCapsular Mechanism in Recurrent Anterior Dislocation ofthe Shoulder. Morphological and
Clinical Studies with Special Reference to the Glenoid Labrum and the Gleno-humeral Ligament. J. Bone and Joint Surg., 44-B: 913-927, Nov.
1962.
39. OSMOND-CLARKE. H.: Habitual Dislocation of the Shoulder. The Putti-Platt Operation. J. Bone and Joint Surg.. 30-B: 19-25, Feb. 1948.
441 OSMOND-CIARKE, H.: Recurrent Dislocation of Shoulder. in Proceedings of the Royal Air Force Orthopaedic Service: Annual Conference. J.
Bone and Joint Surg. .
47-B: 194. Feb. 1965.
41 . PAlMER. IvAR. and WIInN, ANDERS: The Bone Block Method for Recurrent Dislocation of the Shoulder Joint. J. Bone and Joint Surg.. 30-B:
53-58. Feb. 1948.
42. PERTHES: Uher Operationen bei habitueller Schulterluxation. Deutsche Zeitschr. f. Chir., 85: 199-227, 1906.
43. QUIGIEY. 1. B.. andFREEDMAN. P. A.: Recurrent Dislocation ofthe Shoulder. Read at a meeting honoring R. M. Zollinger. Ohio State Univer-
sity. Columbus. Ohio, October 20. 1973.
44. ROWE. C. R.: Prognosis in Dislocations of the Shoulder. J. Bone and Joint Surg., 38-A: 957-977, Oct. 1956.
45. ROWE. C. R.: The Surgical Management of Recurrent Anterior Dislocations of the Shoulder using a Modified Bankart Procedure. Surg. Clin.
North America, 43: 1663-1666, 1963.
46. ROWE. C. R .: The Results of Operative Treatment of Recurrent Dislocations ofthe Shoulder. Surg. Clin. North America. 43: 1667- 1670. 1963.
47. SAHA. A. K.: Recurrent Anterior Dislocation of the Shoulder; A New Concept. Calcutta. Academic Publishers. 1969.
48. SYMEONIDES, P . P .: The Significance of the Subscapularis Muscle in the Pathogenesis of Recurrent Anterior Dislocation ofthe Shoulder. J. Bone
and Joint Surg. . 54-B: 476-483. Aug. 1972.
49. TosvNIEY. C. 0.: The Capsular Mechanism in Recurrent Dislocation of the Shoulder. J. Bone and Joint Surg., 32-A: 370-380, April 1950.
50. TOWNLEY, C. 0.: The Dislocating Shoulder: Analomy, Pathology, Pathogenesis. in Proceedings of The American Academy of Orthopaedic
Surgeons. J. Bone and Joint Surg.. 45.A: 1335-1336. Sept. 1963.
SI . VIEK, PErER. and BElt., B. T.: The Bankart Shoulder Reconstruction. The Use ofPull-Out Wires and Other Practical Details. J. Bone and Joint
Surg. . 41-A: 236-242. March 1959.
52. WATSON-JONF.S, REGINAI I): Note on Recurrent Dislocation of the Shoulder Joint. Superior Approach Causing the Only Failure in Fifty-Two
Operations for Repair of the Labrum and Capsule. J. Bone and Joint Surg. 30-B: ,49-52, Feb. 1948.

ThE JOURNAL OF BONE AND JOINT SURGERY