Abstract
Background: Internal rotation contractures due to external rotation
weakness secondary to brachial plexus birth palsy frequently lead to
glenohumeral deformity and impaired shoulder function. Our surgical approach
to treat these contractures relies on arthroscopic release for young children
(less than three years old) and combines arthroscopic release with latissimus
dorsi transfer for older children. We report the results for the first
thirty-three children followed for a minimum of two years after such
treatment.
Methods: Nineteen children with a mean age of 1.5 years (all younger
than three years of age) underwent arthroscopic contracture release as the
only primary procedure, and fourteen children with a mean age of 6.7 were also
treated with a latissimus dorsi transfer. Passive external rotation with the
arm at the side and passive and active elevation were measured for all
patients preoperatively. Passive and active external rotation, internal
rotation, and elevation were measured for all patients postoperatively.
Magnetic resonance imaging was performed preoperatively and postoperatively to
evaluate the status of the glenohumeral joint.
Results: Preoperative passive external rotation averaged -2° for
the children who underwent arthroscopic contracture release only and -24°
for those who also were treated with a latissimus dorsi transfer. Arthroscopic
release achieved a marked increase in passive external rotation and a centered
position of the glenohumeral joint at the time of surgery in all but the
oldest child in the series, who had severe deformity. The contracture recurred
in four of the younger children who had an isolated release, and this was
treated with a repeat arthroscopic release and a secondary latissimus dorsi
transfer. None of the children who had a primary latissimus dorsi transfer had
recurrence of the contracture.
At the time of follow-up, the mean passive external rotation was increased
by 67° (p < 0.005) in the fifteen children with a successful
arthroscopic release, 81° (p < 0.005) in those treated with a primary
latissimus dorsi transfer, and 78° in the four patients who were treated
with a late latissimus dorsi transfer because the isolated arthroscopic
release failed. The mean active elevation increased 12°, 3°, and
10°, respectively, in the three groups. Internal rotation was not measured
consistently preoperatively, but when it had been it was found to have
decreased substantially postoperatively. Magnetic resonance imaging performed
prior to the surgery showed a pseudoglenoid deformity in eighteen of the
children. At two years, magnetic resonance images were available for fifteen
of those children, and twelve of the images showed marked remodeling of the
deformity.
Conclusions: In children who are younger than three years of age,
arthroscopic release effectively restores nearly normal passive external
rotation and a centered glenohumeral joint at the time of surgery. In most of
these children, external rotation strength is sufficient to maintain this
range of motion and to improve glenoid development when preoperative deformity
was present. The addition of a latissimus dorsi transfer in older children
predictably results in similar improvements. Gains in active elevation are
minimal. All children have a loss of internal rotation, which is moderate in
most of them but is severe in some.
Level of Evidence: Therapeutic Level IV. See Instructions
to Authors for a complete description of levels of evidence.
Internal rotation contracture of the shoulder secondary to weakness of
external rotation is among the most common late sequelae of brachial plexus
birth
palsy1-3
(Fig. 1-A). As a result of
altered musculoskeletal mechanics, glenohumeral deformity develops within the
first two years of life in 60% to 70% of affected
children4-10
(Fig. 1-B). Although there is
now nearly universal agreement that the imbalance of muscle power and
contracture should be treated, there is little consensus regarding how to
accomplish this for children who do not respond to nonoperative measures. The
options are to release the shoulder internal rotators and contracted tissues,
transfer muscles to augment external rotation, or a combination of the two.
For older patients with advanced glenohumeral deformity, most authors have
recommended a rotational osteotomy of the humerus rather than addressing the
soft tissues at the glenohumeral
joint11,12.
However, there are no clear guidelines regarding what constitutes advanced
deformity or the upper limits at which remodeling can occur.
The Brachial Plexus Clinic at our institution began in 1989, and the
management of contractures was initially performed in accordance with the
recommendations of Gilbert et
al.13. Children
less than four years old were treated with an isolated contracture release,
with a release of the subscapularis from its origin. Older children and those
with a severe contracture and weakness were treated with a latissimus dorsi
transfer in addition to a contracture release as the primary operation. With
increasing experience, our understanding of glenohumeral deformities improved
so that we could detect them clinically and associate them with progressive
internal rotation contractures. We also became aware of the limitations of
existing methods of contracture
release5, as have
others14, as we
found that releasing the subscapularis from its origin failed in one of five
children. When we reverted to anterior approaches, in small children with
severe contractures and advanced deformity, we found exposure to be difficult
even after complete release of the pectoralis major and sometimes the
conjoined tendon. The subscapularis was often far posterior and severely
contracted. Z-plasty lengthening was not always possible as externally
rotating the arm distracted the tendon edges beyond a point where they could
be sutured together.
In light of these limitations, in 1999, arthroscopic release of the
internal rotation contracture, with or without a concomitant latissimus dorsi
transfer, became our preferred method, with the aim of preserving the same
indications as were used with open methods. We hypothesized that (1) release
of the internal rotation contracture with an arthroscopic subscapularis
tenotomy and capsular release would reliably restore external rotation at the
time of surgery and after two years of follow-up, and (2) the restoration and
preservation of external rotation achieved by this release, and in some cases
with the addition of a latissimus dorsi transfer, would prevent the
development of glenohumeral deformity if one was not yet present or would
promote remodeling of any existing deformity. We report on the first
thirty-three children whom we have followed for at least two years following
this procedure.
Thirty-seven children with an internal rotation contracture secondary to
brachial plexus birth palsy were treated with an arthroscopic release between
September 1999 and September 2002. Our institutional review board granted
approval for the data collection and statistical analysis of the clinical
information related to this experience. Four children left the health plan
between one and two years after the surgery without providing contact
information, so thirty-three patients were available for our two-year
follow-up analysis (see Appendix). There were fifteen boys and eighteen girls,
and the ages at the time of surgery ranged from ten months to twelve years,
with a mean of 3.7 years. Twenty-six of the children showed involvement of the
C5 and C6 nerve roots, and the other seven had involvement of the C5, C6, and
C7 roots. All of the patient in this series had good hand function.
Surgical release of the contracture was recommended when the patient had
not responded to two to three months of dedicated stretching exercises
supervised by a trained therapist and external rotation was <0° with
the arm at the side or was sufficiently restricted to impair the child's
ability to reach overhead, as evidenced by a bugler's position of the arm as
he or she attempted to do so. These two maneuvers correspond to impairments
with regard to global external rotation and hand-to-mouth function on the
Mallet scale15.
Another absolute indication for surgical release was palpable posterior
displacement of the humeral head that did not reduce with attempted external
rotation. This indication evolved with our experience and has been found to
reliably predict deformity as seen on magnetic resonance imaging.
In accordance with the recommendations of Gilbert et
al.13, an isolated
contracture release was recommended for children who were less than four years
old and a latissimus dorsi transfer in addition to a contracture release was
recommended as the primary operation for older children. When an isolated
contracture release was recommended for a young child, it was our practice to
explain to the parents that another operation (a muscle transfer) could be
necessary in the future. Because some parents wanted to avoid the possibility
of a second operation, three children younger than four years of age had both
a release and transfer as the index procedure. For children of all ages, the
aim of our surgical approach was to restore a centered position of the humeral
head on the glenoid in an effort to promote glenohumeral remodeling. At the
time of this report, no child at any age or with any amount of glenohumeral
deformity was treated with an external rotational osteotomy of the
humerus.
Surgical Technique
Arthroscopy was performed with the patient in a lateral decubitus position
and with use of a 2.7-mm arthroscope. A posterior portal was established
first. Because of the contracture, and often advanced deformity, it was
sometimes necessary to abduct the arm to approximately 90° to pass the
scope across the glenohumeral joint. A surgical assistant maintained the arm
position while applying longitudinal traction. The posterior portal was made
at the posterior glenohumeral joint line about 1 cm below the level of the
posterior part of the acromion. Care was taken to avoid making the portal too
low. A superior position facilitates insertion of the arthroscope over the top
of the humeral head to avoid damaging the articular surface.
An anterior portal was made from outside in, under direct visualization
through the posterior portal. To visualize the entire subscapularis tendon, it
is necessary to release the anterior capsular ligaments, including the middle
glenohumeral ligament and the anterior portion of the inferior glenohumeral
ligament, at their attachment to the glenoid labrum. An electrocautery device
(Valleylab, Boulder, Colorado) set on 20 W was the most useful instrument to
perform the release. Basket forceps were also helpful, especially for
releasing the capsular ligaments. After release of the anterior soft tissues,
the axillary nerve was commonly seen. The muscular portion of the
subscapularis was not released.
In most instances, the contracture can be adequately released by tenotomy
of the subscapularis tendon at its insertion and the overlying joint capsule.
In younger children, this release typically resulted in full external rotation
(70° to 90°) with the arm at the side. In older children and those
with the most severe contractures, it was often also necessary to release the
rotator interval tissue, exposing the base of the coracoid process. A release
was not considered complete unless external rotation of =45° was
achieved while the patient was on the operating table.
For children who were four years of age or older or when the parents
expressed a desire to avoid the possibility of needing a second surgery, a
latissimus dorsi transfer was performed as well. A curved incision measuring
approximately 6 to 8 cm was made in the skin lines, coursing just medial to
the posterior axillary crease toward the midline of the axilla
(Fig. 2). In larger children,
this incision was extended to include the posterior arthroscopic portal. The
latissimus dorsi tendon was then meticulously isolated from the teres major
(which was left in situ), released directly from the humerus, and transferred
under the posterior aspect of the deltoid to the greater tuberosity just
adjacent to the infraspinatus tendon insertion. Four number-2 Ethibond sutures
were used to secure the tendon.
A shoulder spica was applied to hold the arm in adduction and full external
rotation for six weeks following all procedures. This spica was then modified
and used as a night splint for an additional six weeks.
Clinical Analysis
At least two examiners performed all of the range-of-motion evaluations.
The relevant preoperative examination was limited by the young age of most of
the patients, but it always included assessment of passive external rotation
with the arm at the side as well as passive and active elevation (overhead
reach)(see Appendix). Passive external rotation at the side was measured in a
standardized manner in the clinic and was then confirmed with the patient
under general anesthesia at the time of the surgery. For consistent
measurements, it is important that the child's elbow be held against his or
her side, with the scapula stabilized and the shoulder relaxed. The arm is
then rotated until a firm end point is reached. Passive elevation was also
measured in the clinic and confirmed with the patient under anesthesia. If
there was a discrepancy, the values measured with the patient under anesthesia
were used.
A young child's ability to perform active elevation against gravity can
only be approximated by compelling the child to reach for objects overhead
(e.g., a lollipop, a toy, or shiny keys). This often required patience and
sufficient time for the child to relax and sometimes more than one clinic
visit. The parent, or one of the examiners, stabilized the child's trunk while
the other examiner introduced compelling objects to the child, urging an
overhead reach.
With increasing experience, the importance and accessibility of two
additional preoperative measures of motion—passive external rotation and
passive internal rotation with the arm elevated to 90° of abduction (or
lateral elevation)—became apparent. These motions were always measured
as a part of our postoperative protocol, but they were measured in only half
(sixteen) of the patients preoperatively. As with all of the preoperative
measurements, they were performed in the clinic by at least two examiners and
then were confirmed with the patient under general anesthesia.
Range-of-motion data were obtained at a minimum of two years
postoperatively and at a maximum of two years and seven months (see Appendix).
As the children were older at the time of follow-up, it became possible to
perform a conventional examination of the range of motion of the
shoulder16-22.
Therefore, for the follow-up examination, we determined passive and active
external rotation with the arm at the side, passive and active external
rotation with the arm elevated 90°, passive and active elevation, passive
internal rotation with the arm elevated 90° from the horizontal, and
active internal rotation (ability to reach up the back). When possible, we
evaluated the child's ability to perform belly-press and lift-off maneuvers,
which have emerged in the literature as tools for evaluating subscapularis
function23,24.
Imaging Studies
Magnetic resonance imaging of the shoulder was performed preoperatively for
all children. The protocol and classification system for defining glenoid
morphology have been published
previously4,5.
Arthrography was also performed at the time of surgery, and the findings were
described in the same
reports4,5.
Magnetic resonance images made at two to three years after the surgery were
available for twenty-seven children. Glenoid deformities were classified as
concentric, concentric/posterior, or a pseudoglenoid of varying severity
(mild, moderate, or
severe)5,6.
For statistical analysis, this more extensive classification
system5 was
consolidated to include only concentric (for concentric and
concentric/posterior deformities) and nonconcentric (for pseudoglenoids of any
severity).
Statistical Methods
Only three measures were available preoperatively and postoperatively for
all patients: passive external rotation with the arm at the side, passive
elevation, and active elevation. Two other measures—passive external
rotation and passive internal rotation with the arm abducted
90°—were available for half of the patients. The surgery was
classified as one of three types: subscapularis release only, subscapularis
release with a primary latissimus dorsi transfer, and failure of subscapularis
release with a subsequent latissimus dorsi transfer. The joint type was
classified as concentric or nonconcentric. The significance of the overall
paired differences between the preoperative and postoperative measures
according to the surgery type and the joint type within each surgery group was
assessed with the Wilcoxon signed-rank test. Comparisons of the differences in
these paired changes between the joint types within each surgery group were
performed with use of the Mann-Whitney U test. All statistical analyses were
performed with SAS version 9.1.3 software (SAS Institute, Cary, North
Carolina).
Arthroscopic Findings
The arthroscopic findings have been reported
previously5,25.
There was a spectrum ranging from a normal-appearing concentric, conforming
joint with a round humeral head that was well centered on a concave glenoid to
a markedly deformed joint with posterior displacement of a flattened humeral
head articulating with the posterior aspect of a bifurcated, convex
glenoid.
Release Only
Nineteen children had a release as the only primary procedure, and four of
them required a late latissimus dorsi transfer because the internal rotation
contracture recurred within the follow-up period. The mean age of the fifteen
children who did not undergo a subsequent latissimus dorsi transfer was 1.4
years old, which was not significantly different from the mean age of 1.8
years old for the other four children. Eight of these fifteen children had a
pseudoglenoid deformity at the time of the surgery, and seven had a concentric
(either concentric/posterior or concentric) deformity. Three of the children
who required a late latissimus dorsi transfer had a mild pseudoglenoid, and
one had a concentric/posterior glenoid.
Following the fifteen successful releases, passive external rotation with
the arm by the side increased by an average of 67° (p < 0.005)
(Table I). This increase was
greater in the children who had a nonconcentric glenoid (76°) than it was
in those with a concentric glenoid (56°). However, with the numbers
available, none of the differences in the range of motion between the joint
types within the surgical groups were significant. Postoperatively, active
external rotation averaged 61° with the arm at the side and 83° with
the arm in 90° of abduction (Fig.
3). Although active external rotation was not assessed
preoperatively, these measurements clearly reflect gains in comparison with
the average preoperative passive ranges of -2° and 43°,
respectively.
Active elevation increased by an average of 12° (p = 0.05), with a
greater increase in the patients with a concentric glenohumeral joint
(19°) than in those with a nonconcentric joint (6°). Passive external
rotation in 90° of abduction in the nine patients for whom that value was
available increased by an average of 45° (p < 0.005). Passive internal
rotation in 90° of abduction decreased by an average of 37° in the
patients in whom it was measured, suggesting that the range of motion shifted
its centering point rather than increasing absolutely.
The ability to reach up the back was not measured preoperatively, but it
was clearly restricted at the time of followup; on the average, the children
were able to reach only between the posterior superior iliac spine and the
sacrum. The belly-press and lift-off maneuvers were not always testable even
at the time of follow-up, as some children could not comprehend these
commands. Most of the children who were tested were unable to perform these
maneuvers.
Latissimus Dorsi Transfer with Release
Fourteen children had a latissimus dorsi transfer with an arthroscopic
release as their primary procedure (see Appendix). The mean age of these
patients was 6.7 years old. Passive external rotation with the arm by the side
averaged -24° preoperatively. As demonstrated by magnetic resonance
imaging, seven of the children had a nonconcentric joint and seven had a
concentric joint at the time of the surgery.
Passive external rotation with the arm by the side increased by an average
of 81° (p < 0.005) (Table
I and Appendix). This increase was greater in the children with a
nonconcentric glenoid (90°) than it was in those with a concentric glenoid
(71°) (p < 0.05 for both groups), but the difference between the
subgroups was not significant. Postoperatively, active external rotation
averaged 50° with the arm at the side and 74° with the arm in 90°
of abduction. Although active external rotation was not assessed
preoperatively, the postoperative measurements clearly reflect gains in
comparison with the average preoperative passive ranges of -24° and
47°, respectively.
Active elevation increased by an average of only 3°. In the five
patients in whom it was measured, passive external rotation in 90° of
abduction increased an average of 43°, which was nearly significant (p =
0.06). Passive internal rotation in 90° abduction in this group decreased
by an average of 42°. The ability to reach up the back was not measured
preoperatively, but it was clearly restricted at the time of follow-up, with
the children only able to reach between the sacrum and L5 on the average.
Release and Late Transfer
Four children had a repeat arthroscopic release and a latissimus dorsi
transfer as a second surgical procedure because either the contracture
recurred or external rotation weakness was demonstrated after the arthroscopic
release (see Appendix). The mean age at the time of the arthroscopic release
was 1.8 years, and the initial passive external rotation with the arm by the
side averaged -15°. Three of these four children were seen to have a
nonconcentric glenohumeral deformity on magnetic resonance imaging at the time
of the first operation. The mean age at the time of the tendon transfer was
3.0 years old.
Following the tendon transfer, passive external rotation with the arm by
the side increased by an average of 78°
(Table I and Appendix). Because
of the small sample size, meaningful statistical analysis of this gain was not
possible. Postoperatively, active external rotation averaged 63° with the
arm at the side and 91° with the arm in 90° of abduction. Although
active external rotation was not assessed preoperatively, the postoperative
measurements reflect gains in comparison with the average preoperative passive
ranges of -15° and 45°, respectively.
Active elevation increased by an average of 10°. Data for passive
external and internal rotation in 90° of abduction were available for only
two patients, one with a concentric joint and the other with a nonconcentric
joint. Passive external rotation in 90° of abduction increased by an
average of 63°, from a preoperative mean of 45°. Passive internal
rotation in 90° of abduction decreased by an average of 88°, from a
preoperative mean of 53°. This decrease was largely due to a loss of
130° of internal rotation in the child with a pseudoglenoid deformity. The
ability to reach up the back was not measured preoperatively, but it was
restricted at the time of follow-up, with the children only able to reach
between the sacrum and L5 on the average.
Imaging Studies
Preoperative magnetic resonance imaging identified four concentric and
eleven concentric/posterior glenoids. There were thirteen mild, four moderate,
and one severe nonconcentric pseudoglenoid deformities. Concentric and
nonconcentric glenoids were evenly distributed between the two treatment
groups (primary release only and primary latissimus dorsi transfer). Follow-up
images made after two years were available for twenty-seven children.
Twelve of the fifteen children with a concentric joint at the time of the
surgery were available for follow-up imaging. All twelve children showed
preservation of joint concentricity, and eight of the nine
concentric/posterior joints had normalized to fully concentric joints (Figs.
4-A and 4-B). Fifteen of the
eighteen children with a nonconcentric joint at the time of the surgery were
available for follow-up imaging. Eleven of these children showed extensive
glenohumeral remodeling to a concentric joint, and one had remodeling to a
concentric/posterior joint (Figs. 5-A and
5-B). Of these twelve children, five had had release only, five
had had a latissimus dorsi transfer, and two had been treated with a late
latissimus dorsi transfer after a primary release failed. Three children
showed persistent glenoid deformity at the time of follow-up: one had had a
release only and the other two had had a release and a latissimus dorsi
transfer as the primary operation.
Complications and Failures
The arthroscopic release achieved =45° of passive external rotation
at the time of the surgery in all but the oldest child in the series, a
twelve-year-old with a severe deformity. In this patient, the convex shape of
the glenoid and posterior displacement of the humeral head made visualization
difficult. Preoperative external rotation was -60°, and it was not
released beyond 0° after tenotomy of the subscapularis. An open release
through an extended deltopectoral approach was necessary to effectively
release the anterior aspect of the capsule.
All of the children who were treated with a latissimus dorsi transfer
retained a positive range of external rotation, and in all but one (again, the
oldest child, who had severe deformity), the glenohumeral joint maintained a
centered position. The four children who had initially had an isolated release
but did not retain the range of external rotation were treated with a repeat
arthroscopic release. These patients were otherwise not distinguished by any
particular defining characteristics; they were young (0.8 to 2.9 years old)
and had typical internal rotation contractures ranging from 0° to
-20°.
One child lost 40° of active elevation despite a 90° improvement in
passive external rotation. The reasons for this were not entirely clear as the
deltoid was functional but weak. This loss may have been due to either a
partial axillary nerve injury or some other mechanical compromise of shoulder
elevation resulting from loss of the subscapularis function. No other patient
had evidence of an axillary nerve injury.
The average loss of internal rotation was 42° after a latissimus dorsi
transfer combined with a release compared with 37° after an isolated
release. Loss of internal rotation created a functional problem when the value
for postoperative passive internal rotation in 90° of abduction was
negative (four patients). In the most severe case (Case 33; see Appendix), the
change between the preoperative and postoperative values was -130°.
The four children who left the health plan and were lost to follow-up were
all young and all were treated with a release only. All were seen for at least
one year after the surgery, and all but one, in whom the contracture recurred,
were doing well.
In our earliest imaging studies of children with internal rotation
contractures4, we
observed glenohumeral deformities with arthrography in 72% of the patients,
but only when external rotation with the arm at the side was limited to
neutral or less. Deformity was not seen when external rotation was >0°.
A subsequent study with magnetic resonance
imaging5 confirmed
this association but also showed that mild deformities (concentric/posterior)
may appear normal on arthrography and that severe deformities can be seen with
greater detail on magnetic resonance imaging. In the present study, we add to
these observations striking examples of glenohumeral remodeling seen when the
range of external rotation had been restored and was maintained following
arthroscopic release and, in some cases, latissimus dorsi transfer. It is for
this reason that restoration and preservation of the range of external
rotation is the primary objective of our surgical protocol. This objective is
further supported by a recent study by Waters and Bae that showed failure of
glenohumeral remodeling after latissimus dorsi and teres major muscle
transfers when external rotation at the side had not been the primary focus of
treatment26.
Although Waters and Bae used a different radiographic classification system,
our results are roughly comparable with theirs in that their grades of III and
IV probably represented pseudoglenoids of lesser and greater severity.
Many surgical methods have been described for the release of contractures
resulting from brachial plexus birth palsy. In our experience, arthroscopic
release provided early results comparable with those of open methods and had
some distinct advantages. The incisions are more cosmetically acceptable than
those used for the extensile open anterior approach through the deltopectoral
interval. Although the open approach can be done, by skilled surgeons, through
an axillary incision, we have found that severe contractures with posterior
displacement of the glenohumeral joint have required partial or complete
release of additional muscles such as the pectoralis major and conjoined
tendons in order to gain adequate exposure. The other open method of
subscapularis release (from its origin) does not address contractures within
the capsular ligaments, and a second anterior incision may be required to
complete the release.
One of the previous concerns about performing a subscapularis tenotomy, or
any extensive anterior release, relates to the historic problems with
iatrogenic external rotation
contractures3,11,27.
Described problems have ranged from loss of functional internal rotation to
iatrogenic anterior dislocations. The arthroscopic method and surgical
protocol that we described here differ from previously described open
techniques in several ways. As stated previously, early open techniques often
required releasing essentially all structures anterior to the glenohumeral
joint, beginning with the skin and multiple surrounding muscles, in order to
gain exposure and achieve a release. The arthroscopic procedure releases only
the capsule and subscapularis tendon, leaving the muscular portion of the
subscapularis intact. Additionally, we transferred only the latissimus dorsi
tendon, in contrast to techniques in which both the latissimus dorsi and the
teres major are transferred together (with sacrifice of part or all of the
pectoralis major). Transferring two powerful internal rotators to a position
where they function as external rotators and sacrificing as many or more
internal rotators, as part of a surgical procedure that also releases the
internal rotation contracture, may excessively tip the balance between
external rotation and internal rotation power.
Nonetheless, the surgical procedure described here also resulted in a loss
of internal rotation. This was evidenced by the inability of most of the
children who were tested to perform lift-off and belly-press maneuvers
postoperatively and by a loss of passive internal rotation with the arm in
90° of abduction. Again, only the latter can be reliably measured
preoperatively and postoperatively in young children. We anticipate a loss of
motion in this direction even in the most successful cases. Loss of internal
rotation can be expected from any procedure that promotes motion in the
opposite direction (i.e., external rotation), especially when the joint is
incongruous. In the presence of a pseudoglenoid, an effective surgical release
was seen to cause the humeral head to relocate anteriorly on the convex
glenoid. This can certainly result in substantial stiffness in external
rotation, which sometimes persists. For most patients, the decreased internal
rotation is within a functional range and the loss is far outweighed by the
benefits of function in the opposite direction and the opportunity for
glenohumeral remodeling. However, the four patients with a negative value for
postoperative passive internal rotation in 90° of abduction had severe
functional loss of internal rotation, and this prompted consideration of
additional intervention, such as an internal rotational osteotomy. At the time
of writing, only one child has had such a procedure and that patient had the
desired improvement in arm position and function in the early postoperative
period.
Of the nineteen children who underwent an isolated release, four did not
retain an adequate range of external rotation. These children had a secondary
latissimus dorsi transfer, and the only apparent drawback was the need for an
additional operation and a second period of postoperative immobilization.
Furthermore, the fifteen children who had a successful release as an isolated
procedure avoided the need for a latissimus dorsi transfer, with no compromise
of the outcome. As the treatment is essentially an effort to restore a
physiologic balance between the external and internal rotators, the minimum
amount of surgery necessary to achieve this end seems optimal. We therefore
think that the stepwise approach has merit, and the need for additional
surgery in some cases is acceptable.
It can be argued that a latissimus dorsi transfer done early enough may
obviate the need for a release, thereby sparing the subscapularis. However, it
is not possible to control when patients present for treatment, so precise
timing of a prophylactic latissimus dorsi transfer is impossible. Children
often present for the first time with a severe contracture and accompanying
deformity of equal severity. Performing a latissimus dorsi transfer as an
isolated procedure in these children is not only unlikely to achieve a good
clinical result but would be technically difficult unless a position of
external rotation could be achieved at the time of surgery. In these
instances, a release must be performed by some means. We offer the
arthroscopic approach as one that is minimally invasive.
The technical challenges involved in performing arthroscopy on small,
contracted, and deformed shoulders are considerable. Inserting the
arthroscope, even a small one, risks damaging the articular surfaces.
Visualization may be difficult as a result of the limited ability to maneuver
in the contracted joint. The proximity of the axillary nerve further
complicates the procedure. Competence in adult shoulder arthroscopy is highly
recommended, and an understanding of the deformities that are commonly
encountered in patients with this condition are requisite for surgeons
undertaking this
surgery5.
The age that constitutes the upper limit at which remodeling of a
glenohumeral deformity can occur is unknown. We observed remodeling of the
glenohumeral joint at all ages, but the most impressive reorganization of the
glenohumeral anatomy occurred in the younger children (less than four years
old). We still favor a soft-tissue release of the contracture until the age of
eight years, but in the future we may give greater consideration to external
rotation osteotomies for children who are older than eight, particularly when
they have a pseudoglenoid
deformity28.
This study and our treatment protocol have limitations, most of which we
believe are common to previously described methods. Clinical examination of
infants and young children is difficult. The child's fear of the examination,
inability to comprehend instructions, and lack of coordination due to
undeveloped motor function are all challenges faced by the examiner. A
realistic assessment of full active motor function is rarely possible before
the age of four years and is consistently possible only after the age of six
years. Thus, the examination only approximates a complete motor examination,
depending on the child's age and the above-mentioned factors. Young children,
especially infants, can often be compelled to reach for objects in certain
directions, but this is certainly not the same as directing a coordinated
patient who can follow commands to move the arm maximally in a specific
direction. For example, it would be nearly impossible for the majority of
two-year-olds to carry out a command to reach maximally in external rotation
while the elbow is held at the side. Yet this is the function of the
infraspinatus, the most commonly compromised muscle in patients with this
condition.
The Active Movement Scale is a valuable tool that will help advance the
understanding of brachial plexus birth
palsy29,30.
We find it particularly useful in the evaluation of newborns who are being
considered for nerve reconstruction surgery, as it offers a means of
quantifying motor strength of weakened muscles. It does not, however, address
the primary concern of this paper—internal rotation contracture and the
secondary development of glenohumeral dysplasia. In the study of this problem,
meticulous attention must be paid to the passive range of motion. Nonetheless,
our study may have been improved had we used the Active Movement Scale to more
fully understand the functional ability of each child. However, we were
unaware of the scale until we were far into the clinical study.
Among the strengths of this study is the confirmation of the passive range
of external rotation with the child under anesthesia, which is an indisputably
reliable measurement, as confounding variables such as the child's cooperation
and relaxation are eliminated. Unfortunately, we did not perform preoperative
measurements of passive external and internal rotation in 90° of abduction
for all of our patients. We began doing so halfway through the study, as we
became increasingly aware of its accessibility and importance. It is
instructive that essentially all of these children were found to have some
passive external rotation in abduction that they did not use (hence, the
bugler's position with arm elevation). Furthermore, passive internal rotation
in abduction, which is often normal preoperatively, is reduced considerably by
these procedures. The literature does not describe an informative and
consistent method with which to quantify loss of internal rotation, a measure
that may offer a means with which to compare one method of contracture
treatment with another.
In light of our recent and current experience, we recommend an arthroscopic
release for children who are less than three years of age and have passive
external rotation with the arm at the side of less than neutral (0°), we
recommend an arthroscopic release accompanied by a latissimus dorsi transfer
for children who are older than three years of age and have a similar degree
of contracture, and we recommend a latissimus dorsi transfer without a release
for children who are older than three years of age, have no substantial
internal rotation contracture, and have weakness of external rotation (as
evidenced by a persistent internal rotation posture of the arm in elevation).
It is noteworthy that our present surgical protocol differs from that
recommended by Gilbert et
al.13 as well as
from the one that we originally intended to follow, in that we now recommend
latissimus dorsi transfer in children older than three years of age, as
opposed to older than four years of age. This change evolved from clinical
experience and a desire to clarify our surgical indications.
Despite the inherent limitations involved in accomplishing a good physical
examination of infants and young children, our experience revealed that
arthroscopic release, used as prescribed by our treatment protocol, provides
initial results that are at least equal to those of established methods. Gains
in elevation are modest, but gains in external rotation will improve the
child's ability to reach out to the side and overhead, thereby improving many
functions in these directions. Parents should be informed that limitations in
internal rotation will become more severe. This treatment approach will,
however, predictably improve external rotation and promote remodeling of any
glenohumeral deformity associated with the internal rotation contracture.
Tables showing the preoperative and postoperative data on all patients are
available with the electronic versions of this article, on our web site at
(go to
the article citation and click on "Supplementary Material") and on
our quarterly CD-ROM (call our subscription department, at 781-449-9780, to
order the CD-ROM). ?
Fairbank HAT. Birth palsy: subluxation
of the shoulder joint in infants and young children. Lancet.
1913;1:
1217-23.11217
1913
Hoffer MM, Wickenden R, Roper B.
Brachial plexus birth palsies. Results of tendon transfers to the rotator
cuff. J Bone Joint Surg Am.
1978;60;
691-5.60691
1978
[PubMed]
Sever JW. Obstetrical paralysis.
Surg Gynecol Obstet.
1927;44:
547-9.44547
1927
Pearl ML, Edgerton BW. Glenoid deformity
secondary to brachial plexus birth palsy. J Bone Joint Surg Am.
1998.80: 659-67.
Erratum in: J Bone Joint Surg Am. 1998;80:1555-9.80659
1998
[PubMed]
Pearl ML, Edgerton BW, Kon DS, Darakjian
AB, Kosco AE, Kazimiroff PB, Burchette RJ. Comparison of arthroscopic findings
with magnetic resonance imaging and arthrography in children with glenohumeral
deformities secondary to brachial plexus birth palsy. J Bone Joint Surg
Am. 2003;85:
890-8.85890
2003
Kon DS, Darakjian AB, Pearl ML, Kosco
AE. Glenohumeral deformity in children with internal rotation contractures
secondary to brachial plexus birth palsy: intraoperative arthrographic
classification. Radiology.
2004;231:
791-5.231791
2004
[PubMed][CrossRef]
Waters PM, Smith GR, Jaramillo D.
Glenohumeral deformity secondary to brachial plexus birth palsy. J Bone
Joint Surg Am. 1998;80:
668-77.80668
1998
Zancolli EA. Classification and
management of the shoulder in birth palsy. Orthop Clin North
Am. 1981;12:
433-57.12433
1981
Kozin SH. Correlation between external
rotation of the glenohumeral joint and deformity after brachial plexus birth
palsy. J Pediatr Orthop.
2004;24:
189-93.24189
2004
[PubMed][CrossRef]
van der Sluijs JA, van Ouwerkerk WJ,
Manoliu RA, Wuisman PI. Secondary deformities of the shoulder in infants with
an obstetrical brachial plexus lesions considered for neurosurgical treatment.
Neurosurg Focus. 2004;16:
E9.16E9
2004
[PubMed]
Kirkos JM, Papadopoulos IA. Late
treatment of brachial plexus palsy secondary to birth injuries: rotational
osteotomy of the proximal part of the humerus. J Bone Joint Surg
Am. 1998;80:
1477-83.801477
1998
Dunkerton MC. Posterior dislocation of
the shoulder associated with obstetric brachial plexus palsy. J Bone
Joint Surg Br. 1989;71:
764-6.71764
1989
Gilbert A, Brockman R, Carlioz H.
Surgical treatment of brachial plexus birth palsy. Clinical Orthop
Relat Res. 1991;264:
39-47.26439
1991
Chen L, Gu Y, Xu J. Operative treatment
of medial rotation contracture of the shoulder caused by obstetric brachial
plexus palsy. Chin J Traumatol.
2000;3:
13-7.313
2000
[PubMed]
Mallet J. [Obstetrical paralysis of the
brachial plexus. II. Therapeutics. Treatment of sequelae. Priority for the
treatment of the shoulder. Method for the expression of results]. Rev
Chir Orthop Reparatrice Appar Mot.
1972;58 Suppl 1:
166-8. French.58166
1972
Richards RR, An KN, Bigliani LU,
Friedman RJ, Gartsman GM, Gristina AG, Iannotti JP, Mow VC, Sidles JA,
Zuckerman JD. A standardized method for the assessment of shoulder function.
J Shoulder Elbow Surg.
1994;3:
347-52.3347
1994
[CrossRef]
Pearl ML, Harris SL, Lippitt SB, Sidles
JA, Harryman DT 2nd, Matsen FA 3rd. A system for describing positions of the
humerus relative to the thorax and its use in the presentation of several
functionally important arm positions. J Shoulder Elbow Surg.
1992;1:
113-8.1113
1992
[CrossRef]
Pearl ML, Jackins S, Lippitt SB, Sidles
JA, Matsen FA 3rd. Humeroscapular positions in a shoulder
range-of-motion-examination. J Shoulder Elbow Surg.
1992;1:
296-305.1296
1992
[CrossRef]
Pearl ML, Sidles JA, Lippitt SB,
Harryman DT 2nd, Matsen FA 3rd. Codman's paradox: Sixty years later. J
Shoulder Elbow Surg. 1992;1:
219-25.1219
1992
[CrossRef]
Pearl ML, Wong KA. Shoulder kinematics
and kinesiology. In: Norris TR, editor. Orthopaedic knowledge update.
Shoulder and elbow. Rosemont, IL: American Academy of Orthopaedic
Surgeons; 1997.
1997
Beaton D, Richards RR. Assessing the
reliability and responsiveness of 5 shoulder questionnaires. J Shoulder
Elbow Surg. 1998;7:
565-72.7565
1998
[CrossRef]
Beaton DE, Richards RR. Measuring
function of the shoulder. A cross-sectional comparison of five questionnaires.
J Bone Joint Surg Am.
1996;78:
882-90.78882
1996
[PubMed]
Gerber C, Krushell RJ. Isolated rupture
of the tendon of the subscapularis muscle. Clinical features in 16 cases.
J Bone Joint Surg Br.
1991;73:
389-94.73389
1991
[PubMed]
Hertel R, Ballmer FT, Lombert SM, Gerber
C. Lag signs in the diagnosis of rotator cuff rupture. J Shoulder Elbow
Surg. 1996;5:
307-13.5307
1996
[CrossRef]
Pearl ML. Arthroscopic release of
shoulder contracture secondary to birth palsy: an early report on findings and
surgical technique. Arthroscopy. 2003;
19: 577-82.19577
2003
[PubMed][CrossRef]
Waters PM, Bae DS. Effect of tendon
transfers and extra-articular soft-tissue balancing on glenohumeral
development in brachial plexus birth palsy. J Bone Joint Surg
Am. 2005;87:
320-5.87320
2005
[CrossRef]
Wickstrom J, Haslam ET, Hutchinson RH.
The surgical management of residual deformities of the shoulder following
birth injuries of the brachial plexus. J Bone Joint Surg Am.
1955;37:
27-36,45.3727
1955
[PubMed]
Pagnotta A, Haerle M, Gilbert A.
Long-term results on abduction and external rotation of the shoulder after
latissimus dorsi transfer for sequelae of obstetric palsy. Clin Orthop
Relat Res. 2004;426:
199-205.426199
2004
[CrossRef]
Curtis C, Stephens D, Clarke HM, Andrews
D. The active movement scale: an evaluative tool for infants with obstetrical
brachial plexus palsy. J Hand Surg [Am].
2002;27:
470-8.27470
2002
[PubMed][CrossRef]
Bae DS, Waters PM, Zurakowski D.
Reliability of three classification systems measuring active motion in
brachial plexus birth palsy. J Bone Joint Surg Am.
2003;85:
1733-8.851733
2003
[PubMed]