Abstract
Background:
Rotator cuff ruptures are frequently associated with loss of strength
of the shoulder. However, the characteristics of the rotator cuff
tear that are responsible for the loss of force generation and transmission
have not yet been identified. The purpose of this study was to compare
the effects of supraspinatus tendon detachments, tendon defects,
and muscle retractions on in vitro force transmission by the rotator cuff
to the humerus.
Methods:
The rotator cuff tendons from ten cadaver shoulders were loaded
proportionally to the respective cross-sectional areas of their
muscles. A fiberglass rod was cemented into the medullary canal
of the humerus and connected to a three-component load cell for
the measurement of the forces transmitted by the rotator cuff to
the humerus. This study was performed with the humerus in a hanging
arm position and with various sizes of supraspinatus tendon detachments,
tendon defects, and muscle retractions.
Results:
Detachment or creation of a defect involving one-third or two-thirds
of the supraspinatus tendon resulted in a minor reduction in the
force transmitted by the rotator cuff (≤5%), while detachment
or creation of a defect involving the whole tendon resulted in a
moderate reduction (11% and 17%, respectively). Simulated muscle
retraction involving one-third, two-thirds, and the whole tendon
resulted in losses of torque measuring 19%, 36%, and 58%, respectively.
Side-to-side repair of the one-third and two-thirds defects nearly restored
the force transmission capability, whereas a deficit remained after
side-to-side repair following complete resection.
Conclusions:
Our results support the rotator cable concept and correspond to
the clinical observation that patients with a small rupture of the
rotator cuff may present without a loss of shoulder strength. Muscle
retraction is potentially an important factor responsible for loss
of shoulder strength following large rotator cuff ruptures.
Clinical Relevance:
Supraspinatus muscle retraction diminishes glenohumeral abduction
torque significantly more than either a defect in the tendon or
a simple detachment of the tendon from the tuberosity. In cases
of irreparable defects, side-to-side repair may be worthwhile to
restore muscle tension and the integrity of the rotator cable.
Rotator cuff muscle forces generate torque to move the humerus
and create compressive forces to stabilize the glenohumeral joint.
Therefore, rotator cuff ruptures are frequently associated with
loss of strength and stability of the shoulder. However, the characteristics
of the rotator cuff tear that are responsible for the loss of force
generation and transmission have not been identified yet. Three
features of rotator cuff tears that could decrease abduction torque
are tendon detachment, tendon substance loss, and muscle-tendon
retraction.
It is unclear whether the loss of shoulder strength simply increases
with the size of a rotator cuff tear or if there is a threshold
in dimension. Previous investigations have shown a linear correlation
between the size of the tear and either decreased shoulder strength
preoperatively
1
or residual loss of strength postoperatively
2-5
. On the other hand, Essman et al.
6
did not find any relationship between tear size and functional outcome.
Burkhart et al.
7
introduced the concept of the rotator cable and pointed out that
maintenance or repair of this structure could be sufficient to restore
shoulder function. The biomechanical explanation is based on force
transmission around the cuff defect.
In addition to the size of a rotator cuff tear, its shape could also
influence shoulder strength. While simple detachment of the tendon
from its osseous insertion could decrease shoulder strength, a defect
in the tendon substance may be required before loss of shoulder
strength becomes clinically symptomatic. However, to our knowledge
no one has compared the biomechanical effects of tendon detachment
and tendon substance loss.
Defects of a rotator cuff tendon are frequently associated with retraction
of its muscle. Physiologically, a muscle is most effective in the
midrange of pretension
8
and becomes less powerful with increasing retraction. This is well
understood as the Blix curve
9
. Therefore, retraction could be a causative factor in loss of shoulder
strength following a rotator cuff tear and consequently should be
considered in the performance of a rotator cuff repair. Post et
al.
10
stated that restoration of adequate tension of the rotator cuff tendon
is more important than watertight closure.
Most laboratory investigations of rotator cuff tears have involved
the release or excision of a portion of the tendon. A more clinically
relevant model would permit evaluation of the effects of muscle
retraction. The supraspinatus is the most frequently involved tendon
in rotator cuff tears. The purpose of this study was to compare
the effects of supraspinatus tendon detachments, tendon defects,
and muscle retractions on in vitro force transmission by the rotator
cuff to the humerus.
Specimen Preparation
Ten fresh-frozen cadaver shoulders were studied after macroscopic
evidence of rotator cuff tears and radiographic evidence of glenohumeral
osteoarthritis had been ruled out. During dissection, preparation,
and testing, the specimens were moistened with physiologic saline
solution to prevent dehydration. All soft tissues superficial to
the muscle were removed, leaving the supraspinatus, infraspinatus,
teres minor, and subscapularis muscles with their tendons and the joint
capsule with ligaments. The muscles were elevated from the bone.
Their tendons were bluntly separated from the surrounding soft tissues
to ensure unrestricted movement. The supraspinatus muscle was divided
into three equal portions in line with its fibers up to its musculotendinous
junction. While the anterior margin of the supraspinatus tendon
was easily identified adjacent to the rotator cuff interval, the
posterior border was defined so as to be in line with the spine
of the scapula. Nylon strings were used as suture material and were fixed
directly to each of the three portions of the supraspinatus muscle
belly (and to the other muscle bellies in a similar manner). This
was accomplished with use of a running locking (Krackow) suture,
woven through the center of the muscle, all the way down to the
muscle-tendon junction. The nylon strings stopped at the musculotendinous
junction to avoid affecting the rotator cuff tendons themselves.
A fiberglass rod was cemented into the proximal part of the humeral medullary
canal to control position and transmit force. The scapula was fixed
onto a Plexiglas plate.
Testing Device
The scapula was aligned and rigidly mounted in the shoulder-testing
device (
Fig. 1
) so that the medial margin of the scapula was in line with the vertical
axis of the device. The rod cemented in the medullary canal of the
humerus contacted a vertically unconstrained slide in a simple support
manner to maintain equilibrium against moments generated by the
muscle load.
Forces were measured in the x, y, and z directions (abduction-adduction,
anteversion-retroversion, and vertical movement) by the three-component
load cell (sensitivity, 0.3 N). Rotation was controlled by confirming
that the loads applied through the tendons balanced the humerus
to a resting neutral rotation at the beginning of the experiment
and then maintaining those loads on the rotator cuff tendon throughout
the experiment. Humeral rotation was visually confirmed by fixing
a rod perpendicular to the humeral shaft in the anteroposterior
direction. The absence of major perturbations of the perpendicular rod
following cutting of the supraspinatus tendon confirmed that no
humeral rotation had occurred. The distance from the load cell to
the center of rotation of the glenohumeral joint as well as the
contact point of the rod to the fixation frame on the load cell
were kept constant throughout the experiment. Through pulleys, which
varied in position, the nylon strings connected through the muscles
to the rotator cuff tendons were connected to weights. The positions
of the pulleys were carefully adjusted so that the strings imitated
the line of action by running through the centroid of each muscle
11
.
Testing Sequence
As the supraspinatus initiates abduction
12
, the experiment was performed with the humerus in the hanging arm
position and neutral rotation. The rotator cuff tendons were loaded
proportionally to the respective cross-sectional areas of their
muscles
13
. The magnitude of loading was 25.6 N for the supraspinatus, 46.7
N for the infraspinatus, 14.3 N for the teres minor, and 66.3 N
for the subscapularis. In the experiments with simulated muscle
retraction, one of the three portions of the supraspinatus tendon
was excluded from force transmission by cutting it out. The forces
were transmitted from the muscles directly to the tendons through
the intact musculotendinous junction. The forces transmitted to
the humerus were measured with the three-component load cell under
the following conditions (
Figs. 2-A
,
2-B
, and
2-C
): (1) intact rotator cuff simulating the healthy condition; (2) detachment
of one-third of the width of the supraspinatus tendon at its insertion;
(3) detachment of one-third of the width and incision between that
portion of the supraspinatus tendon and the adjacent cuff tissue
for half of the medial extent of the supraspinatus tendon, simulating
loss of tendon substance without muscle retraction; (4) a defect
involving one-third of the width and the total medial extent of
the supraspinatus tendon, simulating loss of tendon substance with
muscle retraction, with load removed from the retracted portion
of the muscle; (5) side-to-side repair of the defect, simulating incomplete
rotator cuff repair; (6) complete repair of the rotator cuff defect;
(7) detachment of two-thirds of the width of the supraspinatus tendon
at its insertion; (8) detachment of two-thirds of the width and
incision between that portion of the supraspinatus tendon and the
adjacent cuff tissue for half of the medial extent of the supraspinatus
tendon, simulating loss of tendon substance without muscle retraction;
(9) a defect involving two-thirds of the width and the total medial extent
of the supraspinatus tendon, simulating loss of tendon substance
with muscle retraction, with load removed from the retracted portion
of muscle; (10) side-to-side repair of the defect, simulating incomplete
rotator cuff repair; (11) complete repair of the rotator cuff defect;
(12) detachment of the total width of the supraspinatus tendon at
its insertion; (13) detachment of the total width and incision between
that portion of the supraspinatus tendon and the adjacent cuff tissue for
half of the medial extent of the supraspinatus tendon, simulating
loss of tendon substance without muscle retraction; (14) a defect
involving the total width and the total medial extent of the supraspinatus
tendon, simulating loss of tendon substance with muscle retraction,
with load removed from the retracted portion of the muscle; (15)
side-to-side repair of the defect, simulating incomplete rotator
cuff repair; (16) complete repair of the rotator cuff defect; (17)
detachment of the total width of the supraspinatus tendon and 5
mm of the infraspinatus tendon at their insertions; and (18) detachment of
the total width of the supraspinatus tendon and 10 mm of the infraspinatus
tendon at their insertions.
The forces and torques were measured for ten seconds, averaged,
and registered by a commercially available computer controlled by
Labview software (National Instruments, Austin, Texas). To ensure
comparability, the forces transmitted under the different conditions
of the rotator cuff were expressed as percentages of the forces
transmitted by the intact rotator cuff. The effect of each percentage
of detachment of the supraspinatus tendon (one-third, two-thirds,
and the whole tendon) was evaluated separately for each state (incision,
defect, retraction, side-to-side repair, and complete repair) with
use of repeated-measures analysis of variance. When a significant
effect was identified, pairwise comparisons were performed with
use of the Student-Newman-Keuls multiple-comparisons procedure.
For each cut size, paired t tests were used to compare the effects
of the detachment and the side-to-side repair, those of the detachment
and the defect, those of the defect and the retraction, and those
of the side-to-side and complete repairs. All statistical tests
were two-sided, and the threshold of significance was set at alpha
= 0.05. The analysis was conducted with use of SAS version 6.12
(SAS Institute, Cary, North Carolina) on a Sun Ultra II computer (Sun
Microsystems, Palo Alto, California).
Simple detachment of the anterior third of the supraspinatus tendon
resulted in only a slight decrease (<1%) in the transmitted
force, which was not significantly different (p > 0.05) from the
decrease resulting from detachment of the anterior two-thirds (2%)
or from the intact state. Compared with the intact state, detachment
of the whole tendon significantly decreased the transmitted force
by 11% (p < 0.05), detachment of an additional 5 mm of the infraspinatus
tendon significantly reduced it by 27% (p < 0.05), and detachment
of an additional 10 mm significantly decreased it by 47% (p < 0.05).
A defect of the anterior third of the supraspinatus tendon resulted
in only a slight decrease (1%) in the transmitted force, whereas
defects of two-thirds and the full width of the tendon significantly
decreased transmitted force by 5% and 17%, respectively (p < 0.05).
Resection of the supraspinatus tendon and muscle resulted in significant
losses in force transmission of 19% (with involvement of one-third
of the tendon), 36% (with involvement of two-thirds), and 58% (with involvement
of the whole tendon) (p < 0.05).
After side-to-side repair of one-third and two-thirds of the musculotendinous
unit (with the tendon left detached at its insertion site), the
loss of force transmission was only 1% and 3%, respectively, compared
with the intact state. Side-to-side repair after complete isolation
and detachment of the supraspinatus muscle and tendon was followed
by a significant deficit of 9% (p < 0.05).
Following each complete repair of the supraspinatus tendon, transmitted
forces were not significantly different from those in the intact
state (p > 0.05)-i.e., each complete repair restored the full capability
of the tendon to transmit force.
When only one-third of the tendon width was involved, there was
no difference between force transmission following tendon detachment
at the insertion and that following creation of a tendon defect
without muscle retraction. However, with two-thirds and full-tendon
involvement, the difference between those two conditions was significant
(p < 0.05). In addition, the differences between the effects
of simulated retraction and those of the simulated tendon defect
were significant with all percentages of tendon involvement (p < 0.05).
In other words, side-to-side connection with the adjacent tendon
was an important factor in force transmission. After all three side-to-side
repairs (of one, two, or three-thirds of the tendon), the forces
transmitted by the supraspinatus tendon were not significantly different
(p > 0.05) from those after simple detachment of the same percentage
of the tendon. Although the differences between the force transmission
following side-to-side repair and that following complete repair of
one, two, and three-thirds of the supraspinatus tendon were significant
(p < 0.05), the magnitudes of those differences were small.
Detachment
In the present study, we compared the effects of supraspinatus tendon
detachment and substance loss. Detachment of one-third or two-thirds
of the supraspinatus tendon had a minor effect on force transmission;
a substantial decrease occurred only after detachment of the whole
tendon. Obviously, rotator cuff muscle forces are effectively transmitted
around small and medium detachments of the supraspinatus tendon.
Clinically, patients with a small-to-medium detachment of a rotator cuff
tendon frequently present without loss of shoulder strength
5
, whereas large detachments result in reduced shoulder strength;
these observations support our findings. According to Burkhart
14
, the muscle forces are effectively transmitted by the rotator cuff
as long as the rotator cable is intact. Our results indicate that
the threshold for a substantial decrease in force transmission is
detachment of the whole supraspinatus tendon, affecting the rotator
cable.
Defect
The effects of detachment and a defect of one-third of the supraspinatus
tendon were not significantly different; both had a minor effect
on force transmission. Clinically, patients with a small defect
in a rotator cuff tendon tend to have full shoulder strength
5
, which corresponds to our findings. Defects of two-thirds of the
supraspinatus tendon and those of the whole tendon resulted in moderate
reductions in force transmission, which were significantly greater
(p < 0.05) than the reductions following detachments of the
same percentages of the tendon. Clinically, defects of the whole
supraspinatus tendon are associated with symptomatic loss of shoulder
strength
5
, which could not be explained by the moderate reduction in force
transmission shown in this in vitro experiment. This apparent contradiction
may be caused by muscle retraction associated with large tendon
defects in vivo, which is usually neglected in biomechanical rotator
cuff models. As a muscle is most effective in the midrange of pretension
8
, retraction reduces its efficacy.
Retraction
To simulate the effect of muscle retraction in this in vitro experiment,
we released a portion of its tendon from its neighboring cuff tissue
longitudinally, until the tendon-muscle unit could freely retract.
Consequently, this portion of the muscle was excluded from force
generation, and this section of the tendon did not transmit forces
generated by adjacent muscles, simulating the in vivo situation.
The result was a substantial and significantly greater (p < 0.05)
decrease in transmitted force compared with that associated with
the tendon defects. Therefore, muscle retraction is potentially
an important factor in clinically symptomatic loss of shoulder strength
following medium-to-large rotator cuff ruptures. It should be considered in
future in vitro investigations of rotator cuff tears.
Repair
Side-to-side repairs of one-third and two-thirds of the supraspinatus
tendon nearly restored the full capability of the rotator cuff to
transmit force. Even when the side-to-side repair followed complete
excision of the supraspinatus tendon, the rotator cuff transmitted
about 90% of the force measured in the intact state. Therefore,
in cases of irreparable rotator cuff ruptures, side-to-side repair
may be worthwhile to restore muscle pretension and the integrity
of the rotator cable
15
. However, retraction becomes irreversible with time as it is associated
with degeneration of the muscle
16
, and excessive pretension reduces the efficacy of the muscle as
does decreased pretension
8
. Complete repair consistently restored the full capability of the
rotator cuff to transmit force and should therefore always be the
ultimate goal of rotator cuff surgery.
Study Limitations
In our experiment, muscle retraction was simulated by excision
of the respective portion of the muscle and tendon. In vivo, the
process of retraction that causes the decrease in force generation
is gradual, and the extent to which the muscle retracts varies.
Excision of the respective tendon section in the present study compromised
force transmission from adjacent healthy tendon sections across
the dysfunctional region, simulating the in vivo situation of complete
retraction. Consequently, our model simplified the complex pathological process
by assuming a complete loss of function in the respective portion
of the muscle and tendon. However, the in vitro model showed the
importance of muscle retraction as a consequence of rotator cuff
ruptures.
In our study, muscle loading was based on the assumption that maximum
muscle force is proportional to its cross-sectional area, which
was derived from the literature
13
. The transmitted forces were reported as a percentage of those
in the intact state. The strings representing lines of muscle action
were adjusted according to the centroids of the tested muscle
11
. Although strings are merely approximations of the varied lines
of action of a muscle, they are a legitimate biomechanical model
for muscle forces.
Finally, to limit the number of specimens required for this experiment
to a reasonable and feasible number, we progressively enlarged the
size of the tendon defect after complete repair of the preceding,
smaller tendon defect in each specimen. However, we ruled out any
effect of this procedure on our results by proving that there was
no significant (p > 0.05) difference in force transmission between
the intact state and complete repairs in any of the specimens. The
efficacy of the repair was further demonstrated by the fact that
no significant difference was detected between the forces following
detachment and those following side-to-side repair.
Our results correspond with the clinical observation that patients
with a small rotator cuff rupture may have no loss of shoulder strength,
and they support the rotator cable concept. Muscle retraction probably
contributes substantially to loss of shoulder strength following
large rotator cuff ruptures and should be considered in future in
vitro models of rotator cuff tears. In cases of irreparable rotator
cuff defects, side-to-side repair may be worthwhile to restore muscle
pretension and the integrity of the rotator cable.
Gazielly DF, Gleyze P, Montagnon C, Bruyere G,Prallet B. [Functional and anatomical results after surgical treatment
of ruptures of the rotator cuff. 1: Preoperative functional and anatomical
evaluation of ruptures of the rotator cuff]. Rev Chir Orthop Reparatrice Appar Mot,1995;81: 8-16.. French.818
1995
[PubMed]
Postacchini F, Perugia D,Rampoldi M. Rotator cuff tears. Results of surgical repair. Ital J Orthop Traumatol,1992;18: 173-88.. 18173
1992
[PubMed]
Grana WA, Teague B, King M,Reeves RB. An analysis of rotator cuff repair. Am J Sports Med,1994;22: 585-8.. 22585
1994
[PubMed]
Gore DR, Murray MP, Sepic SB,Gardner GM. Shoulder-muscle strength and range of motion following
surgical repair of full-thickness rotator-cuff tears. J Bone Joint Surg Am,1986;68: 266-72.. 68266
1986
[PubMed]
Gschwend N, Ivosevic-Radovanovic D,Patte D. Rotator cuff tear-relationship between clinical and anatomopathological
findings. Arch Orthop Trauma Surg,1988;107: 7-15.. 1077
1988
[PubMed]
Essman JA, Bell RH,Askew M. Full-thickness rotator-cuff tear. An analysis of results. Clin Orthop,1991;265: 170-7.. 265170
1991
[PubMed]
Burkhart SS, Esch JC,Jolson RS. The rotator crescent and rotator cable: an anatomic description
of the shoulder's "suspension bridge.". Arthroscopy,1993;9: 611-6 . [Erratum in 1994;10:239].9611
1993
[PubMed]
Hersche O,Gerber C. Passive tension in the supraspinatus musculotendinous
unit after long-standing rupture of its tendon: a preliminary report. J Shoulder Elbow Surg,1998;7: 393-6.. 7393
1998
[PubMed]
Blix M. Die L�nge und die Spannung des Muskels. Skand Arch Physiol,1891;3: 295-318.. 3295
1891
Post M, Silver R,Singh M. Rotator cuff tear. Diagnosis and treatment. Clin Orthop,1983;173: 78-91.. 17378
1983
[PubMed]
Johnson GR, Spalding D, Nowitzke A,Bogduk N. Modelling the muscles of the scapula morphometric and
coordinate data and functional implications. J Biomech,1996;29: 1039-51.. 291039
1996
[PubMed]
Howell SM, Imobersteg AM, Seger DH,Marone PJ. Clarification of the role of the supraspinatus muscle
in shoulder function. J Bone Joint Surg Am,1986;68: 398-404.. 68398
1986
[PubMed]
Veeger HE, Van der Helm FC, Van der Woude LH, Pronk GM,Rozendal RH. Inertia and muscle contraction parameters for musculoskeletal modelling
of the shoulder mechanism. J Biomech,1991;24: 615-29.. 24615
1991
[PubMed]
Burkhart SS. Fluoroscopic comparison of kinematic patterns in massive rotator
cuff tears. A suspension bridge model. Clin Orthop,1992;284: 144-52.. 284144
1992
[PubMed]
Burkhart SS, Nottage WM, Ogilvie-Harris DJ, Kohn HS,Pachelli A. Partial repair of irreparable rotator cuff tears. Arthroscopy,1994;10: 363-70.. 10363
1994
[PubMed]
Nakagaki K, Ozaki J, Tomita Y,Tamai S. Fatty degeneration in the supraspinatus muscle after rotator cuff
tear. J Shoulder Elbow Surg,1996;5: 194-200.. 5194
1996
[PubMed]