- Email: [email protected]
The effect of rotator cuff tears on reaction forces at the glenohumeral joint
ELSEVIER
Journal of Orthopaedic Research 20 (7002) 439446
Journal of Orthopaedic Research www.elsevie~-.con~/locate/orthres
The effect of rotator cuff tears on reaction forces at the glenohumeral joint I.M. Parsons, IV, Maria Apreleva
I,
Freddie H. Fu, Savio L.-Y. Woo
''
Abstract The rotator cuff muscles maintain glenohumeral stability by compressing the hunieral head into the glenoid. Disruption of the rotator cuff compromises concavity compression and can directly affect the loads on the glenohumeral joint. The purpose of this study was to quantify tlie effect of rotator cuff tears on the magnitude and direction of glenohumeral joint reaction forces during active shoulder abduction in the scapular plane using nine cadaveric upper extremities. Motion of the full upper extremity was simulated using a dynamic shoulder testing apparatus. Glenohumeral joint reaction forces were measured by a universal forcemoment sensor. Five conditions of rotator cuff tears were tested: Intcict, bicoiiiplete Suprcispiiiutus Teur, Coiiiplete SLil)rcis~)iiicitus Teur, Supru.~~~inutuslInfi.u.~piizatirs Tear, and Global Tear. Reaction forces a t the glenohumeral joint were found to steadily increase throughout abduction and peaked at maximum abduction for all conditions tested. There were no significant differences in reaction force magnitude for the intact condition (337 f 88 N) or those involving an isolated incomplete tear (296 f 83 N ) or complete tear (300 f 85 N) of the supraspinatus tendon. Extension of tears beyond the supraspinatus tendon into the anterior and posterior aspect of the rotator cuff led to a significant decrease in the magnitude of joint reaction force (126 f 31 N). Similarly, such tears resulted in a significant change in the direction of the reaction force at the glenohumeral joint. These results suggest that joint reaction forces are significantly affected by the integrity of the rotator cuff, specifically, by tlie transverse force couple formed by the anterior and posterior aspects of the cuff. The quantitative data obtained in this study on the effect of rotator cuff tears on magnitude and direction of the reaction force a t the glenohumeral joint helps clarify the relationship between joint motion, joint compression and stability. 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. A11 rights reserved.
Introduction
Experimental studies have demonstrated that the rotator cuff serves two principle functions at the glenohumeral joint: (1) generation of torque necessary for rotation of the humerus on the glenoid; and (2) compression of tlie liumeral head into the glenoid concavity. This latter function compensates for the lack of inherent bony stability at the glenohumeral joint, serving as the primary stabilizing mechanism for this minimally constrained articulation during the functional range of motion [3,13,23,26]. Termed concavity compression, the stabilizing mechanism of the rotator cuff depends on the integrity of a transverse force couple, formed by the *Corresponding author. Tel.: +I-412-648-2000; fax: +I-412-6482001. E-i?wiI trt1dres.s: [email protected] (S.L.-Y. Woo).
'
Current address: Orthopedic Biomechanics Laboratory. Beth Israel Deaconess Medical Center, Harvard Medical School. Boston, MA, USA.
anatomical arrangement of the anterior (subscapularis) and posterior (infraspinatuslteres minor) rotator cuff tendons as they insert into the proximal humerus [15,16, 23,24,26]. Coiicavity compression also resists the superior pull of the deltoid muscle during abduction and provides a fulcrum for concentric rotation of the humeral head on the glenoid [ 15,161. The glenohunieral joint reaction force counteracts the sum of all shoulder muscle forces transmitted across the articulation, and its magnitude depends on the torque generated by the activity of these muscles in moving tlie upper extremity and resisting loads applied along its length [12,16,18,22,26]. Because the rotator cuff and deltoid muscles are the main abductors and rotators at the glenohumeral joint, tlie magnitude of the joint reaction force during active motion provides an index of the competence of concavity compression [25].Previous studies have demonstrated that disruption of the transverse force couple, as occurs in large and massive rotator cuff tears, leads not only to increased translations or subluxation of tlie liumeral head, but also to changes in
0736-0366/02/$ - see front matter 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved PII: S 0 7 3 6 - 0 2 6 6 ( 0 1 ) 0 0 1 3 7 - 1
the magnitude and direction of the reaction force at the glenohumeral joint [8,10,19,21,25]. Th~is,the degree to which different rotator cuff tear configurations affect the mechanical integrity of the transverse force couple can be studied in terms of their affect 011 the magnitude and direction of the glenohumeral joint reaction force during simulated active motion. To date, the effect of rotator cuff injury on glenohumeral joint mechanics has largely been investigated in terms of changes in articular kinematics; however, the effects of rotator cuff tears on glenohumeral joint reaction forces have not been fully elucidated in quantitative terms. In light of the complex relationship between joint stability and motion as mediated by the function of the rotator cuff, quantitative information addressing the consequences of cuff injury on reaction forces will contribute to a more comprehensive understanding of the association between tear configuration and loss of shoulder function. This information may eventually lead to scientific validation of surgical treatments aiiiied at restoring rotator cuff function. The objective of this study was, therefore, to quantify the effect of different rotator cuff tear configurations on glenohumeral reaction forces during active scapular plane abduction. We hypothesize that simulated postero-superior rotator cuff tears will result in a significant reduction in the reaction force magnitude measured during active motion due to disruption of the transverse force couple. In addition, such tear configurations will result in a change of the reaction force direction such that the vector points toward the deficient side of the cuff indicating a tendency to translate in the direction of the tear. Finally, we hypothesize, that in large tear configurations, in which the transverse force couple is disrupted, tlie reaction force will be directed more superiorly as the inferior vector of the infraspinatus/teres minor complex will no longer counteract the pull of the deltoid. This hypothesis is based on the premise that tear configurations involving the infraspinatus/teres minor and subscapularis disrupt the transverse force couple while isolated supraspinatus tendon tears do not.
Materials and methods This study employed tlie dynamic shoulder testing apparatus (DSTA) to simulate active scapular plane abduction. This testing system consists of a rigid frame, which supports a six-degree of freedom mounting block on which cadaveric upper extremities can be fixed to replicate the anatomical position of the slioulder relative to tlie axial skeleton (Fig. 1). A detailed description of this testing system has been previously published, and prior experimental studies have validated the accuracy of the DSTA in si~nulatinghighly repeatable six degree of freedom glenoh~~meral joint motion [5.26]. Full human upper extremities were used to recreiite the inertinl pi-operties of inass distributed along the arm's length during abduction in tlie scapular plane. Motion was simulated by applying forces to the tendons of the rotator cuff and deltoid muscles through specially designed sinusoidal clamps. which were connected to four servo-actuated hydraulic cylin-
ders through a cable-pulley system. Each cylinder was controlled independently by a computer with custom-designed software. The abduction angle of the glenohumeral joint was measured by a magnetic tracking device (The Bird, Ascension Technologies, Burlington. VT), rigidly mounted t o the hunieral shaft of the specimen. A six-degree of freedom universal force-moment sensor (UFS) (JR3, Woodland. CA) was incorporated into the testing system to allow for direct measurement of glenohumeral joint reaction forces during simulated active abduction (Fig. I). The UFS is able to nieasiire three forces and three moments along and about a Cartesian coordinate system defined with respect to the sensor. and its application toward non-contact force measureinent in soft tissues has been previously validated [6].The UFS was positioned on the testing apparatus so that its coordinate gystem was aligned with the axes of a reference coordinate system defined with respect to the scapula (Fig. 2). The axes of the reference system were defined as follows: the X-axis of the UFS corresponded to the anterior (ANT) direction in the antero-posterior plane of the scapula; the Y-axis corresponded to the superior (SUP) direction in the supero-inferior plane of the scapula; and the Z-axis corresponded to the lateral (LAT) direction in the medio-lateral plane of tlie scapula. Custom data acquisition software was designed to allow simultaneous. real-time nieasurenient of glenohumeral joint reaction forces. We performed preliminary studies to evaluate the repeatability of the dynamic shoulder testing system [I]. Ten trials of simulated active glenohumeral abduction were repeated on the same representative intact specimen, and magnitudes of the joint reaction force were recorded. Fig. 3 shows an average and a standard deviation of the magnitude of the glenohumeral joint reaction force after 10 trials of abduction on the same specimen. We determined that the force magnitudes measured by this DSTAIUFS testing system in our experimental set-up were repeatable to within 2 N when measuring reaction forces up to 500 N. Fourteen fresh-frozen full upper extremities were harvested from 13 human cadavers aged 60-80 years (mean age of 70 years) and stored at -20 "C prior to testing. After speciinens were thawed at room temperature, the skin and subcuta~ieoustissues proximal to the glenohumeral joint were removed, exposing the muscular envelope around the scapula. Remnants of the scalene, latissiinus dorsi, pectoralis major and minor, teres major and trapezius were sharply removed exposing the deltoid origin and rotator cuff m~iscles.A longitudinal skin incision was then made over the deltoid tuberosity of tlie humerus to expose the deltoid insertion. The biceps tendon and structures of the rotator interval were preserved during the dissection. and in all cases the biceps tendon was directly observed to be in continuity with its insertion on the superior glenoid labrum. Specimens with a history of trauma or prior surgery to the shoulder were excluded. Gross observation of the rotator cuff and deltoid tendons, further excluded any specimens with iiijury to or attrition of tlie tendons consistent with a rotator c u r tear o r rotator cuff tendinopathy. Five specimens from tlie original set of 14 specimens were excluded due to various pathological findings prior to o r during testing. This resulted in nine specimens: six inale and three female specimens from six male and two female cadavers (two specimens stemming from one female cadaver). Specially designed sinusoidal clamps were firmly secured to each of the rotator cuff and deltoid tendons. Since the infraspinatus and teres minor have a similar line of action and function similarly as external rotators at the shoulder, these muscles were combined to form one posterior tendon complex. To eliminate tlie eHect of distal motion on the inertial properties of the upper extremity. threaded pins were used to immobilize the wrist in neutral llexionlextension and neutral pronatioidsupination. Similarly the elbow was pinned i n full extension. The scapula was potted in an epoxy putty block and the specimen was positioned on tlie apparatus so that the glenoid was facing upwards 10" froin vertical and the scapular body was oriented 30" anteriorly relative to the frame of the DSTA. This position approximates tlie anatomical orientation of the scapula relative to tlie bony thorax when the upper extremity is positioned at the side. An antero-posterior radiograph of the mounted specimen was then taken to confirm the glenoid inclination. Once tlie scapula was appropriately positioned. cables from tlie servo-hydraulic cylinders were attached to the tendon clamps through a system of' pulleys, which could be adjusted to approximate tlie line of action ofeiicli muscle (Fie. 1 ) [14]. Great care WDS taken to assure that
44 1
Subscapularis
I
Direction of muscle forces used to simulate GH motion
Fig. I . Pittsburgh dynamic shoulder testing apparatus. (A) Universal force-moment sensor (UFS) attached to the scapular niount. (B) Cable-pulley system. (C) Load cell. (D) Hydraulic cylinder. (E) LVDT.
Scapular Plane
. .......
0 1 0
20
40
...
60
.
......
80
. .
100
Abduction Angle, deg
Fig. 2 . The coordinate system of the UFS is oriented on the testing apparatus to correspond to the scapular reference coordinate system. X-axis of the UFS corresponds to the anterior (ANT) direction, Y-axis corresponds to the superior (SUP) direction. and Z-axis represents lateral direction (LAT). Elevation, 0, is the angle that the resultant reaction force, R. makes with the transverse plane. Deviation. (p, is the angle that the resultant force component in the transverse plane, RAL. makes with the scapula plane.
the cables did not come into contact with either the scapula or the putty block. This ensured that the only forces measured by the load cell were the ones applied to the glenoid.
Fig. 3. An average and standard deviation of the magnitude of the glenohumeral joint reaction force after 10 trials of abduction repeated 011 the same representative intact specimen.
The line of action of each cable was verified with an antero-posterior radiograph of the testing system. Throughout the testing protocol, specimens were kept moist with physiologic saline solution. Prior to each test, 5 N of force was applied to each tendon to center the hunieral head on the glenoid, defining the “initial position” as 0” of abduction. The position of “maximum abduction” was then achieved by applying equal forces to each tendon at constant rate of 20 Nls until the upper extremity reached approxiniately 90” of scapular plane
abduction. This ratio of forces was based on prior experimental studies in our laboratory using tlie DSTA. Specimens were cycled through the range of maxinium abduction 25 times to minimize tlie effect of softtissue viscoelasticity on glenohumeral motion and reaction forces. For all subsequent testing conditions, forces were applied a t the same rate and custom-designed data acquisition software provided real-time monitoring of reaction forces. abduction angle. and glenohunieral kinematics throughout the range of motion. T o allow a inore direct comparison of tlie effect of rotator cuff injury between specimens of different sizes, we chose to base rotator cuff tear configuration on the number of involved tendons rather than tlie actual tear size. This description of tear configuration also correlates more closely witli observed clinical differences in tlie effect of tear size on shoulder function [27]. All rotator cuff tears were surgically simulated by incising the full thickness of tlie tendon at its insertion into tlie proximal liunierus. For each specimen, five rotator cuff tear testing conditions were defined a s follows: Iiifticr: no disruption of tlie rotator cuff; fircuiirplc~ieSzrprri.spbto/us(SS) Terii: disruption of the anterior half of the supraspinatus tendon: Curiiplete Siq>rrrspinri/us Tear: coniplete disruption of the supraspinatus tendon; Su/>irrs/>iri~r/u.s/fr~~~~iJiri(//iis (SSIfS) Terir: advancement of the tear 1.5 cin into tlie infraspinatuslteres minor tendon unit; and Gluhril Terir: advancement of tlie tear 1.5 cni anteriorly into tlie tendinous insertion of the subscapularis. We assumed that a full thickness disruption of a tendon from its bony attachment to the proximal huinerus would result in a complete dissociation of the applied force from the hunieral head. I n light of this assumption, no force was applied to those tendons in which a coniplete tear was surgically simulated. In addition. only 5 N of force was applied to prevent extension of simulated tears in conditions wliere only part of a tendon was disrupted such as in the Irtcuii~p/t./eSS and SSIIS Terir conditions. Consequently. the intact posterior fibers of the supraspinatus in the fiicoriiplete SS Tecrr configuration allowed transmission of applied force through the effect of tenodesis. This passive effect was not possible in the case of coinplete tendon disruption. For tlie Globol tear condition, the deltoid was tlie only tendon to which a force above 5 N was applied. Each testing condition was repeated three times to ensure repeatability and quality of data. For each condition, tlie magnitude and direction of tlie resultant reaction force at tlie glenohuineral joint were recorded. At the conipletion of the testing protocol, all speciniens were disarticulated at the glenoliunieral joint and tlie labruni and articular surfaces were inspected for evidence of a labral tear, degenerative joint disease or other f o r m of joint pathology. The magnitude of the resultant joint reaction force, R, was determined by the following equation:
where R A N T ,RSuP, and RLnr represent vector components of the reaction force in the anterior-posterior, superior-inferior, and mediallateral directions. recorded by tlie UFS (Fig. ?). The following conventions were defined to describe the direction of the joint reaction force with respect to the glenoid surface: s c r r / d o r plrrrrt: defined by the SUP-axis and LAT-axis of tlie scapula; mizswrst p/urw: defined by ANT-axis and LAT-axis of the scapula. Elevation, 0, is the angle that the resultant reaction force, R, makes witli the transverse plane. A negative 0 corresponds to tlie superior direction of the force (positive SUP-axis). and a positive 0 corresponds to tlie inferior direction of the force (negative SUP-axis). Deviation, cp, is tlie angle that tlie resultant force component in the transverse plane, RAL. makes with the scapula plane. A negative elevation angle cp corresponds to the anterior direction of the force (positive ANT-axis), and a positive cp corresponds to tlie posterior direction of tlie force (negative ANT-axis). Power analysis of tlie initial data was performed to estimate tlie required sample size and ensure 80%) power (c( = 0.05,p = 0.2). A Shapiro-Wilks' W test confirmed the normal distribution of all parameters of interest. A one-factor repeated-measures analysis of variance (ANOVA) was used to evaluate the effect of each rotator cuff tear condition on the magnitude of the resultant glenohumeral reaction force. Similarly, a one-factor ANOVA was used to determine tlie effect of each rotator cuff tear condition on the direction of the glenohu-
ineral joint reaction force. Multiple contrasts assessed differences between individual testing conditions. Statistical significance was set at p < 0.05. Statistica 5.0 software (StatSoft. Tulsa. OK) was used to perform all of the above analyses.
Results
The average weight of full upper extremities used in this investigation measured 3.9 f 1.0 kg (mean f S.D.). The maximum abduction angle achieved for the Infact condition was 85 f 10". This was not significantly different than the abduction angle achieved in the Incornplete SS Teur (86 i 15") and Coritplete SS Teur (87 i 16") conditions. For the simulated SSIIS Tear and GIobul Tear conditions, the arm was unable to abduct past 41 f 1 lo, and 27 f 1 lo, respectively. This 53% difference in abduction angle represented a statistically significant decline ( p < 0.05) compared to the maximum abduction achieved in tear conditions in which the transverse force couple was not disrupted. Fig. 4 shows the change in magnitude of the glenohumeral joint reaction force throughout abduction for the Intcict, Incomplete SS Teur, Complete SS Teur, SSIIS Teur, and Global Tear conditions. Data for the Incoiiiplete SS Tear condition did not differ significantly from the Intuct and Coinplete SS Teur conditions (p > 0.05). Similarly, data for the Global Tear condition did not differ significantly from the SSIIS Tear condition ( p > 0.05). For each rotator cuff tear condition, the joint reaction force increased with increasing abduction angle and peaked at maximum abduction. At maximum abduction, the joint reaction forces for the five testing conditions measured as follows: Intact = 337 f 88 N; Iticoriiplete SS Tear. = 296 k 83 N; Coriiplete SS Tear = 300 & 85 N; SS/IS Tear = 149 f 15 N; Global Tear = 126 k 31 N. Disruption of the transverse force couple resulted in a statistically significant decrease in reaction
0
20
40
60
80
100
Abduclion Angle, deg
Fig. 4. Change in niiignitude of the glenoliume~aljoint reaction force throughout abduction for tlie hi/trc/, fricboinplcrcSS Tcw. C C J I H ~SS ~CVC Tcai, SSIfS Tetrr, and Gluhrrl Terir conditions (mean f S.D.).
*
01
24
2 d
2
rza
.0
Abduction Angle, degrees r-7 I
Incomplete SS Tear
50
E Complete SS Tear
40
SS1ISTear
GlobalTenr
Y
2 30
BN
20
=
E!
I0
zg
o
i%
---
60
40 I
-
-20
{
6
M
-30
.-
-40
ze
T T
TT
60
a0
I00
-
I P
. d
P
5
Gi
-50
A
--
-60
Testing Conditions
Intact lncomplcle SS Tear Complete SS Tear
-70
Fig 5. Magnitude of the glenohumeral joint reaction force norinallzed to the weight of the upper extremity foi each of the testing conditions at maximum abduction (niean f S D.)
force magnitude as occurred in the SSIIS Tear and Globd Tear conditions. To account for the effect of specimen weight on the mass moment of inertia of tlie upper extremity during active motion, the values for the reaction force magnitudes were normalized to the specimen weight [22].Fig. 5 demonstrates magnitudes of the reaction at the glenohumeral joint at maximum abduction normalized to the weight of the upper extremity for each of the five testing conditions. As with abduction angle, tlie magnitudes of the reaction force at the glenohumeral joint did not differ significantly for the Intact, Incoriiplete SS Tecir and Complete SS Tear conditions, in which the transverse force couple was preserved. Normalized force magnitudes for these conditions measured 44 8 N/kg, 3 1 f 7 N/kg and 38 & 8 N/kg for the Iiitrrct, Iiicoriiplete SS Tetir and Coniplete SS Tear conditions, respectively. With the disruption of the transverse force couple, such as in the SSlIS Tear and Globul Tear conditions, the magnitude of the reaction force significantly decreased at maximum abduction (p < 0.05), measuring 19 f 5 N/kg and 12 ZIZ 4 N/kg, respectively. These decreases, relative to the hitact condition, were 61% and 73%) for the SSIIS Tear and for the Globol Tear conditions, respectively. However, there were no statistically significant differences between the SSIIS Tenr and Global Tear conditions (p > 0.05). The elevation angle, 0, of the glenohumeral reaction force demonstrated a continuous decrease throughout abduction reflecting a progressive change in the lines of action of each muscle active during simulated motion (Fig. 6). There were no significant differences in elevation angle found between the hitcict, hiconiylete SS Tear and Coriiplete SS Terw conditions (p > 0.05). The line of action of the supraspinatus muscle changes very little in the scapular plane throughout abduction, and thus, exclusion of the applied force in the supraspinatus tendon has little effect on the elevation angle throughout abduction [9,10,22]. This is evidenced by the absence of a
*
4
SS/lSTear
Fig. 6. Change in elevation angle of the joint reaction force, 0, during the raiige of abduction for the Inioc.1, hicoriiplete SS Tetrr. Coriiplc~tc~ SS Teor, and SSIIS Tear conditions (mean f S.D.).
significant change in elevation angle for tear configurations, which involved either a incomplete or complete supraspinatus tear in comparison with the Intact condition. The elevation angles for the SSIIS Tetir and Global Tear conditions were significantly different than for the three other testing conditions (p < 0.05), but not significantly different from each other 0)> 0.05), and thus, data for the Global Tear conditions is omitted from the graph. At the initiation of abduction, the line of action of the joint reaction force was directed more superiorly with respect to the glenoid articular surface, with the average elevation angle measuring 0 = -61 & 3" for all testing coiiditioiis (Fig. 6). As abduction angle increased, the line of action of the reaction force in the scapular plane became more perpendicular to the glenoid face. At maximum abduction, the average elevation angle for the Iiitiict, Inconipiete SS Teor and Coinplete SS Terir conditions measured 8 = -34 f 3", with no significant difference between individual conditions (p > 0.05). However, the elevation angle of the reaction force at maximum abduction for the SSIIS Tecir and GloDal Tear conditions, which measured to be 0 = -19 f 6", was significantly smaller than for those conditions in which the integrity of the transverse force couple was maintained 0, < 0.05). Fig. 8(A) demonstrates the change in elevation angle from initial to maximum abduction for tlie SSIIS Tear condition. The deviation angle, q , of the glenohunieral joint reaction force did not change significantly throughout the range of abduction for any rotator cuff testing condition 0) > 0.05). Fig. 7 shows the change in deviation angle for the hitact, Coniplete SS tear and SSIIS Tear conditions. The average deviation angle for the Iritrict, Iiiconiplete SS Tear and Coniplete SS Tear condition was cp = -10 & 6" and was not significantly different between these individual conditions (p > 0.05).
A
--
i
Discussion
Intact
Incomplete SS Tear Complete ss Tear
+
SSlISTear
I
100
Abduction Angle, degrees
Fig 7. Change i n deviation angle. ip, during the range of abduction for tlie Ii?Irrc/, Otcorii/~lereSS Teur. Complete S S Tctrr, and SSIIS Twr. conditions (mean f S.D.).
For these three conditions, the reaction force was directed anteriorly from the posterior aspect of the glenoid. For the SSIIS Teur and Global Teut.conditions, the deviation angle was significantly different from the previous conditions (cp = 2 & 6'1, but not different from each other, and thus, data for the Globnl Tear condition is omitted from the graph (JJ< 0.05). The positive values of cp indicate that the reaction force was directed posteriorly from the anterior side of the scapula. The superior view of the scapula in Fig. 8(B) shows the deviation angle of the reaction force during abduction for the Ii7tuc.t and SSIIS TWI.conditions at maximum abduction.
.................................
e =-w \ -
Maxinium abduction for SS/IS Tear (-50')
(A)
= -620\/
Initial abduction for S S l l S Tear
Fig. S. (A) Antero-posterior view of the scapula showing the change in elevation mple. 0, from initial to iiiaximtim abduction for tlie SSIIS T t w condition. ( B ) Superior view of the scapula showing the difference in deviation angle, cp. between tlie Irtrtrcf aiid SSIIS T t w conditions at tiinximum abduction. Scapular landmai-ks are labeled 11s follows: ( I ) acromion: ( 2 ) corncoid process; (3) iiiedial border of scapula.
This study investigated the effect of rotator cuff tears on the magnitude and direction of reaction forces at the glenohumeral joint during active motion. A direct realtime measurement of glenohumeral joint reaction forces was performed using a six-degree of freedom universal force sensor, incorporated into the DSTA. We found that both maximum abduction angle and magnitudes of the peak reaction force declined by over 50% for the SSI IS Terw and Global Tear conditions. As the joint reaction force represents the force, which counters the action of muscles active across the articuiation, this significant drop in force magnitude can be partially explained by the decrease in applied force used in our protocol. Individual peak reaction force magnitudes, however, do not solely reflect differences in the sum of applied force. For example, peak force magnitudes for the Ilzconiplete SS and Complete SS Teur configuration differed from the I n t m f condition by only 1 1% and 12Y4 respectively, despite a 25% difference in applied force. Similarly, extension of the tear into the infraspinatus tendon resulted in an additional 50% drop in reaction force magnitude over the Coniplete SS Teur condition despite a decrease in applied force of only 25%. Furthermore, despite a 25%)reduction in the applied force in the hzconipleze SS and Contplete SS Teur conditions, preservation of the transverse force couple of the subscapularis and infraspinatuslteres minor allowed the remaining rotator cuff and middle deltoid to produce sufficient force to overcome the mass moment of inertia of the full upper extremity and achieve a maximum abduction angle similar to the bzruc.t condition. Previous studies published by our laboratory and by others support these results and show that the small isolated supraspinatus tears often have minimal effect on glenohumeral joint motion and function [2,11,16,25,26]. The concavity compression through the action of the remaining intact rotator cuff was found to be sufficient to provide a fixed fulcrum for concentric rotation of the glenohumeral joint. These results suggest that the effect of rotator cuff tear configuration on the inagnitude of glenohumeral joint reaction forces becomes mechanically apparent when propagation of the tear disrupts the transverse force couple. For tear configurations in which the transverse force couple was disrupted, the deltoid force was unable to achieve maximum abduction. Thus, loss of effective concavity compression prevented arm elevation to 90" by the absence of a stable fulcrum for glenohumeral rotation. These finding are in agreement with both clinical and experimental studies which have shown that balanced transverse moments are necessary for a stable rotation fulcrum that permits abduction of the humerus [2,1 1,16,25,26]. The importance of concavity compression is also clinically apparent in the physical exam of paticnts with a rotator cuff tear. In patients with a
complete supraspinatus tear, motion and strength are more limited by pain than actual tendon dysfunction [17]. However, patients with large and massive tears involving the posterior aspect of the cuff demonstrate marked loss of motion, and strength is often insufficient to overcome the force of gravity [17]. Just as magnitude of the reaction force was significantly affected by disruption of the transverse force couple, the direction of the reaction force also differed significantly as simulated tears were extended into the posterior and anterior aspects of the rotator cuff. Loss of the centering effect of concavity compression resulted in a decrease in the inferiorly directed force vector. Reaction force directions in the transverse plane also differed significantly when the transverse force couple was disrupted. Posterior extension of simulated rotator cuff tears shifted a larger proportion of the resultant force to the component generated by the action of the anterior rotator cuff, thereby, changing the direction of the resultant force to become more posterior. Because the surface area of articulation between the humerus and the glenoid is small, subtle alterations i n the muscular control of shoulder movement due to rotator cuff injuries may consequently exaggerate changes in load transmission across the glenohumeral joint. Shoulder motion is a synthesis of the combined motion of the glenohuineral, scapulothroacic, acroniioclavicular and sternoclavicular joints. The complexity of this motion in combination with the involvement of more than 20 muscles, which govern it, poses challenges for developing an accurate model, which have thus far defied a comprehensive solution. Our current model was not able to reproduce scapular rotation, which changes glenoid inclination throughout the range of abduction. This drawback is shared by all simplified shoulder models. Scapular rotation allows the upper extremity to achieve 90" of abduction without maximizing the deltoid moment arm and serves to reduce joint reaction forces at a position when the moment of inertia of the upper extremity is at its greatest [19-211. Inability to account for scapular rotation in this study likely resulted in magnitudes of the joint reaction force, which overestimate in vivo conditions in an absolute sense. Nevertheless, the constancy of the experimental protocol between rotator cuff tear conditions does allow coinparison of force values between conditions in which the tear configuration is varied to disrupt the transverse force couple. This study provides quantitative data on the effect of rotator cuff tears on magnitude and direction of the reaction force at the glenohunieral joint. Our results confirm the original hypothesis and suggest that the integrity of the transverse force couple significantly affects both the magnitude and direction of these reaction forces. We believe that this information provides scientific basis to clarify the significance of concavity coin-
pression and the relationship between joint motion, joint compression and stability. Our future studies will include evaluation of the reaction forces at the glenohunieral joint following shoulder arthroplasty using various component designs and fixation techniques. This procedure has become a common practice for humeral or scapular fractures as well as for final stage of glenohumeral arthritis, however, it has been shown that changes in the reaction forces at the glenohunieral joint are consistent with the glenoid component loosening and detachment and that component design and fixation are critical to the longevity of the implant [4,7].
Acknowledgements The technical assistance of Jamie Pfaeffle, Theodore Mansoii and Mark Knaub as well as financial support of Ms. Erin McGurk, Orthopaedic Research Laboratories Alumni Council, and University of Pittsburgh Medical Center are gratefully acknowledged.
References [ I ] Apreleva M. Experimental and computational approach to sttidy contact niechanics at the g~enohunieraljoint. In: Bioengineei-ing department. Pittsburgh: University of Pittsburgh; 1999. p. 182. [2] Apreleva M, Parsons IM, Warner JJ, FLIFH, Woo SL. Experimental investigation of reaction forces at the glenohumeral joint during active abduction. J Shoulder Elbow SuIg 2000:9(5):409-17. [3] Bassett RW, Browne AO. Morrey BF. An K N . Glenoliumerd muscle force and iiioinent mechanics in a position of shoulder instability. J Biomech 1990;23:405-15. [4] Collins D, Tencer A, Sidles J, Matsen Fd. Edge displacement and deformation of glenoid components in response to eccentric loading. The effect of preparation of the glenoid bone. J Bone Joint Surg Am Vol 1992:74(4):501-7. [5] Debski RE, McMahon PJ. Thompson WO, Woo SL, Warner JJP. Fu FH. A new dynamic testing appal-atus to study glenohtimeral motion. J Biomech 1995;28:869-74. [6] Ftjie H. Livesay GA, Woo SL-Y. The use of a universal forcemoment sensor to determine in-situ forces i n ligaments: a new methodology. J Biomech Eiig 1995:117:1-7. [7] Hawkins RJ, Bell R H , Jallay B. Total shoulder arthroplasty. Clin Ortliop 1989242: 188-94. [8] Howell SM. Gnliiiat BJ, Reiizi AJ. Marone PJ. Normal and abnormal mechanics of the gleiiohiinieral joint in the horizontal plane. J Bone Joint Surg Am Vol 1988:70:227-32. [9] Howell SM. Iinobersteg AM, Seger DH, Marone PJ. Clarification of the role or the supraspinatus nitiscle in shoulder function. J Bone Joint Surg Am Vol 1986;68:398-404. [lo] Howell SM, Kraft TA. The role of' the supraspinatiis and infraspinatiis muscles in glenohuiiieral kineniatics of anterior should instiibility. C h i Orthop 1991;263:128-34. [ I I ] Hsu HC. Luo ZP. Stone JJS. An K N . Importance of rotator cuR balance on glenohumeral stability and degeneration. Trans Orthop Res Soc 1996:21:232. [I?] Inman VT. Sotinders M, Abbott LC. Observations on the function of the shoulder joint. J Bone Joint Surg A m Vol 1944:42:1-30.
[ I 31 Karduna AR. Kinematics of the glenoliumeral joint: influences of muscle forces, ligainentous constraints, and articular geometry. J Orthop Res 1996;14(6):986-93. [14] Klein AH, McKeriian DJ, Harner CD, Davis PL, Fu FH. A study of rotator cuff force vectors across tlie glenohumeral join! using magnetic resonance imaging. Trans Orthop Res SOC1989:230. [15] Lippitt S, Matsen F. Mechanisms of glenohumeral joint stability. C h i Orthop 1993;(291):20-8. [16] Lippitt SB. Vanderhooft JE, Harris SL, Sidles JA, Harrynian DT, Matsen I11 FA. Glenohumeral stability from concavity-compression: a quantitative analysis. J Shoulder Elbow Surg 1993;2:27-34. 1171 Matsen FA, Arntz CT, Lippitt SB. Rotator cuff. In: Matsen RA. editor. The shoulder. Philadelphia: W.B. Saunders Company; 1998. p. 755-839. [I81 McMahon PJ, Debski RE, Thompson WO, Warner JJP, Fu FH, Woo SL. Shoulder muscle forces and tendon excursions during glenohumeral abduction in the scapular plane. J Shoulder Elbow Surg 1995;4: 199-208. 1191 Morrey BF. Itoi E, An KN. Biomechanics of the shoulder. I n : Rockwood Jr. CA. Matsen I11 FA, editors. The shoulder. Philadelpliia: W.B. Saunders Company; 1998. p. 20845. [20] Nobuhara K. Ikeda H, Shiba R. Analysis of shoulder motion and related forces. In: Chao EYS, Ivins JC, editors. Tumor prosthesis for bone and joint reconstruction: the design and application. New York: Tlieinie-Stratton; 1983. p. 415-51. [21] Poppen NK, Walker PS. Normal and abnormal motion of the shoulder. J Bone Joint Surg Am Vol 1976;58:195-201.
[22] Poppen NK, Walker PS. Forces at the glenohumeral joint in abduction. Clin Orthop 1978;135:165-70. [23] Saha AK. Dynamic stability of the glenoliumeral joint. Acta Orthop Scand 1971:42:491-505. [24] Soslowsky LJ, Flatow EL, Bigliani LU, Mow VC. Stabilization of the glenohumeral joint by articular contact and by contact in the subacromial space. In: Matsen 111 FA, Fu FH, Hawkins RJ, editors. The shoulder: a balance of niobility and stability. Rosemont, IL: American Academy of the Orthopaedic Surgeons; 1993. p. 107-26. [25] Thompson WO, Debski RE, Boardman I11 ND, Taskiran E, Fu FH, Woo SL-Y, et a). A bioniechanical analysis of rotator cuff deficiency in a cadaveric model. Am J Sports Med 1996;24: 28692. [26] Warner JJP, Bowen MK, Deng XH, Torzilli PA, Warren RF. Effect of joint compression on inferior stability of the glenohumeral joint. J Shoulder Elbow Surg l999;8( l):3l-6. [27] Warner JJP, Gerber C. Massive tears of the postero-superior rotator cuff. In: Warner JJP, Iannotti JP, Gerber C, editors. Complex and revision problems in shoulder surgery. Philadelphia: Lippincott-Raven; 1997. p. 177-201. [28] Woo SL. Debski RE, Patel PR. Biomechanics of the full upper extremity in simple abduction: application of the Pittsburgh dynamic shotilder testing apparatus. In: J-J W, editor. Shoulder surgery The Asian prospective. Taiwan: Orthopaedic department, Veterans General Hospital-Taipei: Taipei; 1995. p. 23-8.
Journal of Orthopaedic Research 20 (7002) 439446
Journal of Orthopaedic Research www.elsevie~-.con~/locate/orthres
The effect of rotator cuff tears on reaction forces at the glenohumeral joint I.M. Parsons, IV, Maria Apreleva
I,
Freddie H. Fu, Savio L.-Y. Woo
''
Abstract The rotator cuff muscles maintain glenohumeral stability by compressing the hunieral head into the glenoid. Disruption of the rotator cuff compromises concavity compression and can directly affect the loads on the glenohumeral joint. The purpose of this study was to quantify tlie effect of rotator cuff tears on the magnitude and direction of glenohumeral joint reaction forces during active shoulder abduction in the scapular plane using nine cadaveric upper extremities. Motion of the full upper extremity was simulated using a dynamic shoulder testing apparatus. Glenohumeral joint reaction forces were measured by a universal forcemoment sensor. Five conditions of rotator cuff tears were tested: Intcict, bicoiiiplete Suprcispiiiutus Teur, Coiiiplete SLil)rcis~)iiicitus Teur, Supru.~~~inutuslInfi.u.~piizatirs Tear, and Global Tear. Reaction forces a t the glenohumeral joint were found to steadily increase throughout abduction and peaked at maximum abduction for all conditions tested. There were no significant differences in reaction force magnitude for the intact condition (337 f 88 N) or those involving an isolated incomplete tear (296 f 83 N ) or complete tear (300 f 85 N) of the supraspinatus tendon. Extension of tears beyond the supraspinatus tendon into the anterior and posterior aspect of the rotator cuff led to a significant decrease in the magnitude of joint reaction force (126 f 31 N). Similarly, such tears resulted in a significant change in the direction of the reaction force at the glenohumeral joint. These results suggest that joint reaction forces are significantly affected by the integrity of the rotator cuff, specifically, by tlie transverse force couple formed by the anterior and posterior aspects of the cuff. The quantitative data obtained in this study on the effect of rotator cuff tears on magnitude and direction of the reaction force a t the glenohumeral joint helps clarify the relationship between joint motion, joint compression and stability. 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. A11 rights reserved.
Introduction
Experimental studies have demonstrated that the rotator cuff serves two principle functions at the glenohumeral joint: (1) generation of torque necessary for rotation of the humerus on the glenoid; and (2) compression of tlie liumeral head into the glenoid concavity. This latter function compensates for the lack of inherent bony stability at the glenohumeral joint, serving as the primary stabilizing mechanism for this minimally constrained articulation during the functional range of motion [3,13,23,26]. Termed concavity compression, the stabilizing mechanism of the rotator cuff depends on the integrity of a transverse force couple, formed by the *Corresponding author. Tel.: +I-412-648-2000; fax: +I-412-6482001. E-i?wiI trt1dres.s: [email protected] (S.L.-Y. Woo).
'
Current address: Orthopedic Biomechanics Laboratory. Beth Israel Deaconess Medical Center, Harvard Medical School. Boston, MA, USA.
anatomical arrangement of the anterior (subscapularis) and posterior (infraspinatuslteres minor) rotator cuff tendons as they insert into the proximal humerus [15,16, 23,24,26]. Coiicavity compression also resists the superior pull of the deltoid muscle during abduction and provides a fulcrum for concentric rotation of the humeral head on the glenoid [ 15,161. The glenohunieral joint reaction force counteracts the sum of all shoulder muscle forces transmitted across the articulation, and its magnitude depends on the torque generated by the activity of these muscles in moving tlie upper extremity and resisting loads applied along its length [12,16,18,22,26]. Because the rotator cuff and deltoid muscles are the main abductors and rotators at the glenohumeral joint, tlie magnitude of the joint reaction force during active motion provides an index of the competence of concavity compression [25].Previous studies have demonstrated that disruption of the transverse force couple, as occurs in large and massive rotator cuff tears, leads not only to increased translations or subluxation of tlie liumeral head, but also to changes in
0736-0366/02/$ - see front matter 0 2002 Orthopaedic Research Society. Published by Elsevier Science Ltd. All rights reserved PII: S 0 7 3 6 - 0 2 6 6 ( 0 1 ) 0 0 1 3 7 - 1
the magnitude and direction of the reaction force at the glenohumeral joint [8,10,19,21,25]. Th~is,the degree to which different rotator cuff tear configurations affect the mechanical integrity of the transverse force couple can be studied in terms of their affect 011 the magnitude and direction of the glenohumeral joint reaction force during simulated active motion. To date, the effect of rotator cuff injury on glenohumeral joint mechanics has largely been investigated in terms of changes in articular kinematics; however, the effects of rotator cuff tears on glenohumeral joint reaction forces have not been fully elucidated in quantitative terms. In light of the complex relationship between joint stability and motion as mediated by the function of the rotator cuff, quantitative information addressing the consequences of cuff injury on reaction forces will contribute to a more comprehensive understanding of the association between tear configuration and loss of shoulder function. This information may eventually lead to scientific validation of surgical treatments aiiiied at restoring rotator cuff function. The objective of this study was, therefore, to quantify the effect of different rotator cuff tear configurations on glenohumeral reaction forces during active scapular plane abduction. We hypothesize that simulated postero-superior rotator cuff tears will result in a significant reduction in the reaction force magnitude measured during active motion due to disruption of the transverse force couple. In addition, such tear configurations will result in a change of the reaction force direction such that the vector points toward the deficient side of the cuff indicating a tendency to translate in the direction of the tear. Finally, we hypothesize, that in large tear configurations, in which the transverse force couple is disrupted, tlie reaction force will be directed more superiorly as the inferior vector of the infraspinatus/teres minor complex will no longer counteract the pull of the deltoid. This hypothesis is based on the premise that tear configurations involving the infraspinatus/teres minor and subscapularis disrupt the transverse force couple while isolated supraspinatus tendon tears do not.
Materials and methods This study employed tlie dynamic shoulder testing apparatus (DSTA) to simulate active scapular plane abduction. This testing system consists of a rigid frame, which supports a six-degree of freedom mounting block on which cadaveric upper extremities can be fixed to replicate the anatomical position of the slioulder relative to tlie axial skeleton (Fig. 1). A detailed description of this testing system has been previously published, and prior experimental studies have validated the accuracy of the DSTA in si~nulatinghighly repeatable six degree of freedom glenoh~~meral joint motion [5.26]. Full human upper extremities were used to recreiite the inertinl pi-operties of inass distributed along the arm's length during abduction in tlie scapular plane. Motion was simulated by applying forces to the tendons of the rotator cuff and deltoid muscles through specially designed sinusoidal clamps. which were connected to four servo-actuated hydraulic cylin-
ders through a cable-pulley system. Each cylinder was controlled independently by a computer with custom-designed software. The abduction angle of the glenohumeral joint was measured by a magnetic tracking device (The Bird, Ascension Technologies, Burlington. VT), rigidly mounted t o the hunieral shaft of the specimen. A six-degree of freedom universal force-moment sensor (UFS) (JR3, Woodland. CA) was incorporated into the testing system to allow for direct measurement of glenohumeral joint reaction forces during simulated active abduction (Fig. I). The UFS is able to nieasiire three forces and three moments along and about a Cartesian coordinate system defined with respect to the sensor. and its application toward non-contact force measureinent in soft tissues has been previously validated [6].The UFS was positioned on the testing apparatus so that its coordinate gystem was aligned with the axes of a reference coordinate system defined with respect to the scapula (Fig. 2). The axes of the reference system were defined as follows: the X-axis of the UFS corresponded to the anterior (ANT) direction in the antero-posterior plane of the scapula; the Y-axis corresponded to the superior (SUP) direction in the supero-inferior plane of the scapula; and the Z-axis corresponded to the lateral (LAT) direction in the medio-lateral plane of tlie scapula. Custom data acquisition software was designed to allow simultaneous. real-time nieasurenient of glenohumeral joint reaction forces. We performed preliminary studies to evaluate the repeatability of the dynamic shoulder testing system [I]. Ten trials of simulated active glenohumeral abduction were repeated on the same representative intact specimen, and magnitudes of the joint reaction force were recorded. Fig. 3 shows an average and a standard deviation of the magnitude of the glenohumeral joint reaction force after 10 trials of abduction on the same specimen. We determined that the force magnitudes measured by this DSTAIUFS testing system in our experimental set-up were repeatable to within 2 N when measuring reaction forces up to 500 N. Fourteen fresh-frozen full upper extremities were harvested from 13 human cadavers aged 60-80 years (mean age of 70 years) and stored at -20 "C prior to testing. After speciinens were thawed at room temperature, the skin and subcuta~ieoustissues proximal to the glenohumeral joint were removed, exposing the muscular envelope around the scapula. Remnants of the scalene, latissiinus dorsi, pectoralis major and minor, teres major and trapezius were sharply removed exposing the deltoid origin and rotator cuff m~iscles.A longitudinal skin incision was then made over the deltoid tuberosity of tlie humerus to expose the deltoid insertion. The biceps tendon and structures of the rotator interval were preserved during the dissection. and in all cases the biceps tendon was directly observed to be in continuity with its insertion on the superior glenoid labrum. Specimens with a history of trauma or prior surgery to the shoulder were excluded. Gross observation of the rotator cuff and deltoid tendons, further excluded any specimens with iiijury to or attrition of tlie tendons consistent with a rotator c u r tear o r rotator cuff tendinopathy. Five specimens from tlie original set of 14 specimens were excluded due to various pathological findings prior to o r during testing. This resulted in nine specimens: six inale and three female specimens from six male and two female cadavers (two specimens stemming from one female cadaver). Specially designed sinusoidal clamps were firmly secured to each of the rotator cuff and deltoid tendons. Since the infraspinatus and teres minor have a similar line of action and function similarly as external rotators at the shoulder, these muscles were combined to form one posterior tendon complex. To eliminate tlie eHect of distal motion on the inertial properties of the upper extremity. threaded pins were used to immobilize the wrist in neutral llexionlextension and neutral pronatioidsupination. Similarly the elbow was pinned i n full extension. The scapula was potted in an epoxy putty block and the specimen was positioned on tlie apparatus so that the glenoid was facing upwards 10" froin vertical and the scapular body was oriented 30" anteriorly relative to the frame of the DSTA. This position approximates tlie anatomical orientation of the scapula relative to tlie bony thorax when the upper extremity is positioned at the side. An antero-posterior radiograph of the mounted specimen was then taken to confirm the glenoid inclination. Once tlie scapula was appropriately positioned. cables from tlie servo-hydraulic cylinders were attached to the tendon clamps through a system of' pulleys, which could be adjusted to approximate tlie line of action ofeiicli muscle (Fie. 1 ) [14]. Great care WDS taken to assure that
44 1
Subscapularis
I
Direction of muscle forces used to simulate GH motion
Fig. I . Pittsburgh dynamic shoulder testing apparatus. (A) Universal force-moment sensor (UFS) attached to the scapular niount. (B) Cable-pulley system. (C) Load cell. (D) Hydraulic cylinder. (E) LVDT.
Scapular Plane
. .......
0 1 0
20
40
...
60
.
......
80
. .
100
Abduction Angle, deg
Fig. 2 . The coordinate system of the UFS is oriented on the testing apparatus to correspond to the scapular reference coordinate system. X-axis of the UFS corresponds to the anterior (ANT) direction, Y-axis corresponds to the superior (SUP) direction. and Z-axis represents lateral direction (LAT). Elevation, 0, is the angle that the resultant reaction force, R. makes with the transverse plane. Deviation. (p, is the angle that the resultant force component in the transverse plane, RAL. makes with the scapula plane.
the cables did not come into contact with either the scapula or the putty block. This ensured that the only forces measured by the load cell were the ones applied to the glenoid.
Fig. 3. An average and standard deviation of the magnitude of the glenohumeral joint reaction force after 10 trials of abduction repeated 011 the same representative intact specimen.
The line of action of each cable was verified with an antero-posterior radiograph of the testing system. Throughout the testing protocol, specimens were kept moist with physiologic saline solution. Prior to each test, 5 N of force was applied to each tendon to center the hunieral head on the glenoid, defining the “initial position” as 0” of abduction. The position of “maximum abduction” was then achieved by applying equal forces to each tendon at constant rate of 20 Nls until the upper extremity reached approxiniately 90” of scapular plane
abduction. This ratio of forces was based on prior experimental studies in our laboratory using tlie DSTA. Specimens were cycled through the range of maxinium abduction 25 times to minimize tlie effect of softtissue viscoelasticity on glenohumeral motion and reaction forces. For all subsequent testing conditions, forces were applied a t the same rate and custom-designed data acquisition software provided real-time monitoring of reaction forces. abduction angle. and glenohunieral kinematics throughout the range of motion. T o allow a inore direct comparison of tlie effect of rotator cuff injury between specimens of different sizes, we chose to base rotator cuff tear configuration on the number of involved tendons rather than tlie actual tear size. This description of tear configuration also correlates more closely witli observed clinical differences in tlie effect of tear size on shoulder function [27]. All rotator cuff tears were surgically simulated by incising the full thickness of tlie tendon at its insertion into tlie proximal liunierus. For each specimen, five rotator cuff tear testing conditions were defined a s follows: Iiifticr: no disruption of tlie rotator cuff; fircuiirplc~ieSzrprri.spbto/us(SS) Terii: disruption of the anterior half of the supraspinatus tendon: Curiiplete Siq>rrrspinri/us Tear: coniplete disruption of the supraspinatus tendon; Su/>irrs/>iri~r/u.s/fr~~~~iJiri(//iis (SSIfS) Terir: advancement of the tear 1.5 cin into tlie infraspinatuslteres minor tendon unit; and Gluhril Terir: advancement of tlie tear 1.5 cni anteriorly into tlie tendinous insertion of the subscapularis. We assumed that a full thickness disruption of a tendon from its bony attachment to the proximal huinerus would result in a complete dissociation of the applied force from the hunieral head. I n light of this assumption, no force was applied to those tendons in which a coniplete tear was surgically simulated. In addition. only 5 N of force was applied to prevent extension of simulated tears in conditions wliere only part of a tendon was disrupted such as in the Irtcuii~p/t./eSS and SSIIS Terir conditions. Consequently. the intact posterior fibers of the supraspinatus in the fiicoriiplete SS Tecrr configuration allowed transmission of applied force through the effect of tenodesis. This passive effect was not possible in the case of coinplete tendon disruption. For tlie Globol tear condition, the deltoid was tlie only tendon to which a force above 5 N was applied. Each testing condition was repeated three times to ensure repeatability and quality of data. For each condition, tlie magnitude and direction of tlie resultant reaction force at tlie glenohuineral joint were recorded. At the conipletion of the testing protocol, all speciniens were disarticulated at the glenoliunieral joint and tlie labruni and articular surfaces were inspected for evidence of a labral tear, degenerative joint disease or other f o r m of joint pathology. The magnitude of the resultant joint reaction force, R, was determined by the following equation:
where R A N T ,RSuP, and RLnr represent vector components of the reaction force in the anterior-posterior, superior-inferior, and mediallateral directions. recorded by tlie UFS (Fig. ?). The following conventions were defined to describe the direction of the joint reaction force with respect to the glenoid surface: s c r r / d o r plrrrrt: defined by the SUP-axis and LAT-axis of tlie scapula; mizswrst p/urw: defined by ANT-axis and LAT-axis of the scapula. Elevation, 0, is the angle that the resultant reaction force, R, makes witli the transverse plane. A negative 0 corresponds to tlie superior direction of the force (positive SUP-axis). and a positive 0 corresponds to tlie inferior direction of the force (negative SUP-axis). Deviation, cp, is tlie angle that tlie resultant force component in the transverse plane, RAL. makes with the scapula plane. A negative elevation angle cp corresponds to the anterior direction of the force (positive ANT-axis), and a positive cp corresponds to tlie posterior direction of tlie force (negative ANT-axis). Power analysis of tlie initial data was performed to estimate tlie required sample size and ensure 80%) power (c( = 0.05,p = 0.2). A Shapiro-Wilks' W test confirmed the normal distribution of all parameters of interest. A one-factor repeated-measures analysis of variance (ANOVA) was used to evaluate the effect of each rotator cuff tear condition on the magnitude of the resultant glenohumeral reaction force. Similarly, a one-factor ANOVA was used to determine tlie effect of each rotator cuff tear condition on the direction of the glenohu-
ineral joint reaction force. Multiple contrasts assessed differences between individual testing conditions. Statistical significance was set at p < 0.05. Statistica 5.0 software (StatSoft. Tulsa. OK) was used to perform all of the above analyses.
Results
The average weight of full upper extremities used in this investigation measured 3.9 f 1.0 kg (mean f S.D.). The maximum abduction angle achieved for the Infact condition was 85 f 10". This was not significantly different than the abduction angle achieved in the Incornplete SS Teur (86 i 15") and Coritplete SS Teur (87 i 16") conditions. For the simulated SSIIS Tear and GIobul Tear conditions, the arm was unable to abduct past 41 f 1 lo, and 27 f 1 lo, respectively. This 53% difference in abduction angle represented a statistically significant decline ( p < 0.05) compared to the maximum abduction achieved in tear conditions in which the transverse force couple was not disrupted. Fig. 4 shows the change in magnitude of the glenohumeral joint reaction force throughout abduction for the Intcict, Incomplete SS Teur, Complete SS Teur, SSIIS Teur, and Global Tear conditions. Data for the Incoiiiplete SS Tear condition did not differ significantly from the Intuct and Coinplete SS Teur conditions (p > 0.05). Similarly, data for the Global Tear condition did not differ significantly from the SSIIS Tear condition ( p > 0.05). For each rotator cuff tear condition, the joint reaction force increased with increasing abduction angle and peaked at maximum abduction. At maximum abduction, the joint reaction forces for the five testing conditions measured as follows: Intact = 337 f 88 N; Iticoriiplete SS Tear. = 296 k 83 N; Coriiplete SS Tear = 300 & 85 N; SS/IS Tear = 149 f 15 N; Global Tear = 126 k 31 N. Disruption of the transverse force couple resulted in a statistically significant decrease in reaction
0
20
40
60
80
100
Abduclion Angle, deg
Fig. 4. Change in niiignitude of the glenoliume~aljoint reaction force throughout abduction for tlie hi/trc/, fricboinplcrcSS Tcw. C C J I H ~SS ~CVC Tcai, SSIfS Tetrr, and Gluhrrl Terir conditions (mean f S.D.).
*
01
24
2 d
2
rza
.0
Abduction Angle, degrees r-7 I
Incomplete SS Tear
50
E Complete SS Tear
40
SS1ISTear
GlobalTenr
Y
2 30
BN
20
=
E!
I0
zg
o
i%
---
60
40 I
-
-20
{
6
M
-30
.-
-40
ze
T T
TT
60
a0
I00
-
I P
. d
P
5
Gi
-50
A
--
-60
Testing Conditions
Intact lncomplcle SS Tear Complete SS Tear
-70
Fig 5. Magnitude of the glenohumeral joint reaction force norinallzed to the weight of the upper extremity foi each of the testing conditions at maximum abduction (niean f S D.)
force magnitude as occurred in the SSIIS Tear and Globd Tear conditions. To account for the effect of specimen weight on the mass moment of inertia of tlie upper extremity during active motion, the values for the reaction force magnitudes were normalized to the specimen weight [22].Fig. 5 demonstrates magnitudes of the reaction at the glenohumeral joint at maximum abduction normalized to the weight of the upper extremity for each of the five testing conditions. As with abduction angle, tlie magnitudes of the reaction force at the glenohumeral joint did not differ significantly for the Intact, Incoriiplete SS Tecir and Complete SS Tear conditions, in which the transverse force couple was preserved. Normalized force magnitudes for these conditions measured 44 8 N/kg, 3 1 f 7 N/kg and 38 & 8 N/kg for the Iiitrrct, Iiicoriiplete SS Tetir and Coniplete SS Tear conditions, respectively. With the disruption of the transverse force couple, such as in the SSlIS Tear and Globul Tear conditions, the magnitude of the reaction force significantly decreased at maximum abduction (p < 0.05), measuring 19 f 5 N/kg and 12 ZIZ 4 N/kg, respectively. These decreases, relative to the hitact condition, were 61% and 73%) for the SSIIS Tear and for the Globol Tear conditions, respectively. However, there were no statistically significant differences between the SSIIS Tenr and Global Tear conditions (p > 0.05). The elevation angle, 0, of the glenohumeral reaction force demonstrated a continuous decrease throughout abduction reflecting a progressive change in the lines of action of each muscle active during simulated motion (Fig. 6). There were no significant differences in elevation angle found between the hitcict, hiconiylete SS Tear and Coriiplete SS Terw conditions (p > 0.05). The line of action of the supraspinatus muscle changes very little in the scapular plane throughout abduction, and thus, exclusion of the applied force in the supraspinatus tendon has little effect on the elevation angle throughout abduction [9,10,22]. This is evidenced by the absence of a
*
4
SS/lSTear
Fig. 6. Change in elevation angle of the joint reaction force, 0, during the raiige of abduction for the Inioc.1, hicoriiplete SS Tetrr. Coriiplc~tc~ SS Teor, and SSIIS Tear conditions (mean f S.D.).
significant change in elevation angle for tear configurations, which involved either a incomplete or complete supraspinatus tear in comparison with the Intact condition. The elevation angles for the SSIIS Tetir and Global Tear conditions were significantly different than for the three other testing conditions (p < 0.05), but not significantly different from each other 0)> 0.05), and thus, data for the Global Tear conditions is omitted from the graph. At the initiation of abduction, the line of action of the joint reaction force was directed more superiorly with respect to the glenoid articular surface, with the average elevation angle measuring 0 = -61 & 3" for all testing coiiditioiis (Fig. 6). As abduction angle increased, the line of action of the reaction force in the scapular plane became more perpendicular to the glenoid face. At maximum abduction, the average elevation angle for the Iiitiict, Inconipiete SS Teor and Coinplete SS Terir conditions measured 8 = -34 f 3", with no significant difference between individual conditions (p > 0.05). However, the elevation angle of the reaction force at maximum abduction for the SSIIS Tecir and GloDal Tear conditions, which measured to be 0 = -19 f 6", was significantly smaller than for those conditions in which the integrity of the transverse force couple was maintained 0, < 0.05). Fig. 8(A) demonstrates the change in elevation angle from initial to maximum abduction for tlie SSIIS Tear condition. The deviation angle, q , of the glenohunieral joint reaction force did not change significantly throughout the range of abduction for any rotator cuff testing condition 0) > 0.05). Fig. 7 shows the change in deviation angle for the hitact, Coniplete SS tear and SSIIS Tear conditions. The average deviation angle for the Iritrict, Iiiconiplete SS Tear and Coniplete SS Tear condition was cp = -10 & 6" and was not significantly different between these individual conditions (p > 0.05).
A
--
i
Discussion
Intact
Incomplete SS Tear Complete ss Tear
+
SSlISTear
I
100
Abduction Angle, degrees
Fig 7. Change i n deviation angle. ip, during the range of abduction for tlie Ii?Irrc/, Otcorii/~lereSS Teur. Complete S S Tctrr, and SSIIS Twr. conditions (mean f S.D.).
For these three conditions, the reaction force was directed anteriorly from the posterior aspect of the glenoid. For the SSIIS Teur and Global Teut.conditions, the deviation angle was significantly different from the previous conditions (cp = 2 & 6'1, but not different from each other, and thus, data for the Globnl Tear condition is omitted from the graph (JJ< 0.05). The positive values of cp indicate that the reaction force was directed posteriorly from the anterior side of the scapula. The superior view of the scapula in Fig. 8(B) shows the deviation angle of the reaction force during abduction for the Ii7tuc.t and SSIIS TWI.conditions at maximum abduction.
.................................
e =-w \ -
Maxinium abduction for SS/IS Tear (-50')
(A)
= -620\/
Initial abduction for S S l l S Tear
Fig. S. (A) Antero-posterior view of the scapula showing the change in elevation mple. 0, from initial to iiiaximtim abduction for tlie SSIIS T t w condition. ( B ) Superior view of the scapula showing the difference in deviation angle, cp. between tlie Irtrtrcf aiid SSIIS T t w conditions at tiinximum abduction. Scapular landmai-ks are labeled 11s follows: ( I ) acromion: ( 2 ) corncoid process; (3) iiiedial border of scapula.
This study investigated the effect of rotator cuff tears on the magnitude and direction of reaction forces at the glenohumeral joint during active motion. A direct realtime measurement of glenohumeral joint reaction forces was performed using a six-degree of freedom universal force sensor, incorporated into the DSTA. We found that both maximum abduction angle and magnitudes of the peak reaction force declined by over 50% for the SSI IS Terw and Global Tear conditions. As the joint reaction force represents the force, which counters the action of muscles active across the articuiation, this significant drop in force magnitude can be partially explained by the decrease in applied force used in our protocol. Individual peak reaction force magnitudes, however, do not solely reflect differences in the sum of applied force. For example, peak force magnitudes for the Ilzconiplete SS and Complete SS Teur configuration differed from the I n t m f condition by only 1 1% and 12Y4 respectively, despite a 25% difference in applied force. Similarly, extension of the tear into the infraspinatus tendon resulted in an additional 50% drop in reaction force magnitude over the Coniplete SS Teur condition despite a decrease in applied force of only 25%. Furthermore, despite a 25%)reduction in the applied force in the hzconipleze SS and Contplete SS Teur conditions, preservation of the transverse force couple of the subscapularis and infraspinatuslteres minor allowed the remaining rotator cuff and middle deltoid to produce sufficient force to overcome the mass moment of inertia of the full upper extremity and achieve a maximum abduction angle similar to the bzruc.t condition. Previous studies published by our laboratory and by others support these results and show that the small isolated supraspinatus tears often have minimal effect on glenohumeral joint motion and function [2,11,16,25,26]. The concavity compression through the action of the remaining intact rotator cuff was found to be sufficient to provide a fixed fulcrum for concentric rotation of the glenohumeral joint. These results suggest that the effect of rotator cuff tear configuration on the inagnitude of glenohumeral joint reaction forces becomes mechanically apparent when propagation of the tear disrupts the transverse force couple. For tear configurations in which the transverse force couple was disrupted, the deltoid force was unable to achieve maximum abduction. Thus, loss of effective concavity compression prevented arm elevation to 90" by the absence of a stable fulcrum for glenohumeral rotation. These finding are in agreement with both clinical and experimental studies which have shown that balanced transverse moments are necessary for a stable rotation fulcrum that permits abduction of the humerus [2,1 1,16,25,26]. The importance of concavity compression is also clinically apparent in the physical exam of paticnts with a rotator cuff tear. In patients with a
complete supraspinatus tear, motion and strength are more limited by pain than actual tendon dysfunction [17]. However, patients with large and massive tears involving the posterior aspect of the cuff demonstrate marked loss of motion, and strength is often insufficient to overcome the force of gravity [17]. Just as magnitude of the reaction force was significantly affected by disruption of the transverse force couple, the direction of the reaction force also differed significantly as simulated tears were extended into the posterior and anterior aspects of the rotator cuff. Loss of the centering effect of concavity compression resulted in a decrease in the inferiorly directed force vector. Reaction force directions in the transverse plane also differed significantly when the transverse force couple was disrupted. Posterior extension of simulated rotator cuff tears shifted a larger proportion of the resultant force to the component generated by the action of the anterior rotator cuff, thereby, changing the direction of the resultant force to become more posterior. Because the surface area of articulation between the humerus and the glenoid is small, subtle alterations i n the muscular control of shoulder movement due to rotator cuff injuries may consequently exaggerate changes in load transmission across the glenohumeral joint. Shoulder motion is a synthesis of the combined motion of the glenohuineral, scapulothroacic, acroniioclavicular and sternoclavicular joints. The complexity of this motion in combination with the involvement of more than 20 muscles, which govern it, poses challenges for developing an accurate model, which have thus far defied a comprehensive solution. Our current model was not able to reproduce scapular rotation, which changes glenoid inclination throughout the range of abduction. This drawback is shared by all simplified shoulder models. Scapular rotation allows the upper extremity to achieve 90" of abduction without maximizing the deltoid moment arm and serves to reduce joint reaction forces at a position when the moment of inertia of the upper extremity is at its greatest [19-211. Inability to account for scapular rotation in this study likely resulted in magnitudes of the joint reaction force, which overestimate in vivo conditions in an absolute sense. Nevertheless, the constancy of the experimental protocol between rotator cuff tear conditions does allow coinparison of force values between conditions in which the tear configuration is varied to disrupt the transverse force couple. This study provides quantitative data on the effect of rotator cuff tears on magnitude and direction of the reaction force at the glenohunieral joint. Our results confirm the original hypothesis and suggest that the integrity of the transverse force couple significantly affects both the magnitude and direction of these reaction forces. We believe that this information provides scientific basis to clarify the significance of concavity coin-
pression and the relationship between joint motion, joint compression and stability. Our future studies will include evaluation of the reaction forces at the glenohunieral joint following shoulder arthroplasty using various component designs and fixation techniques. This procedure has become a common practice for humeral or scapular fractures as well as for final stage of glenohumeral arthritis, however, it has been shown that changes in the reaction forces at the glenohunieral joint are consistent with the glenoid component loosening and detachment and that component design and fixation are critical to the longevity of the implant [4,7].
Acknowledgements The technical assistance of Jamie Pfaeffle, Theodore Mansoii and Mark Knaub as well as financial support of Ms. Erin McGurk, Orthopaedic Research Laboratories Alumni Council, and University of Pittsburgh Medical Center are gratefully acknowledged.
References [ I ] Apreleva M. Experimental and computational approach to sttidy contact niechanics at the g~enohunieraljoint. In: Bioengineei-ing department. Pittsburgh: University of Pittsburgh; 1999. p. 182. [2] Apreleva M, Parsons IM, Warner JJ, FLIFH, Woo SL. Experimental investigation of reaction forces at the glenohumeral joint during active abduction. J Shoulder Elbow SuIg 2000:9(5):409-17. [3] Bassett RW, Browne AO. Morrey BF. An K N . Glenoliumerd muscle force and iiioinent mechanics in a position of shoulder instability. J Biomech 1990;23:405-15. [4] Collins D, Tencer A, Sidles J, Matsen Fd. Edge displacement and deformation of glenoid components in response to eccentric loading. The effect of preparation of the glenoid bone. J Bone Joint Surg Am Vol 1992:74(4):501-7. [5] Debski RE, McMahon PJ. Thompson WO, Woo SL, Warner JJP. Fu FH. A new dynamic testing appal-atus to study glenohtimeral motion. J Biomech 1995;28:869-74. [6] Ftjie H. Livesay GA, Woo SL-Y. The use of a universal forcemoment sensor to determine in-situ forces i n ligaments: a new methodology. J Biomech Eiig 1995:117:1-7. [7] Hawkins RJ, Bell R H , Jallay B. Total shoulder arthroplasty. Clin Ortliop 1989242: 188-94. [8] Howell SM. Gnliiiat BJ, Reiizi AJ. Marone PJ. Normal and abnormal mechanics of the gleiiohiinieral joint in the horizontal plane. J Bone Joint Surg Am Vol 1988:70:227-32. [9] Howell SM. Iinobersteg AM, Seger DH, Marone PJ. Clarification of the role or the supraspinatus nitiscle in shoulder function. J Bone Joint Surg Am Vol 1986;68:398-404. [lo] Howell SM, Kraft TA. The role of' the supraspinatiis and infraspinatiis muscles in glenohuiiieral kineniatics of anterior should instiibility. C h i Orthop 1991;263:128-34. [ I I ] Hsu HC. Luo ZP. Stone JJS. An K N . Importance of rotator cuR balance on glenohumeral stability and degeneration. Trans Orthop Res Soc 1996:21:232. [I?] Inman VT. Sotinders M, Abbott LC. Observations on the function of the shoulder joint. J Bone Joint Surg A m Vol 1944:42:1-30.
[ I 31 Karduna AR. Kinematics of the glenoliumeral joint: influences of muscle forces, ligainentous constraints, and articular geometry. J Orthop Res 1996;14(6):986-93. [14] Klein AH, McKeriian DJ, Harner CD, Davis PL, Fu FH. A study of rotator cuff force vectors across tlie glenohumeral join! using magnetic resonance imaging. Trans Orthop Res SOC1989:230. [15] Lippitt S, Matsen F. Mechanisms of glenohumeral joint stability. C h i Orthop 1993;(291):20-8. [16] Lippitt SB. Vanderhooft JE, Harris SL, Sidles JA, Harrynian DT, Matsen I11 FA. Glenohumeral stability from concavity-compression: a quantitative analysis. J Shoulder Elbow Surg 1993;2:27-34. 1171 Matsen FA, Arntz CT, Lippitt SB. Rotator cuff. In: Matsen RA. editor. The shoulder. Philadelphia: W.B. Saunders Company; 1998. p. 755-839. [I81 McMahon PJ, Debski RE, Thompson WO, Warner JJP, Fu FH, Woo SL. Shoulder muscle forces and tendon excursions during glenohumeral abduction in the scapular plane. J Shoulder Elbow Surg 1995;4: 199-208. 1191 Morrey BF. Itoi E, An KN. Biomechanics of the shoulder. I n : Rockwood Jr. CA. Matsen I11 FA, editors. The shoulder. Philadelpliia: W.B. Saunders Company; 1998. p. 20845. [20] Nobuhara K. Ikeda H, Shiba R. Analysis of shoulder motion and related forces. In: Chao EYS, Ivins JC, editors. Tumor prosthesis for bone and joint reconstruction: the design and application. New York: Tlieinie-Stratton; 1983. p. 415-51. [21] Poppen NK, Walker PS. Normal and abnormal motion of the shoulder. J Bone Joint Surg Am Vol 1976;58:195-201.
[22] Poppen NK, Walker PS. Forces at the glenohumeral joint in abduction. Clin Orthop 1978;135:165-70. [23] Saha AK. Dynamic stability of the glenoliumeral joint. Acta Orthop Scand 1971:42:491-505. [24] Soslowsky LJ, Flatow EL, Bigliani LU, Mow VC. Stabilization of the glenohumeral joint by articular contact and by contact in the subacromial space. In: Matsen 111 FA, Fu FH, Hawkins RJ, editors. The shoulder: a balance of niobility and stability. Rosemont, IL: American Academy of the Orthopaedic Surgeons; 1993. p. 107-26. [25] Thompson WO, Debski RE, Boardman I11 ND, Taskiran E, Fu FH, Woo SL-Y, et a). A bioniechanical analysis of rotator cuff deficiency in a cadaveric model. Am J Sports Med 1996;24: 28692. [26] Warner JJP, Bowen MK, Deng XH, Torzilli PA, Warren RF. Effect of joint compression on inferior stability of the glenohumeral joint. J Shoulder Elbow Surg l999;8( l):3l-6. [27] Warner JJP, Gerber C. Massive tears of the postero-superior rotator cuff. In: Warner JJP, Iannotti JP, Gerber C, editors. Complex and revision problems in shoulder surgery. Philadelphia: Lippincott-Raven; 1997. p. 177-201. [28] Woo SL. Debski RE, Patel PR. Biomechanics of the full upper extremity in simple abduction: application of the Pittsburgh dynamic shotilder testing apparatus. In: J-J W, editor. Shoulder surgery The Asian prospective. Taiwan: Orthopaedic department, Veterans General Hospital-Taipei: Taipei; 1995. p. 23-8.