Journal o j Orthopaedic Research 10:187-197 Raven Press, Ltd., New York 63 1992 Orthopaedic Research Society

Tensile Properties of the Inferior Glenohumeral Ligament Louis U. Bigliani, Roger G. Pollock, Louis J. Soslowsky, Evan L. Flatow, Robert J. Pawluk, and Van C. Mow Orthopaedic Research Laboratory, Depurtment of Orthopaedic Surgery, Coiumbiu University, New York, New York, U.S.A.

Summary: The tensile properties of the inferior glenohumeral ligament have been determined in 16 freshly frozen cadaver shoulders. The inferior glenohumeral ligament was divided into three anatomical regions: a superior band, an anterior axillary pouch, and a posterior axillary pouch. This yielded 48 hone-ligament-bone specimens; which were tested to failure in uniaxial tension. The superior hand was consistently the thickest region, averaging 2.79 mm. The thickness of the inferior glenohumeral ligament decreased from antero-superiorly to postero-inferiorly. The resting length of all three anatomical regions was not statistically different. Total specimen strain to hilure for all bone-ligament-bone specimens averaged 27%. Variations occurred between the three regions, with the anterior pouch specimens failing at a higher strain (34%) thar, those from the superior band (24%)or the posterior pouch (23%). Strain to failure for the ligament midsubstance (11%) was found to be significantly less than that for the entire specimen (27%). Thus, larger strain must occur near the insertion sites of the inferior glenohumeral ligament. Stress at failure was found to be nearly identical for the three regions of the ligament, averaging 5.5 MPa. These values are lower than those reported for other soft tissues, such as the anterior cruciate ligament and patellar tendon. The anterior pouch was found to be less stiff than the other two regions, perhaps suggesting that it is composed of more highly crimped collagen fibers. Three failure sites were seen for the inferior glenohumeral ligament: the glenoid insertion (40%), the ligament substance (35%). and the humeral insertion (25%). In addition, significant capsular stretching occurred before failure, regardless of the failure mode. Key Words: Ligament-Biomechanics-Inferior glenohumeral ligament-Material testing-Instability-Shoulder.

Recurrent anterior glenohumeral instability (dislocation or subluxation) is a common clinical problem that can often lead to significant disability. Maintaining glenohumeral stability is a complex phenomenon that depends on the interaction of dynamic muscular forces and static capsulo-ligamentous restraints. Many investigators have at-

tempted to determine the relative importance of these dynamic and static stabilizers, and the underlying mechanisms and pathology responsible for glenohumeral instability (6,8,11,16-18,21,36). Recent studies have shown that with the arm in the position of anterior apprehension (90" abduction and full external rotation), the inferior glenohumeral ligament serves as the primary static restraint against anterior instability of the shoulder (12,25, 37). Although the anatomy and function of the anterior static stabilizers of the shoulder have been de-

Received March 4, 1991; accepted September 26, 1991. Address correspondence and reprint requests to Dr. Louis U. Bigliani at 161 Fort Washington Avenue. Akhley Pavilion, Room 245, New York, NY 10032, U.S.A.

187

188

L . U . BIGLIANI ET AL.

scribed, few reports have focused on the mechanical and structural properties of these tissues. For example, Reeves determined the tensile strength of the glenoid attachment (labrum), the antero-inferior capsule, and the subscapularis tendon on intact shoulder specimens from various age groups. He observed that in young specimens the glenoid attachment site is the weakest, whereas in older specimens the anterior capsule and subscapularis are weaker (29). Kaltsas tested the anterior capsule as a whole with the shoulder in 90" of abduction and demonstrated that the antero-inferior capsule always ruptured first during tensile testing. A maximum peak force of 2,000 N was observed before failure occurred (14). However, these studies did not specify strain rates or the degree of externalhternal rotation of the shoulder during testing. Turkel et al. and others have reported that tension shifts among different regions of the shoulder capsule with varying arm positions, suggesting some degree of specificity of biomechanical function among the different regions of the joint capsule (12,25,37). The success of operative repairs for shoulder instability depends on the accurate identification and correction of existing pathology. Previous clinical reports concerning the etiology of glenohumeral instability have emphasized either capsular stretching or capsular detachment (3-5,13,17,20,21,32,35,36). However, capsular detachment is not always present and thus cannot be the only mechanism for glenohumeral instability. It is difficult to assess capsular stretching in situ, whether or not there is a capsular detachment. At surgery, impressions can be formed about the degree of capsular laxity present in a particular patient's shoulder. This capsular laxity, if present, is dependent on the initial length of the capsular structures, which act as restraints to subluxations. Little information exists concerning the initial length of the different regions of the capsule, thereby making it difficult to accurately assess pathologic situations. Furthermore, no data are available for the mechanical behavior of the capsular tissues, and in particular for the inferior glenohumeral ligament. If reconstructive procedures are to restore normal anatomy and biomechanical function, it is important to know these simple and fundamental properties. Thus, the purpose of the present study is to obtain information about the anatomical dimensions of the inferior glenohumeral ligament, as well as its mechanical behavior in tension. Specifically, the anatomic and material

J Orthop Res, Vol. 10,No. 2,1992

parameters to be determined in this study are the regional thicknesses, widths, and lengths of the inferior glenohumeral ligament, total specimen strain to failure for the entire bone-ligament-bone specimen (total specimen strain or percentage elongation), strain to failure for the ligament substance alone (midsubstance strain), ultimate tensile stress (failure stress), tensile modulus (from the linear portion of the stress-strain curve), and the material parameters obtained from an exponential stressstrain law. MATERIALS AND METHODS

Sixteen freshly frozen human shoulders were obtained from nine male and seven female cadavers. These specimens ranged in age from 56 to 87 years and represented 10 right and six left shoulders. All specimens were stored at - 20°C and thawed for 12 h before dissection. During storage, dissection, preparation, and geometric measurement, all specimens were moistened using a physiologic saline enzyme-inhibitor solution to retard dehydration and specimen degradation. Initially, each cadaver shoulder was dissected to the level of the capsular ligaments. Using both sharp and blunt dissection, all musculature, including the rotator cuff, was carefully removed without disturbing the underlying capsule or bony insertion sites. The inferior glenohumeral ligament was consistently separated from the subscapularis tendon and muscle without difficulty. Shoulder joints were further isolated anatomically by transecting the humeral shaft at the surgical neck 5-6 cm below the capsular insertion, then by transecting the scapula -2-3 cm medial to the capsular insertion region, and finally by partially excising the acromion and coracoid processes. The superior glenohumeral ligament, biceps tendon, and postero-superior capsule were removed to allow visualization of both the inner and outer surfaces of the middle and inferior glenohumeral ligaments. The humeral head was excised at the anatomical neck, superior to the capsular insertion, clearly exposing the inferior border of the middle glenohumeral ligament. A microsagittal saw was used to ensure precise location of all bony transections about the joint capsule insertion sites. The inferior glenohumeral ligament was divided into three anatomical regions, as previously described by Turkel et al. (Fig. 1): the superior band, the anterior portion of the axillary pouch, and the postcrior portion of the pouch (37). The superior

TENSILE PROPERTIES OF THE INFERIOR GLENOHUMERAL LIGAMENT

FIG. 1. Schematic of the glenohumeral ligaments as viewed from posterior with the joint capsule partially opened. The inferior glenohumeral ligament is composed of three regions: a superior band, an anterior axillary pouch, and a posterior axillary pouch. Reproducedwith permission from ref. 19.

band was a consistently palpable thickening that was distinct from the anterior pouch. However, direct palpation was not useful in differentiation between the anterior and posterior pouch. Therefore, a coordinate system for the humeral head and glenoid was established to further define the three capsular subregions, thus allowing consistent dissections from specimen to specimen (Fig. 2). With re-

90" 0

80"

189

spect to the glenoid surface, the posterior pouch spanned from 210" to 270", the anterior pouch from 270" to 330" and the thick, palpable superior band from 330" to 30". On the humeral surface, the coordinates of the attachment sites for these three capsular regions are shifted clockwise (left shoulder) between 20" and 30". For example, the posterior pouch attachment on the humeral side would be from 190" to 250". For the right shoulder, 20" to 30" is added in the counterclockwise direction. The inferior glenohumeral ligament capsular regions were sharply divided longitudinally between their glenoid and humeral bony attachments. The microsagittal saw was used to extend these cuts through the humeral neck and glenoid to avoid damage at the ligamentous insertion sites. This dissection technique allowed all 16 inferior glenohumeral ligament specimens to be reproducibly divided into three bone-ligament-bone specimens (Fig. 3). Before tensile testing, the width and thickness of the ligament substance were measured to allow calculation of cross-sectional areas and average stress. A Bausch and Lomb stereo-zoom microscope, adapted with an Olympus precision X-Y translation stage, and coupled to a Microcode digital micrometer was used for width measurement (31). These measurements were taken at a minimum of Seven locations along the length of the specimen and averaged to obtain an average specimen width (Table I). Specimen thickness was measured with a Leitz electrooptical micrometer ( I ) and was recorded at a

0"

FIG. 2. Coordinate system developed for defining the insertions of the three regions of the inferior glenohumeral ligament. Typically, the posterior axillary pouch inserts on the glenoid from 210" to 270", the anterior axillary pouch from 270" to 330",and the superior band from 330" to 30".On the humeral surface, the coordinates are shifted clockwise (for a left shoulder) between 20" and 30".

FIG. 3. The inferior glenohumeral ligament sectioned into three bone-ligament-bone specimens, each representing an anatomical region of the ligament, for mechanical testing. A: Superior band. B: Anterior pouch. C: Posterior pouch.

J Orthop Res, Vol. 10, N o . 2, 1992

L. U . BIGLlANI ET AL.

190

TABLE 1. Stress-strain dala at failure

Geometric dala ~ ~~

Stress (MPa)

Strain"

(mm)

Length (mm)

(%I

Strainb (%)

13.33 t 2.66 12.61 i 3.05 10.86 5 2.94 12.27 i 2.88

41.3 i 4.5 39.8 2 5.6 41.0 f 4.2 40.7 i 4.7

5.2 i 2.7 5.5 i 2.0 5.6 2 1.9 5.5 i 2.2

24.0 i 6.2 34.0 2 10.5 23.1 f 4.6 27.0 2 8.9

8.3 f 3.2 15.1 +- 5.7 9.9 ? 5.3 10.9 5 5.5

Width

Ligament region

Thickness (mm)

Superior (n = 16) Anterior (n = 16) Posterior (n = 16) Average (n = 48)

2.79 f 0.49 2.34 2 0.43 1.70 t 0.55 2.28 i 0.66

Data are means +- SD. a Total specimen strain. Video dimensional analyzer.

minimum of seven locations along the ligament length and width, and averaged to obtain an average specimen thickness (Table 1). All specimens were noted to be relatively flat with rectangular cross sections. Therefore, cross-sectional area was calculated as the product of the average specimen width and the average specimen thickness. The glenoid and humeral components of a boneligament-bone specimen were positioned within disposable acrylic cylinders and embedded in fastcuring polymethylmethacrylate cement (PMMA) (Fig. 4). The ligament-bone interface was maintained 2 5 mm above the cement surface to avoid thermal damage to the ligament insertion sites during PMMA polymerization. The acrylic cylinders were then inserted into specially designed aluminum specimen holders, which were mounted on an MTS model 858 biaxial servo-hydraulic machine. Each specimen holder was attached to ball joints providing 3" of freedom. The upper holder assembly was rigidly connected to a 5,000 N load cell fixed to the MTS actuator. The lower specimen holder was mounted on an X-Y slider, providing 2 dJI and attached to the MTS base plate. The final degree of freedom, translation in the Z direction, was provided by the axial motion of the MTS actuator. This specimen mounting assembly was designed with 9 df to facilitate accurate positioning of the boneligament-bone specimen. With knowledge of inferior glenohumeral ligament fiber orientation (191, the prevailing ligament fiber bundles could now be aligned parallel to the tensile loading direction by adjustment of the ball joints and the X-Y translation stage. During embedding and handling. preloading of the specimen was avoided by maintaining the ligament in a lax configuration. The load cell was set to zero with the aligned specimen in the lax state. The ligament was then minimally loaded in the MTS and the ligament resting length measured using a rule placed between the

I Orthop Res, Vol. 10, N o . 2 , 1992

bony insertion sites (Table 1 ) . Using a simple preloading protocol as has been advocated by other investigators for loading to failure (39), the ligament was slowly stretched to a 1.5 N tare load and allowed to stress relax for 5 min. Two experimental

t

1

To Upper Specimen

Holder, Load Cell and MTS

I I t PMMA

Acrylic cylinders

To Lower Specimen Holder, X-Y translation and MTS base

FIG. 4. The glenoid and humeral components of the boneligament-bone specimens are embedded in polyrnethylmethacrylate within acrylic cylinders. These cylinders are then inserted into the specimen holders, which are mounted on the MTS servo-hydraulic machine. During loading in uniaxial tension, the length of theentire specimen [LGG(t)],as well as the length of a small-gauge section of the ligament substance [LvDn(t)], are measured in order to calculate total specimen strain and midsubstance strain, respectively.

TENSILE PROPERTIES OF THE INFERIOR GLENOHUMERAL LIGAMENT

groups, each comprising 24 bone-ligament-bone specimens, were then tested at constant elongation rates of either 0.04 mm/s (-0.001 sC1) or 0.004 mm/s (-0.0001 s - ' ) . To avoid ligament dehydration during the tensile test, each specimen was immersed in a chamber containing room temperature physiologic saline with protease inhibitors. During tensile testing, ligament strains were measured using two independent techniques. First, percentage elongation or total specimen strain, defined as the change in bone-ligament-bone specimen length divided by its original length, was calculated for the entire bone-ligament-bone specimen (Fig. 4). The change in bone-ligament-bone specimen length was measured from the axial motion of the MTS actuator because all other components in series with the specimen were designed to displace negligibly when compared with tissue elongation (>40:1) for the maximum loads applied. Second, midsubstance ligament strain was measured with an optical video dimensional analyzer (Instrumentation for Physiology and Medicine, Inc., San Diego, CA, U.S.A.) (40,42). For each specimen, a midsubstance gauge section was established by making two thin, parallel India ink lines, spaced -8 mm apart. During tensile testing, changes in gauge section length, averaged over the specimen's central region, were electronically tracked by the video dimensional analyzer (40,42), This method provided ligament midsubstance strain data that were measured independent of strains at the insertion sites. For each structural and material parameter obtained in this study, statistical comparisons were made between the three regions of the inferior glenohumeral ligament. Specifically, the one-way analysis of variance (ANOVA ONEWAY) was used to determine how the independent classification variable affects the data, and the Student-NewmanKeuls multiple range test (SIMULTANEOUS SNK) was used to compare a number of independent samples simultaneously to determine where significant differences lie between the groups.

RESULTS Anatomy Significant differences in thickness were found among the three regions of the inferior glenohumeral ligament. The superior band was significantly thicker than the anterior pouch (p < 0.025) and the posterior pouch (p < 0.001). The anterior pouch

191

was also thicker than the posterior pouch (p < 0,001). No significant differences were found in resting lengths or cut widths for these three inferior glenohumeral ligament regions. The values for these anatomical measurements are shown in Table 1 in the form of means -t standard deviations. Elongation Rate

N o significant differences in midsubstance failure strain, total specimen failure strain, failure stress, failure modes, or curve-fitting parameters were found between specimens tested at 0.04 mmis and those tested at 0.004 mmis. Therefore, data from these two groups were combined in all analyses. Failure Strains and Stresses Significant differences in total specimen strain for the three inferior glenohumeral ligament regions were determined from the ligament extension measurements (Table l). The anterior pouch failed at a higher total specimen strain than either the superior band (p < 0.001) or the posterior pouch (p < 0.001). Midsubstance ligament strain was obtained from video dimensional analyzer measurements for 40 of the 48 specimens tested. Again, significant differences were noted for midsubstance ligament strains to failure, which paralleled the results for the total specimen strain measurements. The anterior pouch failed at a higher strain than either the superior band (p < 0.00s) or the posterior pouch (p < 0.01). Average midsubstance failure strains, from video dimensional analyzer measurements, represented only 3545% of the total specimen strain at failure, indicating strain variation? along the length of the inferior glenohumeral ligament and that considerable strain must exist in the region near the inferior glenohumeral ligament bone insertion site. The variation of the failure strains for the three regions of the inferior glenohumeral ligament demonstrated consistent trends, independent of the failure mode. More specifically, strain values for specimens failing at midsubstance were consistent with values obtained for specimens failing at the insertion sites. Therefore, it is clear that even in specimens that ultimately failed at their insertion sites, significant ligament midsubstance strain was present. Failure strains, regardless of failure site, for the anterior pouch were greater than for the superior band or posterior pouch. However, the stress at failure was found to be relatively uniform for all

J Orthop Res, Vol. 10,No. 2 , 1992

192

L . U.BIGLIANI ET AL.

three regions of the inferior glenohumeral ligament (Table I ) .

where u is the tensile stress and E is the engineering strain given by AUl, (A1 is the change in length and 1, is the original length of the specimen), and A and B are material coefficients. From this exponential stress-strain representation, the derivative or tangent modulus (stiffness at any stress u) of the material is given by

Failure Modes Three modes of failure were observed in this study: at the glenoid insertion site in 19 of 48 specimens (40%), in the ligament midsubstance in 17 of 48 (35%), and at the humeral insertion site in 12 of 48 (25%). Specifically, seven superior band specimens failed at the glenoid insertion site (44%), four at ligament midsubstance (25%), and five at the humeral insertion site (31%). Eight anterior pouch specimens failed at the glenoid (50%); six at ligament midsubstance (38%) and two at the humeral insertion (12%). Finally, four posterior pouch specimens failed at the glenoid (25%), seven at midsubstance (44%), and five at the humeral insertion (31%).

Thus this tangent modulus is a linear function of the stress u, indicating a stiffening phenomenon with increasing stress. This is the generally accepted model for fibrous soft biological tissues (34,38). The parameter B represents the rate of change of the tangent modulus with respect to the stress u and may be thought of as the rate of fiber recruitment parameter for stretching of collagenous fibrous tissues (31,3841). The product AB ( = C) represents the tangent modulus of the tissue as the stress approaches zero (i.e., u -+ 0). Using a least-square nonlinear regression curve-fitting procedure we have obtained the material constants A and B (Table 2). Our curve-fitting results from the exponential stress-strain law (A, B, and C parameters) and Young's modulus E are displayed in Table 2 in the form of means + standard deviations. The material constant A is significantly greater for the posterior pouch than for the superior band (p < 0.025) or anterior pouch (p < 0.01). The material constant B is significantly greater for the superior band than for the anterior pouch (p < 0.005), with the posterior pouch falling in between. The material constant C (i.e., the tangent modulus in the limit u + 0 in equation 2) also varies with individual inferior glenohumeral ligament regions, with that of the posterior pouch being greater than that of either the anterior pouch (p < 0.005) or the superior band (p < 0.025). Furthermore, regional variations in Young's modulus E paralleled the variations in the tangent modulus C, namely, E and C were greatest for the posterior band and least for the anterior band.

Tensile Modulus and the Stress-Strain Relationship

Two independent curve-fitting methods were used to assess regional inferior glenohumeral ligament stiffness using total specimen strain data obtained from each bone-ligament-bone specimen. The first method used linear regression to determine the tensile (Young's) modulus from the nearlinear portion of the stress-strain curve. Although these tensile moduli did not differ from region to region in a significant manner, the modulus for the anterior pouch tended to be lower than for the posterior pouch or for the superior band (Table 2). The second method for assessing inferior glenohumeral ligament stiffness involved modeling the entire nonlinear stress-strain curve from toe region to failure. This may be done using an exponential stress-strain law proposed previously and applied to a variety of soft tissues, including skin, cartilage, and ligaments (10,30,31,41). This simple exponential stress-strain law is given by u

=

A(eB' - 1)

TABLE 2. Ligament region

A (MPa)

Superior (n = 16) Anterior (n = 16) Posterior (n = 16) Average (n = 48) Data are means

0.43 0.55 0.92 0.64 f

? f f ?

standard deviations.

J Orthop Res, Vol. 10, No. 2 , 1992

0.24 0.52 0.60 0.52

B

C (MPa)

14.15 f 4.17 8.84 5 3.03 11.49 f 4.99 11.44 2 4.60

5.50 & 2.73 4.38 f 3.32 8.42 5 3.88 6.11 f 3.72

E (MPa) 38.74 k 30.33 k 41.91 2 36.92

*

18.09 10.58 12.50 14.55

TENSILE PROPERTIES OF THE INFERIOR GLENOHUMERAL LIGAMENT

DISCUSSION This study is the first to examine the mechanical properties of the inferior glenohumeral ligament in three separate anatomic regions (27,28). The role of the inferior glenohumeral ligament as a significant restraint against anterior glenohumeral instability has been discussed in previous clinical and biomechanical studies (3,4,8,12,17,18,25,36,37).Townley demonstrated the importance of the capsular mechanism by showing that it functioned as an effective barrier against anterior instability, even after the labrum had been excised (36). Turkel et al. divided the inferior glenohumeral ligament into three anatomical regions, and several studies have demonstrated that tension varies in the individual regions according to arm position (12,25,37). As abduction in the scapular plane increases, tension is shifted from the more superior structures to the more inferior ones. In this study, both structural and mechanical differences were noted among the three anatomical regions of the infcrior glenohumeral ligament, reflecting perhaps the variability in composition and specificity of function among these three regions. Structurally, the superior band of the inferior glenohumeral ligament was consistently the thickest region of the ligament. It was easily distinguishable, both visually and palpably, from the middle glenohumeral ligament above and the axillary pouch below. When the arm is in the clinical position of instability (i.e., abducted and externally rotated), the superior band was found to span the joint anteriorly, which may be why the capsule is the most massive in this region. The decrease in the thickness of the inferior glenohumeral ligament from an-

0

h

n

5

tero-superiorly to postero-inferiorly occurred gradually, rather than in an abrupt, steplike fashion. Although manipulating the capsule to simulate various clinical positions can introduce folds in the axillary pouch, our quantitative measurements did not demonstrate a distinct posterior thickening in the inferior glenohumeral ligament as shown by others (24). The use of two simultaneous methods of strain measurement, the video dimensional analyzer technique and the total specimen strain technique, allowed comparison of the ligament midsubstance data to the bone-ligament-bone specimen data, including the determination of failure modes. The inferior glenohumeral ligament exists in vivo as a tissue with complicated bony insertions (18,19), one of which, the attachment to the glenoid, is often implicated as the region of clinical failure (3,4,17,32). Thus, the response of the entire bone-ligamentbone specimen, as well as that of the ligament midsubstance, is clinically significant. Two important clinical concepts exist regarding the relationship between total specimen strain and midsubstance strain when the bone-ligament-bone specimen is tested to failure. First, the ability of the inferior glenohumeral ligament to stretch considerably before ligament or insertion failure suggests that lateral translation of the humeral head is possible under loads that could allow the head to override the glenoid rim. These data support the clinical findings that certain patients can sublux without disruption of either the capsule or its insertion sites. Second, the occurrence of considerable capsular stretching, be it the result of a single traumatic event or of a repetitive nature, also presents the possibility that the ligament may be permanently stretched without insertion site failure. Measuring

FIG. 5. Typical stress-strain curves are displayed for the inferior glenohurneral ligament. A: The abrupt failure pattern usually s e e n when failure occurred at the glenoid insertion. B: The steplike failure pattern often seen with ligament substance failure.

I In

Strain

193

Strain

. I Orthop Res, Vol. 10, NO,2, 1992

194

L. U . BIGLIANI ET AL.

both midsubstance strain and simultaneously recorded total specimen strain provides a method to analyze a mechanism of injuries for clinical shoulder pathology in future in vivo total joint studies. The elongation rates used in this study (0.04 and 0.004 mm/s) are among the slowest that have been reported for ligament testing (22,26). These rates allow for the exclusion of viscoelastic effects seen at higher strain rates (39). Although variation in elongation rates did not produce significant differences, it is possible that our chosen rates were too close to produce such differences. The stress-strain curves obtained before failure (Fig. 5 ) for the inferior glenohumeral ligament were sigmoidally shaped and similar to those reported for other soft tissues (9,22,40). However, strain to failure differed significantly among the three anatomical inferior glenohumeral ligament regions, indicating mechanical heterogeneity in the ligament. These differences suggest variability in the composition, and perhaps micro-organization, and function of the three inferior glenohumeral ligament regions. However, stress at failure was found to be nearly identical for the three regions of the inferior glenohumera1 ligament. Whereas greater tensile forces were needed to rupture the more massive superior band specimens, when cross-sectional area (particularly thickness) differences were accounted for, the three regions behaved similarly with respect to failure stress. This would suggest a fairly homogeneous nature for the collagen fibers of the three regions of the inferior glenohumeral ligament. Taken together, these results appear conflicting: the strain data highlight mechanical and structural differences among the anatomical regions of the ligament, whereas the failure stress data imply uniformity of structure. Our nonlinear tensile stress-strain data for the inferior glenohumeral ligament suggest a hypothesis to explain this apparent paradox. If the number and density of the collagen fibers in the three inferior glenohumeral ligament regions are fairly consistent, then these regions will possess similar ultimate stress values. If the microorganization of the collagen fiber in these three regions are different, then the stress-strain curve will be different despite similar material composition. Such differences in micro-organization might well explain the differing ultimate strain values and differing shapes of the stress-strain curves from region to region. Specifically, the anterior pouch may be composed of more highly crimped collagen fibers and, therefore, this region requires greater strain before reaching the

J Orthop Res, Vol. 10, No. 2 , 1992

same level of stress. This is observed in the longer toe region obtained for typical anterior pouch stress-strain curves. This hypothesis is demonstrated in the variation of the parameters B and C presented in Table 2 and typical stress-strain curves displayed in Fig. 6 . Because parameter B measures the rate of change of the tangent modulus with respect to stress and parameter C represents the initial slope of the stress-strain curve, the hypothesis of a more highly crimped collagen pattern in the anterior pouch region appears likely. However, additional detailed histological studies of the inferior glenohumeral ligament are needed to verify this hypothesis. Whereas statistically significant differences in various structural and mechanical parameters were found among the three regions of the ligament, it is also important to point out that some variability existed in our data. For example, whereas the superior band was usually the thickest region, in three shoulders the anterior pouch was slightly thicker. Likewise, whereas the anterior pouch generally showed the highest strain to failure, in one shoulder its ultimate strain was the lowest of the three regions. Thus, in comparisons among the three regions of the same shoulder, there were exceptions to the general statements, although these were infrequent. However, when comparisons between specimens from different shoulders were made, far more variability was found. For example, the posterior band of the most massive specimen (2.4 mm) was almost 2.5 times thicker than the corresponding band of the least massive (1.0 mm). Similarly,

0.1

0.0

0.2

0.3

Strain & FIG. 6. Stress-strain data were modeled using an exponential stress and strain law. Results indicate that the anterior pouch region is significantly less stiff than the other two regions of the inferior glenohumeral ligament. The equations derived for the stress-strain relationship of each region were the following: 0.43(exp(14.15~)- 1) for the superior band, 0.55(exp(8.84~) 1 ) f o r the a n t e r i o r p o u c h and 0.92(exp(l1.49~)- 1) for the posterior pouch (see Table 2). ~

TENSILE PROPER71ES OF THE INFERIOR GLENOHUMERAL LIGAMENT

whereas the superior band usually failed at 24% strain, one specimen failed at 36% strain, a value even higher than the average for the anterior pouch. Clinically, such variability is not at all unexpected. Structural differences, specifically in thickness, are commonly seen among shoulders of patients of different sizes and weights. Moreover, the structural and mechanical variability in glenohumeral ligaments probably contributes significantly to our perception of clinically loose or tight shoulders, with the most dramatic variations being the frozen shoulder and the multidirectionally unstable shoulder. Our values for ultimate stress are lower than those reported for other ligaments previously studied, such as the anterior cruciate ligament and the patellar tendon (7,15). Butler et al. reported maximum stresses of 37.8 and 58.3 MPa for the anterior cruciate ligament (ACL) and central one third of the patellar tendon, respectively (7). Each of the three regions of the inferior glenohumeral ligament was found to have an ultimate stress of only 5.5 MPa. Similarly, our energy absorption values for this ligament averaged only 0.9 Nm, whereas Woo et al. recently reported 11.6 Nm for a subgroup of human ACL specimens (43). Thus, our values for ultimate stress and particularly the energy absorption are only 1&15% of those reported for the ACL. However, we note that our data are for discrete regions of a single ligament (i.e., only a small portion of the entire capsule), whereas the ACL data are for an entire ligament. These low values suggest that each of these ligament regions-indeed, the entire inferior glenohumeral ligament alone-is probably not strong enough to stabilize the glenohumeral joint independently. However, the cumulative energy absorbed by the entire capsule is certainly higher. It is likely then, that the inferior glenohumeral ligament serves as one important component of a capsulo-ligamentous mechanism that works in concert with the dynamic muscle stabilizers to provide stability for the glenohumeral joint (8,ll). In addition, specimen age probably contributed to these low ultimate stress and energy absorption values. Noyes and Grood reported that boneligament complexes become weaker with increasing age, demonstrating threefold reductions in maximum stress and energy absorption in older cadavers (23). Woo et al. also reported that the ultimate load for ACLs from younger donors (60 yr) (43). Similarly, the energy absorbed decreased from 11.6 to 1.8 Nm with increasing age in their study. The

195

average age of our specimen donors was 72 years, and the youngest was 56 years. If we were to use the age-related data established for the ACL to normalize our data for the inferior glenohumeral ligament, in young adults we might expect ultimate stress and energy absorption values of 18 MPa and 5.8 Nm, respectively. Age may also have had an effect on the relative frequency of failure modes in this study. Other investigators, testing ACLs, have found that younger specimens fail primarily in midsubstance, whereas older adult specimens fail by avulsion (23). However, Reeves, testing the entire shoulder capsule, found that specimens from younger donors usually fail at the glenoid attachment site, and those from older donors in the capsular substance (29). Three different failure sites for the inferior glenohumeral ligament were seen in the present study. Although failure at the humeral insertion site is uncommon clinically, it has been reported (2), and we have also seen capsular failure at this site in several patients. Both glenoid insertion site failure and midsubstance failure are more common and correlate with many pre-existing clinical reports (3,4,17,32,33,35,36). Failure at the glenoid insertion site (Fig. 5A) was typically more abrupt than failure at the ligament midsubstance (Fig. SB). The abrupt peeling away of the glenoid insertion suggested a clinical parallel with the Bankart lesion, in which failure is seen at the labral attachment of the inferior glenohumeral ligament. This mode of failure has been widely described in clinical series and is regarded by some as the principal etiology of recurrent glenohumeral instability (3,4,32,33). Our results suggest that glenoid insertion site failure may not be the only factor in producing traumatic subluxation or dislocation. In this study, capsular rupture was seen almost as frequently as glenoid insertion failure as the mode of inferior glenohumeral ligament failure. This may have occurred because our biomechanical model exerts a pure tensile force on the ligament. However, during the clinical event of dislocation or subluxation, the impact of the dislocating humeral head and/or the impact of the greater tuberosity on the glenoid rim may also be factors leading to capsular detachment from the bony insertion. More importantly, perhaps, is the fact that even in cases where the inferior glenohumeral ligament failed at its glenoid insertion, such failure occurred only after significant elongation of the inferior glenohumeral ligament specimen. This raises the possibility that permanent

J Orthop Res, Vol. 10. No. 2 , 1992

196

L . U . BIGLIANI ET AL.

plastic deformation may occur in the ligament midsubstance, even if avulsion does not occur. This capsular stretching may be responsible for some of the more subtle glenohumeral joint instability observed clinically. The results obtained in this study emphasize the clinical importance of understanding the degree to which capsular stretching can take place, in addition to the effects of a deficient labrum or inferior glenohumeral ligament insertion failure, in the production of traumatic glenohumeral instability. It would also suggest that the repair of recurrent instability should consider the possibility of capsular laxity produced by the initial traumatic event, as well as the detachment of the glenoid insertion of the inferior glenohumeral ligament. Such a repair would restore both the functional length of the ligament and the integrity of its glenoid insertion, minimizing the chances of recurrence and providing improved functional stability for the shoulder. Acknowledgment: This research was sponsored in part by a Bristol-Myers SquibbiZimmer Grant for Excellence in Orthopaedic Research to Columbia University, and an

12. 13. 14. 15.

16. 17. 18. 19.

20. 21.

OREF Career Development Award for E.L.F. 22.

REFERENCES 1. Akizuki S, Mow VC, Muller F, Pita JC, Howell DS, Mani-

2. 3.

4.

5.

6. 7.

8. 9.

10.

11.

court DH: Tensile properties of knee joint cartilage: I. lafluence of ionic condition, weightbearing, and fibrillation on the tensile modulus. J Orthop Res 4:379-392, 1986 Bach BR, Warren RF, Fronek J: Disruption of the lateral capsule of the shoulder. A cause of recurrent dislocation. J Bone Joint Surg [Br] 70:274-276, 1988 Bankart ASB: The pathology and treatment of recurrent dislocation of the shoulder joint. Br J Surg 26:23-29. 1938 Bankart ASB: Recurrent or habitual dislocation of the shoulderjoint. Br Med J 2:1132-1133, 1923 Bigliani LU, Kurzweil PR, Schwartzbach CC, Flatow EL, Wolfe I: Inferior capsular shift procedure for anteriorinferior shoulder instability in athletes. Orthop Trans 13560, 1989 Bost FC, Inman VC: The pathological changes in recurrent dislocation of the shoulder: a report of Bankart’s operative procedure. J Bone Joint Surg 23595-613, 1942 Butler DL, Grood ES, Noyes FR: Effects of structure and strain measurement technique on the material properties of young human tendons and fascia. J Biomech 17:579-596, 1984 Cain PR, Mutschler T, Fu FH, Lee SK: Anterior stability of the glenohumeral joint. A dynamic model. Am J Sports Med 15:144-148, 1987 Frank CB, Woo SL-Y, Andriacchi TP, et al.: Structure, function and composition. In: Znjury and repair of the musculoskeletal soft tissues, ed by SL-Y Woo, JS Buckwalter, Park Ridge, IL: AAOS, 1987, pp 113-166 Fung YCB: Elasticity of soft tissues in simple elongation. Am J Physiol213:1532-1544, 1967 Glousman R, Jobe FW, Tibone JE, Moynes D, Antonelli D, Perry J: Dynamic electromyographic analysis of the throw-

J Orthop Res. Vol. 10, N o . 2 , 1992

23.

24. 25. 26.

27.

28. 29. 30. 31. 32.

ing shoulder with glenohumeral instability. J Bone Joinr Surg [Am] 70:220-226, 1988 Hammon OJ, France EP, Terry GC: Stabilizing function of passive shoulder restraints. Trans Orthop Res Soc 12:77, 1987 Jobe FW, Giangama CE, Glousman RE, Kvitne RS: Anterior capsulo-labral reconstruction in throwing athletes. Orthop Trans 13:23&231, 1989 Kaltsas DS: Comparative study of the properties of the shoulder joint capsule with those of other joint capsules. Clin Urthop Re1 Res 173:2&26, 1983 Kennedy JC, Hawkins RJ,Willis RB, Danylchuk KD: Tension studies of human knee ligaments: yield point, ultimate failure and disruption of the cruciate and tibia1 collateral ligaments. J Bone Joinf Surg [Am] 58:350-355, 1976 Magnuson PB, Stack JK: Recurrent dislocation of the shoulder. JAMA 123:889-892, 1943 McLaughlin HL: Recurrent anterior dislocation of the shoulder. 1. Morbid anatomy. Am J Surg 99:628-632, 1960 Moseley HE, Overgaard B: The anterior capsular mechanism in recurrent anterior dislocation of the shoulder. JBone Joint Surg [Br] 44:913-927, 1962 Mow VC, Bigliani LU, Flatow EL, et al.: The role of joint instability in joint inflammation and cartilage deterioration: a study of the glenohumeral joint. In: Sports-induced injammation, ed by WB Leadbetter et al., Park Ridge, IL: AAOS, 1990, pp 337-355 Neer CS 11, Fithian TE, Hansen PE, Ogawa K, Brems JJ: Reinforced cruciate repair for anterior dislocation of the shoulder. Orthop Trans 9 : 4 4 4 5 , 1985 Neer CS 11, Foster CR: Inferior capsular shift for involuntary inferior and multidirectional instability of the shoulder. J Bone Joint Surg [Am] 62:897-908, 1980 Noyes FR, DeLucas JL, Torvik PJ: Biomechanics of anterior cruciate ligament failure: an analysis of strain-rate sensitivity and mechanisms of failure in primates. J Bone Joint Surg [Am] 56:236-253, 1974 Noyes FR, Grood ES: The strength of the anterior cruciate ligament in humans and rhesus monkeys: age-relatcd and species-related changes. J Bone Joint Surg [Am] 58: 10741082, 1976 O’Brien SJ, Neves M, Rozbruch R, et al.: The anatomy and histology of the inferior glenohurneral complex. Am J Sports Med 18:449456, 1990 Ovesen J , Nielsen S: Stability of the shoulder joint: cadaver study of stabilizing structures. Acfa Orfhop Scand 56:149151, 1985 Petersen RH. Gomez MA, Woo SL-Y: The effect of strain rate on the biomechanical properties of the medial collateral ligament: a study of immature and mature rabbits. Trans Orthop Res Soc 12:127, 1987 Pollock RG, Bigliani LU, Flatow EL, Soslowsky LJ, Mow VC: The mechanical properties of the inferior glenohumeral ligament. The American shoulder and elbow surgeons, 6th open meeting, New Orleans, Paper No. 32, 1990 Pollock RG, Soslowsky LJ, Bigliani LU, Flatow EL, Pawluk RJ, Mow VC: The mechanical properties of the inferior glenohumeral ligament. Trans Orthop Res Soc 15510, 1990 Reeves B: Experiments on the tensile strength of the anterior capsular structures of the shoulder in man. J Bone Joint Surg [Br] 50:858-865, 1968 Ridge MD, Wright V: The description of skin stiffness. Biorheology 2:67-74, 1964 Roth V, Mow VC: The intrinsic tensile behavior of the matrix of bovine articular cartilage and its variation with age. J Bone Joint Surg [Am] 62:1102-1117, 1980 Rowe CR: The surgical management of recurrent anterior dislocations of the shoulder using a modified Bankart procedure. Surg Clin North Am 43: 1663-1666, 1963

TENSILE PROPERTIES OF THE INFERIOR GLENOHUMERAL LIGAMENT 33. Rowe CR, Patel D, Southmayd WW: The Bankart procedure: a long-term end-result study. J Bone Joint Surg [Am] 60: 1-16, 1978 34. Stouffer DC, Butler DL, Hosny D: The relationship between crimp pattern and mechanical response of human patellar tendon-bone units. J Biomech Eng 107:15&165, 1985 35. Thomas SC, Matsen FA 111: An approach to the repair of avulsion of the glenohumeral ligaments in the management of traumatic anterior glenohumeral instability. J Bone Joint Surg [Am] 71:506513, 1989 36. Townley CO: The capsular mechanism in recurrent dislocation of the shoulder. J Bone Joinf Surg [Am] 32:370-380, 1950 37. Turkel SJ, Panio MW, Marshall JL, Girgis FG: Stabilizing mechanisms preventing anterior dislocation of the glenohumeral joint. J Bone Joinf Surg [Am] 63:1208-1217, 1981 38. Viidik A: A rheological model for uncalcified parallel-fibred collagenous tissue. J Biomech 1:3-11, 1968

197

39. Viidik A: Structure and function of normal and healing tendons and ligaments. In: Biomechanics of diarrhrodial joints, ed by VC Mow, A Ratcliffe, SL-Y Woo, New York, NY, Springer-Verlag, 1990, pp 1-38 40. Woo SL-Y: Mechanical properties of tendons and ligaments. Biorheology 193385-396, 1982 41. Woo SL-Y, Akeson WH, Jemmott GF: Measurements of nonhomogeneous, directional mechanical properties of articular cartilage in tension. J Biomech 9:785-791, 1976 42. Woo SL-Y, Gomez MA, Seguchi Y, Endo CM, Akeson WH: Measurement of mechanical properties of ligament substance from a bone-ligament-bone preparation. J Orfhop R e s 1:22-29, 1983 43. Woo SL-Y, Hollis JM, Adams DJ, Lyon RM, Takai S : Tensile properties of the human femur-anterior cruciate ligament-tibia complex. Am J Sports Med 19:217-224, 1991

J Orthup Rcs, Vol. 10, No. 2, 1992