Comparative Medicine Copyright 2009 by the American Association for Laboratory Animal Science

Vol 59, No 5 October 2009 Pages 465475

The Anatomy of the Glenoid Labrum: A Comparison between Human andDog

Martin Sager,1,* Monika Herten,2 Stefanie Ruchay,2 Josef Assheuer,3 MartinKramer,4 andMarcusJger2 The anatomy of the glenohumeral joint in humans is characterized by static and dynamic stabilizing structures. In particular the glenoid labrum (GL), the proximal attachment of the joint capsule and the lateral glenohumeral ligament, is an important passive stabilizer in the human shoulder. Although canine animal models are used frequently to investigate the complex biomechanics of the shoulder, few data regarding the microstructure of the canine GL are available. In this study, the anatomy of the canine GL and related structures (n = 20) was investigated and compared with the human anatomic situation (n = 36). In both human and beagle joints, the GL consisted of 3 zonesthe transition zone, shifting zone, and meniscoid fold, but not all 3 zones were present in all joint segments from canine joints. In particular the peripheral parts of the GL showed rich vascularization in both species. The height and width of the GL in the histologic specimens indicated that the GL is of less importance as a passive stabilizer in dogs. Additional differences between the human and canine CL include the joint ligaments, tendons of the shoulder joint, and lack of rotator cuff. The structural and biomechanical characteristics of the joints of quadrupedal animals raise the question of their appropriateness for shoulder research.

In the fields of orthopedic and trauma research, animals are used frequently to model human pathologic conditions. The most important criteria for the choice of the appropriate animal model are anatomic and physiologic comparability. These fundamentals contribute to the validity of a model, whereas missing or incomplete knowledge may lead to results that cannot be extrapolated to the human condition. The comparability of the biomechanical joint function of quadrupedal laboratory animals with that of upright-ambulating humans is questionable. For instance, in contrast to that in humans, the shoulder joint in quadrupedal animals has a crucial weight-bearing function .32,35 However, the anatomical structure and composition of analogous joints from humans and quadrupeds have a great deal in common. In particular, similarities in the alterations and lesions of the shoulder joint give rise to the possibility of strong correlation between the anatomical design of the joint and the patterns of disease in these species. In a 1980 article,10 the GL is described as the proximal attachment of the joint capsule and the lateral glenohumeral ligament. In humans, the humeral head is nearly 4 times larger than the glenoid cavity, and the GL enlarges the joint surface.30,38 In comparison, the canine shoulder joint (scapulohumeral joint, articulatio humeri) is a ball-and-socket joint. The canine humeral head is more curved in the sagittal than in the dorsal plane and articulates with the obviously smaller glenoid cavity of the scapula.32 One study4 reported a surface relation of the glenoid cavity to humeral head of about 1:3 in dogs, whereas another19 described
Received: 27 Sep 2008. Revision requested: 25 Nov 2008. Accepted: 25 Jan 2009. 1 Central Animal Research Facility and 2Orthopaedic Department, Medical School of the Heinrich-Heine-University, Dsseldorf. Germany; 3Institute for Magnetic Resonance Imaging, Kln, Germany; 4Small Animal Clinic, Veterinary School of the Justus-LiebigUniversity, Giessen, Germany. * Corresponding author. Email: Martin.Sager@med.uni-duesseldorf.de

a contact area between scapula and humerus of 47% in the flexed joint and of 62% in normal weight-bearing condition. The human shoulder joint is highly mobile, typically allowing for abduction of 90, adduction of 20, anteversion of 90, and retroversion of 30. Normal values of internal and external rotation of the humerus are each 70. However, the large range of movement of the human shoulder is increased further by contribution from the scapulothoracal, sternoclavicular, and the acromioclavicular joints, leading to abduction of 180 and anteflexion of approximately 170.37 In contrast, the range of motion of the shoulder joint in quadrupeds is limited because of the adjacent muscles and tendons, which lead the joint to function as a roller joint.22 Extension and flexion in carnivores typically is greater than 120, with external rotation of up to 45, but internal rotation is usually less than 35.41 In a recent study8 angles for abduction in dogs without history of shoulder instability were between 31 and 33, according to goniometric measurements and image analysis. The human shoulder has a wide joint capsule with reserve folds, which elapse during certain movements. The biceps tendon is covered by a tendon sheath and situated in the superior portion of the joint cavity.30 In all companion animals, the joint capsule of the shoulder is spacious. In carnivores in particular, the joint space bulges into 2 cranial and an expanded caudolateral portion.22 The proximal attachment of the capsule is the GL, the distal insertion is located only a few millimeters distal to the joint surface of the humeral head, extending into the periosteum of the humeral neck.11 In carnivores, the proximal part of the tendon of the biceps muscle is surrounded by the joint capsule to the intertubercular groove. In humans, ligaments strengthen the anterior wall of the shoulder joint capsule in the superior, medial, and inferior directions.13,17,18,38 In dogs these structures run mainly medially and

465

Vol 59, No 5 Comparative Medicine October 2009

Figure 1. Macroscopic view of the glenoid cavity, human (left) and dog (right). The small arrows mark the macroscopic visible part of the glenoid labrum. BT, biceps tendon; IGHL, inferior glenohumeral ligament; JC, joint capsule (JC); LGHL, lateral glenohumeral ligament; MF, meniscoid fold; and MGHL, medial glenohumeral ligament.

laterally. The ligamentum coracohumerale, which extends from the coracoid process to the greater tubercle, forms a solid reinforcement strap in the human shoulder joint. It expands from the coracoid process to the greater tubercle. Likewise, the coracoacromial ligament runs also from the coracoid process extending to the acromion covering the shoulder joint like a roof. The glenohumeral ligaments are differently developed in humans and dogs: there are 3 ligaments in humans6,13,17,18,38 but only 2 in dogs.10,29,32,42 The GL is noted in the Nomina Anatomica Veterinaria29 but is mentioned only rarely in the broader veterinary literature. According to Kujat21 the canine glenoid is similar to the human. He describes the labrum as a triangular structure in cross section, which is attached to the rim of the glenoid without being firmly connected to it. In addition, the GL of the dog has a close connection to the tendon of the subscapularis muscle medially, to the tendons of the supraspinatus, infraspinatus, and teres minor muscles laterally, as well as to the tendon of the triceps brachii muscle caudally.21 Lesions of the human GL are well described, and their clinical importance is well known.34 In contrast, few data regarding the canine GL are available and mainly involve contrast arthrograms,24,36 arthroscopic findings,39 reports of the anatomic structure of the GL,3,23 and detailed descriptions of labrum injuries by magnetic resonance imaging.27,28

Several detailed investigations describe the vascularization of the GL in humans.1,9 In contrast, the single microangiographic study in dogs21 indicated a rich vascularization of the canine GL, with exception of its free edge. The extent of vascularization of the GL suggests the prognosis for healing after labrum injuries. The aim of the present study was a comparative macroscopic and microscopic description (including vascularization pattern) of the GL in humans and dogs. A limitation of our study is that we studied beagles, a medium-size breed, and therefore the results cannot be extrapolated to all canine breeds without further review. However, our results yield insight into the options for and limits to extrapolating data from laboratory animal studies on the GL to the situation in humans.

Study design. Thirty-six shoulder joints of 9 female and 9 male human cadavers (age: mean, 73.6 y; range, 51 to 86 y) and 20 joints from 10 male and 10 female beagle dogs [mean age: 6 mo; weight: mean SE, 9.26 1.67 kg; range, 7.0 to 13.4 kg) were investigated. The human joints were prepared in the Anatomy Department of the University of Wrzburg (Germany). The canine joints were obtained immediately after euthanasia from dogs bound for necropsy as controls in toxicologic experiments. The study was approved by the institutional ethics committee.

Materials and Methods

466

Comparison of human and canine shoulder joints

Figure 2. Histomorphometric analysis of the canine GL, lateral segment (segment VI). The line of reference is the calcification zone. The height was determined by the distance starting at a right angle to the reference line to the limit of the articular cartilage. The width was measured at the middle of this distance at a right angle. The distance from the point of the most proximal fibers insertion of the labrum measured perpendicular to the reference line was defined as the depth of the labrum. Bar, 100 m.

Figure 3. Nomenclature regarding the right shoulder joint.

Histologic preparation. The specimens were fixed in 4% neutral buffered formalin. After removal of soft tissue (cutis, subcutaneous fat, muscles, nerves, vessels), the joint capsule was incised circularly from the middle to proximal part. The anatomy of all shoulder joints was analyzed macroscopically by measurement of distinct parameters (longitudinal and transverse diameter) and documented by photography (Figure 1). For histologic preparation, the joints were decalcified in 10% buffered EDTA and embedded in paraffin or methacrylate (Technovit 7100, Heraeus Kulzer, Weinheim, Germany). Each joint socket was divided into 7 segments, with the proximal biceps tendon insertion serving as orientation at 12:00 and progressing clockwise for right shoulder joints but mirror-inverted for left joints. Segments were cut into 3-m circular slices by using a rotating microtome (Autocut 2055, ReichardtJung AG, Heidelberg, Germany). The number of slices per segment was 54 for human joints and 6 for canine. Histochemical staining. Slices were stained with either hematoxylin and eosin or azan according to Heidenhain15 or, for distinction between reticular collagenous soft tissue and muscle, according to Richardson.26 Both the Heidenhain and Richardson staining methods were suitable for polarization microscopy. Immunohistochemistry. Canine joints were investigated immunohistochemically for collagen types I, II, and III and for transglu-

taminase II as a marker of tissue vascularization.12 Tissue sections were deparaffinized in xylol, dehydrated through an ethanol series, and rehydrated in PBS. Antigen unmasking was performed by incubating the slides for 15 min in trypsin (0.05% in PBS; PAA Laboratories, Pasching, Austria) at 37 C. After slides were rinsed with PBS, the activity of endogenous peroxidase was quenched with 0.9% hydrogen peroxide in PBS for 10 min at room temperature, specimens were washed, and nonspecific binding sites were blocked by incubation in blocking solution (DakoCytomation, Hamburg, Germany) for 30 min. Primary antibodies included mouse monoclonal antibodies to collagen I and II (1:100 dilution, Chemicon, Hofheim, Germany) and transglutaminase II (1:40 dilution, Acris Antibodies GmbH, Hiddenhausen, Germany), rabbit polyclonal antibody to collagen III (1:100 dilution, Chemicon), and mouse IgG1 and rabbit IgG (as negative controls) and were applied to tissue sections in a humidified chamber and incubated overnight at 8 C. Slides were washed in PBS and incubated with secondary biotinylated antimouse antibody (1:50 dilution, Dako) for 60 min at room temperature. Slides again were washed in PBS, and antibodyantigen complexes were visualized by using a streptavidinperoxidase solution (1:250 dilution, Dako) and 3-amino-9-ethylcarbazole as the chromogen (Dako). Slides were counterstained with hemalaun (Merck, Darmstadt, Germany). Histomorphometric analysis. Histomorphometrical analyses and microscopic observations were performed by a single experienced investigator blinded to the specific experimental conditions. For image acquisition, a color camera (Color View III, Olympus, Hamburg, Germany) was mounted on a binocular light microscope (Olympus BX50, Olympus). Digital images (original magnification, 200) were evaluated by using a software program (analySIS FIVE docu, Soft Imaging System, Mnster, Germany). For measuring the size of the labrum glenoidale, the calcified area (Figure 2) was used as a reference. The height of this area was determined as the distance at a right angle between the reference line to the peripheral limit of the articular cartilage; width was obtained by taking the distance between the first centrally visible demasked fibers to the outermost peripheral point of the labrum; and depth was the distance from the most proximal fibers of the insertion of the labrum in a right angle to the reference line.

In the following description of the labrum glenoidale, differences in nomenclature between humans and dogs must be considered. In the present study, the origin of the biceps tendon was taken as the reference point for the characterization of segments. The terms used with the incisura in medial or anterior position are defined in Figure 3. Macroscopic description of the human shoulder joint. In 35 investigated human specimens, the labrum did not encircle the entire glenoid, and marked differences in size, shape, structure and attachment were present (Figure 1); 1 specimen could be evaluated because of beginning of autolytic changes. In all specimens, the long biceps tendon was present in the superior segment, but in 34 of 35 joints, it could not be distinguished from the labrum and formed a labrumtendon complex. This complex seemed to be relocatable and fixed in the proximal part at the glenoid, creating a recess (a bulge in the joint capsule) toward the glenoid. The fibers of the biceps tendon were connected to the labrum in different ways; in most (83%) cases, the tendon was anchored by 2 reins, which led to the posterior and anterior labrum. In the

Results

467

Vol 59, No 5 Comparative Medicine October 2009

Figure 4. Summary of findings from the current study and literature regarding differences between the glenohumeral ligaments of humans6,13,17,18,38 and dogs.10,29

remaining 17% of specimens, the tendon seemed to be anchored only posteriosuperiorly, without the additional anterior rein. An anteriosuperior sublabral recess was present in 50% of specimens, due to partial or complete ablation of the labrum from the glenoid. The structure of the human labrum was weaker posteriorly than inferiorly, where it was characterized by recesses between the glenoid and labrum mainly in the superior and anteriosuperior regions. Inferiorly and posteriorly, the labrum predominantly was fixed by the glenoid, but numerous bulges in the joint capsule were present between the glenohumeral ligaments (Figure 4). Macroscopic description of the canine shoulder joint. The biceps tendon served as a central orientation point during preparation and macroscopic description and was used for the arrangement of the glenoid cavity in cranial direction (Figure 1). The origin of the tendon from the supraglenoid tubercle could not be detected macroscopically. Although the biceps tendon was adjacent to capsular tissue and surrounded by lateral rims at the side opposite the glenoid, it was uncovered near the articular cavity. Only the proximal rim was covered by a sliding zone and separated from the cranial pole of the glenoid cavity (Figure 1).

At the edge of the glenoid cavity of the scapula, the labrum (if present) diverged dramatically in size, shape, and attachment among segments. In some segments, the joint lips could not be separated macroscopically because there was no distinct delimitation between the adjacent joint capsule and the glenohumeral tendons. The first segment contained a gap between the biceps tendon and the cranial pole of the glenoid cavity; this gap was filled with a gelatinous-like tissue. This sliding zone was prominent in 11 of the 20 investigated joints, protruding above the hyaline cartilage and flattening toward the rim of the biceps tendon. In 8 of 20 joints, the sliding zone was narrow, presenting a cavity between the cranial glenoid rim and the biceps tendon. In the remaining joint, this cavity was enlarged but lacked any soft transitional tissue. The craniomedial part of the joint (segment II) seemed to be rather flexible and not firmly connected to the glenoid. The medial glenohumeral tendon predominated, and a labrum structure could not be defined. Some cavities (recessus subscapularis) could be detected.

468

Comparison of human and canine shoulder joints

The third segment was unique in that the border of the cartilage surface facing the labrum did not seem to be circumscribed precisely but rather had retractions resembling incisures. The labrum extended into the joint space like a meniscoid fold and could be detached from the glenoid by using forceps (Figure 1). The extent of the overlapping zones varied among the investigated joints and seemed to be frayed in some areas of some joints. The 2 caudal segments (segments IV and V) were very homogenous in appearance. There was a smooth transition between the hyaline edge of the cavity and joint capsule, and the connection seemed to be fixed. The capsule wall was very thin, especially caudolaterally, and translucent at the adjacent muscles. In the sixth segment, the labrum was easily discernable and appeared to have a tough, firm structure that was attached to the hyaline joint area. The labrum arose distinctly from the cartilage surface and deepened the cavity, which was only minimally developed in this area (Figure 1). In the seventh segment, the lateral glenohumeral ligament was inserted craniolaterally, reaching from caudal and inclining toward the labrum. Both structures united in a V cranially in a rough strand that precluded macroscopic distinction between labrum and ligament (Figure 1). The joint socket was enlarged and deepened. Size and form of the glenoid cavity. The longitudinal diameter of the glenoid cavity (mean 1 SD) was 2.0 0.18 cm in dogs and 3.2 0.41 cm in humans, and the transverse diameter at the narrowest point (at the incisura glenoidalis) was 0.6 0.1 cm in dogs and 2.4 0.36 cm in humans. The area of the articulating joint surface in dog was determined by using the means of the transverse diameters at the maximal and narrowest points and assuming an ellipsoidal joint surface. For human joints, the areas were 1.6 cm2 for the glenoid cavity and 4.0 cm2 for the caput humeri, resulting in a ratio of 1:2.5. When defined according to Anetzberger,2 the canine joint cavity (Figure 2) was droplike (type Ia) in shape in 12 dogs (60%), lacked incision (type 1b) in 2 (10%), and was oval (type II) in 6 (30%). Microscopic investigation of human and canine joints. The labrum glenoidale was divided among as many as 3 segments (Figure 5). In the first segment, the transition zone, the labrum was connected to the hyaline joint cartilage (Figures 3 and 5). This zone consisted of filamentous cartilage, in which the collagen fibers were readily apparent due to their low amount of matrix (in contrast to hyaline cartilage). In cross section, the fibers displayed a lattice structure. In human joints, this zone was present predominately in the anterior and anteriosuperior segments, whereas in dogs it spanned the cranial, craniomedial, and craniolateral segments. The second segment consisted of collagen fibers encircling the glenoid (Figure 5). This zone often was connected with and difficult to differentiate from the first segment. In addition, the GL in the second segment often was associated at the bone with Sharpey fibers, which interlaced in an obtuse angle with subchondral bone. The third segment was created by a meniscoid fold (Figure 5), consisting predominantly of well-vascularized synovial tissue (Figure 6), which was connected with collagen fibers to the residual part of the labrum. Comparison of joint segments between humans and dogs. Segment I. In contrast to the other segments, the superior and anteriosuperior segments of the human joint were highly variable

Figure 5. Differentiation of the human GL into 3 characteristic zones: human anterosuperior segment (segment II). Left, hyaline cartilage (CA); bottom, subchondral bone (SCB); middle, transition zone (TZ); top, fibrous zone with circular fibers (CF). Right, meniscoid fold (MF). Type I variation: folding grid. All 3 zones in this segment occurred in 44.4% of specimens. Polarization microscopy; magnification, 12.5.

Figure 6. Medial segment (segment III) of the canine GL, visualization of vascularization. CA, articular cartilage; CF, fibrous zone with circular fibers; C-MGHL-C, complex of joint capsule and medial glenohumeral ligament; MF, meniscoid fold. Immunohistochemistry for transglutaminase II; bar, 200 m.

morphologically. In 75% of all investigated cases, the superior portion comprised a labrumbiceps complex whereas in the remaining cases both structures could be differentiated clearly (Figure 7). The anchorage zone could be detected in only 2 of 35 joints. In addition, a deep sublabral recess, a bundle of circular fibers with bony attachment through Sharpey fibers, and a pronounced meniscoid fold dominated. These findings varied dramatically from those in the first segment of the dogs, in which the anchorage zone was well demarcated and the attachment of the glenoid was nearby. The anchorage zone in canine joints was marked by a substance resembling cartilage fibers, in which the texture of the collagen fibers differed from the parallel pattern of the biceps tendon and displayed marked crossing of compo-

469

Vol 59, No 5 Comparative Medicine October 2009

Figure 7. Human labrumbiceps tendon complex, superior segment (segment I). Left, hyaline cartilage (CA); right, labrumbiceps tendon complex (LBTC) with a visible meniscoid fold (MF). Polarization microscopy; magnification, 12.5.

Figure 9. Transition zone of the canine GL and origin of the biceps tendon, cranial segment (segment I). Immunohistochemistry for collagen type II. Inset, vascularization in the shifting zone (SZ). Immunohistochemistry for transglutaminase II. BT, biceps tendon; CA, articular cartilage; SB, subchondral bone; TZ, transition zone. Bar, 200 m.

Figure 8. Overview of the cranial segment (segment I) of the canine GL. CA, articular cartilage; BT, biceps tendon; JC, joint capsule; SCB, subchondral bone; TZ, transition zone. Hematoxylin and eosin stain; bar, 2 mm.

Figure 10. Anterosuperior segment (segment II) of the human GL. Left, hyaline cartilage (CA). Right, circular fibers in the labruminferior glenohumeral ligamentcomplex (L-IGHL-C). This variation (type II), with its large sublabral recess (SL-REC; also known as a sublabral hole) and lack of a transition zone, accounted for 47.2% of specimens. Polarization microscopy; magnification, 12.5.

nent strands. Immunohistochemical staining for collagen II was positive in this area (Figure 9). The anchorage zone did not exceed the hyaline cartilage layer, in contrast to the next higher zone, which overtopped the cartilage level and was macroscopically identical to the sliding zone (Figures 8 and 9). This zone was virtually triangular in shape and was located between the labrum and the biceps tendon, with the unattached leg extended into the joint space. Immunohistochemical staining for the endothelial cell marker transglutaminase II revealed that the anchorage zone lacked blood vessels, unlike the sliding zone which displayed numerous positively stained vessels at the rim from the joint space to the biceps tendon (Figure 8). Segment II. In the second segment of the human joint, 2 predominant types of labrum could be identified. The first type (grid type) was present in 44.4% of specimens and was located next to the transition zone and comprised a prominent bundle of circular fibers and meniscoid fold (Figure 5). The second type of

labrum occurred in 47.2% of specimens and was characterized by a big sublabral recess (the sublabral hole), and the anchorage zone and the bundle of circular fibers were missing (Figure 10). A minor, third type of labrum in humans (8.4%) was described as miscellaneous. The situation in dogs was similar to the first type of labrum in humans, because the transition zone and bundle of circular fibers were present in parallel but without any meniscoid fold. The appearance of a glenohumeral ligament arising distinctly from other structures was typical in the second segment in dogs. This ligament seemed to be attached by fatty mesentery-like tissue and was placed like a tongue between the glenoid and the final ligament of the subscapular muscle, separating the subscapular recess into 2 parts (Figure 11). The medial glenohumeral tendon of dogs consisted mostly of parallel collagen fibers, which were

470

Comparison of human and canine shoulder joints

Figure 11. Craniomedial segment (segment II) of the canine GL. CA, articular cartilage; MGHL, medial glenohumeral ligament; REC, recess; SCB, subchondral bone; TZ, transition zone. Hematoxylin and eosin stain; bar, 1 mm.

Figure 13. Anterior segment (segment III) of the human GL. Left, hyaline cartilage (CA); center, transition zone (TZ) and meniscoid fold (MF); right, fibrous zone with circular fibers (CF). Polarization microscopy; magnification, 12.5.

Figure 12. Craniomedial segment (segment II) of the canine GL. CA, articular cartilage; CF, fibrous zone with circular fibers (CF); MGHL, medial glenohumeral ligament; REC, recess; SCB, subchondral bone; TZ, transition zone. Immunohistochemistry for collagen type II; bar, 200 m.

Figure 14. Caudomedial segment (segment IV) of the canine GL. Broadly based origin of the triceps muscle. CA, articular cartilage; CF, fibrous zone with circular fibers; JC, joint capsule. Hematoxylin and eosin stain; bar, 2 mm.

aligned into bundles displaying an oval shape in cross-section (Figure 11). In the craniomedial segment, the canine labrum had a 2-layered assembly: the first part consisted of an anchorage zone characterized by positive staining for collagen II (Figure 12) and a grid-like alignment of the fibers and the second part comprising of circular arranged collagen fibers, laterally overlying the anchorage zone. Segment III. The medial segment of the canine GL displayed circular collagen fibers without an anchorage area and crossing collagen fiber bundles. Some collagen fibers protruded from the fiber bundle and proceeded from lateral to mediosuperficial, creating the base of the third zone, which presented a triangular shape in cross-section and extending into the joint space (Figure 6). Immunohistochemical staining revealed that this structure was exceedingly vascularized. This so-called meniscoid fold was most

prominent in the third segment of dogs and declined caudally. Similarly in humans, this meniscoid fold was present in the anterior segment in 70% of all specimens (Figure 13). The human labrum was very wide in the third and fourth segments, averaging 5.1 and 4.5 mm, respectively. Segments IV and V. In the fourth segment in dogs, the broad, tendon-like muscular origin of the caput longum of the triceps brachii muscle predominated. In the caudal segments, the GL had similar characteristics to segment III. The circular fiber bundles were proximal to the calcification area but did not exceed the level of the hyaline cartilage layer (Figure 14). The collagen fibers protruded into a bulged fold but had lost their triangular shape and proceeded only slightly convexly into the joint space. In the fourth segment, the joint wall still displayed signs of the caudal leg of the medial glenohumeral ligament but did not show any branching caudolaterally.

471

Vol 59, No 5 Comparative Medicine October 2009

Figure 15. Craniolateral segment (segment VII) of the canine GL. Left, the transition zone (TZ) in the area of insertion of joint capsule and lateral glenohumeral ligament is shown. CA, articular cartilage; C-LGHL-C, capsuleligament complex; SCB, subchondral bone. Bar, 1 mm. Right, Use of increased magnification (insets A and B) enabled visualization of demasked fibers. Hematoxylin and eosin stain; bar, 100 m (inset A), 20 m (inset B).

In humans, both inferior segments had the typical assembly of the labrum but with a less-developed transition zone without recess formation and only very slightly developed meniscoid folds. The inferior glenohumeral ligament (IGHL) inserted inferiorly at the prominent circular fiber bundle. Segments VI and VII. In dog, the caudolateral part of the glenoid was indicated by a pronounced GL. The circular fiber bundles increased in height and width and were protruding from the hyaline cartilage surface. The extension above the calcification zone was greater than the depth of the subchondral anchorage. The capsule wall was pervaded by numerous crossing fiber bundles, which continued cranial toward the lateral glenohumeral ligament. The strength of the ligament and its entire integration into the joint capsule contributed to a bulky and laterally overhanging appearance. The labrum was only weakly developed in the sixth and the seventh segment in humans. Superiorly, the circular fiber bundle was superimposed by the intruding fibers of the labrumbiceps tendon complex, and the bony base of the labrum was expanded. Between the labrum and glenoid, a recess frequently was found in the posterior and posteriosuperior segment in humans, but small meniscoid folds were only occasionally present.

In dogs, craniolaterally in seventh segment the circular fiber bundle disappeared into the background and was replaced by a strong anchorage zone. The fibers were grid-like and microscopically showed the presence of chondrocytes (Figure 15). The labrum was closely interwoven with the structures of the lateral glenohumeral ligament (Figure 15). The maximal and minimal widths of the GL did not occur in the same segments in human and dogs (Figure 16). In human shoulders, the highest values were found in segments I and III whereas the maximal widths in dogs were located in segments II and IV. The lowest width for humans was in segment VII and for dogs in segments III and IV. The height of the labrum (Figure 17) in humans varied less between segments than did the width. In humans, this measure corresponded to the height of the circular fiber bundle, which in all segments was more developed in this direction than was the transition zone. In dogs, the maximal height of the labrum occurred in segment VI. Pertaining to these measurements, the fixation procedure led to shrinkage in all specimens. Vascular supply In dogs, the transglutaminase method revealed pronounced vascularization of the shifting zone, particularly at the free edge of the joint space, at the biceps tendon of the cranial segment, and in the medial segment in the meniscoid fold.

472

Comparison of human and canine shoulder joints

Figure 16. Mean width (mm) above the calcification zone of segments I through VII of the human (h) and canine (c) GL.

Figure 17. Mean height (mm) above the calcification zone of segments I through VII of the human (h) and canine (c) GL.

Development of the labrum in the different segments. Comparing the characteristic findings of the human and canine GL revealed that this anatomic structure consisted in general of 3 different zones in both species, but with distinct differences in the segments of these zones. The transition zone was identified in all segments of the human glenoid but could only be distinguished in the cranial, craniomedial, and craniolateral segments of the canine labrum. The second zone with its characteristic circular fibers was present in all human segments but missing from the cranial and craniolateral canine segments. The third zone, the meniscoid fold, was present in dogs at an incidence of 70%, mostly in the medial segment. The incidence for the human joints was 100% in the medial and 83% in the superior segment, whereas this zone was not found in the corresponding canine cranial segment. In this segment, human specimens had a superior labrum biceps tendon complex, in which the biceps tendon was woven into the posterior and interior glenoid in different patterns and had an osseous origin at the supraglenoid tubercle.6,14,38 Other authors40 have described 4 types of insertion in humans, which were very often accompanied by a recess within the superior area. Within the breed limits of our study, the origin of the biceps tendon of dogs showed predominantly an osseous fixation with only sporadic anchorage fibers into the GL. This situation could correlate with the reduced range of movement of the roller joint in dogs.8,22,37,41 Vascular supply. We assessed vascularization of the canine and human GL by use of primary mouse monoclonal antibodies to transglutaminase II. Immunohistochemical studies in guinea pigs12 revealed that transglutaminase is expressed by various

Discussion

types of cells, including the endothelium of arteries, veins, and lymph vessels. This enzyme also is found in the mesothelium of the pleura, pericardium, and peritoneum. The presence of transglutaminase in the vascular system has been used for the detection of microvessels during healing processes.7 In dogs, a rich vascular supply was revealed in the shifting zone, particularly at the free edge adjacent the joint space, at the biceps tendon of the cranial segment, and in medial segment in the meniscoid fold. These findings coincide with descriptions for human joints,16 in which blood vessels were located between bundles of highly fibrous avascular connective tissue in the transition zone. Other authors9 have described a vascular supply only of the peripheral zone of the labrum; that study revealed scant vascularization in the superior and anteriosuperior segments but rich vascularization in the posteriosuperior and inferior segments. The possibility of vessels entering the labrum from subchondral bone was excluded. Age may alter the degree of vascularization of the GL, in that older subjects showed less vascularization of the labrum in one study.25 Size and shape of articulation surfaces. In the present study, the ratio of the joint surface of the canine glenoid cavity to that of the humeral head was approximately 1:2.5. In comparison, the ratio in humans is approximately 1:4,30,38 indicating high incongruence in this regard between the shoulders of medium-sized canines (that is, beagles) and humans. That is, the surface area of the human glenoid is 4 to 5 times greater than that of beagle dogs. Further, we were able to confirm 3 known variations in the shape of the canine glenoid cavityin humans, these variations have been described as droplike in shape with incision (type Ia) or without incision (type Ib) and oval (type II).2 The relative influence of these various shapes of the glenoid cavity on the stability of the shoulder joint or fixation of the GL is unknown.38 Support structures. In regard to the positions of the shoulder ligaments, the human superior and medial glenohumeral ligaments were equivalent in position to the 2-component canine medial glenohumeral ligament. The inferior glenohumeral ligament, which was not found in dogs, strengthened the entire caudal capsule wall, which was scarcely developed in dogs. In contrast, the lateral wall of the canine joint was strengthened primarily by the lateral glenohumeral ligament, whereas only a thin capsule is present in the posterior segment of human shoulders.13 Similarly to the fixation of the canine lateral glenohumeral ligament, the human glenohumeral ligaments spread into the GL and only in a few cases were accompanied by additional bony fixation.18 In addition, the origin of the joint capsule at the scapula was variable in humans, depending on the insertions of the ligaments and biceps tendon. The shoulder lacks collateral ligaments outside of the joint; this function is attributed to tendons that work as contractile ligaments.22,42 In this regard, the tendon of the subscapularis muscle was present medially, and those of the infraspinatus muscle and a lateral portion of the supraspinatus muscle were located laterally.32 These muscles form the rotator cuff in humans and have great importance as active stabilization mechanisms of the canine joint.4,5 Additional muscles supporting the stability of the human joint include the teres major, teres minor, biceps, triceps, coracobrachialis, and deltoideus muscles.10 For dogs, these muscles were not only involved in extension and flexion but also in limited rotation, adduction, and abduction of the leg22.

473

Vol 59, No 5 Comparative Medicine October 2009

Biomechanical aspects. One critical feature that must be considered is that the shoulder joint of quadrupedal animals has a weight-bearing function unlike that of humans.35 In this regard, instabilities that would otherwise be minor in the human shoulder likely would have a major influence on gait in dogs.33 The predominant function of locomotion of the front limb of dogs is associated with limited range of motion in favor of increased joint stability. Experiments involving canine models. Several studies have attempted to identify appropriate animal models to answer various questions regarding the human shoulder. Immobilized shoulder joints of dogs were unsuccessful in providing insight into the pathogenesis of the stiffness of the human joint.31 In another study,35 33 animal species (including the dog) were evaluated to identify a model for rotator cuff disease in humans. The rotator cuff forms a caplike roof and is assembled from 4 muscles (the supraspinatus, infraspinatus, subscapularis, and teres minor) and their associated tendons extending from the scapula to the greater and lesser tubercle.37 Apart from some primate species, which may be less desirable for experimental purposes for various reasons, only the anatomic structure of laboratory rats showed sufficient analogy to the human shoulder joint: the skeletal structure of rats shows an almost identical development of acromion and clavicula which, together with the acromioclavicular ligament and coracoids, formed a closed arch over the underlying supraspinatus tendon.35 However, the dog is used in research addressing the blood supply of the human rotator cuff.21 According to an ultrasonographic description of the shoulder anatomy,20 the canine shoulder joint lacks a true rotator cuff; therefore comparison of these structures between humans and dogs is not possible. In using the GL of the dog as a model for humans, macroscopic and microscopic differences of this component of the shoulder must be considered together with species-specific differences in the overall structure of the shoulder joint. Any of these differences could mitigate the utility of the model and the validity of subsequent comparison.

The authors would like to thank the research group of U Knig, T Barthel, F Gohlke, and JF Lhr (Bayerische Julius Maximilians, Universitt Wrzburg, Germany) for providing us the excellent specimens of human glenoid labrum (Figures 1, 5, 7, 10, and 13).

Acknowledgment

1. Andary JL, Petersen SA. 2002. The vascular anatomy of the glenohumeral capsule and ligaments: an anatomic study. J Bone Joint Surg Am 84-A:22582265. 2. Anetzberger H, Putz R. 1996. The scapula: principles of construction and stress. Acta Anat (Basel) 156:7080. 3. Bardet JF. 1998. Diagnosis of shoulder instability in dogs and cats: a retrospective study. J Am Anim Hosp Assoc 34:4254. 4. Bardet JF. 2002. Shoulder diseases in dogs. Vet Med 97:909918. 5. Bardet JF. 2002. Shoulder instability and joint pain in dogs and cats. First World Orthopaedic Veterinary Congress, 58 Sep 2002, Munich, Germany. 6. Barthel T, Konig U, Bohm D, Loehr JF, Gohlke F. 2003. Anatomy of the glenoid labrum. Orthopade 32:578585. [Article in German]. 7. Buemi M, Galeano M, Sturiale A, Ientile R, Crisafulli C, Parisi A, Catania M, Calapai G, Impala P, Aloisi C, Squadrito F, Altavilla D, Bitto A, Tuccari G, Frisina N. 2004. Recombinant human erythropoietin stimulates angiogenesis and healing of ischemic skin wounds. Shock 22:169173.

References

8. Cook JL, Renfro DC, Tomlinson JL, Sorensen JE. 2005. Measurements of angles of abduction for diagnosis of shoulder instability in dogs using goniometry and digital image analysis. Vet Surg 34:463468. 9. Cooper DE, Arnoczky SP, OBrien SJ, Warren RF, DiCarlo E, Allen AA. 1992. Anatomy, histology, and vascularity of the glenoid labrum. An anatomical study. J Bone Joint Surg Am 74:4652. 10. Craig E, Hohn RB, Anderson MS, Anderson WD. 1980. Surgical stabilization of traumatic medial shoulder dislocation: dogs and cats. Kleintierpraxis 25:329338. [Article in German]. 11. Evans HE. 1993. The skeleton. In: Evans HE, editor. Millers anatomy of the dog. Philadelphia (PA): WB Saunders. 12. Gendek-Kubiak H, Gendek EG. 2004. Expression of tissue transglutaminase in blood and lymphatic vessel endothelia and in mesothelium. Rocz Akad Med Bialymst 49 Suppl 1:195197. 13. Gohlke F, Essigkrug B, Schmitz F. 1994. The pattern of the collagen fiber bundles of the capsule of the glenohumeral joint. J Shoulder Elbow Surg 3:111128. 14. Harzmann HC, Burkart A, Wortler K, Vaitl T, Imhoff AB. 2003. Normal anatomical variants of the superior labrum biceps tendon anchor complex. Anatomical and magnetic resonance findings. Orthopade 32:586594.[Article in German]. 15. Heidenhain M. 1915. ber die Mallorysche Bindegewebsfrbung mit Karmin und Azokarmin als Vorfarben. Z wiss Mikr 32:361 372. 16. Hertz H, Weinstabl R, Grundschober F, Orthner E. 1986. Macroscopic and microscopic anatomy of the shoulder joint and the limbus glenoidalis. Acta Anat (Basel) 125:96100.[Article in German]. 17. Huber WP, Putz RV. 1997. Periarticular fiber system of the shoulder joint. Arthroscopy 13:680691. 18. Knig U. 1998. Das Labrum glenoidale: eine anatomischhistologische Studie unter besonderer Bercksichtigung des Kollagenfaserverlaufs. Wrzburg (Germany): Bayerische Julius Maximilians Universitt. 19. Korvick D, Athanasiou K. 1997. Variations in the mechanical properties of cartilage from the canine scapulohumeral joint. Am J Vet Res 58:949953. 20. Kramer M, Gerwing M. 1994. Sonography of the shoulder joint in dogs and its surrounding soft tissues. Part A: sonographical anatomy of the shoulder joint and its soft issues. Kleintierpraxis 39:7180. [Article in German]. 21. Kujat R. 1986. The microangiographic pattern of the glenoid labrum of the dog. Arch Orthop Trauma Surg 105:310312. 22. Liebich HG, Maierl J, Knig HE. 2004. Verbindungen der Knochen der Schultergliedmae In: Knig HE, Liebich HG, editors. Anatomie der Haussugetiere: Lehrbuch und Farbatlas fr Studium und Praxis. Stuttgart (Germany): Schattauer. 23. Mitchell RAS, Innes JF. 2000. Lateral glenohumeral ligament rupture in three dogs. J Small Anim Pract 41:511514. 24. Morgan JP, Silverman S. 1993. Arthrography. In: JP Morgan JP, Silverman S, editors. Techniques of veterinary radiography. Ames (IA): Iowa State University Press. 25. Prodromos CC, Ferry JA, Schiller AL, Zarins B. 1990. Histological studies of the glenoid labrum from fetal life to old age. J Bone Joint Surg Am 72:13441348. 26. Richardson KC. 1960. Studies on the structure of autonomic nerves in small intestine, correlating the silver impregnated image in light microscopy with permanganate-fixed ultrastructure in electron microscopy. J Anat 94:457472. 27. Sager M, Assheuer J. 2002. Injuries of the biceps tendon in dogs: an MRI study. First World Orthopaedic Veterinary Congress, 58 Sep 2002, Munich, Germany. 28. Sager M, Assheuer J. 2005. The glenoid labrum: an important detail in the pathology of the scapulohumeral joint in the dog. Twelfth Annual Conference of the European Association of Veterinary Diagnostic Imaging, 58 Oct 2005, Naples, Italy.

474

Comparison of human and canine shoulder joints

29. Schaller O. 1992. Arthrologia: articulationes membri thoracici; In: Schaller O, editor. Illustrated veterinary anatomical nomenclature. Stuttgart (Germany): F Enke. 30. Schiebler TH, Schmidt W, Zilles K. 2003. Rumpfwand und Extremitten: Schultergrtel und obere Extremitt. In: Schiebler T, Schmidt W, Zilles K, editors. Anatomie: Zytologie, Histologie, Entwicklungsgeschichte, makroskopische und mikroskopische Anatomie des Menschen. Berlin (Germany): Springer. 31. Schollmeier G, Uhthoff HK, Sarkar K, Fukuhara K. 1994. Effects of immobilization on the capsule of the canine glenohumeral joint. A structural functional study. Clin Orthop Relat Res3742. 32. Seiferle E, Frewein J. 2004. Eigenmuskulatur der Schultergliedmae der Fleischfresser. In: Nickel R, Schummer A, Seiferle E, editors. Lehrbuch der Anatomie der Haustiere, Bewegungsapparat. Berlin (Germany): P Parey. 33. Sidaway BK, McLaughlin RM, Elder SH, Boyle CR, Silverman EB. 2004. Role of the tendons of the biceps brachii and infraspinatus muscles and the medial glenohumeral ligament in the maintenance of passive shoulder joint stability in dogs. Am J Vet Res 65:1216 1222. 34. Snyder SJ, Karzel RP, Del Pizzo W, Ferkel RD, Friedman MJ. 1990. SLAP lesions of the shoulder. Arthroscopy 6:274279. 35. Soslowsky LJ, Carpenter JE, DeBano CM, Banerji I, Moalli MR. 1996. Development and use of an animal model for investigations on rotator cuff disease. J Shoulder Elbow Surg 5:383392.

36. Suter PF, Carb AV. 1969. Shoulder arthrography in dogsradiographic anatomy and clinical application. J Small Anim Pract 10:407413. 37. Tillmann B. 2003. Obere Extremitt. In: Tillmann B, Tndury G, Zilles K, editors. Anatomie des Menschen Rauber Kopsch. Stuttgart (Germany): Thieme. 38. Tischer T, Putz R. 2003. Anatomy of the superior labrum complex of the shoulder. Orthopade 32:572577. [Article in German]. 39. Van Ryssen B, Van Bree H, Whitney WO, Schulz KS. 2003. Small animal arthroscopy. In: Slatter HD, editor. Textbook of small animal surgery. Philadelphia (PA): WB Saunders. 40. Vangsness CT Jr, Jorgenson SS, Watson T, Johnson DL. 1994. The origin of the long head of the biceps from the scapula and glenoid labrum. An anatomical study of 100 shoulders. J Bone Joint Surg Br 76:951954. 41. Vollmerhaus B, Waibl H, Roos H. 1994. Gelenkknorpel, Gelenkkapsel, Gelenke der Schultergliedmae. In: Frewein J, Vollmerhaus B, editors. Anatomie von Hund und Katze. Berlin (Germany): Blackwell. 42. Wnsche A, Budras KD. 2004. Synoviale Einrichtungen (Synovialstrukturen) der Schultergliedmae. In: Budras KD, Fricke W, Richter R, editors. Atlas der Anatomie des Hundes Lehrbuch fr Tierrzte und. Studierende. Hannover (Germany): Schltersche Verlagsanstalt

475