Detailed knowledge of individual stress distribution of the glenohumeral joint can be important not only for the basic scientist but also for the clinician. For example, to explain the high incidence of secondary arthritis after surgical reconstruction in patients with glenohumeral instability, individual stress distributions of the glenohumeral joint would be helpful.4,6,7,17 If stress distribution varies with different situations one might use the distributions to select between treatment options. The clinical consequence would be that in such cases, not only an isolated surgical technique but a combination of a Bankart repair and a Neer shift may be necessary. However, analysis of individual stress distribution in glenohumeral instability is problematic because many interactive factors of the complex stabilizing mechanism such as individual capsuloligamentous laxity, muscle strength, or joint congruity are difficult to assess in vivo and in vitro.
Therefore, among various possible approaches, measurement of subchondral bone mineralization by computed tomography (CT) osteoabsorptiometry seems to be a good method to analyze individual stress distribution in the unstable glenohumeral joint.1,3,12 The rationale of this method is that subchondral bone mineralization can be used as a morphologic marker for the individual loading history of a joint in humans. A relationship exists between a change of mechanical stress on a joint surface and a change of subchondral bone mineralization.5,10,12 This biologic response to stress distribution, which depends on all factors influencing individual joint mechanics, can be described by mineralization patterns and quantified indirectly by localization of representative density maxima.12,18 To date, most of the studies evaluating glenoid subchondral bone mineralization using CT osteoabsorptiometry were done on anatomic specimens. To evaluate glenohumeral instabilities, however, an in vivo approach is required because selection depends not only on morphologic but often on clinical data to identify traumatic and atraumatic causes. Recent data on glenoid subchondral bone mineralization of macroscopically normal shoulder specimens established a density distribution with a constant anterior maximum located close to the glenoid notch in 100% and a second frequent maximum located close to the posterior glenoid rim in 82% of cases, which indicates a constant, bicentric glenohumeral stress distribution.18 In contrast, shoulders with supraspinatus tendon ruptures frequently have a third density maximum of central position and a superior and anterior shift of the posterior glenoid maximum of varying degrees, whereas the anterior maximum remains unchanged.3
This means that defined mechanical situations of the glenohumeral joint develop characteristic patterns of subchondral bone mineralization that probably are attributable to a decrease of mineral density at the normal site and an increase at a new site. It also shows that changes of density maximum position correspond with the direction of superior humeral head displacement and, according to the size of tendon rupture, correspond to the degree of humeral head displacement.3 However, in anterior glenohumeral instability, which may have varying causes, increased humeral head translation is typical in the anterior and the inferior directions. According to dynamic imaging studies, glenohumeral translation mainly alters during elevation and external rotation when compared with stable shoulders.9,20 These kinematic studies also observed a relationship between the degree and the direction of humeral head translation and different causes leading to instability. Therefore, in anterior glenohumeral instability a change of glenoid stress distribution attributable to altered glenohumeral force transmission has to be expected.
Therefore, two hypotheses were tested: (1) that increased anterior and inferior humeral head translation in patients with anterior glenohumeral instability leads to an anterior and inferior shift of the glenoid density maximum position; and (2) that the position of glenoid density maxima is different between traumatic and atraumatic glenohumeral instability.
MATERIALS AND METHODS
Computed tomography data sets from patients, who were treated at our clinic for chronic anterior glenohumeral instability (n = 13), were grouped retrospectively for analysis of subchondral bone mineralization of the glenoid. The diagnostic CT arthrographs, which were done before specific treatment was started, were available from each selected, symptomatic shoulder. Patients were included only if instability was not combined with a rupture of the rotator cuff that would influence glenohumeral stress distribution.3 A minimum of 1 year of symptoms attributable to anterior glenohumeral instability was considered necessary to ensure an adequate time for the biologic response at the subchondral bone plate.1,12 In addition, patients were subgrouped into categories according to the underlying disease into traumatic or atraumatic glenohumeral instabilities on the basis of the medical documentation and CT studies. According to the criteria of Thomas and Matsen,19 major trauma leading to the first dislocation, unilateral involvement with a stable contralateral shoulder, an anteroinferior capsulolabral avulsion confirmed by CT arthrographs (transverse sections between 2 and 4 mm perpendicular to the glenoid), and the surgical reports after an open Bankart procedure were classified as instability of traumatic origin. In contrast a minor trauma leading to the first dislocation, bilateral involvement with positive sulcus sign, and indices of general capsuloligamentous laxity without evidence of a capsulolabral detachment were classified as instability of atraumatic origin. In five patients surgical reports after a Neer shift also were available whereas one patient was treated nonoperatively with a physical rehabilitation program. Because of the retrospective character of the study, patient consent was not necessary.
Computed tomography scans (transverse sections, 2 mm perpendicular to the glenoid) were taken from age-matched, healthy shoulder specimens (n = 13) that served as control group (Institute for Forensic Medicine at our university). None of these specimens had any deformity evident for a fracture or signs of previous surgeries. Macroscopic exclusion criteria were defined as signs of pathologic changes or the presence of a special anatomic variant at the anterior capsulolabral complex such as a Buford complex or a sublabral foramen. Therefore, each shoulder specimen was harvested in toto and inspected in a standardized manner. After confirmation of an intact bursal side of the rotator cuff an anterior arthrotomy of the glenohumeral joint close to the humeral head was done to rule out disease of the articular side of the rotator cuff, tears of the biceps tendon or at its insertion into the superior glenoid, detachment or fraying of the capsulolabral insertion at the glenoid circumference, avulsions of the glenohumeral ligaments at the humeral side, and signs of cartilage destruction of both articular surfaces.
Computed Tomography Osteoabsorptiometry
To do CT osteoabsorptiometry, the data set from each CT arthrograph scan was loaded on a workstation. The density distribution of the subchondral bone mineralization of the glenoid was calculated in a standardized manner.3,18 All data were processed with a special software program (Analyze®, Mayo Foundation, Rochester, MN). After three-dimensional reconstruction of the scapula in a true lateral position of the glenoid, the image was rotated 90° to a frontal view of the glenoid. A selective depiction of the glenoid subchondral bone plate was done from each slice and the data were processed by means of a maximum intensity projection to calculate the density distribution of subchondral bone mineralization. For better presentation, the reconstruction of glenoid density distribution was projected as a map over the three-dimensional image of the glenoid and colors were used according to defined levels of hounsfield units (steps of 100; range, 200–1200 HU).
The analysis included descriptive and quantitative evaluations as reported previously.3,18 The combination of diagrammed maxima was recorded and the position of the constant anterior glenoid density maximum (x1,y1) and the second most frequent density maximum localized at the posterior glenoid (x2,y2) were measured in a standardized manner. Therefore, mirror images of left shoulders were used to obtain a presentation of the glenoid surface comparable with right shoulders. Furthermore, a linear extensible coordinate system with an axis length of 20 AU (arbitrary units) was used to exclude individual variations of size. The x axis represents the maximum AP glenoid diameter, and the y axis represents the maximum superoinferior glenoid diameter. The reference point of each maximum was defined as the centroid of the area represented by the two highest density levels. Statistical analysis for comparison of the anterior and posterior maximum positions to normal shoulders was done with the Mann-Whitney U test. Probability values less than 0.05 were considered significant.
A group consisting of patients with recurrent anterior glenohumeral instability (Tables 1 and 2) was compared with a control group consisting of normal glenohumeral specimens.
The mean age of the patients with anterior glenohumeral instability (four women, nine men) was 27 ± 4 years (range, 22–35 years) compared with 32 ± 9 years (range, 18–48 years) in the control group (specimens from two women, 11 men). The left shoulder was examined in eight patients and the right shoulder was examined in five patients whereas the shoulders of the control group included six left and seven right specimens. Seven male patients had a traumatic glenohumeral instability with anteroinferior capsulolabral detachment, and all had surgery (Table 1). Four women and two men had an atraumatic glenohumeral instability without avulsion of the capsulolabral complex. Five patients had surgery whereas one patient was treated nonoperatively (Table 2). The shape of the glenoid2 was teardroplike with an anterior notch in five cases, teardroplike without an anterior notch in seven cases, and oval in one case, whereas in the control group the shape of the glenoid was teardroplike with an anterior notch in eight cases and without an anterior notch in five cases.
The position of both assessed density maxima changed significantly in anterior glenohumeral instability with a shift of varying degrees in the direction of humeral head displacement (Fig 1).
As described previously, different combinations of density maxima on the glenoid surface may be distinguished.3,12,18 Mineralization Type A (typical in normal shoulders) that has an anterior maximum with a posterior maximum (Type A1) or without a posterior maximum (Type A0) was seen in nine shoulders from the specimens in the control group (Type A1, n = 8; Type A0, n = 1) compared with eight shoulders of patients in the instability group (Type Al, n = 5; Type A0, n = 3). An additional centrally positioned maximum (frequent in shoulders with a tear of the supraspinatus tendon) determined as Type B1 (with posterior maximum) or Type B0 (without posterior maximum) was observed in two shoulders of specimens in the control group (Type B1, n = 2) but was not seen in unstable shoulders of patients. An additional anterior but more inferiorly positioned maximum determined as mineralization Type C was found in two of the shoulders from control specimens (Type C1, n = 2) and in five shoulders of patients with instability (Type C1, n = 5).
The anterior glenoid density maximum shifted anterior (px1 = 0.005) and inferior (py1 = 0.004) in shoulders with anterior glenohumeral instability and the posterior maximum shifted anterior (px2 = 0.002; py2 = 0.767). The degree and the direction of the shifted position of anterior density maximum varied between traumatic and atraumatic glenohumeral instabilities. The anterior glenoid density maximum shifted mainly anterior in traumatic instability (px1 = 0.024; py1 = 0.088) and inferior in atraumatic instability (px1 = 0.025; py1 = 0.003). A statistical analysis of the data is shown in Table 3.
Stress distribution of a joint depends on variables such as height and direction of the resulting force, size and position of the contact areas of the corresponding joint surfaces, and the subarticular bone architecture.5,10,12,14–16 All factors together lead to a specific distribution pattern of subchondral bone plate mineralization, which characterizes the individual loading history of a joint.5,12
Computed tomography osteoabsorptiometry may be used as an analytic tool to monitor the stress distribution of unstable joints in an in vitro situation.1,3,12 Because the method is based on the mineral content of bone, it is unlikely that this evaluation will change after death. Therefore, comparison between in vivo and in vitro situations seems valid to us. Nevertheless, some, methodologic limitations in this study on a small cohort of patients such as selection criteria and density maximum localization deserve consideration. Selection of patients with an atraumatic origin is problematic because of subjective clinical criteria and variable labral attachments, whereas selection of specimens by macroscopic inspection cannot rule out atraumatic instability. Therefore any suspect anatomic variant such as a sublabral foramen, a Buford complex, or visible changes such as fraying at the capsulolabral complex also were excluded. However, after considering these limitations, the locations of maximum density between normal and unstable shoulders would have been even more significant and therefore these differences would not alter conclusions. Another point to consider is the method of localization of the anterior density maximum by a rectangular coordinate system, which is a simple and reproducible method, because values of the anterior shift could depend on variations of glenoid shape. However, because glenoid shape did not vary substantially between both groups, a significant influence on data seems unlikely.
Several radiologic and biomechanical studies have investigated the relationship between anterior glenohumeral instability and humeral head translation, suggesting a change of the contact areas of the humeral head with the glenoid and the resultant transarticular force.8,9,13,20 Accordingly, in the current study patients with anterior glenohumeral instability and mineralization of the glenoid subchondral bone showed increased anterior and inferior density when compared with normal joints. Shoulders with anterior glenohumeral instability showed no additional maximum at a central position in contrast to shoulders with a tear of the supraspinatus tendon.3 However, an increased incidence of an additional maximum positioned anteroinferior was observed in anterior glenohumeral instability, which was absent in shoulders with an isolated supraspinatus tear3 and rare in normal shoulders.18 This suggests a morphologic criterion corresponding to the typical direction of humeral head translation which has not yet been reported.
Our data for a small number of patients revealed a significant change of glenoid density maxima positions in anteroinferior glenohumeral instability indicating an altered stress distribution on the glenoid toward anterior and inferior in the direction of humeral head translation.
In addition, CT osteoabsorptiometry enables more specific analysis of glenoid stress distribution, which allows assessment of variations between traumatic and atraumatic glenohumeral instabilities. Therefore, in the traumatic subgroup, which typically is characterized by detachment of the anteroinferior capsulolabral complex, only the anterior shift of the anterior density maximum was significant whereas the inferior shift was not. A possible explanation might be that primarily anterior humeral head translation is affected because of decreased passive anterior stabilization, whereas the passive superior suspension mechanism of the humeral head remains intact. However, in shoulders with atraumatic instability, primarily the inferior shift of the anterior density maximum was increased, whereas the anterior shift was of a minor degree. Because these shoulders typically have an increased laxity of the joint capsule and its ligaments because of weak connective tissue, humeral head suspension and anterior humeral head translation are likely to be pathologically altered. Why glenohumeral instability tends to be more significant toward inferior than anterior might be explained by the uninjured capsulolabral attachment in this subgroup.
Nevertheless, variations in glenoid subchondral bone mineralization of unstable joints also may depend on the individual active stabilizing mechanism, which is in accord with a report analyzing active humeral head stabilization by dynamic magnetic resonance imaging.20 In a kinematic study, Von Eisenhart-Rothe et al20 suggested that different etiology-related alterations of the cavity compression mechanism are likely, as described by Lippitt and Matsen.11
In vivo CT osteoabsorptiometry is an additional tool to biomechanical analysis in investigating complexity of anterior glenohumeral instability. Our data showed a shift of glenoid density maxima positions in the direction of humeral head displacement. This shows that anterior glenohumeral instability often leads to a typical change of subchondral bone mineralization, which indicates more anterior and inferior glenoid stress distribution in comparison with normal joints. Because the degree of anterior and inferior shift of the anterior density maximum position is likely to depend on the cause of glenohumeral instability, this variation of glenohumeral stress distribution should be considered in future therapeutic concepts.
The authors thank Professor Walli for continuing support.
1. Anetzberger H, Müller-Gerbl M, Scherer MA, et al: Change in subchondral bone mineralization after reconstruction of the anterior cruciate ligament of the sheep. Unfallchirurg 97:655–660, 1994.
2. Anetzberger H, Putz R: The scapula: Principles of construction and stress. Acta Anat 156:70–80, 1996.
3. Anetzberger H, Schulz CU, Pfahler M, Refior HJ, Muller-Gerbl M: Subchondral mineralization patterns of the glenoid after tear of the supraspinatus. Clin Orthop 404:263–268, 2002.
4. Brems JJ: Arthritis of dislocation. Orthop Clin North Am 29:453–466, 1998.
5. Carter DR, Orr TE, Fyhrie DP: Relationships between loading history and femoral cancellous bone architecture. J Biomech 22:231–244, 1989.
6. Chapnikoff D, Besson A, Chantelot C, et al: Bankart procedure: Clinical and radiological long-term outcome. Rev Chir Orthop Reparatrice Appar Mot 86:558–565, 2000.
7. Cole BJ, Warner JJ: Arthroscopic versus open Bankart repair for traumatic anterior shoulder instability. Clin Sports Med 19:19–48, 2000.
8. Flatow EL, Ateshian GA, Soslowsky LJ, et al: Computer simulation of glenohumeral and patellofemoral subluxation: Estimating pathological articular contact. Clin Orthop 306:28–33, 1994.
9. Howell SM, Galinat BJ, Renzi AJ, Marone PJ: Normal and abnormal mechanics of the glenohumeral joint in the horizontal plane. J Bone Joint Surg 70A:227–232, 1988.
10. Kummer B: Kausale Histogenese der Gewebe des Bewegungsapparates und funktionelle Anpassung. In Staubesand J (ed). Benninghof. Anatomie 1. München, Urban and Schwarzenberg 199-213, 1985.
11. Lippitt S, Matsen F: Mechanisms of glenohumeral joint stability. Clin Orthop 291:20–28, 1993.
12. Müller-Gerbl M: The subchondral bone plate. Adv Anat Embryol Cell Biol 141:115–116, 1998.
13. Paletta Jr GA, Warner JJ, Warren RF, Deutsch A, Altchek DW: Shoulder kinematics with two-plane x-ray evaluation in patients with anterior instability or rotator cuff tearing. J Shoulder Elbow Surg 6:516–527, 1997.
14. Pauwels F: Eine neue Theorie über den Einfluβ mechanischer Reize auf die Differenzierung der Stützgewebe. Z Anat Entwickl-Gesch 121:478–515, 1965.
15. Radin EL, Paul IL: Does cartilage compliance reduce skeletal impact loads? The relative force-attenuating properties of articular cartilage, synovial fluid, periarticular soft tissues and bone. Arthritis Rheum 13:139–144, 1970.
16. Radin EL, Paul IL, Lowy M: A comparison of the dynamic force transmitting properties of subchondral bone and articular cartilage. J Bone Joint Surg 52A:444–456, 1970.
17. Rosenberg BN, Richmond JC, Levine WN: Long-term followup of Bankart reconstruction: Incidence of late degenerative glenohumeral arthrosis. Am J Sports Med 23:538–544, 1995.
18. Schulz CU, Pfahler M, Anetzberger H, et al: The mineralization patterns at the subchondral bone plate of the glenoid cavity in healthy shoulders. J Shoulder Elbow Surg 11:174–181, 2002.
19. Thomas SC, Matsen III FA: An approach to the repair of avulsion of the glenohumeral ligaments in the management of traumatic anterior glenohumeral instabilities. J Bone Joint Surg 71A:506–513, 1989.
© 2004 Lippincott Williams & Wilkins, Inc.
20. Von Eisenhart-Rothe RM, Jäger A, Englmeier KH, Vogl TJ, Graichen H: Relevance of arm position and muscle activity on three-dimensional glenohumeral translation in patients with traumatic and atraumatic shoulder instability. Am J Sports Med 30:514–522, 2002.