CLINICAL ORTHOPAEDICS AND RELATED RESEARCH Number 448, pp.

2227 2006 Lippincott Williams & Wilkins

Shoulder Kinematics in 25 Patients with Impingement and 12 Controls

E. Hallstrm, MD*; and J. Krrholm, MD, PhD

We used dynamic radiostereometry to study the threedimensional kinematics of the shoulder joint during active abduction. Twenty-five patients experiencing shoulder symptoms (Neer Stage 2) for more than 18 months, without total rotator cuff tears, participated. Eight men and four women without shoulder symptoms constituted controls. The rotation of the humeral head relative to a fixed scapula and the absolute rotation of the humerus (caused by humeral, scapular, and trunk motion) were measured. The rotations were calculated in the order of abduction/adduction (anteroposterior axis), internal/external rotation (longitudinal axis), and flexion/extension (transverse axis). The absolute abduction of the humerus in our patients did not differ from controls, nor did the abduction in the glenohumeral joint. During abduction, the humeral centre displaced medially, proximally, and anteriorly. In the patient group, slightly more (11.5 mm) proximal translation was observed. Presence of impingement syndrome was associated with increased proximal translation of the humeral head center, which occurred in the early phase of the arc of motion. Level of Evidence: Diagnostic Level I. See Guidelines for Authors for a complete description of Levels of Evidence.

Shoulder impingement is thought to be caused by inadequate space for clearance of the rotator cuff tendon as the arm is elevated.7,23,41 Kinematic changes are believed to occur primarily in symptomatic patients and to result in additional decrease of the subacromial space, which could
From the *Department of Orthopaedics, Uddevalla Hospital, Uddevalla, Sweden; and the Department of Orthopaedics, Institute of Surgical Sciences, Sahlgrenska University Hospital, Gothenburg, Sweden. One or more of the authors (EH, JK) received funding from the Ingabritt and Arne Lundberg Research Foundation, Vstra Gtalandsregionen, FoU-rdet FyrBoDal och Gteborgs Lkarfrening. Each author certifies that his institution has approved the human protocol for this investigation and that all investigations were conducted in conformity with ethical principles of research and that informed consent was obtained. Correspondence to: Dr. E. Hallstrm, MD, Department of Orthopaedics, Uddevalla Hospital, S-451 80 Uddevalla, Sweden. Phone: 46-522-92535; Fax: 46-522-93060; E-mail: erling.hallstrom@vgregion.se. DOI: 10.1097/01.blo.0000224019.65540.d5

aggravate the symptoms.7,16,29 Motions that bring the greater tuberosity closer to the coracoacromial arch5 may be particularly problematic. These motions include excessive superior or anterior translation of the humeral head and the glenoid fossa, inadequate lateral (external) rotation of the humerus, and decreased normal scapular upward rotation.31,32,36,38,40 Several studies have recorded shoulder kinematics using biplane radiography with10 or without18 implanted markers. These studies have been restricted to a few subjects and focused on normal shoulder kinematics. More recently, open configuration magnetic resonance imaging (MRI) has evolved as an instrument to further explore the kinematics of the normal and the diseased shoulder joint, though cost and availability may present a problem.8 Some investigators have used methods based on digitizing discrete bony landmarks palpable through the skin.20 Another skin-based approach involves capturing three-dimensional scapular orientation directly with a magnetic tracking device.24 This has been tried statically by coupling a magnetic sensor to an alignment jig15,25 and dynamically by attaching a magnetic sensor directly to the acromion.19,24 Although most of these methods have shown satisfactory reliability, only one study24 has addressed the issue of accuracy. An alternative approach is dynamic radiostereometry.17,33,39 This method is based on fixed skeletal landmarks, has a documented high resolution and also enables recordings during active joint motion. We hypothesized patients with shoulder symptoms would, for a given active abduction, position the humeral head center more proximally than controls. We addressed three questions: (1) How precisely can a patient repeat an active abduction of the shoulder joint? (2) What is the relative contribution of scapular rotation and motions in the glenohumeral joint to an active abduction motion of the arm? (3) Does the detailed three-dimensional motion of the shoulder joint in patients with impingement syndrome without total rupture of the rotator cuff tendon or signs of osteoarthritis differ from controls?
22

Number 448 July 2006

Dynamic RSA of the Shoulder Joint

23

MATERIALS AND METHODS

Sixteen men and 9 women aged an average 51 years (range, 2963 years) with shoulder symptoms (Neer Stage 2)26,27 for more than 18 months and without rotator cuff tears were included. Eight men and 4 women (mean age 32 years; range, 2259 years) without shoulder symptoms constituted controls. All subjects were studied during active shoulder abduction. Four to six spherical tantalum markers ( 0.8 or 1.0 mm) were inserted into the scapula (acromion) and the humeral head under local anesthesia. Two to 6 weeks later, the patients were studied during active abduction using radiostereometric analysis (RSA Biomedical, Ume, Sweden). To enable radiostereometric studies of the glenohumeral joint, a setup including 2 filmexchangers designed for simultaneous exposures was used. The exposure rate was set at 2 per second. The height of the filmexchangers could be adjusted depending on the length of the individual patient. The two film-exchangers were placed side by side. A uniplanar calibration cage specially designed to suit the two film-switchers was constructed and fixed in front of them (Fig 1). Before the radiographic examinations were initiated, each subject (patients and controls) performed several abduction trials to feel as comfortable as possible and to obtain a fairly constant speed. The subjects were taught to maintain the glenohumeral joint within the limits of the aperture (film size 35 x 35 cm). The radiographic examinations were initiated by recording of a starting or reference position. A pair of stereoradiographs was exposed with the arm aligned to the longitudinal axis of the body and the forearm in external rotation with the palm facing forward corresponding to a well-defined anatomic position. All subsequent recordings were related to this position of the arm. Thereafter, the dynamic recordings were started using a speed of 2 simultaneous stereoradiographic exposures per second for 5 to 6 seconds. Because of variable motion speed and ability to synchronize the motions to the film-exchanger, only 5 to 8 (of 10 or 12 available) representative pairs could be included in the final analysis in each patient (median, 7). The roentgen films were scanned at 300 dpi using a flat-bed scanner (Sharp JX 610, Osaka, Japan), and measured using a digital technique.1 To enable measurements of humeral head

translations in a reproducible way, a fictive point corresponding to the humeral head center was constructed. Circular templates were used to find the head center, but only on the two images of the reference position. The position of these centers were measured using the RSA digital software and its three dimensional coordinates were recorded. Thus, this plotting was done once in each shoulder (subject). The position of this point was thereafter transferred to all other subsequent examinations of the same shoulder using its computed position relative to the humeral head markers. These computations were based on presence of stable and sufficiently well-scattered tantalum marker in the humeral head.28 We measured rotations of the humeral head using the scapula as a fixed reference segment. In RSA, rotations are calculated in a specific order; first around the transverse, then around the longitudinal and finally around the anteroposterior (AP) axis. Because the patient performed an active abduction corresponding to rotations around the AP axis, we decided to adjust the position of the cage coordinate system 90 by rotation around the longitudinal axis. This means in this study, rotations were calculated in the order of abduction/adduction (AP axis), internal/external rotation (longitudinal axis), and flexion/extension (transverse axis) (Fig 2). To estimate the contribution of scapular abduction at maximum relative abduction in the glenohumeral joint, we also recorded the absolute rotations of the humerus at maximum around the AP axis. These data were extracted from the same recording used to obtain information about the relative motions. This rotation corresponds to the rotation of the humerus in relation to the cage coordinate system, which, in addition to the relative motions in the glenohumeral joint, also includes rotations proximal to this joint, ie, of the scapula and the trunk. The median values (ranges) of the mean error of rigid body fitting (indicator of marker stability) and condition number (indicator of marker scatter) were 0.117 (range, 0.0300.330) and 125 (range, 80 455). The reproducibility was tested in six patients, who repeated the active abduction after a time interval of 15 minutes.

Fig 1. Two film-exchangers designed for simultaneous exposures were used (exposure rate, 2 per second).

Fig 2. Position of the coordinate axes at the reference or starting position. At the computation of relative motions the scapula maintained this position throughout the examination. In this study, rotations were calculated in order of abduction/adduction (AP axis), internal/external rotation (longitudinal axis) and flexion/extension (transverse axis).

24

Hallstrm and Krrholm

Clinical Orthopaedics and Related Research

Data for each type of motion analyzed were interpolated linearly at 5 intervals of abduction. Statistical analyses (SPSS 13.0 for Windows, Chicago, IL, USA) were based on recordings between 25 and 55 of abduction in the glenohumeral joint using scapulae as a fixed reference segment. The selection of the 25 to 55 interval was to maximize number of observations and meant 25 patients and 12 controls could be included in the statistical analysis. Repeated measure ANOVA was used. The significance level was set at p < 0.05. The study was approved by the local ethical committee.

RESULTS Two pooled standard deviations for flexion/extension and internal/external rotation when performed two times by six of the patients amounted to 3.0 and 4.2, respectively. The corresponding values for medial/lateral, proximal/distal, and anterior/posterior translations of the humeral head centre were 0.6, 0.2 and 0.2 mm, respectively. Patients and controls exhibited similar mean absolute maximum abduction of the humerus (regardless of origin) around the AP axis (139, range, 101168 and 137, range, 126155, respectively). The corresponding mean relative maximum abduction of the humerus with the scapula fixed (in the glenohumeral joint) was also similar (70,range, 5793 and 72, range, 4283, respectively. Thus, only about 50% of the total motion occurred in the glenohumeral joint, regardless of the presence of impingement at maximum abduction of the arm. During the initial phase of active abduction, the humerus was slightly extended compared to its reference (starting position) of the hanging arm and the palm facing forwards. With proceeding abduction, there was an associated flexion in both groups, which reached a mean of 10

Fig 4. Internal/external rotation during active abduction, mean, SE. The distortion of the mean at 010 and 6070 of abduction is partly due to missing observations.

at 60 abduction, despite the presence of impingement (Fig 3). In both groups there was also similar external rotation of the humerus of about 20 and 35 in the patients and controls, respectively (Fig 4). In patients and controls, the humeral head center similarly shifted a mean 1 and 2 mm medially with abduction of the shoulder joint (Fig 5). Small proximal displacements were observed in both groups. Patients with impingement did, however, maintain a position, which was 1 to 1.5 mm more proximal (p 0.04) during the motion (Fig 6). Patients with impingement showed almost no mean AP shift of the head centre. In controls there was a minor mean anterior displacement but without any difference compared to the patient group (p 0.3) (Fig 7).

Fig 3. Flexion/extension of humerus during active abduction, mean, SE. The distortion of the mean at 010 and 6070 of abduction is partly due to missing observations.

Fig 5. Mediolateral translation of humeral head centre during active abduction. The distortion of the mean at 010 and 6070 of abduction is partly due to missing observations.

Number 448 July 2006

Dynamic RSA of the Shoulder Joint

25

Fig 6. Proximal/distal translation of humeral head centre during active abduction, mean, SE. The distortion of the mean at 010 and 6070 of abduction is partly due to missing observations.

DISCUSSION Generally, the rotational center of the humeral head cannot be reliably determined with active or passive motion of the intact glenohumeral joint because the motion will not necessarily be purely rotational, particularly if one assumes mismatch in curvature between the glenoid and the humeral head.12,22,34,35 We simplified the humeral joint area configuration to a circle. This means the point of measurement might not always have been located at the center of rotation. The measuring point was, however, equal in patients and controls to minimize the risk of any systematic differences. Such differences are more likely to occur if

Fig 7. Anteroposterior translation of humeral head centre during active abduction, mean, SE. The distortion of the mean at 010 and 6070 of abduction is partly due to missing observations.

any of the randomly located bone markers or the centers of these markers have been used. We note several limitations to our study. Computation and presentation of three-dimensional motions may be executed in different ways. The mathematical method mainly used in RSA (Euler angles) mimics the way chosen by most orthopaedic examiners to describe motions. We therefore think this method is clinically relevant even if there are alternatives. Another limitation is data only were collected from a single and standardized series of RSA examinations. Dynamic recordings during performance of different activities, including tasks requiring shoulder abduction above the horizontal plane would have been more relevant, but can at present not be performed using RSA, because of ethical and methodological reason. Our studies may, however, provide new and basic information about shoulder joint kinematics during a standardized motion. Recordings of the absolute motions are as accurate as recordings of relative motions, but they are more difficult to interpret. The absolute motions constantly refer to the cage coordinate system which is fixed and, in our study, had one axis parallel to the floor. This means the absolute rotations do not refer to a specific joint, but included the complete abduction of the humerus. This motion most probably constitutes a mixture of spine, thoracoscapular and glenohumeral movements. The glenohumeral joint is regarded as the most moveable joint in the body. There are few skeletal constraints, which enables a wide variability of motions. By asking our subjects to perform a number of trials before the final recording, we tried to obtain as many reproducible arcs of motion as possible. According to our tests of reproducibility, the differences between repeated abductions were surprisingly low. There was, however, considerable scatter of data in both groups. This scatter showed a relatively constant distribution around the mean with increasing abduction, except from flexion/extension motion in the patient group. The standard error of the mean for this parameter increased more than 50% from 25 to 55 of abduction. The significance of this finding is unclear, but might indicate patients with impingement try to find a position in flexion or extension of the shoulder that makes active abduction as easy and painless as possible. There has been a considerable divergence in the past concerning the relative roles of scapular rotation and glenohumeral movement in abduction of the arm.4,11,13 Inman et al13 stated during coronal abduction, scapular rotation and increasing abduction of the glenohumeral joint occur simultaneously throughout most of the movement. Between 30 and 170, the ratio of humeral to scapular rotation was reported to be about 2 to 1. During the first 30 of abduction, they found a great variability of the

26

Hallstrm and Krrholm

Clinical Orthopaedics and Related Research

scapular motions and reported translations in both the medial and lateral direction. Freedman6 reported that for every 3 of glenohumeral movement there were 2 of scapular rotation. This ratio was constant up to the final phase of abduction, when an increase corresponding to relatively more motion in the glenohumeral joint was found. Freedman and Munro6 reported mean total scapular rotation of 65 and mean total glenohumeral movement of 103. According to our data, the relative contribution of the glenohumeral joint to the global abduction of the humerus was smaller, even though we included any rotations of the trunk. The reasons for these discrepancies are unclear, but it seems likely the methods used to measure rotation and their resolutions are important. According to our findings, external humeral rotation is coupled to active abduction of the arm. This rotation could facilitate motion of the tuberosity under the acromion, and thereby avoid impingement of soft tissues. Our observations are not quite on line with those of MacGregor21 concerning internal humeral rotation at the end of the movement. This discrepancy could partly be because many of our patients did not reach 180 of absolute motions. Another important difference is MacGregor studied passive motions in fresh cadavers, which may not correspond to the situation in vivo. Meskers et al25 found the average radius of the humeral head is slightly larger than the glenoid. He assumed the cartilage was relatively deformable, which means in vivo the contact between the two articular areas would become closer. He, like Soslowsky et al,37 declared that at least in normal shoulders the glenohumeral joint behaves like a perfect ball and socket joint, which from a methodological point of view supports our use of the head center to represent humeral translations. Computation of the shortest distance between the humeral head and the acromion would add still more information. Such studies may become possible provided that in addition to the RSA examinations the shoulder joint is examined with computed tomography to map out the skeletal surface and at the same time is used to determine the position of the tantalum markers. Harryman et al9 used cadaver shoulders and found translational motions of the humeral head. These researchers applied pressure on the humerus and the effect of any muscle activation was not studied. Chen et al 2 studied 12 male volunteers (mean age, 27 years) without any shoulder symptoms using conventional radiography. They found essentially no change in the position of the humeral head as the arm was abducted from 0 to 130 (mean, 0.3 mm). After fatigue, the motion did, however, increase to an average of 2.5 mm, perhaps because of a more distal position before the motion was initiated.

We found a slight proximal displacement of about 1 mm during the early phase of abduction probably because of the effect of muscular activation. With the following abduction there was a slight distal displacement past 20, resulting in a position close to the one reported by Chen et al 2 at maximum abduction. In accord with Deutsch et al 3 and Paletta et al,30 we found the presence of cuff dysfunction implied a subtle but measurable increase of the proximal shift of the humeral head. This motion had occurred before 10 of abduction. Thereafter, the head center seemed to maintain a relatively fixed position when studied in the horizontal plane. The pathophysiologic background to the observed pattern of translation in patients with impingement is unclear. It could be caused by an unknown anatomical defect or variation or changed innervations of the muscles. Another and perhaps more probable explanation could be that these patients try to stabilize their shoulders during active abduction to reduce pain. Itoi et al14 used nine fresh frozen cadaver shoulders to simulate standard clinical tests of instability, including anterior and posterior translation and sulcus tests. In all tests the anterior and inferior humeral displacements were substantially restricted if they were performed with the humerus in internal but not in neutral and external rotation. Harryman et al9 postulated that humeral rotation caused asymmetric tightening of the capsule, which could result in translation of the humeral head. If the arm is internally rotated, the posterior capsule becomes tightened and pushes the head anteriorly. They called this the capsular constraint mechanism. We observed external rotation and would, according to this theory, have recorded posterior translation of the humeral head with increasing abduction. Instead, we found a minimum anterior shift, which places this theory into question. Patients with impingement syndrome showed only minor deviations of the glenohumeral kinematics compared with controls. The tendency to increase proximal shift of the humeral head, which occurred in the very early phase of activity, seemed to be followed by decreased humeral translations during the abduction. We think this is an effect of increased muscular activity probably caused by pain or meant to reduce pain during the motion. References
1. Brlin N, Thien T, Krrholm J. The precision of radiostereometric measurements. Manual vs. digital measurements. J Biomech. 2002; 35:6979. 2. Chen SK, Simonian PT, Wickiewicz TL, Otis JC, Warren RF. Radiographic evaluation of glenohumeral kinematics: a muscle fatigue model. J Shoulder Elbow Surg. 1999;8:4952. 3. Deutsch A, Altchek DW, Schwartz E, Otis JC, Warren RF. Radiologic measurement of superior displacement of the humeral head in the impingement syndrome. J Shoulder Elbow Surg. 1996;5: 186193.

Number 448 July 2006

Dynamic RSA of the Shoulder Joint

27

4. Duvall EN. Critical analysis of divergent views of movement at the shoulder joint. Arch Phys Med Rehabil. 1955;36:149154. 5. Flatow EL, Soslowsky LJ, Ticker JB, Pawluk RJ, Hepler M, Ark J, Mow VC, Bigliani LU. Excursion of the rotator cuff under the acromion. Patterns of subacromial contact. Am J Sports Med. 1994; 22:779788. 6. Freedman L, Munro RR. Abduction of the arm in the scapular plane: scapular and glenohumeral movements. A roentgenographic study. J Bone Joint Surg Am. 1966;48:15031510. 7. Fu FH, Harner CD, Klein AH. Shoulder impingement syndrome. A critical review. Clin Orthop Relat Res. 1991;269:162173. 8. Graichen H, Stammberger T, Bonel H, Haubner M, Englmeier KH, Reiser M, Eckstein F. Magnetic resonance-based motion analysis of the shoulder during elevation. Clin Orthop Relat Res. 2000;370: 154163. 9. Harryman DT 2nd, Sidles JA, Clark JM, McQuade KJ, Gibb TD, Matsen FA 3rd. Translation of the humeral head on the glenoid with passive glenohumeral motion. J Bone Joint Surg Am. 1990;72: 13341343. 10. Hgfors C, Peterson B, Sigholm G, Herberts P. Biomechanical model of the human shoulder joint--II. The shoulder rhythm. J Biomech. 1991;24:699709. 11. Hollinshead WH. Anatomy for Surgeons. First Ed. London: Cassell & Co Ltd, 1958. 12. Iannotti JP, Gabriel JP, Schneck SL, Evans BG, Misra S. The normal glenohumeral relationships. An anatomical study of one hundred and forty shoulders. J Bone Joint Surg Am. 1992;74:491500. 13. Inman VT, Saunders M, Abbott LC. Observations on the function of the shoulder joint. J Bone Joint Surg Am. 1944;16:130. 14. Itoi E, Motzkin NE, Morrey BF, An KN. Contribution of axial arm rotation to humeral head translation. Am J Sports Med. 1994;22: 499503. 15. Johnson GR, Stuart PR, Mitchell S. A method for the measurement of three dimensional scapular movement. Clin Biomechanic. 1993; 8:269273. 16. Kamkar A, Irrgang JJ, Whitney SL. Nonoperative management of secondary shoulder impingement syndrome. J Orthop Sports Phys Ther. 1993;17:212224. 17. Krrholm J, Brandsson S, Freeman MA. Tibiofemoral movement 4: changes of axial tibial rotation caused by forced rotation at the weight-bearing knee studied by RSA. J Bone Joint Surg Br. 2000; 82:12011203. 18. Kondo M. Changes of the tilting angle of the scapula following elevation of the arm. Nippon Seikeigeka Gakkai Zasshi. 1986;60: 175185. 19. Ludewig PM, Cook TM. Alterations in shoulder kinematics and associated muscle activity in people with symptoms of shoulder impingement. Phys Ther. 2000;80:276291. 20. Ludewig PM, Cook TM, Nawoczenski DA. Three-dimensional scapular orientation and muscle activity at selected positions of humeral elevation. J Orthop Sports Phys Ther. 1996;24:5765. 21. Macgregor L. Rotation of the shoulder; a critical enquiry. Br J Surg. 1936;24:425438. 22. Maki S, Gruen T. Anthropometric studies of the glenohumeral joint. Trans Orthop Res Soc. 1976;1:173.

23. Matsen FA, Arntz GT. The Shoulder. Philadelphia: WB Saunders, 1990. 24. McQuade KJ, Smidt GL. Dynamic scapulohumeral rhythm. the effects of external resistance during elevation of the arm in the scapular plane. J Orthop Sports Phys Ther. 1998;27:125133. 25. Meskers CGM, Vermeulen HM, de Groot JH, van Der Helm FC, Rozing PM. 3D shoulder position measurement using a six degree of freedom electromagnetic tracking device. Clin Biomech (Bristol, Avon). 1998;13:280292. 26. Neer CS 2nd. Anterior acromioplasty for the chronic impingement syndrome in the shoulder. a preliminary report. J Bone Joint Surg Am. 1972;54:4150. 27. Neer CS 2nd. Impingement lesions. Clin Orthop Relat Res. 1983; 173:7077. 28. Nilsson KG, Krrholm J, Ekelund L. Knee motion in total knee arthroplasty. A roentgen stereophotogrammetric analysis of the kinematics of the Tricon-M knee prosthesis. Clin Orthop Relat Res. 1990;256:147161. 29. Paine RM. Voight M. The role of the scapula. J Orthop Sports Phys Ther. 1993;18:386391. 30. Paletta GA Jr, 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. 1997;6:516527. 31. Poppen NK, Walker PS. Normal and abnormal motion of the shoulder. J Bone Joint Surg Am. 1976;58:195201. 32. Roach KE, Budimen-Mak E. Development of a shoulder pain and disability index. Arthritis Care Res. 1991;4:143149. 33. Saari T, Uvehammer J, Carlsson LV, Herberts P, Regner L, Karrholm J. Kinematics of three variations of the FreemanSamuelson total knee prosthesis. Clin Orthop Relat Res. 2003;410: 235247. 34. Saha AK. Theory of the shoulder mechanism: descriptive and applied. Springfield, IL: Thomas, 1961. 35. Saha AK. Dynamic stability of the glenohumeral joint. Acta Orthop Scand. 1971;42:491505. 36. Solem-Bertoft E, Thuomas KA, Westerberg CE. The influence of scapular retraction and protraction on the width of the subacromial space. An MRI study. Clin Orthop Relat Res. 1993;296:99103. 37. Soslowsky LJ, Flatow EL, Bigliani LU, Mow VC. Articular geometry of the glenohumeral joint. Clin Orthop Relat Res. 1992;285: 181190. 38. Tibone JE, Shaffer B. Shoulder pain: when is it impingement? J Musculo Med. 1995;12:6577. 39. Uvehammer J, Krrholm J, Brandsson S. In vivo kinematics of total knee arthroplasty. Concave versus posterior-stabilised tibial joint surface. J Bone Joint Surg Br. 2000;82:499505. 40. Warner JJ, Micheli JL. Scapulothoracic motion in normal shoulders and shoulders with gleonhumeral instability and impingement syndrome. A study using Moire topographic analysis. Clin Orthop Relat Res. 1992;285:191199. 41. Zuckerman JD, Kummer FJ, Cuomo F, Simon J, Rosenblum S, Katz N. The influence of the coracoacromial arch anatomy on rotatorcuff tears. J Shoulder Elbow Surg. 1992;1:414.