- Email: [email protected]
In vivo gleno-humeral joint loads during forward flexion and abduction
Journal of Biomechanics 44 (2011) 1543–1552
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
In vivo gleno-humeral joint loads during forward flexion and abduction G. Bergmann a,n, F. Graichen a, A. Bender a,b, A. Rohlmann a, A. Halder c, A. Beier c, P. Westerhoff a a b c
Julius Wolff Institut, Charite´—Universit¨ atsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin, Germany Berlin—Brandenburg Center for Regenerative Therapies, Berlin, Germany Klinik f¨ ur Endoprothetik Sommerfeld, Beetz-Sommerfeld, Germany
a r t i c l e i n f o
abstract
Article history: Accepted 25 February 2011
To improve design and preclinical test scenarios of shoulder joint implants as well as computer-based musculoskeletal models, a precise knowledge of realistic loads acting in vivo is necessary. Such data are also helpful to optimize physiotherapy after joint replacement and fractures. This is the first study that presents forces and moments measured in vivo in the gleno-humeral joint of 6 patients during forward flexion and abduction of the straight arm. The peak forces and, even more, the maximum moments varied inter-individually to a considerable extent. Forces of up to 238%BW (percent of body weight) and moments up to 1.74%BWm were determined. For elevation angles of less than 901 the forces agreed with many previous model-based calculations. At higher elevation angles, however, the measured loads still rose in contrast to the analytical results. When the exercises were performed at a higher speed, the peak forces decreased. The force directions relative to the humerus remained quite constant throughout the whole motion. Large moments in the joint indicate that friction in shoulder implants is high if the glenoid is not replaced. A friction coefficient of 0.1–0.2 seems to be realistic in these cases. & 2011 Elsevier Ltd. All rights reserved.
Keywords: Shoulder joint Load Force Moment In vivo Measurement Telemetry Gleno-humeral joint Abduction Forward flexion Elevation Friction
1. Introduction Previously muscle forces and contact loads in the glenohumeral joint were calculated using musculoskeletal models, resulting in widely differing results (Anglin et al., 2000; Buechel et al., 1978; Dul, 1988; Inman et al., 1996; Karlsson and Peterson, 1992; Kessel and Bayley, 1986; Poppen and Walker, 1978; Post et al., 1979; Runciman, 1993; Van der Helm, 1994; Van der Helm and Veeger, 1996). Uncertainties can, among others, be caused by the complex shoulder geometry and by large muscles numbers. Reliable knowledge about shoulder joint loads is essential to improve model predictions (Favre et al., 2009), for pre-clinical test of function, strength, fatigue and fixation of joint and fracture implants, for physiotherapy, and to advise patients. First data from a shoulder implant, measuring the spatial contact forces and moments (Westerhoff et al., 2009b), showed forces higher than 100% body weight (%BW) and high moments during some activities of daily living (Bergmann et al., 2007; Westerhoff et al., 2009a). Very high moments could indicate either high friction coefficients, an eccentric contact force, or additional forces from the surrounding structures. The moments are expected to counteract the momentary rotation in the joint. If
n
Corresponding author. Tel.: þ49 30 450 659081. E-mail address: [email protected] (G. Bergmann).
0021-9290/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2011.02.142
this were not the case, additional loads at the head must again be assumed. The functional outcome of shoulder arthroplasty varies more than for hip and knee joint replacements (Boileau et al., 2002). Furthermore the subjects, investigated now, differed considerably with regard to age and physical abilities. We therefore hypothesized that the loads individually vary much. The goal of this study was to measure in vivo the glenohumeral contact forces and moments in several subjects during abduction and forward flexion. Due to the standardized movements, we expected more uniform loads than during activities of daily living.
2. Methods 2.1. Instrumented implant The shoulder endoprosthesis measures the contact load between glenoid and humeral head. It is based on the BIOMODULAR implant (Biomet, Germany). Implant neck and stem are equipped with a 9-channel telemetry, 6 strain-gages and an inductive power supply (Westerhoff et al., 2009b). The inner electronics are connected to the antenna by a heart-pacemaker feedthrough. Extensive mechanical and electrical tests were performed to guarantee the patient’s safety. Customized hard- and software is used for measurements, pre-processing the signals, controlling the power supply and transferring the signals (Graichen et al., 2007). The loads are monitored in real time and stored with the subject’s video images for detailed analyses.
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
2.2. Subjects All 6 right-handed subjects (Table 1) received the implant due to osteoarthritis via a deltoid-pectoral approach. The rotator cuffs were functional enough for preserving the glenoids. At 27 months post-operatively (mpo) S2R complained about shoulder and neck pain; a rotator cuff tear was suspected by his orthopaedist. The Ethics Committee of our hospital approved implantation of the instrumented implants. Before surgery, the procedure was explained to the subjects, and they gave their written consent in regard to implantation, measurements, and publication of their images and videos. 2.3. Activities The following ways of performing an exercise (‘parameters’) were assumed to influence the contact loads: movement direction (abduction vs. forward flexion), speed, a weight in the hand and height of arm elevation.
Slow 0 kg Fast 0 kg
Data obtained in the implant’s coordinate system was transformed to a humerus based system (Wu et al., 2005) using CT data. This system is fixed at the centre of the implant head (Fig. 1) and moves with the arm. The þx-axis points anteriorly, parallel to the forearm axis at 901 elbow flexion; it is the rotation axis for abduction. Forward flexion is performed around the þ z-axis, which points in the lateral direction. The þ y-axis is parallel to the humeral axis. Loads measured in a left shoulder are transferred to the right joint. Measured are 3 force components Fx, Fy, Fz and 3 moment components Mx, My, Mz, acting at the humeral head (Fig. 1). The moments turn right around the positive axes. From these components, the resultant forces F and resultant moments M are calculated. The moments can be caused by friction only and are then dependent on the force magnitudes. However, they can additionally be caused by forces that act eccentrically or from outside of the glenoid. The moments are therefore expected to vary more than the forces. Higher moment than force variations are an indirect sign for eccentric movements between humeral head and glenoid. Friction occurs only in the plane of movement. During forward flexion we therefore expected only a moment Mz, and during abduction only Mx.
S1R
S2R
S3L
S4R
Slow 2 kg Fast 2 kg
1544
S5R
S8R
Fig. 1. Peak forces and moments. Forward flexion 901, without and with a 2-kg weight in the hand, slow and fast speed. Median peak forces Fpm (top) and peak moments Mpm (bottom) plus the individual ranges of Fp, Mp from 6 subjects. X ¼ no data. Inset: coordinate system of right joint.
Table 1 Subjects investigated and motion ranges: active ranges without brackets, passive ranges with brackets (not determined for internal rotation). Subject
S1R
S2R
S3L
S4R
S5R
S8R
Gender Age (years) Weight (kg) Height (m) Implant head size (mm) Replaced joint Post-operative time(s) (months)
Male 70 101 1.61 48 Right 9 and 12
Male 63 91 1.68 44 Right 0.9 and 27
Female 71 73 1.67 48 Left 5 and 23
Female 81 50 1.63 44 Right 0.9 and 15
Female 68 103 1.73 48 Right 6 and 22
Male 73 83 1.71 50 Right 3
Range of motion Elevation Abduction External rotation Internal rotation
1201(1401) 901(1101) 301 (401) 901
1601(1701) 1101(1201) 301 (401) 901
1501(1601) 1301(1501) 401 (501) 901
1101(1201) 901 (1001) 301 (351) 801
1101(1251) 901 (1001) 251(351) 801
1201(1301) 1201(1301) 401(451) 901
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
1545
The subjects always stood upright with the arms hanging down. They performed the one-sided exercises 5 times. Compensating movements like trunk rotation, abnormal scapula motion or elbow flexion could not always be prevented perfectly. Some patients could not perform strenuous tasks as fast abduction with a 2-kg weight. Arm positions were determined from the synchronous videos. Due to lacking equipment and experience to measure the scapula motion, and because of interferences between induction coil and EMG, no motion data or EMG were recorded. First measurements were taken when the subjects were able to perform the exercises; they were repeated (except for S8R) later on (2 times in Table 1). The data from the two time points were averaged together. During forward flexion and abduction the arm was elevated slowly to 901 and then lowered back at a non-prescribed but slow speed without holding it at 901. Some subjects did that also at a higher speed and to larger angles. These exercises were investigated with and without 2 kg in the hand. 2.3.1. Forward flexion All 6 subjects elevated the arm to 901 at a higher speed without weight and 5 of them with weight. Three subjects could perform forward flexion with and without weight to more than 901. 2.3.2. Abduction Slow and fast 901abduction without weight was performed by all subjects. With 2 kg, slow abduction to 901 was investigated in 5 and fast abduction in 4 subjects. Three subjects additionally could abduct the arm higher than 901 without weight. 2.4. Data evaluation and nomenclature Forces and moments are presented in %BW and %BWm. Conversion to N or Nm requires multiplication with 1% of the patient’s body weight in Newton (Table 1). This nomenclature was used: F, M Time-dependent resultant force and moment during one trial. Fp, Mp Peak values of F, M within one trial. Fpm, Mpm Median values of Fp, Mp from several trials of the same subject. Fpmx, Mpmx Arithmetic mean of Fpm, Mpm from 3 to 6 subjects (‘average’ subject). Median values and ranges are reported here for the single persons (Fpm and Mpm) as well as the values for the ‘average’ subject (Fpmx and Mpmx). Using data from the ‘average’ subject, the influences of the 3 parameters weight, speed and elevation angle on the loads were calculated, using the corresponding values from the ‘standard’ exercise (slow, no weight, 901) as reference values. Significances were only determined if the compared groups consisted of at least 4 subjects. Changes of Fpm, Mpm in one and the same subject but with different parameters were regarded as ‘distinct’ if their ranges did not overlap for the same subject. A ‘variation’ of forces or moments was defined as their range in percent of their median value.
3. Results 3.1. Forward flexion 3.1.1. Forward flexion 901, no weight, slow speed, 6 subjects (Fig. 1) The force Fpmx was 73%BW (range of Fpm: 55–87%BW). The moment Mpmx lay at 0.26%BWm and varied even more (0.16–0.41%BWm). 3.1.2. Forward flexion 901, no weight, fast speed, 6 subjects (Fig. 1) Fpmx decreased to 59%BW. Only in S3L did Fpm increase, but she unintentionally lifted the arm higher than 901 during fast motion. Mpmx was 0.24%BWm. The influence of higher speed on Mpm was not uniform. In 3 subjects Mpm fell, in 2 it stayed constant and in S3L it increased. 3.1.3. Forward flexion 901, 2-kg weight, slow speed, 5 subjects (Fig. 1) Fpmx rose to 122%BW due to the additional weight. This increase was distinct in all subjects because the ranges of Fpm for 0 kg and 2 kg never overlapped. Mpm also always increased to
Fig. 2. Moments during 901 forward flexion and abduction, 2-kg weight, slow speed. Typical single trials. Top ¼Moment Mz during forward flexion. Mz should be negative during upward movement and positive downwards. Bottom¼ Moment Mx during abduction. Mx should be positive during upward movement (inlay) and negative downwards. Symbols: start of downward movement.
Mpmx ¼0.49%BWm. The patterns of all moments (Fig. 2) varied individually to a great extent. The moment Mz was expected to be negative during the upward movement and positive downwards, but this was only observed in S1R and S8R (Fig. 2). The directions of high forces F in the frontal zy-plane were relatively constant throughout the whole motion cycle and similar between the subjects (Fig. 3, left frame). The average angle a was 31.71 (range 20–391). In the sagittal plane the angle b lay in the range of 26–351 in S1R, S5R, S8R, was only 111 in S2R, and totally different in S3L ( 131).
3.1.4. Forward flexion 901, 2-kg weight, fast speed, 5 subjects (Figs. 1 and 4) As observed without the 2-kg weight, the higher speed let Fpm decrease in all subjects. Fpmx fell to 101%BW. The moment Mpmx declined slightly to 0.48%BWm. This change was not uniform, however; in S3L Mpm even increased. The time courses of F during elevation were very different for slow and fast speeds (S1R, S2R in Fig. 4). At slow speed, F continuously increased up to flexion angles of 901. At high speed the maximum was already reached shortly before 451. Then F fell until the highest arm position was reached. It must be emphasized that the arm was not held in this position, but immediately lowered again.
1546
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
Fig. 3. Force directions. Forward flexion and abduction 901, 2-kg weight, slow speed. Force vectors and maximum force during one loading cycle. Forces below 20% of Fp are not displayed. Typical examples from 5 subjects. Magnitude and orientation of peak forces are indicated. Load directions a and b are averaged for all cycles of single subjects. Note the different scales! Left frame ¼forward flexion, right frame ¼ abduction. Left columns in frames ¼ frontal plane zy, right columns¼ sagittal plane xy.
3.1.5. Forward flexion4901, no weight, slow speed, 3 subjects (Fig. 5) The horizontal lever arm between the arms’ centre of gravity and the joint centre (analogous to the insert in Fig. 4) decreases for elevation angles larger than 901. Therefore, one would expect F to fall, but it increased further in all 3 subjects (Fig. 5). Peak forces Fp always acted at the highest arm position. They were 70%, 121% and 88%BW in S2R, S3L and S8R, respectively, compared to 62%, 77% and 82%BW when moving the arm upwards through the 901 position and only 52%, 58% and 56%BW at 901 when moving it downwards through 901. Again the directions of high forces relative to the humerus remained nearly constant throughout the whole motion cycle. The variation of a was only 5–71 for forces higher than 80% of the peak forces. As under other conditions (Section 3.1.3, Fig. 2) the time patterns of the moments differed considerably between the
subjects (Fig. 5). The expected negative/positive signs of Mz during upward/downward movement of the arm were only observed in S3L and S8R.
3.1.6. Forward flexion4901, 2-kg weight, slow or fast speed, 3 subjects (Fig. 6) Typical examples of load components and force directions are shown in Fig. 6. Fp was 123%, 238% and 212%BW in S2R, S3L and S8R, respectively. This was 76%, 97% and 141% more than that without weight. In the case of fast upward motion of S2R and S8R, both subjects reached their force maximum before the highest elevation angle, similar as during fast forward flexion up to 901 (Fig. 2). S3L carried out the exercise slowly and her force–time curve increased continuously up to the 901 position. She reached the absolutely highest forces and moments (1.74%BWm) among
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
1547
3.2.5. Abduction4901, no weight, slow speed, 3 subjects (Fig. 8) Similar to forward flexion, F still increased above 901 abduction. Fp reached 88, 121 and 88%BW for S2R, S3L and S8R in the examples of Fig. 8. Intra-individually the time courses and force directions were similar as during forward flexion. Resultant moments had values of 0.21%BWm, 0.89%BWm and 0.30%BWm. The expected positive/negative signs of Mx when lifting/lowering the arm were observed in S2R and S3L, but not in S8R. 3.2.6. Postsurgical course For most subjects, no statements can be made yet about the postsurgical course due to the short observation time. In subject S2R the forces decreased slightly by about 20% between 4 and 7mpo, but this was not a distinct change. The moments varied considerably: 0.9 and 27mpo they were 3 times higher than 4mpo. Fig. 4. Resultant forces. Forward flexion and abduction 901, 2-kg weight, slow and fast speed. Typical single trials. Values in symbols ¼ elevation angles. During slow movement (oval symbols) F increases steadily up to 901. During fast movement (rectangular symbols) F reaches its smaller peak values already before 451 and then falls up to 901. Selected examples of subjects S1R, S2R (forward flexion) and S5R (abduction). Inlay shows lever arm l of the centre of gravity (COG), depending on the lifting angle a.
all exercises. Again, the expectations about the signs of Mz were not met.
3.2. Abduction 3.2.1. Abduction 901, no weight, slow speed, 6 subjects (Fig. 7) Fpmx was 81%BW, which is slightly higher than that during forward flexion (73%BW). However, abduction caused less uniform forces (compare Figs. 1 and 7); the individual range of Fpm (46–115%BW) was much larger. Mpmx lay at 0.44%BWm. The individual variation of Mpm (0.07–1.18%BWm) was larger than that of the forces.
3.2.2. Abduction 901, no weight, fast speed, 6 subjects (Fig. 7) Fpmx was reduced to 65%BW by the faster motion; but the changes of Fpm were not very uniform. Fpm decreased by 41% in S4R but increased slightly in S2R, who did the task extremely fast and ambitiously. Mpmx fell to 0.36%BWm, which is lower than at slow speed. Again the individual differences of Mpm were large. As for Fpm, S4R showed the largest reduction.
3.2.3. Abduction 901, 2-kg weight, slow speed, 5 subjects (Fig. 7) Fpmx increased to 129%BW and Mpmx to 0.69%BWm due to the additional weight. The average forces were nearly the same as during forward flexion. Again, the force directions a in the frontal plane (Fig. 3, right frame) were very constant throughout the whole motion (average 28.61) and similar to those during forward flexion. This was also true for the sagittal plane, except for S3L who showed a different behaviour between forward flexion and abduction. As expected, Mx was positive during upwards movement of the arm and negative downwards in all subjects (Fig. 2, bottom). In S3L the peak value of Mx was several times higher than in all other subjects.
3.2.4. Abduction 901, 2-kg weight, fast speed, 4 subjects (Fig. 7) The faster speed let Fpmx decreased to 101%BW. The moments Mpmx fell to 0.61%BWm. All subjects reported that the fast task was easier to perform.
3.3. Influence of movement direction, speed, weight and elevation angle (Fig. 9) 3.3.1. Forces Fig. 9 shows the load changes due to variations of the 4 parameters. Load changes in percent and the significance of load differences were always calculated from identical subjects in both groups. No significant differences in the peak forces Fpmx were found between standard forward flexion and abduction (bold ‘‘0’’ in Fig. 9). Increasing the speed lowered Fpmx by 18–22%. The decrease was independent of the weight carried, and significant except for abduction without a weight. Carrying 2 kg increased Fpmx by 51–75% at slow and fast speed and for different elevation angles. These changes were always significant for 901 elevation. Elevating the arm higher than 901 let Fpmx rise by 21–40%. The investigated number of only 3 subjects prohibited any calculation of significances. 3.3.2. Moments No significant differences of Mpmx were found between standard forward flexion and abduction. The impact of the 4 parameters on the moments Mpmx was less pronounced than it was on the forces. Only the additional weight caused a significant moment rise by 40–53%, both during slow and fast movement.
4. Discussion This study is the first one reporting shoulder joint loads measured in vivo during abduction and forward flexion for a group of patients. The anatomic changes caused by the prosthesis are probably smaller than if the glenoid had additionally been replaced. Nevertheless the data cannot directly be transferred to healthy subjects, because soft tissue alterations and other factors may affect the loading. Several important observations were made in the majority of measurements and subjects; due to restricted cohort sizes not all of them could be proven statistically, however. 1. An additional weight of 2 kg when lifting the arm increases the force by 51–75%. 2. Fast arm lifting without stopping in the highest position causes about 20% lower forces than slow lifting. 3. The directions of high forces relative to the humeral head are very constant in most subjects and during different exercises. 4. In 3 of the subjects, the forces are distinctly higher during standard abduction than they were for forward flexion. In the
1548
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
Fig. 5. Forces and moments. Forward flexion 4901, no weight, slow speed. Three subjects. Upper diagrams ¼ resultant force F and components in % of body weight (%BW). Lower diagrams ¼resultant moment M and components in %BWm. Elevation angles are indicated. Vector plots ¼ force directions in frontal and sagittal plane during whole motion. Subject S2R holds a roll of paper in his hand.
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
1549
Fig. 6. Forces and moments. Forward flexion4901, 2-kg weight, slow or fast speed. Three subjects. Fast speed for S2R and S8R, slow speed for S3L. For more information see Fig. 5.
combined data from all subjects no statistical difference could be found, however. 5. Moments in the joint vary more than the forces and their high magnitudes indicate that the implant head is often loaded eccentrically or from outside of the glenoid.
During activities of daily living, the highest joint forces were determined when steering a car with one hand, setting down 1.5 kg on a table with a straight arm, and laying down 2 kg in a high shelf (Westerhoff et al., 2009a). The forces now observed at high arm elevations with additional weight were in the same range.
S2R
S3L
S4R
Slow 2 kg Fast 2 kg
Slow 0 kg Fast 0 kg
S1R
Slow 2 kg Fast 2 kg
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
Slow 0 kg Fast 0 kg
1550
S5R
S8R
Fig. 7. Peak forces and moments. Abduction 901, without and with 2-kg weight, slow and fast speed. Median peak forces Fpm (top) and peak moments Mpm (bottom) plus ranges of Fp, Mp from 6 subjects. X ¼no data.
Anglin et al. (2000) listed the forces predicted by various musculo-skeletal models. For 901 abduction without additional weight and 75 kg BW, forces of 50–90%BW were calculated; values as measured in 4 of our subjects. The higher forces in the other 2 subjects may possibly be explained by antagonistic muscle activities that are difficult to simulate exactly. Our observation that forces always increased when the arm was lifted higher than 901 was predicted only very recently (Favre et al., 2009). In other models (Poppen and Walker, 1978; Terrier et al., 2008; van der Helm, 1994) the force then decreased. The effect is presumably due to co-contractions, stabilizing the arm position (Steenbrink et al., 2009). High joint loads during overhead work may explain shoulder problems of certain professional groups (Herberts et al., 1981). Moments in the shoulder joint varied even more than did the forces, and were remarkably high. In S3L they were up to 4 times higher than in the other subjects (Figs. 1, 2 and 7). Assuming a constant friction coefficient of 0.1 in the joint and a force of Fp ¼120%BW, a purely friction-induced moment M, acting at a 48 mm head, would be 0.29%BWm. This is the range found in S2R, S4R and S5R. For S1R and S8R a coefficient of even 0.2 must be assumed. It should be tested whether an inclusion of friction in musculo-skeletal models could in any way improve the predictions. Extreme moments, as in S3L, cannot be caused by friction alone, unless one would assume a coefficient of up to 0.5. Such moments are most probably explained by additional forces, transferred outside of the glenoid and the labrum glenoidale, which do not pass the head centre. Such forces could possibly be exerted by the acromion. The observations that the force directions in S3L deviate much from those of the other subjects during forward flexion (Fig. 3) and that the force magnitudes are
highest in S3L when lifting 2 kg (Figs. 1 and 7), support this assumption. The forces were always about 20% lower when elevating the arm quickly (Fig. 4). When the arm is in a low position, the horizontal lever arm l of its COG is short (Fig. 4), and the moment required to accelerate the arm upwards is small. It is therefore energy-saving to accelerate the arm early during the upwards motion, let it swing up nearly passively to the highest position, and finally have it ‘falling down’ (Fig. 4; S2R and S8R in Fig. 6). Obviously all subjects used this tactic. As observed previously for the hip and knee joint (Bergmann et al., 2001; Heinlein et al., 2009) the force directions remained quite constant throughout all activities (Fig. 3). Even between abduction and forward flexion, similarities were seen in the same subject. This indicates that either the force on the glenoid is very variable and/or the scapula motion is adjusted constantly to keep the resultant force direction inside the glenoid. The reported forces and moments are essential to improve shoulder prostheses design, and proximal humeral fracture implants. The data are not representative for the loading of inverse implants, however. Highest peak forces and moments from all subjects during slow motion with 2 kg (Figs. 3 and 7) represent the most severe test conditions, simulating forward flexion/abduction. In the frontal plane the forces can be applied at the reported average angles a (3.1.3/3.2.3). In the sagittal plane, that extreme force angle b (Fig. 3) should be chosen as the angle causing the most critical mechanical condition for the implant and its fixation. Because no generally dominant axis for the largest moments exists, that torque axis should be chosen which results in the most severe mechanical situation. Less demanding tests could be performed under the same conditions but with those decreased force magnitudes which simulate exercises
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
1551
Fig. 8. Forces and moments. Abduction 4901, no weight, slow speed. Three subjects. For more information see Fig. 5.
without weight. These proposed test setups certainly require further discussion. Even more important than the humeral load will be that of the scapula, because glenoid replacement fixation is still unsatisfactory. Synchronous motion capture and load measurements will
deliver the force directions relative to the scapula and the movement between humerus and scapula. This data will also be used as ‘gold standard’ to improve the analytical methods, as previously done with hip load data (Bergmann, 2001; Heller et al., 2001; Stansfield et al., 2003).
1552
Fpmx
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
Mpmx
Subj.
+68%* +92%° fast +57%* +49% *
Fpmx
Mpmx
5
-18%*
-2%°
4
-22%*
-15%*
Subj.
Fast
5 4
Fpmx
+75%* +53%*
Subj.
-19%*
-8%°
6
-20%°
-18%°
6
>90°
5 5
Fpmx
Mpmx
Subj.
Fpmx
Mpmx
+40%n
+151%n
3
+51%n
+19%n
3
---
---
---
---
---
---
0
+2kg
+2 kg Mpmx
Subj.
+69%* +75%°
0
Fpmx
Mpmx
Subj.
+2 kg
Standard Exercise
Fast
Fpmx
Mpmx
73%BW
0.26%BWm
6
81%BW
0.44%BWm
6
0 Type of movement:
Subj.
>90°
Mpmx Subj. n +21% +85%n 3 Fpmx
+23%n
-8%n
3
0 Forward Flexion
Abduction
Fig. 9. Load changes by movement direction, speed, weight and elevation angle. Grey or green boxes: numbers for forward flexion. White or blue boxes: numbers for abduction. ‘‘Standard’’: slow exercise up to 901without a weight. ‘‘Subj.’’: number of investigated subjects. Parameters: ‘‘Fast’’: activity performed at faster speed. ‘‘2 kg’’: activity with 2 kg held in the hand. ‘‘4 901’’: elevation higher than 901. Significances: ‘‘*’’: significant (Student’s t-test p o0.05). ‘‘0’’: not significant. ‘‘n’’: significance test not applicable.
Conflict of interest statement All authors have no conflict of interest that could influence the work presented here.
Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft (BE 804/17-1 and SFB 760-C-6), Berlin Brandenburg Center of Regenerative Therapies, Synthes GmbH Switzerland, and Biomet Merck Deutschland GmbH. We thank all patients for their engaged co-operation! References Anglin, C., Wyss, U.P., Pichora, D.R., 2000. Glenohumeral contact forces. Proc. Inst. Mech. Eng. [H] 214, 637–644. ¨ Berlin. Bergmann, G., 2001. Hip98, Loading of the Hip Joint. Freie Universitat, Bergmann, G., Deuretzbacher, G., Heller, M., Graichen, F., Rohlmann, A., Strauss, J., Duda, G.N., 2001. Hip contact forces and gait patterns from routine activities. J. Biomech. 34, 859–871. Bergmann, G., Graichen, F., Bender, A., Kaab, M., Rohlmann, A., Westerhoff, P., 2007. In vivo glenohumeral contact forces—measurements in the first patient 7 months postoperatively. J. Biomech. 40, 2139–2149. Boileau, P., Krishnan, S.G., Tinsi, L., Walch, G., Coste, J.S., Mole, D., 2002. Tuberosity malposition and migration: reasons for poor outcomes after hemiarthroplasty for displaced fractures of the proximal humerus. J. Shoulder Elbow. Surg. 11, 401–412. Buechel, F.F., Pappas, M.J., DePalma, A.F., 1978. ‘Floating socket’ total shoulder replacement: anatomical, biomechanical and surgical rationale. J. Biomed. Mater. Res. 12, 89–114. Dul, J., 1988. A biomechanical model to quantify shoulder load at the workplace. Clin. Biomech. 3, 124–128. Favre, P., Snedeker, J.G., Gerber, C., 2009. Numerical modelling of the shoulder for clinical applications. Philos. Trans. R. Soc. A: Math. Phys. Eng. Sci. 367, 2095–2118. Graichen, F., Arnold, R., Rohlmann, A., Bergmann, G., 2007. Implantable 9-channel telemetry system for in vivo load measurements with orthopedic implants. IEEE Trans. Biomed. Eng. 54, 253–261. Heinlein, B., Kutzner, I., Graichen, F., Bender, A., Rohlmann, A., Halder, A.M., Beier, A., Bergmann, G., 2009. ESB Clinical Biomechanics Award 2008: complete data
of total knee replacement loading for level walking and stair climbing measured in vivo with a follow-up of 6–10 months. Clin. Biomech. (Bristol, Avon.) 24, 315–326. Heller, M.O., Bergmann, G., Deuretzbacher, G., Durselen, L., Pohl, M., Claes, L., Haas, N.P., Duda, G.N., 2001. Musculo-skeletal loading conditions at the hip during walking and stair climbing. J. Biomech. 34, 883–893. Herberts, P., Kadefors, R., Andersson, G., Petersen, I., 1981. Shoulder pain in industry: an epidemiological study on welders. Acta Orthop. Scand. 52, 299–306. Inman, V.T., Saunders, J.B., Abbott, L.C., 1996. Observations of the function of the shoulder joint. 1944. Clin. Orthop. Relat. Res., 3–12. Karlsson, D., Peterson, B., 1992. Towards a model for force predictions in the human shoulder. J Biomech. 25, 189–199. Kessel, L., Bayley, I., 1986. Clinical Disorders of the Shoulder, 2nd ed. Churchill Livingstone, Edinburgh. Poppen, N.K., Walker, P.S., 1978. Forces at the glenohumeral joint in abduction. Clin. Orthop. Relat. Res., 165–170. Post, M., Jablon, M., Miller, H., Singh, M., 1979. Constrained total shoulder joint replacement: a critical review. Clin. Orthop. Relat. Res., 135–150. Runciman, R.J., 1993. Biomechanical model of the shoulder joint. Bioengineering Unit, Vol. Ph.D. Thesis, University of Strathclyde, Glasgow. Stansfield, B.W., Nicol, A.C., Paul, J.P., Kelly, I.G., Graichen, F., Bergmann, G., 2003. Direct comparison of calculated hip joint contact forces with those measured using instrumented implants. An evaluation of a three-dimensional mathematical model of the lower limb. J. Biomech. 36, 929–936. Steenbrink, F., de Groot, J.H., Veeger, H.E., van der Helm, F.C., Rozing, P.M., 2009. Glenohumeral stability in simulated rotator cuff tears. J. Biomech. 42, 1740–1745. Terrier, A., Vogel, A., Capezzali, M., Farron, A., 2008. An algorithm to allow humerus translation in the indeterminate problem of shoulder abduction. Med. Eng. Phys. 30, 710–716. van der Helm, F.C., 1994. Analysis of the kinematic and dynamic behavior of the shoulder mechanism. J. Biomech. 27, 527–550. van der Helm, F.C., Veeger, H.E., 1996. Quasi-static analysis of muscle forces in the shoulder mechanism during wheelchair propulsion. J. Biomech. 29, 39–52. Westerhoff, P., Graichen, F., Bender, A., Halder, A., Beier, A., Rohlmann, A., Bergmann, G., 2009a. In vivo measurement of shoulder joint loads during activities of daily living. J. Biomech. 42, 1840–1849. Westerhoff, P., Graichen, F., Bender, A., Rohlmann, A., Bergmann, G., 2009b. An instrumented implant for in vivo measurement of contact forces and contact moments in the shoulder joint. Med. Eng. Phys. 31, 207. Wu, G., van der Helm, F.C., Veeger, H.E., Makhsous, M., Van Roy, P., Anglin, C., Nagels, J., Karduna, A.R., McQuade, K., Wang, X., Werner, F.W., Buchholz, B., 2005. ISB recommendation on definitions of joint coordinate systems of various joints for the reporting of human joint motion—Part II: Shoulder, elbow, wrist and hand. J. Biomech. 38, 981–992.
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
In vivo gleno-humeral joint loads during forward flexion and abduction G. Bergmann a,n, F. Graichen a, A. Bender a,b, A. Rohlmann a, A. Halder c, A. Beier c, P. Westerhoff a a b c
Julius Wolff Institut, Charite´—Universit¨ atsmedizin Berlin, Augustenburger Platz 1, 13353 Berlin, Germany Berlin—Brandenburg Center for Regenerative Therapies, Berlin, Germany Klinik f¨ ur Endoprothetik Sommerfeld, Beetz-Sommerfeld, Germany
a r t i c l e i n f o
abstract
Article history: Accepted 25 February 2011
To improve design and preclinical test scenarios of shoulder joint implants as well as computer-based musculoskeletal models, a precise knowledge of realistic loads acting in vivo is necessary. Such data are also helpful to optimize physiotherapy after joint replacement and fractures. This is the first study that presents forces and moments measured in vivo in the gleno-humeral joint of 6 patients during forward flexion and abduction of the straight arm. The peak forces and, even more, the maximum moments varied inter-individually to a considerable extent. Forces of up to 238%BW (percent of body weight) and moments up to 1.74%BWm were determined. For elevation angles of less than 901 the forces agreed with many previous model-based calculations. At higher elevation angles, however, the measured loads still rose in contrast to the analytical results. When the exercises were performed at a higher speed, the peak forces decreased. The force directions relative to the humerus remained quite constant throughout the whole motion. Large moments in the joint indicate that friction in shoulder implants is high if the glenoid is not replaced. A friction coefficient of 0.1–0.2 seems to be realistic in these cases. & 2011 Elsevier Ltd. All rights reserved.
Keywords: Shoulder joint Load Force Moment In vivo Measurement Telemetry Gleno-humeral joint Abduction Forward flexion Elevation Friction
1. Introduction Previously muscle forces and contact loads in the glenohumeral joint were calculated using musculoskeletal models, resulting in widely differing results (Anglin et al., 2000; Buechel et al., 1978; Dul, 1988; Inman et al., 1996; Karlsson and Peterson, 1992; Kessel and Bayley, 1986; Poppen and Walker, 1978; Post et al., 1979; Runciman, 1993; Van der Helm, 1994; Van der Helm and Veeger, 1996). Uncertainties can, among others, be caused by the complex shoulder geometry and by large muscles numbers. Reliable knowledge about shoulder joint loads is essential to improve model predictions (Favre et al., 2009), for pre-clinical test of function, strength, fatigue and fixation of joint and fracture implants, for physiotherapy, and to advise patients. First data from a shoulder implant, measuring the spatial contact forces and moments (Westerhoff et al., 2009b), showed forces higher than 100% body weight (%BW) and high moments during some activities of daily living (Bergmann et al., 2007; Westerhoff et al., 2009a). Very high moments could indicate either high friction coefficients, an eccentric contact force, or additional forces from the surrounding structures. The moments are expected to counteract the momentary rotation in the joint. If
n
Corresponding author. Tel.: þ49 30 450 659081. E-mail address: [email protected] (G. Bergmann).
0021-9290/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jbiomech.2011.02.142
this were not the case, additional loads at the head must again be assumed. The functional outcome of shoulder arthroplasty varies more than for hip and knee joint replacements (Boileau et al., 2002). Furthermore the subjects, investigated now, differed considerably with regard to age and physical abilities. We therefore hypothesized that the loads individually vary much. The goal of this study was to measure in vivo the glenohumeral contact forces and moments in several subjects during abduction and forward flexion. Due to the standardized movements, we expected more uniform loads than during activities of daily living.
2. Methods 2.1. Instrumented implant The shoulder endoprosthesis measures the contact load between glenoid and humeral head. It is based on the BIOMODULAR implant (Biomet, Germany). Implant neck and stem are equipped with a 9-channel telemetry, 6 strain-gages and an inductive power supply (Westerhoff et al., 2009b). The inner electronics are connected to the antenna by a heart-pacemaker feedthrough. Extensive mechanical and electrical tests were performed to guarantee the patient’s safety. Customized hard- and software is used for measurements, pre-processing the signals, controlling the power supply and transferring the signals (Graichen et al., 2007). The loads are monitored in real time and stored with the subject’s video images for detailed analyses.
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
2.2. Subjects All 6 right-handed subjects (Table 1) received the implant due to osteoarthritis via a deltoid-pectoral approach. The rotator cuffs were functional enough for preserving the glenoids. At 27 months post-operatively (mpo) S2R complained about shoulder and neck pain; a rotator cuff tear was suspected by his orthopaedist. The Ethics Committee of our hospital approved implantation of the instrumented implants. Before surgery, the procedure was explained to the subjects, and they gave their written consent in regard to implantation, measurements, and publication of their images and videos. 2.3. Activities The following ways of performing an exercise (‘parameters’) were assumed to influence the contact loads: movement direction (abduction vs. forward flexion), speed, a weight in the hand and height of arm elevation.
Slow 0 kg Fast 0 kg
Data obtained in the implant’s coordinate system was transformed to a humerus based system (Wu et al., 2005) using CT data. This system is fixed at the centre of the implant head (Fig. 1) and moves with the arm. The þx-axis points anteriorly, parallel to the forearm axis at 901 elbow flexion; it is the rotation axis for abduction. Forward flexion is performed around the þ z-axis, which points in the lateral direction. The þ y-axis is parallel to the humeral axis. Loads measured in a left shoulder are transferred to the right joint. Measured are 3 force components Fx, Fy, Fz and 3 moment components Mx, My, Mz, acting at the humeral head (Fig. 1). The moments turn right around the positive axes. From these components, the resultant forces F and resultant moments M are calculated. The moments can be caused by friction only and are then dependent on the force magnitudes. However, they can additionally be caused by forces that act eccentrically or from outside of the glenoid. The moments are therefore expected to vary more than the forces. Higher moment than force variations are an indirect sign for eccentric movements between humeral head and glenoid. Friction occurs only in the plane of movement. During forward flexion we therefore expected only a moment Mz, and during abduction only Mx.
S1R
S2R
S3L
S4R
Slow 2 kg Fast 2 kg
1544
S5R
S8R
Fig. 1. Peak forces and moments. Forward flexion 901, without and with a 2-kg weight in the hand, slow and fast speed. Median peak forces Fpm (top) and peak moments Mpm (bottom) plus the individual ranges of Fp, Mp from 6 subjects. X ¼ no data. Inset: coordinate system of right joint.
Table 1 Subjects investigated and motion ranges: active ranges without brackets, passive ranges with brackets (not determined for internal rotation). Subject
S1R
S2R
S3L
S4R
S5R
S8R
Gender Age (years) Weight (kg) Height (m) Implant head size (mm) Replaced joint Post-operative time(s) (months)
Male 70 101 1.61 48 Right 9 and 12
Male 63 91 1.68 44 Right 0.9 and 27
Female 71 73 1.67 48 Left 5 and 23
Female 81 50 1.63 44 Right 0.9 and 15
Female 68 103 1.73 48 Right 6 and 22
Male 73 83 1.71 50 Right 3
Range of motion Elevation Abduction External rotation Internal rotation
1201(1401) 901(1101) 301 (401) 901
1601(1701) 1101(1201) 301 (401) 901
1501(1601) 1301(1501) 401 (501) 901
1101(1201) 901 (1001) 301 (351) 801
1101(1251) 901 (1001) 251(351) 801
1201(1301) 1201(1301) 401(451) 901
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
1545
The subjects always stood upright with the arms hanging down. They performed the one-sided exercises 5 times. Compensating movements like trunk rotation, abnormal scapula motion or elbow flexion could not always be prevented perfectly. Some patients could not perform strenuous tasks as fast abduction with a 2-kg weight. Arm positions were determined from the synchronous videos. Due to lacking equipment and experience to measure the scapula motion, and because of interferences between induction coil and EMG, no motion data or EMG were recorded. First measurements were taken when the subjects were able to perform the exercises; they were repeated (except for S8R) later on (2 times in Table 1). The data from the two time points were averaged together. During forward flexion and abduction the arm was elevated slowly to 901 and then lowered back at a non-prescribed but slow speed without holding it at 901. Some subjects did that also at a higher speed and to larger angles. These exercises were investigated with and without 2 kg in the hand. 2.3.1. Forward flexion All 6 subjects elevated the arm to 901 at a higher speed without weight and 5 of them with weight. Three subjects could perform forward flexion with and without weight to more than 901. 2.3.2. Abduction Slow and fast 901abduction without weight was performed by all subjects. With 2 kg, slow abduction to 901 was investigated in 5 and fast abduction in 4 subjects. Three subjects additionally could abduct the arm higher than 901 without weight. 2.4. Data evaluation and nomenclature Forces and moments are presented in %BW and %BWm. Conversion to N or Nm requires multiplication with 1% of the patient’s body weight in Newton (Table 1). This nomenclature was used: F, M Time-dependent resultant force and moment during one trial. Fp, Mp Peak values of F, M within one trial. Fpm, Mpm Median values of Fp, Mp from several trials of the same subject. Fpmx, Mpmx Arithmetic mean of Fpm, Mpm from 3 to 6 subjects (‘average’ subject). Median values and ranges are reported here for the single persons (Fpm and Mpm) as well as the values for the ‘average’ subject (Fpmx and Mpmx). Using data from the ‘average’ subject, the influences of the 3 parameters weight, speed and elevation angle on the loads were calculated, using the corresponding values from the ‘standard’ exercise (slow, no weight, 901) as reference values. Significances were only determined if the compared groups consisted of at least 4 subjects. Changes of Fpm, Mpm in one and the same subject but with different parameters were regarded as ‘distinct’ if their ranges did not overlap for the same subject. A ‘variation’ of forces or moments was defined as their range in percent of their median value.
3. Results 3.1. Forward flexion 3.1.1. Forward flexion 901, no weight, slow speed, 6 subjects (Fig. 1) The force Fpmx was 73%BW (range of Fpm: 55–87%BW). The moment Mpmx lay at 0.26%BWm and varied even more (0.16–0.41%BWm). 3.1.2. Forward flexion 901, no weight, fast speed, 6 subjects (Fig. 1) Fpmx decreased to 59%BW. Only in S3L did Fpm increase, but she unintentionally lifted the arm higher than 901 during fast motion. Mpmx was 0.24%BWm. The influence of higher speed on Mpm was not uniform. In 3 subjects Mpm fell, in 2 it stayed constant and in S3L it increased. 3.1.3. Forward flexion 901, 2-kg weight, slow speed, 5 subjects (Fig. 1) Fpmx rose to 122%BW due to the additional weight. This increase was distinct in all subjects because the ranges of Fpm for 0 kg and 2 kg never overlapped. Mpm also always increased to
Fig. 2. Moments during 901 forward flexion and abduction, 2-kg weight, slow speed. Typical single trials. Top ¼Moment Mz during forward flexion. Mz should be negative during upward movement and positive downwards. Bottom¼ Moment Mx during abduction. Mx should be positive during upward movement (inlay) and negative downwards. Symbols: start of downward movement.
Mpmx ¼0.49%BWm. The patterns of all moments (Fig. 2) varied individually to a great extent. The moment Mz was expected to be negative during the upward movement and positive downwards, but this was only observed in S1R and S8R (Fig. 2). The directions of high forces F in the frontal zy-plane were relatively constant throughout the whole motion cycle and similar between the subjects (Fig. 3, left frame). The average angle a was 31.71 (range 20–391). In the sagittal plane the angle b lay in the range of 26–351 in S1R, S5R, S8R, was only 111 in S2R, and totally different in S3L ( 131).
3.1.4. Forward flexion 901, 2-kg weight, fast speed, 5 subjects (Figs. 1 and 4) As observed without the 2-kg weight, the higher speed let Fpm decrease in all subjects. Fpmx fell to 101%BW. The moment Mpmx declined slightly to 0.48%BWm. This change was not uniform, however; in S3L Mpm even increased. The time courses of F during elevation were very different for slow and fast speeds (S1R, S2R in Fig. 4). At slow speed, F continuously increased up to flexion angles of 901. At high speed the maximum was already reached shortly before 451. Then F fell until the highest arm position was reached. It must be emphasized that the arm was not held in this position, but immediately lowered again.
1546
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
Fig. 3. Force directions. Forward flexion and abduction 901, 2-kg weight, slow speed. Force vectors and maximum force during one loading cycle. Forces below 20% of Fp are not displayed. Typical examples from 5 subjects. Magnitude and orientation of peak forces are indicated. Load directions a and b are averaged for all cycles of single subjects. Note the different scales! Left frame ¼forward flexion, right frame ¼ abduction. Left columns in frames ¼ frontal plane zy, right columns¼ sagittal plane xy.
3.1.5. Forward flexion4901, no weight, slow speed, 3 subjects (Fig. 5) The horizontal lever arm between the arms’ centre of gravity and the joint centre (analogous to the insert in Fig. 4) decreases for elevation angles larger than 901. Therefore, one would expect F to fall, but it increased further in all 3 subjects (Fig. 5). Peak forces Fp always acted at the highest arm position. They were 70%, 121% and 88%BW in S2R, S3L and S8R, respectively, compared to 62%, 77% and 82%BW when moving the arm upwards through the 901 position and only 52%, 58% and 56%BW at 901 when moving it downwards through 901. Again the directions of high forces relative to the humerus remained nearly constant throughout the whole motion cycle. The variation of a was only 5–71 for forces higher than 80% of the peak forces. As under other conditions (Section 3.1.3, Fig. 2) the time patterns of the moments differed considerably between the
subjects (Fig. 5). The expected negative/positive signs of Mz during upward/downward movement of the arm were only observed in S3L and S8R.
3.1.6. Forward flexion4901, 2-kg weight, slow or fast speed, 3 subjects (Fig. 6) Typical examples of load components and force directions are shown in Fig. 6. Fp was 123%, 238% and 212%BW in S2R, S3L and S8R, respectively. This was 76%, 97% and 141% more than that without weight. In the case of fast upward motion of S2R and S8R, both subjects reached their force maximum before the highest elevation angle, similar as during fast forward flexion up to 901 (Fig. 2). S3L carried out the exercise slowly and her force–time curve increased continuously up to the 901 position. She reached the absolutely highest forces and moments (1.74%BWm) among
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
1547
3.2.5. Abduction4901, no weight, slow speed, 3 subjects (Fig. 8) Similar to forward flexion, F still increased above 901 abduction. Fp reached 88, 121 and 88%BW for S2R, S3L and S8R in the examples of Fig. 8. Intra-individually the time courses and force directions were similar as during forward flexion. Resultant moments had values of 0.21%BWm, 0.89%BWm and 0.30%BWm. The expected positive/negative signs of Mx when lifting/lowering the arm were observed in S2R and S3L, but not in S8R. 3.2.6. Postsurgical course For most subjects, no statements can be made yet about the postsurgical course due to the short observation time. In subject S2R the forces decreased slightly by about 20% between 4 and 7mpo, but this was not a distinct change. The moments varied considerably: 0.9 and 27mpo they were 3 times higher than 4mpo. Fig. 4. Resultant forces. Forward flexion and abduction 901, 2-kg weight, slow and fast speed. Typical single trials. Values in symbols ¼ elevation angles. During slow movement (oval symbols) F increases steadily up to 901. During fast movement (rectangular symbols) F reaches its smaller peak values already before 451 and then falls up to 901. Selected examples of subjects S1R, S2R (forward flexion) and S5R (abduction). Inlay shows lever arm l of the centre of gravity (COG), depending on the lifting angle a.
all exercises. Again, the expectations about the signs of Mz were not met.
3.2. Abduction 3.2.1. Abduction 901, no weight, slow speed, 6 subjects (Fig. 7) Fpmx was 81%BW, which is slightly higher than that during forward flexion (73%BW). However, abduction caused less uniform forces (compare Figs. 1 and 7); the individual range of Fpm (46–115%BW) was much larger. Mpmx lay at 0.44%BWm. The individual variation of Mpm (0.07–1.18%BWm) was larger than that of the forces.
3.2.2. Abduction 901, no weight, fast speed, 6 subjects (Fig. 7) Fpmx was reduced to 65%BW by the faster motion; but the changes of Fpm were not very uniform. Fpm decreased by 41% in S4R but increased slightly in S2R, who did the task extremely fast and ambitiously. Mpmx fell to 0.36%BWm, which is lower than at slow speed. Again the individual differences of Mpm were large. As for Fpm, S4R showed the largest reduction.
3.2.3. Abduction 901, 2-kg weight, slow speed, 5 subjects (Fig. 7) Fpmx increased to 129%BW and Mpmx to 0.69%BWm due to the additional weight. The average forces were nearly the same as during forward flexion. Again, the force directions a in the frontal plane (Fig. 3, right frame) were very constant throughout the whole motion (average 28.61) and similar to those during forward flexion. This was also true for the sagittal plane, except for S3L who showed a different behaviour between forward flexion and abduction. As expected, Mx was positive during upwards movement of the arm and negative downwards in all subjects (Fig. 2, bottom). In S3L the peak value of Mx was several times higher than in all other subjects.
3.2.4. Abduction 901, 2-kg weight, fast speed, 4 subjects (Fig. 7) The faster speed let Fpmx decreased to 101%BW. The moments Mpmx fell to 0.61%BWm. All subjects reported that the fast task was easier to perform.
3.3. Influence of movement direction, speed, weight and elevation angle (Fig. 9) 3.3.1. Forces Fig. 9 shows the load changes due to variations of the 4 parameters. Load changes in percent and the significance of load differences were always calculated from identical subjects in both groups. No significant differences in the peak forces Fpmx were found between standard forward flexion and abduction (bold ‘‘0’’ in Fig. 9). Increasing the speed lowered Fpmx by 18–22%. The decrease was independent of the weight carried, and significant except for abduction without a weight. Carrying 2 kg increased Fpmx by 51–75% at slow and fast speed and for different elevation angles. These changes were always significant for 901 elevation. Elevating the arm higher than 901 let Fpmx rise by 21–40%. The investigated number of only 3 subjects prohibited any calculation of significances. 3.3.2. Moments No significant differences of Mpmx were found between standard forward flexion and abduction. The impact of the 4 parameters on the moments Mpmx was less pronounced than it was on the forces. Only the additional weight caused a significant moment rise by 40–53%, both during slow and fast movement.
4. Discussion This study is the first one reporting shoulder joint loads measured in vivo during abduction and forward flexion for a group of patients. The anatomic changes caused by the prosthesis are probably smaller than if the glenoid had additionally been replaced. Nevertheless the data cannot directly be transferred to healthy subjects, because soft tissue alterations and other factors may affect the loading. Several important observations were made in the majority of measurements and subjects; due to restricted cohort sizes not all of them could be proven statistically, however. 1. An additional weight of 2 kg when lifting the arm increases the force by 51–75%. 2. Fast arm lifting without stopping in the highest position causes about 20% lower forces than slow lifting. 3. The directions of high forces relative to the humeral head are very constant in most subjects and during different exercises. 4. In 3 of the subjects, the forces are distinctly higher during standard abduction than they were for forward flexion. In the
1548
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
Fig. 5. Forces and moments. Forward flexion 4901, no weight, slow speed. Three subjects. Upper diagrams ¼ resultant force F and components in % of body weight (%BW). Lower diagrams ¼resultant moment M and components in %BWm. Elevation angles are indicated. Vector plots ¼ force directions in frontal and sagittal plane during whole motion. Subject S2R holds a roll of paper in his hand.
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
1549
Fig. 6. Forces and moments. Forward flexion4901, 2-kg weight, slow or fast speed. Three subjects. Fast speed for S2R and S8R, slow speed for S3L. For more information see Fig. 5.
combined data from all subjects no statistical difference could be found, however. 5. Moments in the joint vary more than the forces and their high magnitudes indicate that the implant head is often loaded eccentrically or from outside of the glenoid.
During activities of daily living, the highest joint forces were determined when steering a car with one hand, setting down 1.5 kg on a table with a straight arm, and laying down 2 kg in a high shelf (Westerhoff et al., 2009a). The forces now observed at high arm elevations with additional weight were in the same range.
S2R
S3L
S4R
Slow 2 kg Fast 2 kg
Slow 0 kg Fast 0 kg
S1R
Slow 2 kg Fast 2 kg
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
Slow 0 kg Fast 0 kg
1550
S5R
S8R
Fig. 7. Peak forces and moments. Abduction 901, without and with 2-kg weight, slow and fast speed. Median peak forces Fpm (top) and peak moments Mpm (bottom) plus ranges of Fp, Mp from 6 subjects. X ¼no data.
Anglin et al. (2000) listed the forces predicted by various musculo-skeletal models. For 901 abduction without additional weight and 75 kg BW, forces of 50–90%BW were calculated; values as measured in 4 of our subjects. The higher forces in the other 2 subjects may possibly be explained by antagonistic muscle activities that are difficult to simulate exactly. Our observation that forces always increased when the arm was lifted higher than 901 was predicted only very recently (Favre et al., 2009). In other models (Poppen and Walker, 1978; Terrier et al., 2008; van der Helm, 1994) the force then decreased. The effect is presumably due to co-contractions, stabilizing the arm position (Steenbrink et al., 2009). High joint loads during overhead work may explain shoulder problems of certain professional groups (Herberts et al., 1981). Moments in the shoulder joint varied even more than did the forces, and were remarkably high. In S3L they were up to 4 times higher than in the other subjects (Figs. 1, 2 and 7). Assuming a constant friction coefficient of 0.1 in the joint and a force of Fp ¼120%BW, a purely friction-induced moment M, acting at a 48 mm head, would be 0.29%BWm. This is the range found in S2R, S4R and S5R. For S1R and S8R a coefficient of even 0.2 must be assumed. It should be tested whether an inclusion of friction in musculo-skeletal models could in any way improve the predictions. Extreme moments, as in S3L, cannot be caused by friction alone, unless one would assume a coefficient of up to 0.5. Such moments are most probably explained by additional forces, transferred outside of the glenoid and the labrum glenoidale, which do not pass the head centre. Such forces could possibly be exerted by the acromion. The observations that the force directions in S3L deviate much from those of the other subjects during forward flexion (Fig. 3) and that the force magnitudes are
highest in S3L when lifting 2 kg (Figs. 1 and 7), support this assumption. The forces were always about 20% lower when elevating the arm quickly (Fig. 4). When the arm is in a low position, the horizontal lever arm l of its COG is short (Fig. 4), and the moment required to accelerate the arm upwards is small. It is therefore energy-saving to accelerate the arm early during the upwards motion, let it swing up nearly passively to the highest position, and finally have it ‘falling down’ (Fig. 4; S2R and S8R in Fig. 6). Obviously all subjects used this tactic. As observed previously for the hip and knee joint (Bergmann et al., 2001; Heinlein et al., 2009) the force directions remained quite constant throughout all activities (Fig. 3). Even between abduction and forward flexion, similarities were seen in the same subject. This indicates that either the force on the glenoid is very variable and/or the scapula motion is adjusted constantly to keep the resultant force direction inside the glenoid. The reported forces and moments are essential to improve shoulder prostheses design, and proximal humeral fracture implants. The data are not representative for the loading of inverse implants, however. Highest peak forces and moments from all subjects during slow motion with 2 kg (Figs. 3 and 7) represent the most severe test conditions, simulating forward flexion/abduction. In the frontal plane the forces can be applied at the reported average angles a (3.1.3/3.2.3). In the sagittal plane, that extreme force angle b (Fig. 3) should be chosen as the angle causing the most critical mechanical condition for the implant and its fixation. Because no generally dominant axis for the largest moments exists, that torque axis should be chosen which results in the most severe mechanical situation. Less demanding tests could be performed under the same conditions but with those decreased force magnitudes which simulate exercises
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
1551
Fig. 8. Forces and moments. Abduction 4901, no weight, slow speed. Three subjects. For more information see Fig. 5.
without weight. These proposed test setups certainly require further discussion. Even more important than the humeral load will be that of the scapula, because glenoid replacement fixation is still unsatisfactory. Synchronous motion capture and load measurements will
deliver the force directions relative to the scapula and the movement between humerus and scapula. This data will also be used as ‘gold standard’ to improve the analytical methods, as previously done with hip load data (Bergmann, 2001; Heller et al., 2001; Stansfield et al., 2003).
1552
Fpmx
G. Bergmann et al. / Journal of Biomechanics 44 (2011) 1543–1552
Mpmx
Subj.
+68%* +92%° fast +57%* +49% *
Fpmx
Mpmx
5
-18%*
-2%°
4
-22%*
-15%*
Subj.
Fast
5 4
Fpmx
+75%* +53%*
Subj.
-19%*
-8%°
6
-20%°
-18%°
6
>90°
5 5
Fpmx
Mpmx
Subj.
Fpmx
Mpmx
+40%n
+151%n
3
+51%n
+19%n
3
---
---
---
---
---
---
0
+2kg
+2 kg Mpmx
Subj.
+69%* +75%°
0
Fpmx
Mpmx
Subj.
+2 kg
Standard Exercise
Fast
Fpmx
Mpmx
73%BW
0.26%BWm
6
81%BW
0.44%BWm
6
0 Type of movement:
Subj.
>90°
Mpmx Subj. n +21% +85%n 3 Fpmx
+23%n
-8%n
3
0 Forward Flexion
Abduction
Fig. 9. Load changes by movement direction, speed, weight and elevation angle. Grey or green boxes: numbers for forward flexion. White or blue boxes: numbers for abduction. ‘‘Standard’’: slow exercise up to 901without a weight. ‘‘Subj.’’: number of investigated subjects. Parameters: ‘‘Fast’’: activity performed at faster speed. ‘‘2 kg’’: activity with 2 kg held in the hand. ‘‘4 901’’: elevation higher than 901. Significances: ‘‘*’’: significant (Student’s t-test p o0.05). ‘‘0’’: not significant. ‘‘n’’: significance test not applicable.
Conflict of interest statement All authors have no conflict of interest that could influence the work presented here.
Acknowledgements This work was supported by the Deutsche Forschungsgemeinschaft (BE 804/17-1 and SFB 760-C-6), Berlin Brandenburg Center of Regenerative Therapies, Synthes GmbH Switzerland, and Biomet Merck Deutschland GmbH. We thank all patients for their engaged co-operation! References Anglin, C., Wyss, U.P., Pichora, D.R., 2000. Glenohumeral contact forces. Proc. Inst. Mech. Eng. [H] 214, 637–644. ¨ Berlin. Bergmann, G., 2001. Hip98, Loading of the Hip Joint. Freie Universitat, Bergmann, G., Deuretzbacher, G., Heller, M., Graichen, F., Rohlmann, A., Strauss, J., Duda, G.N., 2001. Hip contact forces and gait patterns from routine activities. J. Biomech. 34, 859–871. Bergmann, G., Graichen, F., Bender, A., Kaab, M., Rohlmann, A., Westerhoff, P., 2007. In vivo glenohumeral contact forces—measurements in the first patient 7 months postoperatively. J. Biomech. 40, 2139–2149. Boileau, P., Krishnan, S.G., Tinsi, L., Walch, G., Coste, J.S., Mole, D., 2002. Tuberosity malposition and migration: reasons for poor outcomes after hemiarthroplasty for displaced fractures of the proximal humerus. J. Shoulder Elbow. Surg. 11, 401–412. Buechel, F.F., Pappas, M.J., DePalma, A.F., 1978. ‘Floating socket’ total shoulder replacement: anatomical, biomechanical and surgical rationale. J. Biomed. Mater. Res. 12, 89–114. Dul, J., 1988. A biomechanical model to quantify shoulder load at the workplace. Clin. Biomech. 3, 124–128. Favre, P., Snedeker, J.G., Gerber, C., 2009. Numerical modelling of the shoulder for clinical applications. Philos. Trans. R. Soc. A: Math. Phys. Eng. Sci. 367, 2095–2118. Graichen, F., Arnold, R., Rohlmann, A., Bergmann, G., 2007. Implantable 9-channel telemetry system for in vivo load measurements with orthopedic implants. IEEE Trans. Biomed. Eng. 54, 253–261. Heinlein, B., Kutzner, I., Graichen, F., Bender, A., Rohlmann, A., Halder, A.M., Beier, A., Bergmann, G., 2009. ESB Clinical Biomechanics Award 2008: complete data
of total knee replacement loading for level walking and stair climbing measured in vivo with a follow-up of 6–10 months. Clin. Biomech. (Bristol, Avon.) 24, 315–326. Heller, M.O., Bergmann, G., Deuretzbacher, G., Durselen, L., Pohl, M., Claes, L., Haas, N.P., Duda, G.N., 2001. Musculo-skeletal loading conditions at the hip during walking and stair climbing. J. Biomech. 34, 883–893. Herberts, P., Kadefors, R., Andersson, G., Petersen, I., 1981. Shoulder pain in industry: an epidemiological study on welders. Acta Orthop. Scand. 52, 299–306. Inman, V.T., Saunders, J.B., Abbott, L.C., 1996. Observations of the function of the shoulder joint. 1944. Clin. Orthop. Relat. Res., 3–12. Karlsson, D., Peterson, B., 1992. Towards a model for force predictions in the human shoulder. J Biomech. 25, 189–199. Kessel, L., Bayley, I., 1986. Clinical Disorders of the Shoulder, 2nd ed. Churchill Livingstone, Edinburgh. Poppen, N.K., Walker, P.S., 1978. Forces at the glenohumeral joint in abduction. Clin. Orthop. Relat. Res., 165–170. Post, M., Jablon, M., Miller, H., Singh, M., 1979. Constrained total shoulder joint replacement: a critical review. Clin. Orthop. Relat. Res., 135–150. Runciman, R.J., 1993. Biomechanical model of the shoulder joint. Bioengineering Unit, Vol. Ph.D. Thesis, University of Strathclyde, Glasgow. Stansfield, B.W., Nicol, A.C., Paul, J.P., Kelly, I.G., Graichen, F., Bergmann, G., 2003. Direct comparison of calculated hip joint contact forces with those measured using instrumented implants. An evaluation of a three-dimensional mathematical model of the lower limb. J. Biomech. 36, 929–936. Steenbrink, F., de Groot, J.H., Veeger, H.E., van der Helm, F.C., Rozing, P.M., 2009. Glenohumeral stability in simulated rotator cuff tears. J. Biomech. 42, 1740–1745. Terrier, A., Vogel, A., Capezzali, M., Farron, A., 2008. An algorithm to allow humerus translation in the indeterminate problem of shoulder abduction. Med. Eng. Phys. 30, 710–716. van der Helm, F.C., 1994. Analysis of the kinematic and dynamic behavior of the shoulder mechanism. J. Biomech. 27, 527–550. van der Helm, F.C., Veeger, H.E., 1996. Quasi-static analysis of muscle forces in the shoulder mechanism during wheelchair propulsion. J. Biomech. 29, 39–52. Westerhoff, P., Graichen, F., Bender, A., Halder, A., Beier, A., Rohlmann, A., Bergmann, G., 2009a. In vivo measurement of shoulder joint loads during activities of daily living. J. Biomech. 42, 1840–1849. Westerhoff, P., Graichen, F., Bender, A., Rohlmann, A., Bergmann, G., 2009b. An instrumented implant for in vivo measurement of contact forces and contact moments in the shoulder joint. Med. Eng. Phys. 31, 207. Wu, G., van der Helm, F.C., Veeger, H.E., Makhsous, M., Van Roy, P., Anglin, C., Nagels, J., Karduna, A.R., McQuade, K., Wang, X., Werner, F.W., Buchholz, B., 2005. ISB recommendation on definitions of joint coordinate systems of various joints for the reporting of human joint motion—Part II: Shoulder, elbow, wrist and hand. J. Biomech. 38, 981–992.