ARUP; ISBN: 978-0-9562121-5-3 - CMBBE 2012 - Cardiff University

5. DISCUSSION
A musculoskeletal numerical model of the upper limb was developed. The anatomical data was
taken from the Visible Human Project [15][16]. It includes the three joints participating in humerus
movement. They are spanned over by 28 muscle fibres. Displacement boundary conditions on the
humeral head are released during the movement, to allow its displacement. To solve the mechanical
indetermination of the muscle force distribution, an optimization algorithm is used. Mean square
muscle stresses are minimized. The model is used to simulate an abduction movement in the scapula
plane. Muscle forces, reaction forces in the glenohumeral joint, humeral head translation and
loading patterns on the glenoid cartilage are computed.
The obtained muscle forces are in principle in agreement with results from the literature [5][20].
The anterior parts of the deltoid are active during the entire motion making a bell curve with the
maximum situated around 90◦ absolute abduction. Nevertheless, posterior parts of the deltoid rest
almost inactive. This does not correspond with in-vivo measurements [14]. The loading of the glenohumeral
joint is in agreement with the results found in literature [21]. In-vivo measurements
showed a continuous rise to 60-70% BW from 0° to 90° abduction. Peak forces of 80-90% BW
were measured at maximum abduction angles. Although the local minimum of the computed forces
at 120° is not documented in experimental data, the general trend of rising joint loading during abduction
is visible in the numerical results. The directions of humeral head translations are in agreement
with experimental data [11] but too large in magnitude. Contact pressure on the cartilage correspond
to experimental data [22]. The joint loadings of 440 N used in experiments correspond to
an abduction angle of 70°. During the whole movement cartilage pressure is lower than injuries
joint loadings [23].
The model has still some limitations. Concerning the muscle modelling, it is known that maximal
muscle forces are not only dependent on physiological cross sections, but also on actual length, contraction
rates and contraction/relaxation mode [24]. First approaches in that direction show a significant
change in muscle force distribution with a higher number of active muscles during the whole
simulation. Furthermore, changes of muscle moment arms due to tissue deformation during movements
can not be reproduced with 1D muscle modelling. The large humeral head translations can be
traced back to the fact, that glenohumeral reaction forces are restricted by the direction criteria
throughout the whole movement. In consequence, the humeral head moves along the cartilage border.
This results in the maximum possible translation range. If the direction criteria is strengthened
to less than 50% of the glenoid cartilage surface, humeral head translation can be forced to reproduce
experimental results, but leads to significant higher joint loadings. As the humeral head translations
are too large, the contact zones are distributed along the cartilage border.
Although the humeral head translation seems excessive, the stabilization method of the humeral
head is working. Displacement boundary conditions can be removed during the movement and humeral
head translation can be simulated. Furthermore, glenohumeral joint loadings show promising
results.
We have shown that a direction criteria on reaction forces of the glenohumeral joint can govern the
stability problem, but whether leads to large humeral head translations or large joint loadings. If optimization
routines are used to compute muscle forces, monitoring of humeral head translation
needs to be included into the optimization routine and the computation of rotator cuff muscle activities
might be handled differently.
6. REFERENCES
1. N.K. Poppen and P.S. Walker: Forces at the glenohumeral joint in abduction, Clinical
Orthopedics, 1978, 135, 165-170
2. C. Högfors, G. Sigholm and P. Herberts: Biomechanical model of the human shoulder -
I.Elements, Journal of Biomechanics, 1987, 20(2), 157-166