Presented at NACOB 98: North American Congress on Biomechanics Canadian Society for Biomechanics - American Society of Biomechanics University of Waterloo Waterloo, Ontario, Canada August 14-18, 1998

---

THE EFFECTS OF JOINT CENTER LOCATION ON NET MOMENTS AT THE SHOULDER

D.Lafontaine ^1,2 & D.Gagnon ²
Department of cellular and molecular medicine ¹ , University of Ottawa, Ottawa, Canada
Faculté d'éducation physique et sportive ² , Université de Sherbrooke, Sherbrooke,, Canada

INTRODUCTION

This study compares the effects of four methods of estimating the position of the gleno-humeral joint center on the moments generated about that joint. It also serves as a first step in developping a tool that will be appropriate for evaluating shoulder function more thoroughly.

REVIEW AND THEORY

The shoulder joint invariably plays a role in every materials handling task. Also, injuries to this region are second only to low back pain in frequency of diagnostic for work related injuries (Sommerich et al . 1993). Women are more at risk of developing work related shoulder problems since they occuppy more of the « at risk » jobs for developping these musculo-skeletal diseases. Nonetheless, the loading of this joint or «joint complex» (Chaffin et al. 1991) as it is often refered to, has not, thus far, been the focus of many studies. The reasons for this are multiple. First and foremost is the complexity of the joint complex, which makes model-ling it quite difficult. Second is the fact that the displacements of the clavicle and scapula are not easily tracked with external markers

PROCEDURES

Subjects (n=5) performed multiple trials of lifting tasks in a laboratory setting with markers (n=30) attached over various anatomical landmarks. One of these trials was selected for image clarity and digitized. Each trial was digitized twice with the Peak 5 system. The initial digitization assigned coordinates to all external markers. In the second digitization, the shoulder joint was tracked directly. The four joint center position estimation methods (JCPEM) compared used different combinations of external markers to estimate the 3D position of the shoulder joint. These four methods were: ACTUAL, RODS, 2 MARKERS and DIRECT. The ACTUAL method served as the standard in this study since it had provided valid results in the past. For all methods, the same force files were used for each trial of each subject. The effects of four types of tasks (high/low translation and sagittal/diagonal elevation) on the net moments at the glenohumeral and L5/S1 joints were also examined.

RESULTS AND DISCUSSION

Root mean square (RMS) values comparing all four JCPEM were used to compare joint center position and net moments at the shoulder joint. In many cases, the DIRECT digitizing method provided the results that most differed from those obtained with the other JCPEM. However, all JCPEM provided the most different results in at least one occasion. It has also been noted that results obtained with the various JCPEM differ more between one another than with the standard. The differences differently along the three axes (XYZ). These differences between the methods range from 1 mm (subject 5 task 2 DIRECT vs. RODS along z) to 57 mm (subject 2 task 3 DIRECT vs. RODS along z).

Subject/ task	(T1)	(T2)	(T3)	(T4)
S1 lt.shoul. rt. shoul total	408 499 907	651 592 1243	308 474 782	323 516 839
S2 lt.shoul rt. shoul total	204 202 406	224 245 469	/ / /	211 202 413
S3 lt.shoul rt. shoul total	188 144 332	188 217 405	174 194 368	135 133 268
S4 lt. shoul rt. shoul total	134 159 293	224 237 461	231 177 408	211 137 348
S5 lt. shoul rt. shoul total	147 182 329	174 171 345	257 128 329	131 138 269

(lt.shoul.: left shoulder; rt.shoul: right shoulder)

Table 1: Mean resultant RMS (mm) values for joint center position data

Mean resultant RMS values were used to evaluate the effects that different tasks may have had on the joint center positions (Table 1). The observed differences are quite small between methods for all subjects. Values obtained from subject 1 are up to three times higher than those of other subjects. For all five subjects, the low translation task (T2) provides the greatest differences among JCPEM.

The same comparisons (RMS) between the JCPEM were made for the net moments at the shoulder. The sensitivity of the model that was used to calculate the joint loading is of 5 N·m (Larivière et al., 1996), therefore differences had to be above that threshold to have any significance. The worst case is that of subject 1 low translation task (T2), where a difference of 52 N·m was observed (right shoulder, RODS vs. DIRECT, Transverse axis). For the same subject and task, a difference of only 1 N·m was observed along the Longitudinal axis at the left shoulder. For task 1 of subject 5, all comparisons made between methods are below the sensitivity threshold of the model (Table 2).

Axis	Left shoulder				Right shoulder
	A	R	2	D	A	R	2	D
	0	1	1	1	0	1	1	1
L		0	1	1		0	1	1
			0	1			0	1
	0	2	1	2	0	1	2	2
S		0	1	3		0	2	2
			0	2			0	3
	0	4	3	2	0	3	4	4
T		0	2	4		0	3	3
				3				4

Table 2: RMS differences(N·m) between all four JCPEM, for moments of force values of S5T2 (L: longitudinal axis; S: sagittal axis; T: transverse axis)

As was the case for the joint center position data, the effects of the type of task performed by the subjects on net moment values was also evaluated. For these results as well, a small influence was detected (Table3). The low translation task (T2) providing the largest differences for the net shoulder moment results as well. The shoulder moments obtained with all the methods under study are in accordance with values reported in the litterature (Fleisig et al., 1996).

The effects of the JCPEM under investigation on the net moments at the L5/S1 joint were also evaluated. These results were not compared using RMS values, since there were virtually no differences of interest, given that the sensitivity of the model for moments at L5/S1 is of 10 N·m. The differences between net moments obtained using the various JCPEM for the shoulder joint were even smaller at the lumbosacral joint.

Subject/Task	T1	T2	T3	T4
S1 lt.shoul rt. shoul total	85 222 307	138 205 343	38 135 173	100 194 294
S2 lt.shoul rt. shoul total	37 35 72	77 44 115	/ / /	51 38 89
S3 lt.shoul rt. shoul total	41 39 80	43 41 84	37 28 65	20 25 45
S4 lt.shoul rt. shoul total	34 35 69	42 45 87	47 24 71	50 29 79
S5 lt.shoul rt. shoul total	22 26 48	28 31 59	16 16 32	22 19 41

Table 3: Mean resultant RMS differences (N·m) for net moments at the shoulder joint

After examining the minimal and maximal values obtained, it was decided to forgo the RMS comparisons. The extreme values for all tasks using all four JCPEM proved very similar. Net moment values at L5/S1 were also similar to those found in the litterature (Plamondon et al., 1996).

CONCLUSION

What comes out of the results of this study, is that if joint center position data is not of great importance, then researchers need not use complex marker sets to obtain valid net moment values. This is an interesting finding for those that are interested in conducting in-field studies. Future studies on the biomechanics of the shoulder joint should include EMG since a majority of the reported lesions in this area are muscular in nature. As well, future studies should use women occupying at risk jobs as subjects to measure the loads imposed on their shoulders when performing work specific tasks.

REFERENCES

Sommerich et al ., Ergon., 36(6), 697-717,1993.

Chaffin et al., Occupational Biomechanics, John Wiley & Sons, 1991.

Larivière et. al., Proc.CSB IX, p.294-295, 1996.

Fleisig et al., J.Appl.Biom., 12, 207-224, 1996.

Plamondon et al., Clin.Biom., 11(22), 101-10, 1996.