T. M. Bach, L. J. Barnes, O. M. Evans and I. G. A. Robinson
La Trobe University, Melbourne, Australia

INTRODUCTION

A number of investigators have suggested that the mass and mass distribution of modern transfemoral (TF) prostheses are sub-optimal and that substantial improvement could be obtained by adding mass to the prosthesis. Support for this concept has been provided by a number of empirical studies (Menkveld et al., 1981; Tashman et al., 1985; Skinner and Mote, 1989; Hale, 1990) and computer simulation studies (Mena et al., 1981; Menkveld et al., 1981; Tsai and Mansour, 1986; Bach 1994). Skinner and Mote (1989) attempted to identify an optimal inertial load using an empirical approach but such an approach is time consuming and probably not suited to identification of optimal loads for individual amputees. Bach et al. (1993a) developed a computer simulation of swing phase in human walking which successfully predicted characteristics of amputee gait. A version of the model could be solved to predict optimal inertial loading either by minimizing energy cost or by maximizing gait symmetry but the two approaches resulted in different predictions (Bach, 1994). The purpose of the present study was to test the optimization predictions in a group of amputees and to distinguish between the two approaches to the optimization.

METHOD

Five young active male transfemoral amputees (body mass: 78.5 ± 8.1 kg, stature: 1.79 ±0.04 m, age 36.2 ±6.5 years) were studied. A similar group of five male controls were also tested.

For each amputee subject an experimental prosthesis of modular endoskeletal design with a constant friction knee was fabricated. Prosthetic limb inertial characteristics were measured using standard methods and the inertial characteristics of the stump were estimated using a segmentation model (Bach et al., 1993b). Corrected body mass (CBM) was estimated for each subject by subtracting prosthesis mass and estimated stump mass from measured body mass and assuming the remainder was 0.839 of intact (corrected) body mass. Subject stature, CBM, prosthesis and stump characteristics were used as input data to a computer simulation which predicted an optimal inertial load and load location. Two approaches to the optimization were considered: one approach involved minimizing the predicted mechanical energy cost of walking, the other involved maximizing swing phase symmetry between sound and prosthetic limbs.

Five tests of the experimental prosthesis were performed: no added mass (NAM), optimal energy mass at the optimal energy location (MeLe), optimal symmetry mass at the optimal symmetry location (MsLs) and two crossover conditions (MeLs and MsLe). The added masses consisted of lead weights moulded to fit tightly around the pylon of the prosthesis.

Subjects performed one test wearing their own prosthesis and five tests wearing the experimental prosthesis. Subjects walked on a motorized treadmill at a speed of 1.0 m·s-1. Metabolic energy expenditure was estimated from oxygen consumption measured using a metabolic cart. Kinematic data was obtained using a 5-camera VICON motion analysis system. Temporal characteristics, angular kinematics, swing phase kinetics and mechanical energetics were estimated from kinematic data. Subjective rankings of prosthesis performance for each experimental condition were obtained using a Visual Analog Scale.

A one-way ANOVA for repeated measures was used to examine the effects of experimental treatments. Post-hoc comparisons were made using Tukey's HSD test. Correlations between dependent variables were computed within subjects and averaged using a Fisher z transform.

RESULTS

Metabolic power estimated from oxygen consumption was significantly reduced (F(16,4)=5.02, p=.008) and subjective ratings were significantly higher (F(16,4)=6.09, p=.004) for optimal symmetry loading compared to the unloaded prosthesis. Gait symmetry was significantly higher for the optimal symmetry loading than for the unloaded prosthesis both using an index of swing phase difference (F(16,4)=16.28, p<.001) and using a velocity index (Murray et al., 1980) which expressed the mean trunk velocity during sound single support to that during prosthetic single support (F(16,4)=5.03, p=.008). The energy optimization results were not significantly different from NAM except for swing phase differences.

Total mechanical power measured assuming energy transfers within and between segments (Pierrynowski et al., 1980) did not differ significantly between experimental treatments even though significant differences were found for metabolic power. The correlation between total mechanical power and metabolic power was .271 (p>.10). Metabolic power was significantly correlated with both symmetry indices (swing phase difference r=.707, p<.01; velocity index: r=-.631, p<.01)

Computer simulation predictions of gait symmetry and mechanical energy expenditure were compared with observed values. The correlation between predicted and observed swing phase difference was .903 (p<.01) and between predicted and observed mechanical energy cost was .353 (p>.10).

Angular kinematic patterns, joint moments and segment energy histories provided insights into differences between experimental treatments.

DISCUSSION

The results or this investigation provided strong experimental support for a computer simulation approach to the optimization of TF prosthetic limbs. The optimization approach which maximized gait symmetry resulted in significant increases in gait symmetry and subjective ratings of performance and in significant reductions in metabolic energy expenditure. The experiment did not demonstrate that the predicted symmetry loadings were in fact optimal. Such an assertion would require that many combinations of added mass and mass location were tested. However, the experiment provided strong support for an optimization approach which maximizes gait symmetry.

The failure of the energy optimization approach suggests that assumptions of the model used to predict energy expenditure may be inadequate or erroneous. However, lack of correlation between metabolic and mechanical power suggests that the relationship between these variables is inadequately understood. Although mechanical energy analysis has been promoted as a potent tool for understanding inefficiencies in human gait (Winter, 1979), the data presented here suggests that further development of this method is required before its potential can be realized.

This study provided some empirical support for the common clinical assumption that gait symmetry is related to energy expenditure. However, it should be emphasized that gait symmetry was achieved by addressing the underlying mechanical cause of asymmetry rather than by training subjects to adopt a more symmetrical gait. Differences in gait symmetry which resulted from changes in inertial characteristics demonstrated that the asymmetrical gait patterns adopted by amputees are not necessarily habitual but may be related to mechanical characteristics of the prosthesis.

There can no longer be any doubt that the mass distribution of current prosthetic designs is inappropriate. It is therefore essential that consideration be given to the identification of optimal inertial parameters and the means by which these features can be derived based on the characteristics and requirements of individual amputees. This study demonstrated the feasibility of a computer simulation approach to this problem. Such an approach has the potential to substantially improve amputee management.

REFERENCES

Bach, T. M. et al. (1993a) Proceedings of the IVth International Symposium on Computer Simulation in Biomechanics, Montlignon, France. pp BMR2-BMR5.
Bach, T. M. et al. (1993b) Proceedings of the XIVth Internaional Congress on Biomechanics, Paris, France. pp130-131.
Bach, T. M. (1994) Unpublished PhD Thesis. La Trobe University, Melbourne.
Hale, S. A (1990) Pros. Orth. Int. 14: 125-135.
Mena, D. et al. (1981). J. Biomech. 14:823-832.
Menkveld, S. et al. (1981) Trans. 27th Ann. Meeting, Orthop. Res. Soc.5:51.
Murray M. P. et al. (1980) Bull. Pros. Res. 17:35.
Pierrynowski, M. R. et al. (1980). Ergonomics 23:147-156.
Skinner, H. B. and C. D. Mote (1989) Rehab. Res. Develop. Prog. Reports 26:32.
Tsai, C. S. and J. M. Mansour (1986). J. Biomech. Eng. 108:65-72.
Tashman, S. et al.(1985) Clin. Pros. Orth. 9:23.
Winter, D. A. (1979) Physio. Canada 30:183.