American Academy of Orthotists & Prosthetists

Link to Figures
Show Figures

Article Text
Figures
References

Send Link
Send HTML Article

Quantification of the Effect of Dorsi-/ Plantarflexibility of Ankle-Foot Orthoses on Hemiplegic Gait: A Preliminary Report

Sumiko Yamamoto, PhD
Masahiko Ebina, CPO
Shigeru Kubo
Hideo Kawai, RPT
Takeo Hayashi, CPO
Mitsuo Iwasaki, CPO
Toshio Kubota, MD
Shinji Miyazaki, PhD

ABSTRACT

A series of studies has been done to establish an objective and quantitative method to determine the optimum dorsi-/plantarfiexibility of an ankle-foot orthosis (AFO) for various patients.

Introduction

Ankle-foot orthoses (AFOs) are commonly used in clinical practice to supplement the gait disabilities of hemiplegic patients secondary to cerebrovascular accidents and brain injuries. However, few biomechanical research studies have clarified how this improvement of gait is achieved in relation to AFOs' mechanical properties.

Factors involved in the selection and prescription of AFOs include dorsi-/ plantarflexibility, initial setting of the dorsi-/plantar angle and the length of the sole. Among these factors, dorsi-/ plantarflexibility exerts the most significant effect on patients' gait.

Flexibility usually is controlled by selecting a proper trim line around the malleolar aspect of the AFO. Often a less-flexible AFO is prescribed because of irreversibility of the trimming process and greater endurance, which generally results in so-called overbracing.

Biomechanical analyses of AFOs have been made by Chowaniec et al. and by Lehmann et al., but they have not investigated the quantitative effect of dorsi-/plantarfiexibility (1-4). Both studies also limited measurement to a single step although it is well-known that the step-to-step variation of hemiplegic gait is large. Thus, their data are difficult to interpret in a statistical sense.

The authors have begun a series of studies to establish an objective and quantitative method to determine the optimum dorsi-/plantarfiexibility of an AFO for each patient. As the first step, Yamamoto et al. have measured the flexibility of various types of plastic AFOs when fitted to the human limb (5). Miyazaki et al. have made a mathematical model of gait with an AFO, showing that the total ankle joint moment due to floor reaction forces equals the sum of the moment generated by the muscles around the ankle plus the corrective moment generated by the AFO due to its deformation.

They then developed a system that measured the total ankle joint moment due to floor reaction forces and the corrective moment generated by the AFO, continuously and separately t estimate the ankle joint moment generated by the muscles (6). Based 01 these achievements, the purpose of the study was to demonstrate the feasibility of quantifying the effect of dorsiplantarflexibility on hemiplegic gait in a statistical manner.

Method

Measurement

Measurement was made on a level floor of 20 m in length. The following quantities were measured simultaneously:

• total ankle-joint moment due to floor reaction force

• corrective moment generated by an AFO

• temporal factors

• step length

• ankle angle and knee angle in the sagittal plane

• EMGs of plantarflexors and of dorsiflexors.

The total ankle-joint moment was measured by a device developed by Miyazaki et al. (6). This device uses a 2.5mm capacitive force transducer attached to the sole of the footwear that measures the total moment within an accuracy of +/- 3N-m Temporal factors were extracted from the vertical foot forces measured by the capacitive transducer.

Because of the need to change the dorsi-/plantarfiexibility of the AFO for a short time during the experiment and measure its corrective moment, a specially designed AFO was used (6). Figure 1 shows the experimental AFO, which consists of a plastic shoe, modified medial and lateral double-Klenzak joint, and a metal support. In each Klenzak joint, two coil springs are housed, and the dorsi-/plantarflexibility of the AFO can be controlled by changing the stiffness of the springs. In the study, soft, medium and hard springs were used and are denoted as spring No. 1, No. 2 and No. 3, respectively.

At the time of measurement, the Klenzak joints were adjusted so their centers were aligned to the center of the ankle joint. The ankle joint angle was then measured by a potentiometer attached to the center of the lateral Klenzak joint. This angle was used to retrieve the corrective moment of the AFO from the ankle.

The plastic shoe (PLS), made by cutting a conventional shoehorn-type AFO at the ankle, can be removed from the experimental AFO by releasing the screw bolts of the stirrup plate connecting the Klenzak joints and the PLS. The disassembled PLS was used as experimental footwear to simulate the patients' normal shoe. The medial and lateral walls of the PLS were trimmed to 25-30 mm so that the PLS does not provide excessive medio-lateral stability. The subjects wore similar PLSs on their sound feet during the experiment. The experimental AFO, including the PLS, weighed 1,060 g - much more than the shoehorn-type AFOs usually worn but comparable to the weight of conventional double-metal-bar AFOs.

The step length was measured by a simple device developed by Miyazaki (7). The accuracy of the device is +/- 10 mm. The angle of the affected side in the sagittal plane was measured by the potentiometer attached to the experimental AFO when the subject walked with the AFO, and was measured by a potentiometer-type uni-axial goniometer when the subject walked with the isolated PLS. The knee angle of the affected side in the sagittal plane was also measured by a potentiometer-type uni-axis goniometers in both conditions.

EMGs of plantarflexors and of dorsiflexors of the affected side were measured by pairs of disposable surface electrodes, electrically rectified and low-pass filtered. Whole-body gait pattern was recorded from the lateral side by a VTR for general reference.

Subjects

Two hemiplegic patients volunteered as subjects. Subject A was a 64-yearold male left-hemiplegic patient who suffered from cerebral infarction. He was graded as stage III in the Brunnstrom scale. The measurement was made seven months after the subject's accident. He used a cane and usually wore a shoehorn-type AFO. He is 166 cm tall, and weighs 62 kg.

Subject B was a 38-year-old male right-hemiplegic patient who had suffered a brain injury. He also was graded as stage III. Subject B was measured 13 months after his accident and did not use a cane but usually wore a shoehorn-type AFO. He is 167 cm tall and weighs 64 kg.

Procedure

Measurements were first made with the isolated PLS. The subjects were asked to walk at a comfortable speed. Initial and last portions of gait were discarded, and the steady state gait was measured for a continuous 30 seconds in one path and sampled at 100 Hz. Data recording the two paths was stored for analysis. Before measurement, the subjects were allowed to take a trial walk of the two paths to become accustomed to the walkway and the footwear.

The sequence was repeated with the experimental AFO using different flexibilities: coil spring No. 1 and No. 2 for subject A, and No. 1, No. 2 and No. 3 For subject B. To minimize fatigue, the subjects rested between sessions.

Statistical Analysis

The following 24 quantitative parameters were extracted from raw data for each step or cycle and were analyzed for variance. One-way allocation was assumed, and the flexibility of the AFO was treated as the variation factor.

Temporal factors were cycle time, average walking speed, stance phase of the affected side, single support phase of the affected side, double support phase from the sound side to the affected side, and double support phase from the affected side to the sound side.

Spatial factors were step length from the affected side to the sound side and step length from the sound side to the affected side.

Joint angles (all for the affected side) were maximum plantarfiexion during stance and maximum dorsiflexion during stance, maximum plantarfiexion during swing for the ankle joint, maximum extension during stance, maximum flexion during swing for the knee joint.

Ankle joint moment: 1) maximum total ankle joint moment in the dorsiflex direction where negative value indicates the subject and the AFO generate moment so the ankle joint is dorsiflexed, 2) maximum muscle moment in dorsiflex direction, 3) maximum total ankle joint moment in plantarflex direction, 4) maximum muscle moment in plantarflex direction (all for the affected side), 5) maximum total ankle joint moment in dorsiflex direction, 6) maximum total joint moment in plantarflex direction (both for the sound side).

Integrated EMG (all for the affected side): 1) integrated value of EMG (IEMG) of dorsiflexors during stance, 2) IEMG of dorsiflexors during swing, 3) IEMG of plantarflexors during stance, 4) IEMG of plantarflexors during swing.

Results

Figure 2a shows the relationship between the ankle joint angle and the corrective moment of the experimental AFO assembled with three coil springs: No. 1, No. 2 and No. 3. The horizontal axis represents the ankle angle and the vertical axis the corrective moment. A steeper curve indicates lower flexibility.

Figure 2b shows the relationship between the ankle joint angle and the corrective moment of three typical plastic AFOs (5). The broken loop shows the flexibility of a hemispiral AFO. The two solid loops show the characteristics of shoehorn-type AFOs. The steeper loop represents a very rigid shoehorn-type AFO and the middle loop a flexible shoehorn-type AFO. Note that the experimental AFO with coil spring No. 1, No. 2 and No. 3 roughly corresponds to a hemispiral AFO, a flexible shoehorn-type AFO and a very rigid shoehorn-type AFO, respectively.

Figure 3 shows the typical waveforms of the ankle joint moment. In each graph, the ankle joint moments of the affected side for five consecutive steps are superimposed. The horizontal axis represents time in seconds, and its origin is the timing of foot contact for each step. The vertical axis represents the ankle joint moment; the solid line shows the total ankle joint moment due to floor reaction forces, and the broken line shows the corrective moment generated by the AFO. The results for subject A are shown in the upper row.

The left graph shows the ankle joint moment when the subject walked with the isolated PLS. The middle and the right graphs show the results when the subject walked with the experimental AFO assembled with coil springs No. 1 and No. 2, respectively. Similarly, the results for subject B are shown in the lower row. It is important to note the magnitude of the total ankle joint moment. However, this small correcting moment did generate noticeable differences in the pattern and timing of total ankle moment.

The fact that a slight change in corrective moment resulted in a marked difference in gait pattern is more clearly shown by the results of statistical analysis of variance. In Figure 4 , Figure 5 , Figure 6 , Figure 7 , and Figure 8 , the open circle and the filled triangle represent the results of subject A and subject B, respectively. In each graph, the horizontal axis represents the level of the variation factor in one-way allocation, that is, the flexibility of the AFO (0 indicates that no AFO was used, i.e., isolated PLS). This condition is assumed to be the extreme case in which the flexibility of the postulate AFO is infinitive. Levels 1, 2 and indicate the AFOs assembled with coil springs No. 1, No. 2 and No. 3, respectively. As the number increases, the flexibility decreases.

Results are plotted only when the differences in the main effect are statistically significant at a confidence level of .01. The vertical axis represents the mean at each level. The left vertical scale applies to the results of subject A and the right vertical scale to those of subject B. The number of samples at each level was 25 to 27 for subject A and 26 or 27 for subject B.

For subject A, plantarfiexion during swing decreased monotonously as flexibility decreased (see Figure 6 ). Since one of the major functions of the AFO is to prevent excessive plantarfiexion during swing to assure better limb clearance, degree of flexibility is important. In this sense, the least flexible AFO is the best AFO.

However, such an AFO affects other aspects of gait. It tends to cause forward thrust of the shank, causing knee collapse after heel contact. It also causes back knee in the late stage of stance phase. Therefore, the best or optimum flexibility is a trade-off between various factors.

From this point of view, it is important that all other parameters present either maximum or minimum flexibility level 1, and that most can be interpreted as a sign of better gait function. For example, the minimum cycle time means fastest cadence. Combined with the maximum step length from the affected side to the sound side and from the sound side to the affected side, it results in maximum walking speed (see Figure 4 and Figure 5 ). It is evident that the optimum level of flexibility for subject A in this situation is 1. It is interesting to note that muscle moment in plantarflex direction increases when the subject wears the AFO (see Figure 7 ). For subject B, there is no clear tendency governing all parameters. It may be said that the optimum level of flexibility for this subject is 2 since the walking speed and the step lengths are at maximum at this level (see Figures 4 and Figure 5 ). It must be noted that the maximum extension during stance, i.e., back knee, also occurs at this level (see Figure 6 ). This subject walks at maximum speed with the AFO assembled with coil spring No. 2 at a risk of future complications at the knee. Another interesting observation is that IEMG of the plantarflexors and the muscle moment in plantarflex direction vary considerably with flexibility (see Figure 7 and Figure 8 ).

Discussion

The gait patterns of two hemiplegic patients were quantified and subjected to statistical analysis with the aid of recently developed continuous measurement techniques. Preliminary results demonstrated the basic feasibility of quantifying the effect of dorsi-/plantarfiexibility of an AFO on the gait, thus determining the optimum flexibility for each patient.

Present techniques involve several clinical problems. The first problem is that the experimental AFO employed in this study is too heavy and may not simulate plastic AFOs worn daily. To solve this problem, a new, lighterweight experimental AFO is being designed.

The second problem is the time and labor of the measurement. It took at least 90 minutes to complete the measurement for four levels of flexibility. Most of this time was consumed attaching measuring instruments to the subject and changing the coil springs. Time may be saved by selecting the most critical parameters to reduce the number of measuring instruments. In the experimental AFO, the coil springs housed in the Klenzak joints will be replaced by larger coil springs and placed outside the joints to ease handling. These improvements will help reduce the physical and mental burden on the subject and the number of people necessary for the measurement.

Quantitative measurement and subsequent statistical analysis may not always provide sufficient information to determine optimum flexibility. A good example was given in the case of subject B. He walked at maximum speed when using the AFO assembled with coil springs No. 2, but he also presented the maximum back knee at this level.

Careful examination of the waveform of the knee angle and the images recorded in the VTR revealed that this subject had an unstable knee, and the back knee occurred only at the last stage of the stance phase for a rather short duration. The moderate back knee occurred at flexibility levels of 0, 1 and 3. Intense clinical study must be made to assess how hazardous this maximum back knee at flexibility level 2 is the subject and to determine optimum flexibility.

When analyzing variance in this study, one-way allocation was assumed, and flexibility was treated as the only variation factor. Theoretically, a two-way allocation model should have been used in which flexibility and weight of the AFO were designated as variation factors. However, no essential difference in interpretation would have resulted because weight had the same effect on the mean of each parameter at flexibility levels 1, 2 and 3. In the case of flexibility level 0, i.e., PLS, it had a different effect since the isolated PLS is much lighter than the assembled AFO.

This neglect of the lightweight factor created bias in favor of the PLS because the gait with PLS would have been otherwise impeded by the additional weight. Still, the optimum occurred at flexibility level 1 or 2. Thus, the conclusion would have been the same.

The observation that maximum muscle moment in plantarflex direction varied significantly with the use of the AFO and its flexibility in both subjects is contrary to the common understanding of the function of AFOs. It is generally believed that the active ankle moments generated by paralyzed plantarflexors do not change considerably regardless of the use of AFOs because the active ankle moment is uniquely determined by the degree of paralysis of the muscles.

One of the main functions of AFOs is to supplement the decreased plantarflex moment late in the stance phase. However, the present measurement clearly demonstrated that the corrective supplementary moment generated by the AFO is very small compared with the total ankle moment and the net muscle moment did change significantly (see Figure 3 ). In particular, plantarflex muscle moment increased when subject A walked with the AFO.

In the case of subject B, maximum plantarflex muscle moment occurred when he walked without the AFO. However, considering the weight factor mentioned above, the maximum plantarflex moment might have occurred with the AFO at a flexibility level of 2 in this subject if the AFO had been much lighter.

Although we have not yet investigated the underlying mechanism(s) of this new observation, we have two working hypotheses. One possibility is that the subject activated maximum plantarflexor force because lateral stability was augmented by the AFO. The subject could have been relieved from a fear of lateral instability. The second possibility is that the activity level of the plantarflexors is determined by the dynamic matching between the mechanical factors and human pathological factors although the range of activity is basically limited by the underlying paralysis of the muscles.