The impact of ankle–foot orthoses on toe clearance strategy in hemiparetic gait: a cross-sectional study

Impaired paretic limb advancement is a clearly observable manifestation of gait pathology in individuals with hemiparesis due to stroke [1‐3]. Previous studies have reported specific gait changes following hemiparesis, such as decreased knee flexion, hip flexion, and ankle dorsiflexion during the swing phase, which can negatively influence the achievement of toe clearance [1‐6]. Reduction in toe clearance of the affected limb leads to tripping while walking, which is a major cause of falls [7, 8]. In healthy individuals, toe clearance is mainly achieved by limb shortening, which is affected by hip flexion, knee flexion, and ankle dorsiflexion. On the other hand, to obtain sufficient toe clearance during the swing phase, individuals with hemiparesis often require compensatory strategies that modify the kinematic pattern, including hip hiking and circumduction, which are common gait deviations [3, 9]. These changes during the swing phase have a reciprocal relationship. When the limb shortening is reduced due to paresis, the compensatory movements will be increased to contribute to toe clearance; hence, they are in a trade-off relationship [10].

Ankle–foot orthoses (AFOs) are frequently prescribed to improve walking ability in hemiparetic patients, as they provide passive or dynamic support of ankle movement. AFOs provide support not only during the stance phase of gait by encouraging lateral stability or improving early stance knee moments, but also in the swing phase to maintain ankle dorsiflexion and facilitate toe clearance [11‐17]. The effect of AFOs on the swing phase is additionally reflected in the compensatory movements. Cruz et al. [18] demonstrated that the compensatory pelvic obliquity observed in response to impaired ankle dorsiflexion in hemiplegic patients was minimized when the patients wore an AFO. Improved joint motions and decreased compensatory movement when using AFOs could potentially contribute to an efficient gait and promote walking activity in hemiparetic patients.

Clarification of the mechanical effect of AFOs on these gait parameters, and quantifications of compensatory movements would be helpful for clinical decision-making in rehabilitation clinics. For example, understanding the influence of rehabilitative training and the use of AFOs on gait indices (i.e., ankle angle, knee angle, hip elevation, or toe clearance) would help to determine the best rehabilitative strategy and to identify the need for AFO use in individual patients.

The aim of this study was to clarify the mechanical effect of AFOs and to quantify the impact of AFO use on hemiparetic gait pattern during the swing phase, as this information would be helpful for clinical decision-making in rehabilitation clinics. For example, understanding the influence of rehabilitative training and the AFO and its types on gait indices (i.e., ankle angle, knee angle, hip elevation, or toe clearance) would help to determine the best rehabilitative strategy and to investigate the need for AFO use in individual patients. Based on a prior study showing the relationship between limb shortening and compensatory movements [10], we hypothesized that the AFOs would positively affect functional limb shortening in a way that would consequently impact on toe clearance and compensatory maneuvers, particularly represented by hip elevation. Previous studies have shown the effects of AFOs and a relationship between limb shortening and compensatory movements. In the normal gait pattern, functional limb shortening (representing lower limb joint movement) is a main strategy for toe clearance. However, patients with hemiparesis have impaired lower limb function, and thus require compensatory strategies (e.g., hip hiking, circumduction of the paretic limb) to promote swing phase propulsion [19, 20]. Additionally, the extent of toe clearance is mainly determined by the extent of functional limb shortening and hip elevation as compensatory movements, which are in a trade-off relationship [10]. AFO usage reduces the gait pattern deviation and increases the walking ability, thereby reducing energy costs [21, 22]. In this study, we hypothesized that the AFOs would positively affect functional limb shortening in a way that would consequently impact on toe clearance and compensatory maneuvers, particularly represented by hip elevation. To determine the actual impact of limb shortening and compensatory movements on toe clearance, the vertical component of the movements that comprised toe clearance was calculated using three-dimensional kinematic motion analysis. The changes in joint angles were also investigated.

Patient characteristics and comfortable self-selected speeds used at walking trial both with and without AFO in each participant are presented in Table 1.

Representative trajectories of toe clearance, shortening of hip-toe distance, and paretic hip elevation due to pelvic obliquity are presented in Fig. 6.

The gait indices (mid-swing–mid-stance) are summarized in Fig. 7. The toe clearance was significantly increased by the use of an AFO (Fig. 7a: 33.9 ± 19.6 vs. 37.6 ± 16.9, p = 0.038). The toe clearance could be divided into two parts: the limb shortening and compensatory movements. The compensatory movements were significantly decreased by the use of the AFO (Fig. 7b; 33.0 ± 19.4 vs. 28.1 ± 17.2: p = 0.001), whereas the limb shortening was increased (Fig. 7c; 2.2 ± 25.6 vs. 10.3 ± 23.7: p < 0.001). Among the three components of compensatory movements (pelvic obliquity on paretic side, non-paretic hip elevation, and foot elevation due to circumduction), a significant difference was noted with and without the AFO for the pelvic obliquity on the paretic side (19.8 ± 18.9 vs. 16.4 ± 19.7, p = 0.003), and a weak tendency was noted for a reduction in the non-paretic hip elevation by the AFO (12.6 ± 17.4 vs. 11.2 ± 18.6, p = 0.234). No difference was observed in the foot elevation due to circumduction (Fig. 7b’). The limb lengthening by ankle movement was significantly decreased by the use of AFO (− 13.6 ± 7.4 vs. -7.6 ± 5.4, p < 0.0001), contributing to limb shortening. No significant change was observed in limb shortening due to knee movement (15.8 ± 25.3 vs. 17.9 ± 23.6, p = 0.183; Fig. 7c’).

The joint angle changes on the paretic limb in walking trials performed with and without AFOs was shown in Fig. 8. There were significant quantitative differences between trials performed with AFOs and those performed without AFOs in the decreases in hip flexion (13.6 ± 6.9 (with) vs 14.5 ± 7.3 (without), p = 0.019), knee flexion (25.7 ± 14.3 (with) vs 29.1 ± 15.9 (without), p = 0.001), and ankle plantar flexion (6.1 ± 5.4 (with) vs 10.4 ± 5.4 (without),p = 0.007) during the swing phase.

The subgroup analyses regarding the effect of AFOs on kinematic components revealed no significant differences in any of the measured indices in subacute vs chronic stroke patients, or in tAFOs vs APS-AFOs.

The present study confirmed the positive influence of AFOs on functional limb shortening, with resulting improvements in toe clearance and compensatory movement. The average toe elevation in hemiparetic gait was enhanced by the use of AFOs, irrespective of the stage of hemiparesis and/or the type of AFO used. Using an AFO resulted in an increase in ankle dorsiflexion during the swing phase and a subsequent decrease in compensatory movements, which was consistent with previous studies [17, 18, 34]. Furthermore, the present findings manifested the simple relationship between the changes in limb shortening and compensatory movements to achieve toe clearance, by breaking down each movement into the actual impact on toe clearance. As shown in this study, the toe clearance was increased by the use of AFOs (33.9 vs. 37.6, P = 0.038), due to the increase in vertical gain by limb shortening (2.2 vs. 10.3, p < 0.001). The compensatory movements account for the rest. This type of visualization could be useful for monitoring the beneficial effects of AFOs.

Improved joint motions and decreased compensatory movement when using AFOs could potentially contribute to efficient gait and promote walking activity in hemiparetic patients.

The degree of limb shortening is mainly determined by the degree of knee flexion, especially in the healthy subjects or patients with mild paresis [10]. Conversely, AFOs achieve limb shortening due to the mechanical property that limits ankle plantarflexion, which could be referred to as another compensatory approach to improve limb shortening. The present and previous findings indicate that toe clearance in hemiparetic patients with the use of AFOs could be achieved by the following mechanism. First, decreased knee flexion in hemiparetic gait may be compensated for by the AFO’s effect on decreasing ankle plantar flexion and subsequently maintaining the optimal hip-toe distance. Second, hip hiking (as a compensatory gait pattern) is consequently minimized because of acquired optimal limb shortening, as shown in the present study. The interrelationship of relevant gait indices are hypothesized as shown in Fig. 9.

Decreased knee flexion may be considered a negative effect of AFOs on the hemiparetic gait. The possible explanations for this include the following: First, ankle plantar flexion that has been decreased by AFO decreases limb lengthening, so the patient might not need to flex the knee as much to shorten the limb, because the same toe clearance can be achieved with less effort. In fact, in the present study, the use of AFO decreased limb lengthening due to ankle plantar flexion (Fig. 7c). In addition, the weight of the orthosis might influence knee flexion by making it more difficult for the paretic limb to flex the knee against gravity. Another possibility is the discrepancy between the knee-flexion angle and the effect of the knee flexion, which is seen when the knee-flexion angle is very small. In this case, slight knee flexion does not merit the toe clearance. For example, if the hip-knee-toe markers were aligned on the vertical line (Fig. 10a) and the ankle angle was fixed, the knee flexion could lengthen the hip-toe distance (Fig. 10b). In this situation, a discrepancy would arise between the knee-flexion angle and the limb shortening; a small knee flexion may not have a beneficial effect on limb shortening. The present study identified a conflict between the effect of AFO on reducing knee flexion and the relative increase in limb shortening due to knee movement (Figs. 7c and 8), which might have been influenced by this paradoxical relationship between joint angle and limb shortening.

In addition, the effect of AFO use on knee angles itself could have influenced by the different characteristics of the participant groups. A previous study showed an increase in knee flexion by AFO use in hemiplegic patients; Gatti et al. reported a positive effect of AFO use on knee flexion angle (26.316 without AFO vs. 30.785 with AFO) in the patient group with paresis scores of 4–5 on the Scandinavian Stroke Scale, which is milder paresis than in the present study [35].The effect of AFO use on knee movements should be further investigated, including whether the reduction in knee flexion has a negative impact on the recovery of knee movement.

The use of an AFO had a weak tendency to reduce the elevation of the non-paretic limb. Non-paretic hip elevation may occur in the following situations. First, it can occur if there is unusual elevation of the pelvis at the mid-swing of the paretic limb. This can be achieved either by ankle plantar flexion or by over-extension of the knee at the mid-stance of the non-paretic limb. However, the data did not evidence either of these causes.

Second, this can occur when the vertical position of the pelvis is lowered at the mid-stance of the paretic limb and elevated during the swing phase of the paretic limb. The lowered vertical position of the pelvis can occur due to the knee flexion of the paretic limb or the drop of the non-paretic side of the pelvis (Trendelenburg’s sign). In this case, the non-paretic hip drop showed significant correlation with the extent of the non-paretic hip elevation, while paretic knee flexion did not, indicating that the lowered pelvic position caused by the drop of the non-paretic side of the pelvis during the stance phase of the paretic limb was the main reason for the non-paretic hip elevation during the swing phase of the paretic limb. There was a weak, non-significant tendency of this drop of the pelvis to be decreased by the use of AFO (Additional file 4: Figure S3). The AFO effect on this non-paretic limb elevation must be confirmed in studies that have larger samples.

The present study clarified the real clinical picture of AFO-affected gait patterns, contributing a deepened understanding of the systematically balanced relationship between the enhancement of paretic limb clearance and the ability to reduce compensatory movement during the swing phase. Recognition and quantification of the interrelationships among the changes in joint kinematics, joint displacement, functional limb shortening, and the actual impacts on toe clearance would contribute to the comprehensive understanding of the effects of intervention and help clinicians to develop strategies to guide patients toward improving the advancement of the paretic limb during the swing phase. For example, when toe clearance is insufficient due to paresis, clinicians could consider facilitating improvements in the ability to effect limb shortening; otherwise, they could prepare AFOs for that purpose. If the effects are insufficient, compensatory movements could be encouraged. Conversely, once the clinician can confirm that the patient has gained sufficient toe clearance, the focus of the intervention could be redirected to the reduction of compensatory movements. With the use of gait analysis systems, we can quantify the actual impact and monitor the effect of each intervention and set goals for the patient. This will require further quantitative investigations in future studies for development, but this structured quantification may enable tailor-made rehabilitation training for individual patients. In addition to facilitating toe clearance in the swing phase, AFOs assist in other aspects of gait such as gait stability during the stance phase. Further analysis of the effect of AFOs should be encouraged to further the comprehensive understanding of the holistic effect of AFO use.

This study had a number of limitations. First, all included patients were able to independently ambulate on the treadmill without gait aids or orthoses. The findings may not be representative of those with severe hemiparesis. Further analysis of patients with more impaired gait abilty would facilitate further understanding of the effect of AFOs. Second, there was a relatively small sample population with heterogeneity. However, subgroup analyses found that the type of AFO and stage of hemiparesis did not significantly affect the kinematic parameters. Third, there was a small number of markers compared with commonly-used marker sets such as plug-in gait and Helen Hayes [36, 37], and the joint angles were indirectly estimated from the positions of the joint markers in the sagittal plane. We consider that this would not be critical enough to negate the present results, as the validity of measuring abnormal gait patterns, including the degree of joint angle movements, using this simple marker set has been shown previously [29, 38‐40]. However, the limitations of this simplified method should be further confirmed in future studies.

The present study quantified the impact of AFOs on toe clearance strategy in hemiparetic patients, and investigated the interrelationships between movement parameters. The results offer an insight into the effects of AFOs on the overall biomechanism of hemiparetic gait during the swing phase. Furthermore, these attempts to quantify and visualize the gait strategy could contribute to the clinical judgment and decision-making regarding rehabilitation strategy, which will help achieve better outcomes in rehabilitation practice.