In considering the wealth of investigation that has been done to evaluate the efficacy of AFOs, no population has received more scrutiny than the post-stroke hemiplegia patient population. Unfortunately, the presentations of the study subjects and the orthoses used in these investigations have historically been poorly described, making the value of such investigations to clinical orthotists and others in rehabilitation limited. Knowing the outcomes of poorly defined or controlled interventions on somewhat vaguely described patient populations does little to inform the specific design characteristics that will best benefit an individual with a defined physical presentation. Fortunately, several recent studies have better defined both the characteristics of the AFO interventions and the physical presentations of the study subjects, making their findings more relevant to clinical practitioners. This article focuses on some of these new studies.
A Safe Starting Point
For a population as diverse as the post-stroke hemiplegic population, generalizations are challenging. However, for the sake of a starting point, the majority of individuals within this population who benefit from the provision of an AFO may be characterized as having poor balance, hypertonicity in the affected lower limb, inappropriate and unstable foot and ankle posturing, and instability at the knee, often characterized by stancephase hyperextension. Among the ambulant, walking occurs at a slower velocity with a reduced cadence and is generally characterized by excessive plantarflexion in swing phase, which may be accompanied by compensatory circumduction or increased knee flexion to aid in swing-phase clearance. Systematic reviews confirm that AFOs, across their many variations, improve these basic gait issues; gait velocity and cadence improve, and swingphase equinus is decreased.1 The larger, more controversial question is whether specific design characteristics in the form of joint stops and/or assists, footwear modifications, or footplate techniques are more appropriate for certain presentations within this patient population.
Do patients presenting with variations in available ankle range of motion (ROM) respond differently to different AFO design characteristics? This important question, largely overlooked in previous studies, received critical evaluation in recent work done at the Rancho Los Amigos National Rehabilitation Center, Downey, California.2 In their study, the authors evaluated the effects of three different AFO types on two representative patient populations. The first patient population presented with ankle dorsiflexion mobility beyond 0 degrees, or neutral, with an extended knee (mean 1.2 degrees dorsiflexion). The second patient population presented with plantarflexion contractures between 10 and 15 degrees (mean 10.7 degrees plantarflexion with knee extension, 1.5 degrees dorsiflexion with knee flexion).
By dividing their study group into two distinct populations and eliminating those subjects who would have fallen between the two presentations (i.e., those with a dorsiflexion range between 0 and 10 degrees plantarflexion with a fully extended knee), some valuable insights were gained. First, across the various AFO types, patients with greater ranges of available ankle motion walked much faster than those with moderate plantarflexion contractures (0.597 m/s vs. 0.373 m/s).2 Interestingly, much more performance variability was observed in the group with available dorsiflexion than in the group with limited ROM. Specifically, mean velocity in the first group ranged from 0.528 m/s in a solid AFO set at 0 degrees (providing a tibial inclination of approximately 5 degrees with the heel lift of the shoe) to 0.647 m/s in an articulating AFO made from the same positive plaster model, with a plantarflexion stop set at 0 degrees and no restrictions on dorsiflexion—a difference of 22 percent.2
By contrast, there was much less variability in performance among the second group in which substantial ROM limitations were present. While mean velocities were slowest in the shoes only condition (0.352 m/s), the mean value in the fastest condition, the articulated AFO with a plantarflexion stop, was only 10 percent faster (at 0.388 m/s).2 The difference between the three AFO conditions was even smaller, with the slowest condition, the solid AFO (0.375 m/s), being only 3 percent slower than the articulated AFOs with plantarflexion stops, and only 0.5 percent slower than the third bracing condition in which the AFOs featured dorsiflexion assists through rubber grommets anterior to the ankle joints, dorsiflexion stops through posteriorly attached check straps, and no limitations to plantarflexion (0.0377 m/s).2 The take-home message here is that, assuming the ankle is cast in a neutral sagittal alignment (i.e., 0 degrees dorsiflexion), variations in available ROM in both dorsiflexion and plantarflexion appear to produce a more noticeable effect in patients with greater ankle ROM than in those with stiffer ankles. While this is a fairly intuitive statement, it may explain the lack of treatment effect reported by other AFO studies in this population. Treatment effects are much less likely to be observed in post-stroke hemiplegic patients with stiffer ankles.2
In the next study, the Rancho Los Amigos National Rehabilitation Center researchers investigated a cohort comprised of patients who had at least 10 degrees of passive dorsiflexion with an extended knee and little or no elevated tone. Among other considerations, the authors sought to quantify the effect of AFO ankle angle on knee-flexion angle during loading response.3 Using AFOs that could be fixed in a neutral alignment (defined as a vertical tibia when the foot was flat on the floor in the AFO and the shoe) with 5 degrees of relative dorsiflexion and 5 degrees of relative plantarflexion, the authors measured the resulting peak knee-flexion angles. Their observations confirmed that knee flexion during loading response increased with dorsiflexion and decreased with plantarflexion. However, the differences were generally small and failed to reach statistical significance. Their preliminary findings suggest considerable variability in the amount of knee flexion observed during loading response within this population that can be modestly affected by changes to the angle of the AFO. This supports the current clinical practice of assessing the strength and control of the knee extensors when determining the angle the orthosis. Interestingly, the majority of the study subjects chose to keep the study AFOs in the position of relative dorsiflexion, with a minority preferring the "neutral" alignment, and no subjects selecting the plantarflexed alignment. Unfortunately, the knee extensor strength of the study cohort was not assessed or reported.
In recent years, there has been a growing interest in evaluating and "tuning" the angle of the tibia that results from the combination of an AFO and accompanying footwear. While the related principles are widely discussed, very little in the way of published research has been put forward. For now, the effect of this "tuning" process among individuals with stroke-induced hemiplegia is limited to a single published case study in which a 61-year-old woman presents 15 months post stroke with a left solid AFO cast at 90 degrees.4 Using standard techniques in the gait lab, researchers verified the presence of hyperextension in mid-stance and terminal stance and that the ground reaction force (GRF) was passing excessively anteriorly. Posterior wedges were added to the patient's shoe until her shank was advanced to 14 degrees of forward inclination. This footwear modification reduced the mid-stance hyperextension but failed to resolve the hyperextension at the end of each gait cycle in terminal stance. At this point, a stiff rocker sole was added to the shoe, which, combined with the heel wedges, eliminated the knee hyperextension throughout the gait cycle by moving the GRF posterior through the knee.4 While the findings of case studies are always preliminary, this paper lends some objective documentation to support the underlying principle of affecting knee kinematics through footwear modifications.
Dorsiflexion-assist mechanisms are commonly recommended and provided for patients with stroke-induced hemiplegia, but are they actually strong enough to dorsiflex the hemiplegic foot in swing? This question was evaluated in a recent Thranhardt paper at the Annual Meeting & Scientific Symposium of the American Academy of Orthotists and Prosthetists (the Academy).5 In this study, the gait of 21 subjects an average of six years post stroke was analyzed in a shoes-only condition following two weeks of acclimation to a series of AFOs manufactured from the same positive plaster molds. Two of these AFO types were largely identical, being articulated with plantarflexion stops set at 90 degrees. However, one condition used dorsiflexion-assist joints and the other used standard joints. According to the gait lab observations, the median values for the ankle angle in mid swing and at initial contact were identical between the two conditions.5 Thus, at least for the observed study population, the inclusion of dorsiflexion-assist joints failed to increase swing-phase dorsiflexion values beyond those provided by the plantarflexion stop.5
Within orthotic practice, there are two general approaches to footplate length: the full footplate in which the plastic extends past the toes, and the ¾-length footplate in which the plastic is cut just proximal to the metatarsal heads. The former is advocated in part to prevent clawing of the toes over the edge of the orthosis, while the latter appears to allow more natural extension at the ball of the foot, facilitating an easier third rocker, and, of tremendous importance to some patients, taking up less space in the shoe. Do variations in footplate length ultimately affect the function of the AFO? This question, among several others, was pursued by research conducted at Northwestern University, Chicago, Illinois.6 In their study, 16 subjects with hemiplegia resulting from chronic stroke completed gait lab data collection in articulated AFOs with plantarflexion stops set at 0 degrees of tibial inclination within the shoe, with both the full- and ¾-length footplates.6 Acclimation periods of two weeks were provided in each condition prior to data collection.
The authors found that when incorporated as part of an articulated AFO, footplate length appears to affect several key gait variables. Among their observations were the mean values for peak dorsiflexion angles in stance in both AFO conditions, a shoe-only condition, and among age-matched, non-affected controls. This parameter was found to be significantly higher among controls than the study subjects in the no-AFO condition (approximately 14 and 9 degrees respectively). Of interest, the combination of the articulating AFO and full footplate produced a significant increase in median peak stance-phase dorsiflexion, presumably by impeding heel rise such that motion occurred at the second rocker rather than the third rocker of gait. When the same subjects were examined in the same AFOs following two weeks of acclimation to the shortened, ¾-length footplate, median peak dorsiflexion values in stance returned to the levels seen in the no-AFO condition.6 Similar trends were observed with respect to the median peak plantarflexion moments in stance. These were significantly higher in controls than the study subjects in the no-AFO condition. They increased significantly in the presence of the articulated AFO/ full-footplate condition and returned to those observed in the no-AFO condition when the footplate length was reduced.6
While statistical significance was observed, the clinical significance of these observations remains unclear though footplate length certainly appears to affect both the second and third rockers of gait in this population. Importantly, these observations were all associated with articulating AFOs. The impact of footplate length on solid AFO designs may be significantly different and requires its own evaluation.
The techniques used to modify the plantar surface of an AFO for hemiparetic patients following their strokes are quite variable. At one extreme, their plantar surface is relatively flat, representing the partial weight bearing curves of the foot against the floor. At the opposite extreme are the "tone-reducing" modifications, where specific modifications are applied to the plantar surface of the foot in an attempt to reduce spasticity of the limb. Citing the failure of prior studies to adequately separate the neurophysiologic effects (i.e., tone reducing) from biomechanical effects (i.e., orthoses), a research team from La Trobe University, Melbourne, Australia, attempted to isolate the effects of these tone-reducing footplate modifications on the excitability of the soleus muscles in individuals with post-stroke hemiplegia.7
The research team achieved its goal by measuring the ratios of reflex and response muscle amplitudes as a measurement of the reflex excitability or spasticity of the affected muscle across several conditions. These included baseline measurements prior to any interventions, and measurements in two articulated AFOs made from positive plaster models that were identical with the exception of the plantar-surface modifications. One AFO had the tone-reducing modifications within the footplate, while the other featured a more traditional, flat footplate. The muscle ratios used to measure spasticity across the various conditions found that, on average, the tone-reducing features had no effect on spasticity levels.7 In fact, for the two subjects who demonstrated a change in muscle excitability with the use of the tone-reducing AFOs, the effect was one of increased excitability, suggesting the muscles would be more prone to spasticity. Thus, the ability of these footplate modifications to reduce spasticity in post-stroke hemiplegia can be reasonably questioned.
While the basic benefits of AFOs for patients with hemiplegia due to stroke have been fairly well established, the specific design variables of the orthoses and presentation characteristics of the subjects have historically been poorly described and controlled. Fortunately, several recent publications have begun to delve into the details, providing insight into the potential effects of specific AFO design variables in more narrowly defined patient presentations. As this trend continues, the published literature will better guide clinicians as they create their treatment plans for individual patients.
Phil Stevens, MEd, CPO, FAAOP, is in clinical practice with Hanger Clinic, Salt Lake City, Utah. He can be reached at
- Guerra Padilla M., F. Molina Rueda, and I. M. Alguacil Diego. 2011. Effect of ankle-foot orthoses in postural control after stroke: a systematic review. Neurología.
- Mulroy S. J., V. J. Eberly, J. K. Gronely, W. Weiss, and C. J. Newsam. 2010. Effect of AFO design on walking after stroke: Impact of ankle plantarflexion contractures. Prosthetics and Orthotics International 34(3):277–92.
- Silver-Thorn B., A. Herrmann, T. Current, and J. McGuire. 2011. Effect of ankle orientation on heel loading and knee stability for post-stroke individuals wearing anklefoot orthoses. Prosthetics and Orthotics International 35(2):150–62.
- Jagadamma K. C., E. Owen, F. J. Coutts, J. Herman, J. Yirrell, T. H. Mercer, and M. L. Van Der Linden. 2010. The effects of tuning an ankle-foot orthosis footwear combination on kinematics and kinetics of the knee joint of an adult with hemiplegia. Prosthetics and Orthotics International 34(3):270–76.
- Fatone S., R. Stine, and S. Gard. 2010. Randomized crossover study of AFO ankle components in adults with poststroke hemiplegia. The Academy TODAY 6(4):A4–A5.
- Fatone S., S. A. Gard, and B. S. Malas. 2009. Effect of ankle-foot orthosis alignment and foot-plate length on the gait of adults with poststroke hemiplegia. Archives of Physical Medicine and Rehabilitation 90(5):810–18.
- Ibuki A., T. Bach, D. Rogers, and J. Bernhardt. 2010. An investigation of the neurophysiologic effects of tonereducing AFOs on reflex excitability in subjects with spasticity following stroke while standing. Prosthetics and Orthotics International 34(2):154–65.