Biomechanics of propulsion: Implications for AFOs

For many patients, the ability of an ankle foot orthosis to enhance propulsion is key to improving gait efficiency and reducing fatigue. But experts are only beginning to understand the biomechanical complexities that influence propulsion, which start with push-off but don’t end there.

By Cary Groner, Lower Extremity Review January 2016

In assessing and treating conditions that affect patients’ ability to walk, it makes sense for researchers and clinicians to pay attention to the propulsive aspect of the gait cycle. Increasingly, however, experts have concluded that much of what we thought we understood about propulsion is oversimplified or just plain wrong.

Moreover, interventions to improve propulsion, such as energy-returning carbon fiber ankle foot orthoses (AFOs) or functional electrical stimulation (FES), are far more effective in some patients than others—a variability that may have more to do with the individuals themselves than with their diagnoses. It’s crucial that clinicians be able to sort out such factors if they’re to help their patients, but often they don’t even agree on terminology, and assessing outcomes is a subjective art.

Disputed definitions

Propulsive force generation comprises two primary factors: ankle moment and the position of the center of pressure relative to the body’s center of mass.[1] These, in turn, can encompass a variety of other complex biomechanical factors including ankle dorsiflexion and plantar flexion, knee extension and flexion moments, timing and magnitude of activation of the gastrocnemius and other plantar flexors, trailing limb angle, and energy consumption.

Each of these factors plays a greater or lesser role depending on the condition under consideration. Poststroke hemiplegia patients have different propulsion issues than kids with cerebral palsy (CP), or teenagers with Charcot-Marie-Tooth disease (CMT), or adults with multiple sclerosis—and therefore will require different therapeutic approaches. Disagreements over terminology and biomechanics aren’t just theoretical—clinicians will have a hard time optimizing propulsion if they don’t understand its complexities—and a few experts have recently begun an effort to clarify matters.

“People want the spring AFO to mimic the calf muscle, but then there needs to be clarity about what the calf muscle actually does and when it does it,” said Elaine Owen, MSc, MCSP, a pediatric physical therapist at the Child Development Center in Bangor, North Wales, UK.

Clarity, however, can be hard to come by in this area, for a number of reasons.

“Push-off and propulsion may mean something different to a podiatrist than to someone working in a three-D gait lab,” Owen continued.

Owen has recently traced part of the problem to a statement within an important text on gait analysis, David Winter’s 1991 Biomechanics and Motor Control of Human Gait.[2]

“In that book, there’s a graph depicting the gait cycle that, at one phase, shows decreased dorsiflexion that’s described as plantar flexing. It seems that people ever since have passed down that semantic confusion,” Owen explained.

Owen added that further problems can result when researchers and clinicians fail to distinguish terminal stance from pre-swing. In terminal stance, she explained, one limb is contacting the ground and the ankle remains in dorsiflexion; in preswing, which follows terminal stance and technically constitutes push-off, both limbs are contacting the ground and the trailing ankle is moving from dorsiflexion to plantar flexion.

“It’s a big problem in terms of developing rehabilitation strategies and orthosis designs,” she said. “We can’t put ‘normal’ back into pathological gait if we aren’t clear exactly what normal is.”

Owen is a proponent of the gait cycle descriptions published by Jacqueline Perry, MD, of Rancho Los Amigos National Rehabilitation Center in Downey, CA.[3] These include loading response, midstance, terminal stance, and preswing, as well as the point at which true ankle plantar flexion begins.

Despite this reliable source, misunderstandings about propulsion persist, and there is also disagreement about the way push-off occurs.

“There’s an argument that the lengthening muscle in terminal stance acts like a spring tensioner on the [associated] tendon,” Owen said. “When the muscle finally begins to shorten, some of the push-off may actually come from the release of the tendon.”

Evidence exists for this view: Japanese researchers have elucidated a mechanism by which the gastrocnemius medialis tendon may act as a spring from the beginning of single-limb support to toe-off, increasing walking efficiency.[4]

Power & prejudice

Photo courtesy of Kinetic Research.

Increasingly, researchers and clinicians are looking beyond push-off itself and exploring how other phases of the gait cycle contribute to propulsion. Owen said that experts see power generation in gait as important for two primary reasons.

“One is that something has to propel the lower limb into swing phase,” she said. “The other is that it’s important for general energy and momentum within the gait cycle, and for moving the trunk.”

One question that’s plagued researchers, she noted, is whether ankle plantar flexion and knee flexion in preswing merely propel the swing limb, or whether they’re also involved in maintaining trunk momentum. In 2014, findings presented during a Thranhardt Lecture at the annual meeting of the American Academy of Orthotists & Prosthetists supported the first view. The Taiwanese authors offered evidence that the preswing movement of the thigh required power from the hip in addition to power from the ankle, which was consumed by the shank, knee, and thigh.[5]

“They redid the calculations and said that a lot more of the force for propelling the trunk through the gait cycle came from loading response early in the cycle rather than later,” Owen explained. “When you make initial contact, the heel lever forces the foot to the floor and the shank is pulled forward from reclined to vertical; I think it’s massively propulsive, almost like a catapult. If we apply that theory to orthoses, we have to get first rocker and entry into midstance correct.”

Owen has arrived at her own definition of push-off. “I’d say it’s from the time the calf muscles start shortening—forty to forty-five percent through the cycle—through toe-off,” she said. “But that’s with the understanding that in terminal stance, the calf muscles are shortening and the ankle is reducing dorsiflexion. Then, in preswing, the ankle moves from dorsiflexion to plantar flexion.”

Photo courtesy of Cascade DAFO

Bryan Malas, CO, MHPE, director of orthotics/prosthetics at the Ann & Robert H. Lurie Children’s Hospital of Chicago, and assistant professor of physical medicine and rehabilitation at Northwestern University in Chicago, agreed that clinicians need to develop and incorporate a more nuanced view of the way power flows through the propulsive gait cycle.

“Most people focus on the large power burst at the ankle during preswing, but there are smaller bursts as well,” Malas said. “We see one at late loading, when the hip starts to extend; we see one at midstance as the knee extends; and then there’s one at the initiation of swing, when hip flexion occurs. It’s important to understand this, because if we provide an orthosis, we have to be clear what we want it to do.”

Practical considerations

Malas added that the propulsive power generated in the gait cycle is usable primarily because it’s transmitted by mechanisms of the foot—the windlass mechanism, the calcaneal-cuboid locking mechanism, plantar flexion of the first ray, and so forth. If those capa­bilities are compromised, and the foot’s leverage diminished, propulsion suffers—and studies of partial-foot amputees have clearly shown the effect of reduced leverage on propulsive power.[6]

“We see this a lot in kids who have cerebral palsy,” Malas said. “They tend to have a hyperpronated foot and they’ve lost that anterior lever arm, so they really have no push-off. It’s hard for them to take a full step because they’re standing on a compromised base. To compensate they have to use the contralateral limb.”

Owen noted that kids with CP often have issues with the timing of the propulsive phase of the cycle as well.

“They may get the knee extension they need in terminal stance, but they couple this with plantar flexion, when in normal gait there is dorsiflexion,” she said. “We’ll use fixed-ankle AFOs, or plantar flexion-stop AFOs, to keep them from doing that so they get a good stretch on the calf muscle. You get more of a normal gait, and a normal terminal stance.”

Children with spina bifida can’t do any of that, she said, because their lack of calf muscle leads to excessive dorsiflexion in terminal stance. Nevertheless, in some cases the orthotic strategy may be similar to that for kids with CP.

Photo courtesy of Surestep.

“We use AFOs, often fixed ankle with a plantar flexion stop, whether they’re weak and going into excessive dorsiflexion without heel rise, or whether they’re stiff and going into early plantar flexion,” she said. “People sometimes say we’re preventing push-off, but the kids can pull off in preswing, the way a normal limb takes itself through the gait cycle. The hip flexors compensate, and the gains you make in hip and knee extension in terminal stance far outweigh the loss of that five percent of the gait cycle involving an active push.”

“Kids with spina bifida have no plantar flexor strength and tend to walk in a crouched gait,” added Malas. “They’re in no position for any sort of propulsion because the knee is not extended, so they show more hip flexion to try to get the limb from stance phase into swing phase.”

Owen noted that some experts object to the fixed-AFO strategy in children with CP because they feel that it reduces the work of the muscles involved in propulsion, potentially weakening them.

“That’s one reason people want them to have the springy devices, or plantar flexion-free devices, that will allow them to use the calf muscle,” she said. “However, some of these devices spring back into plantar flexion during terminal stance, and then people say that’s OK, in normal gait there’s plantar flexion in terminal stance—but there isn’t! If you start to plantar flex the foot as the knee extends in terminal stance, you lose the calf muscle stretch, which is therapeutic in kids with CP.”

Propulsion—as well as stability—may also be compromised if an AFO is not properly aligned, Malas said.

“If one of our objectives is to have the AFO share the work and create a relatively stable base during single-limb support, then some loading of the orthosis is required,” he said. “If we want it to provide enough stability in single-limb support for the person to feel safe taking a step with the contralateral side, then loading through the orthosis should occur between midstance and terminal stance. However, this may not happen if a solid AFO, set with an ankle angle of ninety degrees, is placed into a shoe with a heel-sole differential that inclines the AFO too far forward. In that case, loading of the orthosis may not occur until late terminal stance or early preswing, and then it may be too late to create enough single-limb stability. The result is decreased step length, cadence, and speed.”

A 2012 study by researchers in Amsterdam reported that energy-returning carbon fiber AFOs helped decrease the energy cost of walking by 9.8% in patients with multiple sclerosis, but not due to an augmented net ankle push-off. Rather, the effect was because the AFO simply took over 60% of the ankle work.[7]

“In that study, the average ankle angle of the AFO was around five degrees of plantar flexion,” said Malas. “So you’re giving the AFO a chance to load the orthosis as the person advances the tibia over the foot. I think that’s one reason there’s so much variation in study conclusions; it’s often not clear what the ankle alignment is in the AFO.”

For Malas, such findings raise bigger questions.

“Is the AFO working because it helps with power generation, or because of the increased stability it provides?” he asked. “I don’t think we know, but if people don’t feel stable, they shorten their step length.”

Knees and ankles

Dutch researchers raised similar points in a recently published paper.[8] They suggested that gait efficiency in children with spastic CP could ideally be maximized by optimizing the tradeoff between the enhanced push-off power (from spring-like AFOs) and normalized knee flexion in stance (via rigid AFOs). They assessed kids with ventral shell spring-hinged AFOs, with the hinge set to a rigid, stiff, or flexible setting; the spring-like properties were eliminated in the rigid version. Ultimately they concluded that, in terms of energy cost, an AFO’s effect on knee kinetics and kinematics may be more important than its effect on push-off.

Lead author Yvette Kerkum, PhD, who coauthored the paper as part of her doctorate at Vrije University Medical Center in Amsterdam, told LER the findings are more descriptive than prescriptive.

“The results varied a lot between patients despite the fact that we included a homogeneous group of children,” she said. “Some children responded best to rigid devices, others to more flexible ones. Ankle power was preserved by the more flexible AFOs, but that didn’t seem critical to energy costs. I think the main problem is that you can preserve the child’s remaining push-off, but that’s still only half of normal. If you really want to make a difference in energy costs by affecting ankle power, you have to add power.”

Individual attention

Photo courtesy of Allard USA.

In any case, lower extremity clinicians emphasize careful assessment of each patient.

“Your strategy depends on what outcomes you’re trying to achieve,” Elaine Owen said. “For example if you’re treating an adult after a stroke—an acquired disability—the patient will have a normal, mature skeleton and you’ll make different decisions than you would for a child who’s skeletally immature, growing and trying to get to age sixteen with minimal surgeries.”

Karen Nolan, PhD, a research scientist at the Kessler Foundation in West Orange, NJ, had a similar take.

“Prescription decisions are complicated,” she said. “We need orthotic options that match the abilities of each patient but that function in terms of long-term rehabilitation goals. I’m not sure anyone has truly connected all the dots between propulsion, regaining and maintaining function, and increasing stability.”

Nolan, who often deals with patients who have poststroke hemiplegia, added that the patient’s desires are crucial to the process.

“The patient may choose a device that’s not in line with what would be to their best advantage mechanically,” she said. “That doesn’t mean they’re making a mistake. It’s based on what’s important to them, what most enhances their quality of life. Their primary goal may not be to increase walking speed.”

Energy-return systems seem to offer particular promise in patients with CMT, according to experts.

“Ideally you’ll combine propulsion with stability, so the person feels comfortable in single-limb support,” said Bryan Malas. “We’ve put younger kids with CMT into energy-storing designs and seen them run for the first time. If we’re loading the system and slowing down the tibia, I think that’s having a more proximal effect—allowing the thigh to come over the knee and getting that limb closer to knee extension. The ankle is in controlled dorsiflexion, and all of these are prerequisites for limb advancement on the contralateral side.”

For David Misener, CPO, who practices in Albany, NY, discussions of CMT are personal; he has the condition, as does his son, and he’s found significant benefits from energy-return AFOs.

“If someone with CMT doesn’t have enough strength to get up on their toes, they’re missing the propulsive stage of gait, and that’s where an AFO makes a tremendous difference,” he said. “You’re constantly working your plantar flexors and dorsiflexors just to keep your balance, and when you have CMT and they’re not working to capacity, you have to use more proximal muscles. All those little bits of energy add up. After switching to energy-storing AFOs, I have much more energy at the end of the day, and my patients have told me the same thing.”

Research supports Misener’s experience. For example, a 2014 study in Gait & Posture found that CMT patients walked faster in custom AFOs versus unbraced.[9] Another found that three AFOs reduced foot drop and removed the need for some proximal compensatory actions in CMT patients, and that two of the three AFOs did so with no decrease in peak ankle power generation.[10]


In some patients, particularly adults with poststroke hemiplegia, researchers are exploring FES as a way to augment propulsive gait. Nolan and her colleagues at the Kessler Foundation have investigated dorsiflexion-assist FES devices and reported results equivalent to those obtained with AFOs.[11]

At the University of Delaware in Newark, researchers have been exploring FES devices that stimulate both plantar flexors and dorsiflexors, and reported they help correct poststroke gait deficits at the ankle and the knee during both swing and stance phases of gait.[12] They have also added fast treadmill walking to the FES approach and reported the combination improved gait parameters including peak anterior ground reaction force, and trailing limb angle.[13]

Lead author Trisha Kesar, PT, PhD, told LER the team was interested because propulsion affects gait speed, which is an important rehabilitation target in stroke patients.[14]

“Able-bodied people get fifty percent of their propulsion from each leg, whereas those who’ve had a stroke might get as little as five percent from the paretic side,” said Kesar, now an assistant professor of physical therapy at the Emory University School of Medicine in Atlanta.

Propulsive gains appeared relatively minor, however—propulsion in the paretic limb rose from 28.8% to 33.1% of the total—raising questions about clinical efficacy.[12]

“It was statistically significant, but was it meaningful; did the patients perceive a difference?” Kesar wondered.

Some did and, some didn’t, as it turned out.

“In the short [test] span of thirty seconds to a minute, in one session, it was only perceptible to them if they were focused and engaged,” she continued. “In the long-term study just being published,[15] where subjects had three sessions a week for twelve weeks, results are much more marked in terms of what people perceive.”

The studies have produced some unexpected results, Kesar added. For example, the team found that dorsiflexion-assist FES alone worsened swing-phase knee flexion, whereas adding plantar flexion FES overcame that effect and brought things back to normal. The researchers are still investigating what caused those effects.

“As of now, the plantar stimulation is more of a motor learning tool,” she said. “We use it to teach the nervous system how to use muscles after a stroke, not as a neuro-prosthetic, which is an important difference between our intervention and commercial stimulators. Hopefully, if you practice enough, and the stim teaches you correctly, you’ll walk outside with more propulsion and decreased foot drop, without needing the stim.”

Even so, Kesar acknowledged that, eventually, she’d like something patients could use on an ongoing basis.

“If there were an over-the-counter device for plantar- and dorsiflexor stimulation, I’d send my patients home with it,” she said. “They wouldn’t be limited to the clinic rehab and could get much more practice in that good-quality movement. The technology is there; now people are just working to make it portable.”

Cary Groner is a freelance writer in the San Francisco Bay Area.

Source Lower Extremity Review

  1. Mechanisms to increase propulsive force for individuals poststroke, Hsiao H, Knarr BA, Higginson JS, Binder-Macleod SA. J Neuroeng Rehabil. 2015 Apr 18;12:40. doi: 10.1186/s12984-015-0030-8. Full text, PDF
  2. The biomechanics and motor control of human gait: Normal, elderly and pathological, Winter DA. University of Waterloo, Canada: Waterloo Biomechanics; 1991.
  3. Gait analysis: Normal and pathological function, 2nd ed., Perry J, Burnfield J. Thorofare, NJ: Slack Inc.
  4. In vivo behaviour of human muscle tendon during walking, Fukunaga T, Kubo K, Kawakami Y, Fukashiro S, Kanehisa H, Maganaris CN. Proc Biol Sci. 2001 Feb 7;268(1464):229-33. doi: 10.1098/rspb.2000.1361. Full text, PDF
  5. Ankle push-off and mechanical energy flow in human gait, Chen HB, Chang LW. Presented at the American Academy of Orthotists and Prosthetists annual meeting, Chicago, February 2014. PDF
  6. Deliberations about the functional benefits and complications of partial foot amputation: do we pay heed to the purported benefits at the expense of minimizing complications? Dillon MP, Fatone S. Arch Phys Med Rehabil. 2013 Aug;94(8):1429-35. doi: 10.1016/j.apmr.2013.03.023. Epub 2013 Apr 10.
  7. Spring-like Ankle Foot Orthoses reduce the energy cost of walking by taking over ankle work, Bregman DJ, Harlaar J, Meskers CG, de Groot V. Gait Posture. 2012 Jan;35(1):148-53. doi: 10.1016/j.gaitpost.2011.08.026. Epub 2011 Nov 1. Full text, PDF
  8. The Effects of Varying Ankle Foot Orthosis Stiffness on Gait in Children with Spastic Cerebral Palsy Who Walk with Excessive Knee Flexion, Kerkum YL, Buizer AI, van den Noort JC, Becher JG, Harlaar J, Brehm MA. PLoS One. 2015 Nov 23;10(11):e0142878. doi: 10.1371/journal.pone.0142878. Full text, PDF
  9. Functional and dynamic response characteristics of a custom composite ankle foot orthosis for Charcot-Marie-Tooth patients, Dufek JS, Neumann ES, Hawkins MC, O’Toole B. Gait Posture. 2014;39(1):308-13. doi: 10.1016/j.gaitpost.2013.07.121. Epub 2013 Aug 3.
  10. Foot drop splints improve proximal as well as distal leg control during gait in Charcot-Marie-Tooth disease, Ramdharry GM, Day BL, Reilly MM, Marsden JF. Muscle Nerve. 2012 Oct;46(4):512-9. doi: 10.1002/mus.23348.
  11. Long-Term Follow-up to a Randomized Controlled Trial Comparing Peroneal Nerve Functional Electrical Stimulation to an Ankle Foot Orthosis for Patients With Chronic Stroke, Bethoux F, Rogers HL, Nolan KJ, Abrams GM, Annaswamy T, Brandstater M, Browne B, Burnfield JM, Feng W, Freed MJ, Geis C, Greenberg J, Gudesblatt M, Ikramuddin F, Jayaraman A, Kautz SA, Lutsep HL, Madhavan S, Meilahn J, Pease WS, Rao N, Seetharama S, Sethi P, Turk MA, Wallis RA, Kufta C. Neurorehabil Neural Repair. 2015 Nov-Dec;29(10):911-22. doi: 10.1177/1545968315570325. Epub 2015 Feb 4. Full text, PDF
  12. Functional electrical stimulation of ankle plantarflexor and dorsiflexor muscles: effects on poststroke gait, Kesar TM, Perumal R, Reisman DS, Jancosko A, Rudolph KS, Higginson JS, Binder-Macleod SA. Stroke. 2009 Dec;40(12):3821-7. doi: 10.1161/STROKEAHA.109.560375. Epub 2009 Oct 15. Full text, PDF
  13. Combined effects of fast treadmill walking and functional electrical stimulation on post-stroke gait, Kesar TM, Reisman DS, Perumal R, Jancosko AM, Higginson JS, Rudolph KS, Binder-Macleod SA. Gait Posture. 2011 Feb;33(2):309-13. doi: 10.1016/j.gaitpost.2010.11.019. Epub 2010 Dec 22. Full text, PDF
  14. Self-selected gait speed: A critical clinical outcome, Braden H. LER 2012;4(11):43-48.
  15. Reducing The Cost of Transport and Increasing Walking Distance After Stroke: A Randomized Controlled Trial on Fast Locomotor Training Combined With Functional Electrical Stimulation, Awad LN, Reisman DS, Pohlig RT, Binder-Macleod SA. Neurorehabil Neural Repair. 2016 Aug;30(7):661-70. doi: 10.1177/1545968315619696. Epub 2015 Nov 30. Full text, PDF
  Further reading

The mechanisms and mechanical energy of human gait initiation from the lower-limb joint level perspective, Zhao G, Grimmer M, Seyfarth A. Sci Rep. 2021 Nov 18;11(1):22473. doi: 10.1038/s41598-021-01694-5. Full text, PDF

A unified perspective on ankle push-off in human walking, Zelik KE, Adamczyk PG. J Exp Biol. 2016 Dec 1;219(Pt 23):3676-3683. doi: 10.1242/jeb.140376. Full text


The importance of being earnest about shank and thigh kinematics especially when using ankle-foot orthoses, Owen E. Prosthet Orthot Int. 2010 Sep;34(3):254-69. doi: 10.3109/03093646.2010.485597. Full text, PDF

Mobility Menu

follow us in feedly

Call 403-240-9100