Quantifying the Spring-Like Properties of Ankle-Foot Orthoses (AFOs) : JPO: Journal of Prosthetics and Orthotics

Secondary Logo

Journal Logo


Quantifying the Spring-Like Properties of Ankle-Foot Orthoses (AFOs)

Novacheck, Tom F. MD; Beattie, Cammie PT; Rozumalski, Adam MS; Gent, George CO; Kroll, Gary BOCO, CO, RTO

Author Information
JPO Journal of Prosthetics and Orthotics: October 2007 - Volume 19 - Issue 4 - p 98-103
doi: 10.1097/JPO.0b013e31812e555e
  • Free

Some children with cerebral palsy and most children with myelomeningocele have some degree of weakness of the ankle plantarflexors (gastrocnemius and soleus). Typically the ankle plantarflexors contract eccentrically to restrict dorsiflexion and tibial advancement in second rocker (midstance phase of the gait cycle)1,2 and concentrically to provide approximately 50% of the power for push-off in third rocker (just before toe-off).2,3

When plantarflexors fail to function adequately in midstance, excessive dorsiflexion results and is accompanied by excessive knee flexion, or “crouch gait.”1,2,4 When the activity of the gastrocnemius is inadequate, insufficient power for push-off results in decreased clearance in swing, reduced stride length, and decreased walking speed.5

For many children with neurologic dysfunction, it is desirable to use an ankle-foot orthosis (AFO), which is stiff enough to provide support in midstance, yet flexible enough to allow energy return in third rocker, or push-off. A posterior leaf spring (PLS) style AFO is often prescribed for this purpose and may be fabricated in a variety of designs. A PLS AFO consists of a calf cuff that tapers to a narrow band behind the ankle (the “leaf”) and widens back out to capture the heel and extend to the tip of the toes. The material used, the width of the ankle area, the arc of the leaf (defined as the curved contour in the leaf spring portion of a PLS AFO) all affect the stiffness of the AFO.6–8 Some materials and designs are intended to facilitate the storage of energy: that is, the ability to capture the energy that develops as the device deforms into dorsiflexion, store it until the limb is starting to unload in terminal stance, and then return that energy for push-off. The newer designs (chevron and carbon fiber) were created to have a greater capacity for energy storage than the single layer PLS AFO.9

Currently there is no way, before AFO fabrication and fitting, to use gait data to deduce the properties of an AFO that are most appropriate to improve an individual’s gait. The orthotist uses his/her expertise to adapt the stiffness of the AFO through design and material variables to comply with the physician’s prescription. There are several studies which have addressed the contributions of stiff AFOs to the gait of patients who are plantarflexor deficient.10–12 The angle versus moment characteristics (stiffness) of both an individual’s ankle and an AFO can be measured. However, there is no way for the physician or orthotist to quantify the amount of stiffness required by a particular child to compensate for reduced or absence of second and third rockers in gait.

We tested three types of PLS AFOs: standard single layer, dual layer chevron thermoplastic PLS, and carbon fiber PLS (“chevron” describes the shape of the posterior reinforcement of the dual layer AFO). The chevron and carbon fiber designs have claimed to be energy storing (gaining potential energy by deformation through second rocker, which is released at third rocker).9 We wanted to investigate whether they lead to improvements not only in second rocker (energy absorption) but also in power return for push-off (energy return). A critical preliminary step is to measure the individual stiffness characteristics of all three types of AFOs.

Several studies have addressed methods of assessing the stiffness of an AFO.13–16 Cappa et al.15 and Sumiya et al.13 both developed devices that are independent of motion capture equipment and can measure angles and moments during deflection of an AFO. The device described by Katdare16 can also measure angles and moments and was developed to interface with Gillette’s motion analysis system. The device has been validated against both shell and beam bending models and found no significant difference when the same AFO was tested on different days. It was therefore deemed the most practical for use in this study. The goal of this study was to ensure that the Katdare orthotics stiffness tester could quantify and differentiate the biomechanical properties of stiffness and energy return of these three AFO designs.


Three different designs of PLS AFOs were fabricated in three different sizes with three different stiffnesses for each size. The chevron and standard AFOs were fabricated in-house whereas the carbon fiber AFOs were made by Chladek Orthotics and Prosthetics (Des Moines, IA, USA). Each AFO was individually tested in the Center for Gait and Motion Analysis at Gillette Children’s Specialty Healthcare.

All of the AFOs in the standard PLS design category had identical spring curvature (arc of the leaf). This was meant to reduce the variability in structural stiffness. All carbon fiber AFOs had a flat leaf design. This was done to reduce the variability that a curved radius has on the mechanics of the leaf spring. All chevron PLS designs had flat leaf springs to match the carbon fiber design. At the point of flexion the width of the chevron and carbon fiber leaf springs were identical according to each size group.

Before fabricating our standard PLS AFOs, we contacted several manufacturers of off-the-shelf (OTS) PLS AFOs. Each was asked to explain how they calculated their product’s arc of the leaf. In all instances, the arc of the leaf was random, without any clinical or scientific data to support their product’s final design. For this study we selected the Orthomerica OTS AFO (Newport Beach, CA) as the arc contour that we would duplicate. The duplication process began by selecting the OTS AFOs that best fit our three sizes of plaster models. A generous layer of liquid plaster was applied to the posterior aspect of each model, followed quickly by applying the OTS AFOs and securing them with pressure sensitive tape. After the plaster had set, the OTS AFOs were removed, the models were finish sanded, and prepped for vacuum forming.

Precautions taken to limit the variables of the carbon fiber fabrication process were as follows:

  • All carbon fiber AFOs were made of a 300 g preimpregnated carbon fiber composite utilizing a 29% epoxy resin impregnation.
  • A consistent lay-up was used including a breather material to ensure noxic release during the curing cycle.
  • During the fabrication process, the AFO was placed under pressure in a “pressure vessel” to 32 psi and then heated at a rate of 1° per minute up to 250° held for 15 minutes.
  • The temperature was then decreased 1° per minute down to 150° (removed from the pressure vessel) and absent from stress for 24 hrs.
  • The curing cycle followed the prime curing cycle recommended by the engineer from Chladek Orthotics and Prosthetics (Des Moines, IA).
  • The stiffness varied with the number of layers of carbon fiber as well as the orientation of fibers between each layer.

The thickness of the leaf in the standard and chevron PLS designs was dependent on the thickness of the materials used, temperature of the plastic, and the amount of stretch applied when draping the plastic for vacuum forming. Precautions taken to limit these variables were:

  • A handheld infrared temperature sensor was used to check the surface temperature of each sheet of plastic before it was removed from the oven. The plastic was tested in the center and at the four corners until it reached the manufacturer’s recommended working temperature for that specific plastic.
  • The same two orthotists performed all the required vacuum forming to minimize variability in the amount of stretch that occurs during the draping process of vacuum forming.

The orthotics stiffness tester (as described by Katdare16) was bolted directly to a forceplate on the floor of the motion lab (Fig. 1). The AFOs were secured to the device using a system of screws and clamps. The AFOs were then put through a range of motion in the sagittal plane. The device allowed each AFO to be put through the same range of motion. The motion capture system was then able to determine the angles and forces induced in each AFO.

Figure 1.:
A picture of the orthotics stiffness testing device designed by Katdare.

Each AFO was tested five times and a multivariate analysis of variance used to assess if make, size, assigned stiffness, or speed make a difference in measured stiffness and energy dissipation. Three AFOs were tested twice to assess the repeatability of the protocol.

Calculations were then based on these measurements. Stiffness was calculated as the average slope in the plot of angle versus moment (Fig. 2). Because of the viscoelastic nature of plastic (and to some extent carbon fiber), the force required to bend the AFO to a certain position is not the force generated when the AFO is released. This is seen as a hysteresis loop in the angle versus moment plot. Energy dissipation was calculated as the area inside the hysteresis loop. A basic user interface was developed in Matlab (The Mathworks, Natick, MA) to facilitate data processing.

Figure 2.:
The viscoelastic nature of the AFOs caused a hysteresis in the angle versus moment curve. The red curve represents dorsiflexion, blue is plantarflexion, and green is the line of best fit.

Each AFO was tested at three different speeds of deformation, which were meant to match three different gait velocities: very slow (30 steps/minute), slow (40 steps/minute), and typical (60 steps/minute). A metronome was used to standardize the speed.

A total of 45 AFOs were tested. The following combinations of materials, design, and plastics were used:

  • Nine carbon fiber PLS AFOs (provided by Chladek Orthotics & Prosthetics, Des Moines, IA, USA).
    • Three large, three medium, three pediatric with each size group containing a flexible, moderate, and firm leaf spring.
  • Eighteen standard PLS AFOs (fabricated in the Assistive Technology Department at Gillette Children’s Specialty Healthcare, St. Paul, MN, USA).
    • Six large, six medium, and six pediatric single layer thermoplastic PLS AFOs.
    • Each size group of six AFOs was fabricated from one of each thickness: 5/32”, 3/16”, and 1/4” in both copolymer and polypropylene.
  • Eighteen chevron PLS AFOs (fabricated in the Assistive Technology Department at Gillette Children’s Specialty Healthcare, St. Paul, MN, USA).
    • Twelve large, 12 medium, and 12 pediatric chevron PLS AFOs were fabricated [although 36 chevron AFOs were fabricated, only 18 were tested. The total thickness of chevron plus outer shell was calculated and only the thickest AFO, thinnest AFO, and AFO of median thickness were tested. This was done to facilitate comparison to the standard and carbon fiber groups (Table 1)].
    • T1-3
      Table 1:
      Plastic thickness for PLS AFOs
    • Within each of the three size groups, six had a 5/32” copolymer outer shell and six had a 3/16” copolymer outer shell.
    • Each quantity of six had either a polypropylene or copolymer posterior reinforcement (chevron pattern) in one of each thickness: 5/32”, 3/16”, and 1/4” (Table 1).


Repeatability testing of the device revealed a significant difference between sessions for each of the three braces (p < 0.05). However, that difference was less than 8% of the mean stiffness for each brace. This difference was thought to be less than the manufacturing precision of AFOs. Therefore, the device was deemed to have adequate repeatability for the sake of this study.

The difference in the means of the energy measurements were more than 40% of the mean and thus the energy measurements were not deemed repeatable. Therefore, only the calculated stiffness results are presented.

The results of the analysis of variance show that the speed at which the orthoses were tested did not significantly affect the stiffness calculations (Table 2). Size was a significant factor for all orthoses except the chevron poly AFO.

Table 2:
Significance of various factors on calculated stiffness within a given orthotic using ANOVA

The orthotists manufactured each AFO to correspond to each stiffness category (flexible, moderate, stiff) using standard practices. Stiffness testing showed that the orthotists were predictably able to satisfy these criteria. Stiffness differences between categories were very significant for each of the brace designs, i.e., firm was found to be stiffer than those classified as moderate, etc. Example data for the chevron poly orthoses are shown in Figures 3–5. The effect of design within each size group was also tested and found to be significant with p < 0.001. (Fig. 6)

Figure 3.:
The effect of size on calculated stiffness for the chevron poly PLS. *Significant difference at the p < 0.05 level.
Figure 4.:
The effect of speed of dorsiflexion on calculated stiffness for the chevron poly PLS. *Significant difference at the p < 0.05 level.
Figure 5.:
The effect of manufacturer stiffness on calculated stiffness for the chevron poly PLS. *Significant difference at the p < 0.05 level.
Figure 6.:
The effect of orthosis type on calculated stiffness within a given size.

In testing the difference between the stiffnesses of individual braces, 990 comparisons were made (n × (n − 1)/2). Of those comparisons, 930 were significantly different (p < 0.05). This shows that the device is sensitive enough to detect differences in stiffness in orthoses of several different designs and sizes.


AFOs may be prescribed for many different reasons including improved foot alignment, support for weak musculature, and control of tonal abnormalities. Midstance ankle moment and power generation at toe-off are important aspects of typical gait and contribute to stability, velocity, and energy efficiency.1–3 If an individual has inadequate plantarflexor function to generate the force required for stance phase support and terminal stance push-off, a properly designed AFO may be able to compensate for some of these deficits.

The interaction between an AFO and its wearer is complex and inadequately studied. One of the first steps in taking a scientific approach to the study of these interactions is finding a reliable way to measure the mechanical properties of the AFO. Once these properties can be quantified, one can then study how they can be manipulated by altering the design of the AFO. Then, the more clinically significant question of matching AFO properties to a patient’s specific needs can be addressed.

The study by Sumiya et al.7 used polypropylene PLS AFOs with a thickness of 3 mm. Measurements were taken under static conditions at several dorsi- and plantarflexion angles for nine different AFOs of various trim lines. Calculating stiffnesses from their results gives a range from 0.2 to 1.3 Nm/degree. The polypropylene PLS AFOs tested in this study had stiffnesses that ranged from 0.4 to 1.7 Nm/degree. Lateral malleolus height was not measured in this study; therefore, the results cannot be directly compared with that of Sumiya et al.7 However, when taking into consideration the differences in thickness, trim lines, and test conditions, these measurements are very close and help to cross-validate both studies.

In the current study, three different designs were used. The Katdare stiffness tester16 was clearly able to detect the difference between AFOs built to be stiff and those built to be moderate and so on. We expected this difference and, when combined with the agreement in previously published results, confirmed the fact that the device was sensitive enough for the current study.

For each AFO design, we intended the orthotists to fabricate orthoses of various stiffnesses. We found that for each design, the orthotists were successful. We also found significant differences in the stiffness of the AFOs dependent on material used. No attempt was made to match stiffness between designs.

Even though the device was able to reliably measure stiffness, it failed to repeatably measure energy storage. Several possible explanations for this exist. As this was a prototype machine, its inherent design may be inadequate for measuring energy storage. The speed and range of motion were controlled manually and the foot and shank surrogates were unsophisticated. In addition, standardization of ankle joint placement relative to the brace was difficult. Because energy storage and return remain important clinical concerns, further work is planned to address this question. Stiffness tester design modifications may improve its accuracy and sensitivity allowing more insight into the energy storage and return characteristics of orthoses.

Future studies are planned to assess stiffness deficiency caused by weak ankle plantarflexors utilizing kinetic data obtained from instrumented gait analysis. The goal will be to match stiffness deficit to an appropriately stiff AFO to normalize the midstance plantarflexor moment. If improvements in the design of the orthotics tester are effective, power generation at push-off can also be assessed. This information could provide quantitative guidelines to physicians to order and orthotists to fabricate an AFO that is best suited to a patient’s needs.


This study demonstrates the usefulness of the Katdare orthotics stiffness tester for quantifying the stiffness of three different types and sizes of PLS AFOs. It also confirms the ability within an orthotics laboratory to consistently fabricate an orthosis to general stiffness specifications (flexible, moderate, stiff) within a design type. The Katdare orthotics stiffness tester was not able to reliably quantify energy storage during testing of these AFOs in dorsi- and plantarflexion.


1. Perry J. Gait Analysis, Normal and Pathological Function. Thorofare, NJ: Slack Inc.; 1992.
2. Gage J. The Treatment of Gait Problems in Cerebral Palsy. London, England: Mac Keith Press; 2004.
3. Winter DA. The Biomechanics and Motor Control of Human Gait: Normal, Elderly, and Pathological. 2nd ed. Waterloo, Ontario: University of Waterloo Press; 1991.
4. Kepple TM, Siegel KL, Stanhope SJ. Relative contributions of the lower extremity joint moments to forward progression and support during gait. Gait Posture 1997;6:1–8.
5. Perry J, Mulroy SJ, Renwick SE. Relationship between lower extremity strength and stride characteristics in post-polio syndrome. Arch Phys Med Rehabil 1993;74:165–169.
6. Convery P, Grieg RJ, Ross RS, Sockalingham S. A three center study of the variability of ankle foot orthoses due to the fabrication and grade of polypropylene. Prosthet Orthot Int 2004;28:175–182.
7. Sumiya T, Suzuki Y, Kasahara T. Stiffness control in posterior-type plastic ankle-foot orthoses: effect of ankle trimline. Part 2: Orthosis characteristics and orthosis/patient matching. Prosthetics and Orthotics International 1996;20(2):132–7.
8. Nagaya M. Shoehorn-type ankle-foot orthoses: prediction of flexibility. Archives of Physical Medicine & Rehabilitation 1997;78(1):82–4.
9. Wolf S, Knie I, Rettig O, et al. Carbon fiber spring AFOs for active push-off. Paper presented at the Gait and Clinical Motion Society 10th Annual Meeting, Portland, Oregon, April 6–9, 2005.
10. Duffy CM, Graham HK, Cosgrove AP. The influence of ankle foot orthoses on gait and energy expenditure in spina bifida. Journal of Pediatric Orthopedics 2000;20(3):356–61.
11. Thompson JD, Ounpuu S, Davis RB, DeLuca PA. The effects of ankle-foot orthoses on the ankle and knee in persons with myelomeningocele: an evaluation using three-dimensional gait analysis. Journal of Pediatric Orthopedics 1999;19(1):27–33.
12. Hullin MG, Robb JE, Loudon IR. Ankle-foot orthosis function in low-level myelomeningocele. Journal of Pediatric Orthopedics 1992;12(4):518–21.
13. Sumiya T, Suzuki Y, Kasahara T. Stiffness control in posterior-type plastic ankle-foot orthoses: affect of trimline. Part 1; a device for measuring ankle moment. Prosthetics and Orthotics International 1996;20:129–131.
14. Syngellakis S, Arnold MA, Rassoulian H. Assessment of the non-linear behaviour of plastic ankle-foot orthoses by the finite element method. Proceedings of the Institution of Mechanical Engineers. 2000; Vol 214, Part H.
15. Cappa P, Patane F, Pierro MM. A novel device to evaluate the stiffness of ankle foot orthosis devices. Journal of Biomechanical Engineering 2003;125:913–7.
16. Katdare K. The non-linear stiffness of ankle-foot orthoses: measurement and prediction [dissertation]. Minneapolis: The Biomedical Engineering Graduate Program, University of Minnesota, 1999.

AFO; stiffness; energy storage; cerebral palsy; myelomeningocele

© 2007 American Academy of Orthotists & Prosthetists