Gait & Posture 30 (2009) 356–363

Contents lists available at ScienceDirect

Gait & Posture

journal homepage: www.elsevier.com/locate/gaitpost

Kinematics and kinetics with an adaptive ankle foot system during stair
ambulation of transtibial amputees
Merkur Alimusaj *, Laetitia Fradet, Frank Braatz, Hans J. Gerner, Sebastian I. Wolf
Department of Orthopaedic Surgery, University of Heidelberg, Schlierbacher Landstr. 200 a, 69118 Heidelberg, Germany

A R T I C L E I N F O A B S T R A C T

Article history: Conventional prosthetic feet cannot adapt to specific conditions such as walking on stairs or ramps.
Received 8 December 2008 Amputees are therefore forced to compensate their prosthetic deficits by modifying the kinematics and
Received in revised form 2 June 2009 kinetics of their lower limbs. The Proprio-FootTM (Ossur) intends to reduce these compensation
Accepted 12 June 2009
mechanisms by automatically increasing dorsiflexion during stair ambulation thanks to an adaptive
microprocessor-controlled ankle. The present investigation proposes to analyze the biomechanical
Keywords: effects of the dorsiflexion adaptation in transtibial (TT) amputees during stair ambulation.
Transtibial amputee
Sixteen TT amputees and sixteen healthy controls underwent conventional 3D gait analysis.
Gait analysis
Prosthesis
Kinematics and kinetics of the lower limbs were compared during stair ascent and descent performed by
Stair patients with the prosthetic foot set to a neutral ankle angle and with an adapted dorsiflexion ankle angle
Microprocessor ankle of 48. Norm distance as well as minimum and maximal values of sagittal kinematics and kinetics were
calculated for comparisons between patients and control subjects.
For both stair ascent and descent, an improvement of the knee kinematics and kinetics could
particularly be noticed on the involved side with an increase of the knee flexion and an increase of the
knee moment during stance.
Therefore, despite its additional weight compared to a conventional prosthetic ankle, the Proprio-
FootTM should be beneficial to active TT amputees whose knee musculature strength does not constitute
a handicap.
ß 2009 Elsevier B.V. All rights reserved.

1. Introduction
Stair ambulation increases the kinetic demand compared with
level walking [1–4] and emphasizes motor deficits. Tiedemann
et al., in their investigation on the factors required for stability and
safety on stairs in elderly persons [5], indeed noticed that not only
strength but also reduced sensation, balance, pain, and increased
fear of falling could lead to functional deficits during stair
ambulation. For groups like amputees suffering from loss of
muscle strength and loss of joint mobility, balance, or propriocep-
tion, stair ambulation becomes specifically challenging [6–9]. For
instance, amputees are characterized by a greater slowness, stance
asymmetry, and increased muscular effort than controls [6,8].
Stair ambulation with a prosthesis constitutes a particular
problem, since patients are totally dependent on the functional
features and the technical components of the artificial limb. The
recent development of a computerized knee for transfemoral (TF)
amputees led to better support when descending stairs and ramps
[7]. Conversely, the design and the function of prosthetic feet have
* Corresponding author.
E-mail address: merkur.alimusaj@ok.uni-heidelberg.de (M. Alimusaj).
not been considerably modified and are still commonly based on
the use of a combination of elastic carbon fibre springs,
polyurethane keels, and simple mechanical axes. The ankle
function is therefore always limited by the design and the
elasticity of the artificial joint that is usually adjusted for walking
on level ground. Due to the lack of recognition of the specificities
related to stair ambulation and the prosthetic shortcomings,
transtibial (TT) amputees are particularly handicapped and subject
to different compensation mechanisms to deal with the neutral
and fixed ankle position as seen in traditional prosthetic ankles.
During stair ascent, TT amputees use a particular compensation
mechanism that could be a result of a strategy favouring knee
stability on the involved side as during level walking [10]. They
produce a strong hip moment to elevate the body during the stance
phase on their involved side, conversely to non-disabled subjects
and TF amputees who mainly rely on a knee moment [7,10]. The
preparation of the next foot contact is also particularly problematic
for TT amputees and this applies to both sides [7]. When preparing
step contact for the sound limb, the missing active plantarflexion
on the involved side leads to an insufficient vertical position of the
body centre of mass (CoM). When preparing step contact for the
involved limb, the missing dorsiflexion on the involved side
reduces the possibility of clearing the foot directly prior to the

0966-6362/$ – see front matter ß 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2009.06.009
 M. Alimusaj et al. / Gait & Posture 30 (2009) 356–363
support phase. Both problems are compensated by the sound limb
through an increased knee flexion during late swing phase and an
increased plantar flexion during late stance phase [7].
During stair descent, amputees adopt a specific landing strategy
on their involved side, with a CoM positioned directly over the
landing limb at initial contact. It probably ensures that the ground
reaction force is directed anterior of the knee joint centre to keep
the knee fully extended or locked [11]. During loading response,
the amount of knee flexion is also restricted, possibly as a result of
the reduced dorsiflexion [11] or by the lack of neuromuscular
control of stance stability, due to loss of the tibialis anterior and
triceps surae [7]. Regarding the sound side, Schmalz and colleagues
noticed that ground contact is initiated with increased plantar
flexion at the ankle probably to compensate for the lack of
dorsiflexion favouring the body lowering, which causes the
amputees to ‘‘fall’’ onto their sound limb.
According to the aforementioned findings, stair ambulation is
always limited in TT amputees due to the shortcomings of the
actual prosthetic feet and their neutral dorsi/plantarflexion. With
the perspective of improving the ambulation of TT amputees on
stairs, quasi-passive prosthetic ankles have recently been devel-
oped. These new prosthetic ankles are based on active damping or
spring-clutch mechanisms to adjust the ankle angle according to
the ground surface [12,13] (Ossur, 2002–2008; US Patent 6443993,
2002). The ankle foot system Proprio-FootTM (Ossur hf, Iceland)
constitutes one of them (Fig. 1).
Among different features, this prosthetic foot adapts to stairs by
readjusting the ankle angle of the unloaded foot in swing phase
thanks to a step motor driven by a microprocessor that processes
signals of integrated accelerometers and angular sensors.
A pilot study performed in our laboratory with six TT amputees
confirmed this feature of stair recognition and angle adaptation
and showed that patients walked with significantly reduced knee
extension during stair ambulation [14]. The aim of the current
study is to further test the effects of the dorsiflexion/plantarflexion
adaptation on the kinematics and kinetics in TT amputees during
stair ambulation. It is hypothesized that knee and hip kinematics
and kinetics will be closer to normal when walking with an
adapted, i.e., dorsiflexed prosthetic ankle compared to walking
with the same prosthetic foot fixed in neutral position. In
particular, it is expected that the prosthetic ankle adjusted in
dorsiflexion could favour a more physiologic knee flexion on the
involved side during the loading response and mid-stance both
Fig. 1. Proprio-FootTM with prosthetic socket. 1 = prosthetic socket;
2 = Microprocessor-controlled ankle unit; 3 = Defined ankle axis of rotation;
4 = Carbon fibre foot module (Variflex LP);.
357
during stair ascent and stair descent. Further, it could decrease the
need for compensation by plantarflexion on the sound side in stair
ascent and reduce the ‘‘fall’’ phenomenon observed at initial
contact to compensate for the lack of dorsiflexion.
2. Material and methods
2.1. Participants
Sixteen healthy control subjects (10 male; 6 female; mean age: 31.1  10.3 years;
height 173.4  8.2 cm; mass 73.9  13.8 kg) without gait deficits and 16 TT amputees
(16 males) recruited in the outpatient clinic of our department (age 50.3  12 years;
height 178.4  7.1 cm; mass 83.7  15.0 kg; time since amputation 25.3  20.9 years,
number of traumatic cases 13, sarcoma 3) participated in this study, approved by the
local ethics committee.
Technically, the TT amputees used prosthetic foot sizes between 25 and 29 cm
and had to have at least 18.5 cm clearance under the socket. Clinically they had to
fulfill the following inclusion criteria:
1. Nonvascular adult K3-K4 amputee;
2. Able to negotiate stairs and ramps without any additional walking aids;
3. No stump problems during the past 3 months;
4. No comorbidities in the limbs;
5. Able and willing to handle the new electronic device.
2.2. Protocol
During their first visit to the laboratory, patients were provided with the Proprio-
FootTM and got a 3 h practical and theoretical introduction to it. In particular, they
were instructed on how to use the different features of this new device according to
the manufacturer’s recommendations. The alignment of the prostheses was set for
each patient following the manufacturer’s technical guidelines verified using the
L.A.S.A.R posture (Otto-Bock, Germany) [15]. The target angle for stair adaptation of
the Proprio-FootTM was predefined for all participants to 48 ankle dorsiflexion. After
this visit, all patients had at least 14 days to adapt to the Proprio-FootTM in their
usual environment, e.g., at home, at work, or during their free-time activities.
Following this trial period, patients underwent an instrumented 3D gait analysis
using an optical 3D motion capture system (VICON, United Kingdom) working at
120 Hz with reflective markers of 14 mm diameter. The standard gait model [16] was
applied with the knee axis defined according to the estimations of Nietert [17] by
using a knee axis gauging tool (Kniedrehpunktlehre, Otto Bock, Germany). All patients
and control subjects used the same type of shoes (Deichmann, Germany), and markers
were placed on the shoe at the heel and forefoot accordingly. At least 8 stair ascents
and descents were monitored on a custom-made 80 cm-wide staircase consisting of 5
steps of 15 cm height and a step distance of 32 cm. Two out of five steps were
equipped with force plates (type 9865B, Kistler Instruments, Switzerland) [18].
The adaptive prosthetic ankle was tested fixed at neutral position to simulate a
conventional prosthetic ankle and with an adapted mode of 48 dorsiflexion, later
referred to as ‘‘neutral’’ and ‘‘adapted’’ ankle, respectively. It was indeed chosen to test
the adaptation mode and not the total unit, which explains why patients did not walk
with a real conventional prosthetic ankle. This choice also avoids inaccuracy due to
two marker placements required by two distinct measurement sessions. The walking
tasks (stair ascent/descent) were tested in a randomized order across all patients.
2.3. Data analysis
Data of kinematics and kinetics were computed using PlugInGait [16] and were
normalized to the gait cycle (101 data points). The average and standard deviation
between trials were calculated for each subject and condition with MatLab 6.5.1
(Mathworks Inc. Natick, MA, USA).
The effect of the new prosthetic ankle was first analyzed by comparing the
kinematics and kinetics time-series of the two conditions with the respective data
of the control subjects using the formalized norm distance (ND) [19] defined by:
jU½k  U norm ½kj
ND½k ¼ ; (1)
s norm ½k
where U[k] represents the time-series (e.g. the hip angle), Unorm[k] the reference
time series of control subjects, s norm[k] the corresponding standard deviation, and
where k indicates the sample index of a time normalized stride (typically in
percentage gait cycle). Hence, this ND yields the difference between individual gait
data and the average value of the reference data weighted by the variation within
this reference. Here, the ND across the gait cycle was calculated for sagittal
kinematics and kinetics of the hip, knee, and ankle and compared for the two
conditions ‘‘neutral’’ and ‘‘adapted.’’
In addition, gait speed and gait features that related to the hypotheses
formulated in the introduction and to important characteristics of stair ambulation
emphasized in previous studies were compared.
Normal distribution of all data was confirmed using the Shapiro-Wilkinson test.
Consequently, t-tests were further employed for statistical analysis. Patients’ intra-
 358
M. Alimusaj et al. / Gait & Posture 30 (2009) 356–363
Fig. 2. Sagittal kinematics and kinetics during stair ascent for controls and patients’ involved side. The diagrams are normalized to gait subphases. For controls, only standard deviations are given, for patients, means and standard
deviations are given. IC: initial contact; oIC: opposite initial contact; FO: foot-off; oFO: opposite foot-off.
 M. Alimusaj et al. / Gait & Posture 30 (2009) 356–363 359

individual differences across conditions were evaluated using paired t-tests, and
differences to data of control subjects were analyzed via unpaired t-tests. For all the
statistical tests, the level of significance was set to p < 0.05.
3. Results
3.1. Stair ascent
Fig. 2 illustrates the variation of the kinematics and the kinetics
data during stair ascent for control subjects and the involved side
of patients, including both ‘‘neutral’’ and ‘‘adapted’’ walking
conditions. For clarity, the pattern of the sound side that did not
deviate significantly from those presented in the literature [7,10] is
not represented in this figure but is provided as on-line material. It
can be seen clearly that on average the adapted ankle increased
dorsiflexion throughout the movement on the involved side, which
was confirmed statistically at initial contact (p < 0.001) and for the
maximal dorsiflexion (p < 0.001).
The mean values, standard deviation, and significance of the
statistical tests are presented in Table 1. The differences between the
sound side and controls, as well as the differences between the
involved side and controls, were similar to those obtained in
previous papers, regardless of the condition ‘‘neutral’’ or ‘‘adapted.’’
When patients walked with the adaptive ankle, the sagittal
kinematics and kinetics were closer in tendency to the control data,
Table 1
Mean  standard deviation for the kinematics and kinetics data during stair ascent (angles are in degrees, moments in N m kg1, power in W kg1, and speed in m s1).
both in the involved and in the sound side. On the involved side, this
is shown by the norm distances that were significantly smaller with
the adapted ankle than with the neutral ankle for the hip flexion/
extension moment (p < 0.001) and power (p < 0.001) as well as for
the knee flexion/extension moment (p < 0.01). Further, patients had
a smaller hip flexion at initial contact (p < 0.01) and a smaller
maximal hip flexion (p < 0.001) when walking with the adapted
ankle than when walking with the neutral ankle. They also had a
smaller knee extension (p < 0.001) with the adapted ankle than with
the neutral ankle but also, in terms of kinetics during stance phase, a
smaller hip power generation (p < 0.001) and a greater knee
extension moment (p < 0.01) and power generation (p < 0.05).
On the sound side, the norm distance did not change even
though patients had a smaller maximal hip extension (p < 0.01)
and a smaller maximal plantarflexion (p < 0.01) when walking
with the adapted ankle than when walking with the neutral one.
Patients were significantly slower than the control subjects
(p < 0.001 for both conditions) and the prosthetic ankle did not
affect this speed (p > 0.05).
3.2. Stair descent
Fig. 3 shows the variation of the kinematics and the kinetics
data during stair descent for control subjects and the involved side
of patients for the two conditions. A figure representing the

Parameter Controls Patients’ sound side Patients’ involved side

Neutral Adapted Neutral Adapted

***
Hip flex @IC 62  7 60  5 59  6 69  4 68  4**,oo
Hip flex min 86 14  5** 13  5**,o 86 95
Hip flex max 64.9  7.3 63.2  5.1 62.6  5.4 73.1  4.2*** 72.5  4.3***,ooo

Knee flex @IC 61  4 59  4 58  5 49  8*** 49  7***

Knee flex min 75 54 54 5  6***(ext) 0  6***,ooo
Knee flex max 86  5 81  5* 82  5 80  8* 81  9

Dorsiflex @IC 21.4  3.0 22.1  3.9 21.6  3.8o 5.9  3.3*** 9.5  3.1***,ooo
Plantarflex max 26 16  5*** 14  6***,oo 3  3** 8  3***,ooo

Hip ext moment max 0.70  0.14 0.97  0.27*** 1.03  0.28*** 1.15  0.2*** 1.16  0.23***

Knee ext moment max 1.03  0.14 1.03  0.25 1.09  0.288 0.24  0.18*** 0.34  0.18***,oo
Knee flex moment max 0.37  0.14 0.47  0.18 0.48  0.18 0.46  0.15 0.28  0.16ooo

Plantarflex moment max1 0.79  0.26 0.83  0.23 0.87  0.27 0.64  0.25 0.58  0.24*,o
Plantarflex moment max2 1.53  0.18 1.60  0.25 1.6  0.27 0.68  0.19*** 0.67  0.20***

Hip Power generation max1 1.33  0.29 1.28  0.53 1.30  0.53 2.54  0.56*** 2.27  0.49***,ooo
Hip Power generation max2 0.86  0.34 0.57  0.31* 0.56  0.3* 0.70  0.22 0.73  0.22

Knee Power generation max1 2.34  0.64 1.72  0.50** 1.71  0.56** 0.46  0.33*** 0.68  0.40***,ooo
Knee Power generation max2 0.94  0.23 0.52  0.26*** 0.52  0.24*** 0.50  0.16*** 0.42  0.18***,o

Plantarflex power generation max 2.87  0.78 4.25  1.50** 4.13  1.59** 0.41  0.23*** 0.32  0.22***,oo

ND hip flex/ext 0.82  0.26 0.84  0.29 1.09  0.34 1.03  0.33
ND hip flex/ext moment 1.45  0.37 1.4  0.36 1.52  0.34 1.44  0.29ooo
ND hip flex/ext power 0.97  0.36 0.93  0.26 1.38  0.32 1.27  0.22ooo

ND knee flex/ext 1.55  0.57 1.55  0.6 2.57  0.94 1.98  0.74
ND knee flex/ext moment 1.37  0.42 1.39  0.38 2.1  0.53 1.77  0.31o
ND knee flex/ext power 1.13  0.2 1.16  0.26 1.39  0.35 1.19  0.2

ND dorsi/plantarflex 1.15  0.2 1.09  0.22 3.04  0.73 2.29  0.55oo

ND dorsi/plantarflex moment 2.34  1.03 2.17  0.92 1.86  0.48 1.78  0.47oo
ND dorsi/plantarflex power 2.49  1.36 2.16  1.21 1.15  0.16 1.23  0.15

Speed 0.57  0.05 0.44  0.06*** 0.45  0.07***

@IC: angle at initial contact; ND: norm distance.

*
p < 0.05.
**
p < 0.01.
***
p < 0.001 significance of the t-test during the comparison with controls.
o
p < 0.05.
oo
p < 0.01.
ooo
p < 0.001 significance of the t-test during the comparison with the neutral ankle.
 360
M. Alimusaj et al. / Gait & Posture 30 (2009) 356–363
Fig. 3. Sagittal kinematics and kinetics during stair descent for controls and patients’ involved side. The diagrams are normalized to gait subphases. For controls, only standard deviations are given, for patients, means and standard
deviations are given. IC: initial contact; oIC: opposite initial contact; FO: foot-off; oFO: opposite foot-off.
 M. Alimusaj et al. / Gait & Posture 30 (2009) 356–363 361

kinematics and the kinetics for the sound side is provided as on-
line material. The patterns of the angles, moments, and powers
were similar to those described in the literature for control
subjects [1,2] and TT amputees [7].
Fig. 3 confirms that the adapted ankle successfully increased
dorsiflexion on the involved side during stair descent, which leads
to statistical differences for the dorsiflexion at initial contact
(p < 0.001) and for the maximal dorsiflexion (p < 0.001).
In stair descent, compared to the neutral ankle, the adapted
ankle had reduced differences with control subjects both in the
involved and in the sound side (Table 2). For the involved side, the
norm distance for the hip flexion/extension angle (p < 0.05),
moment (p < 0.01), and power (p < 0.001) as well as the knee
flexion/extension moment (p < 0.01) were indeed smaller when
walking with the adapted ankle than when walking with the
neutral one. The improvement of the knee moment pattern with
the adapted ankle compared with the neutral ankle was partially
explained by the reduction of the knee extension at initial contact
(p < 0.01) and during mid-stance phase (p < 0.001), which
coincides with an increase of the first extension moment
(p < 0.001) for the adapted ankle.
For the sound side, the norm deviation was not influenced
significantly, since only the dorsi/plantarflexion moment was
modified across the two conditions and was significantly
decreased when walking with the adapted ankle (p < 0.05).
Nevertheless, in patients the maximal hip flexion was closer to
that measured for control subjects when walking with the adapted
ankle than when walking with the neutral ankle (p < 0.05).
Additionally, the maximal knee flexion (p < 0.01) was also greater
with the adapted ankle than with the neutral ankle, whereas the
dorsiflexion power absorption was smaller with the adapted ankle
than with the neutral one (p < 0.01).
The adapted ankle did not modify the patients’ speed (p > 0.05),
such that patients were always slower than controls (p < 0.001 for
both conditions).
4. Discussion
The present paper proposed to investigate the stair ambulation
of TT amputees and to test whether an adaptation of the prosthetic
ankle dorsiflexion can help amputees during stair ascent/descent.
4.1. Stair ascent
As in previous studies, the TT amputees adopted a ‘‘hip strategy’’
on their involved side during stair ascent. Compared with controls,

Table 2
Mean  standard deviation for the kinematics and kinetics data during stair descent (angles are in degrees, moments in N m kg1, power in W kg1, and speed in m s1).

Parameter Controls Patients’ sound side Patients’ involved side

Neutral Adapted Neutral Adapted

Hip flexion @IC 22  8 17  6 18  6 22  6 22  5

Hip flexion min 13  6 13  6 13  5 11  10 13  9
Hip flexion max 42  8 37  8 39  78 43  6 44  5

Knee flexion @IC 13  4 5  4*** 4  4*** 5  5*** 6  5***,oo

Knee flexion max 90  4 81  5*** 82  5***,88 79  8*** 80  8***
Knee flexion @mid-stance 29  5 23  7** 22  7***,ooo 8  7*** 14  7***,ooo

Plantarflexion @IC 10  5 17  11* 17  11* 4  3*** 8  3***,ooo

Dorsiflexion max 36  4 39  4 38  4 15  6*** 20  4***,ooo

Hip extension moment max1 0.45  0.22 0.31  0.19 0.20  0.11***,8 0.49  0.18 0.54  0.18o
Hip extension moment max2 0.01  0.10 0.01  0.19 0.01  0.18 0.43  0.25*** 0.44  0.18***

Knee extension moment max1 0.84  0.18 0.81  0.38 0.72  0.33 0.28  0.15*** 0.38  0.17***,ooo
Knee extension moment max2 0.72  0.15 0.97  0.22*** 0.95  0.2*** 0.43  0.20*** 0.48  0.18***

Plantarflexion moment max1 1.24  0.15 1.53  0.45* 1.49  0.45* 0.67  0.29*** 0.65  0.24***
Plantarflexion moment max2 1.31  0.18 1.54  0.19** 1.52  0.19** 1.00  0.27*** 1.07  0.15***

Hip Power generation max 0.61  0.31 0.43  0.20 0.42  0.20 0.35  0.14** 0.37  0.18*
Hip Power absorption max 0.34  0.16 0.36  0.24 0.27  0.16 0.72  0.32*** 0.64  0.21***

Knee Power absorption max1 1.61  0.76 1.62  1.11 1.19  0.80o 0.28  0.21*** 0.43  0.30***,o
Knee Power absorption max2 2.55  0.57 2.26  0.52 2.27  0.57 1.61  0.64*** 1.79  0.66**

Plantarflex power absorption max 4.29  1.12 6.96  2.86*** 6.35  2.64**,oo 0.69  0.19*** 0.70  0.18***
Plantarflex power generation max 2.67  0.63 2.63  0.88 2.59  1.08 0.73  0.30*** 0.73  0.25***

ND hip flex/ext 0.85  0.45 0.79  0.35 1.23  0.40 1.13  0.45o
ND hip flex/ext moment 0.97  0.40 0.96  0.40 1.72  0.53 1.73  0.42
ND hip flex/ext power 0.64  0.22 0.66  0.26 1.54  0.49 1.50  0.34

ND knee flex/ext 1.51  0.64 1.46  0.65 2.94  0.65 2.61  0.59oo
ND knee flex/ext moment 1.46  0.29 1.36  0.33 2.20  0.34 1.97  0.24oo
ND knee flex/ext power 0.94  0.22 0.87  0.22o 1.35  0.19 1.31  0.20

ND dorsi/plantarflex 1.09  0.41 1.14  0.40 3.27  0.73 2.57  0.42ooo

ND dorsi/plantarflex moment 1.34  0.54 1.24  0.42 2.06  0.42 2.03  0.28
ND dorsi/plantarflex power 1.25  0.60 1.13  0.44 1.49  0.11 1.42  0.10ooo

Speed 0.68  0.11 0.48  0.08*** 0.47  0.08***

@IC: initial contact; LR: loading response; MSt: mid-stance; TSt: terminal stance; PSw: pre-swing; ND: norm distance.
*
p < 0.05.
**
p < 0.01.
***
p < 0.001 significance of the t-test during the comparison with controls.
o
p < 0.05.
oo
p < 0.01.
ooo
p < 0.001 significance of the t-test during the comparison with the neutral ankle.
362 M. Alimusaj et al. / Gait & Posture 30 (2009) 356–363
patients had their involved hip more flexed at initial contact,
whereas the knee was generally more extended throughout the
movement, especially during terminal stance. At the end of the
loading response, a smaller knee extension moment and power
generation and a greater hip extension moment and power
generation were also measured. This hip strategy found for the
involved side could be used to ensure limb stabilization [10] or a
sufficient vertical position of the centre of gravity compromised by
the lack of plantarflexion when preparing the foot contact for the
sound side [7].
The increased dorsiflexion of the adapted ankle had the
tendency to diminish this ‘‘hip strategy’’ and to result in a
generally more physiological behaviour on the involved side. This
was shown by the norm distance, the parameter that estimates
how different a time-series is from a reference. This norm distance
was indeed smaller for the parameters characterizing the kinetics
of the involved hip, i.e., the hip moment and power and for the knee
moment when walking with the adapted ankle than when walking
with the neutral ankle. As hypothesized, on the involved side, the
adapted ankle favoured knee flexion during loading response and
mid-stance and reduced the need for hip flexion at initial contact.
In addition, the maximal hip extension power generation slightly
decreased, while the maximal knee extension power generation
increased while still being far from normal. This result suggests
that TT amputees do not fully use the available knee power of their
involved side and this is mainly due to a lack of dorsiflexion rather
than due to stabilization reasons.
Interestingly, the greater dorsiflexion of the adapted ankle near
foot-off did not lead to compensation mechanisms on the involved
side, although it could be thought to be less favourable for toe
clearance, namely, to avoid the next step. In fact, no statistical
differences were found in the kinematics or kinetics after stance-
swing transition. This phase, so important in terms of safety issues,
should be further analyzed.
On the sound side, as described in the literature, the hip was less
extended toward the end of the stance phase and the ankle was a
lot more plantarflexed compared with control subjects, probably to
elevate the body and position the opposite limb for the following
strike [7], which also resulted in a reduction of the hip and knee
flexion power generation. With the adapted ankle, this compensa-
tion mechanism decreased modestly even significant, since the
sound side of patients generated less plantarflexion with the
adapted ankle. However, it did not decrease the need for
plantarflexion power generation. In fact, several patients didn’t
show any change in their plantarflexion, which explains the
relative small evolution of the mean.
4.2. Stair descent
Noteworthy differences similar to those mentioned in the
literature were established between the patients’ involved side
and the control subjects, regardless of the prosthetic ankle worn.
Patients tended to land on a more extended involved limb as
reflected by the greater knee extension found at initial contact.
Patients might adopt this strategy to position their CoM directly over
their landing limb, probably due to the absence of plantarflexion
and/or to reduce the moment required by the knee extensors [11].
This specific position coincided with smaller knee extension
moment and flexion power absorption during loading response.
The knee and ankle didn’t create a considerable power absorption in
this limb position. However, it has been shown that amputees do not
load their involved limb as much as their sound limb during level
walking [20]. This might also be the case during stair descent, which
would reduce the demand for power absorption on patients’
involved side. The hip flexion power generation found simulta-
neously to knee power absorption might be used to help stabilizing
the body during this phase. For the control of the second body-
lowering phase following the opposite initial contact, patients also
had the tendency to develop a ‘‘hip strategy’’ on their involved side.
The peaks of the knee extension moment and flexion power
absorption appearing at the end of the stance phase were indeed
reduced as compared to control subjects. Moreover, remarkable hip
extension moments and hip flexion power absorption were found
for the patients’ involved limb, whereas contrarily hip flexion
moment and power generation were measured for control subjects.
To walk with the adapted ankle led to kinematics and kinetics
closer to physiologic patterns for the involved side as attested by the
decrease of the norm distance for the hip angle, moment, and power,
as well as for the knee moment. Further, the adapted ankle reduced
the aforementioned adaptations, mainly at the beginning of the
stance phase. As hypothesized, the involved knee flexion was more
pronounced during mid-stance, which results in a greater knee
extension moment and power absorption to control the body weight
acceptance in the first moments of the stance phase. Conversely, the
‘‘hip strategy’’ also adopted to control body lowering at the end of the
stance phase was not significantly modified with the adapted ankle
since no parameters were statistically different between the two
conditions during this phase.
While during stair ascent the lack of dorsiflexion on the involved
side caused a reduced ability to elevate the CoM, it caused a reduced
ability to lower the CoM during stair descent. This limitation was
compensated on the sound side, whatever the prosthetic ankle, by a
more stretched limb at initial contact, as attested by the greater
plantarflexion and the greater knee extension measured compared
with control subjects. It also provoked greater energy absorption for
weight acceptance of the body lowering as attested by the greater
plantarflexion moment and greater plantarflexion power absorption
measured during loading response on the patients’ sound side
compared with the controls.
This phenomenon of fall described by Schmalz and colleagues
[7] was generally reduced when walking with the adapted ankle.
However, it did not modify the kinematics and kinetics on the
patients’ sound side to a noteworthy extent. Only the norm
distance of the dorsi/plantarflexion moment was indeed signifi-
cantly reduced and only the maximal dorsiflexion power absorp-
tion was significantly smaller when walking with the adapted
ankle. To explain this last result, the greater dorsiflexion provided
by the adapted ankle on the involved limb probably enabled the
CoM to be lower.
4.3. Outcomes of the new prosthetic ankle
The Proprio-FootTM is one of the new quasi-passive prosthetic
ankles that are able to ‘‘actively’’ change the prosthetic ankle angle
in the unloaded swing phase thanks to a microprocessor when
walking on stairs or on ramps. It is passive in the sense that no
power is being generated through the ankle in stance. In the
present study, the effect of the adaptation of the prosthetic ankle
angle during stair ambulation was tested. Both during stair ascent
and descent, the most noticeable improvements provided by the
increased dorsiflexion of the adapted ankle were related to knee
flexion kinematics and kinetics of the involved side during stance.
This improved knee function which partially compensates for the
absence of ankle motion and proprioception can be regarded as a
substantial benefit in the patient group investigated.
The increase in weight of the actual Proprio-FootTM, compared
to a conventional device such as the Vari-Flex-LP1 (995 g versus
405 g), is not seen as being really critical. In the present study, we
did not see any negative effect in the data during swing phase,
namely when the weight would be mainly relevant [21] and when
patients could possibly feel increased inertial effects. It does
however raise the relevance of a good socket fitting since for
 M. Alimusaj et al. / Gait & Posture 30 (2009) 356–363
‘‘heavy’’ prosthetic devices pistoning effects are occasionally being
reported. The increased weight might nevertheless constitute a
limit for non-active amputees since energy consumption increases
with foot weight [22]. The Proprio-FootTM might then not be
beneficial to amputees whose endurance and knee musculature
are limited. For an assessment of whether a patient can expect
substantial benefits from innovative prosthetic devices such as the
Proprio-Foot, it is recommended to perform individual tests with
appropriate gait analysis devices.
Other positive aspects were found with the adapted ankle but
since they were relatively small, their benefit to the patients is
questionable. Actually the adaptation angle of 48 in the stair mode
was rather small compared to physiologic ankle angles. With an
increased and a possibly individually adjusted dorsiflexion angle
other improvements might be obtained.
The lack of power generation still limits the benefits of this type
of prosthesis. Some attempts have been made to develop active
prosthetic ankles [23]. However, such prostheses, albeit promising,
always suffer from weight and size problems due to the motor and
the batteries needed to generate power. Thus, until the problems
linked to the feasibility of active prostheses are solved, quasi-passive
ankles such as the Proprio-FootTM constitute a good alternative for
active patients whose life induces a lot of stair ambulation.
5. Conclusion
The increased dorsiflexion provided during stair ambulation by
new prosthetic ankles such as the Proprio-FootTM improves the
knee kinematics and kinetics of transtibial amputees on the
involved side. However, to really appreciate the potential benefits
of such a device for one particular patient, individual tests are
recommended.
Acknowledgement
The authors would like to thank all participants for their
attendance as well as Johannes Siegel and Marieke Dahl for their
contributions to the measurements.
Conflict of interest statement
The present study was partially financially supported by Ossur
hf Europe.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.gaitpost.2009.06.009.
363
Reference
[1] Protopapadaki A, Drechsler WI, Cramp MC, Coutts FJ, Scott OM. Hip, knee,
ankle kinematics and kinetics during stair ascent and descent in healthy young
individuals. Clin Biomech 2007;22:203–10.
[2] McFadyen BJ, Winter DA. An integrated biomechanical analysis of normal stair
ascent and descent. J Biomech 1988;21:733–44.
[3] Reeves ND, Spanjaard M, Mohagheghi AA, Baltzopoulos V, Maganaris CN. The
demands of stair descent relative to maximum capacities in elderly and young
adults. J Electromyogr Kinesiol 2008;18:218–27.
[4] Andriacchi TP, Andersson GB, Fermier RW, Stern D, Galante JO. A study of
lower-limb mechanics during stair-climbing. J Bone Joint Surg Am 1980;62:
749–57.
[5] Tiedemann AC, Sherrington C, Lord SR. Physical and psychological factors
associated with stair negotiation performance in older people. J Gerontol A Biol
Sci Med Sci 2007;62:259–1265.
[6] Schmalz T, Blumentritt S, Jarasch R. Energy expenditure and biomechanical
characteristics of lower limb amputee gait: The influence of prosthetic
alignment and different prosthetic components. Gait Posture 2002;16:
255–63.
[7] Schmalz T, Blumentritt S, Marx B. Biomechanical analysis of stair ambulation
in lower limb amputees. Gait Posture 2007;25:267–78.
[8] Powers CM, Boyd LA, Torburn L, Perry J. Stair ambulation in persons with
transtibial amputation: an analysis of the Seattle LightFoot. J Rehabil Res Dev
1997;34:9–18.
[9] Powers CM, Rao S, Perry J. Knee kinetics in trans-tibial amputee gait. Gait
Posture 1998;8:1–7.
[10] Yack H, Nielsen D, Shurr D. Kinetic patterns during stair ascent in patients with
transtibial amputations using three different prostheses. J Prosth Orthot 1999;
57–62.
[11] Jones SF, Twigg PC, Scally AJ, Buckley JG. The mechanics of landing when
stepping down in unilateral lower-limb amputees. Clini Biomech 2006;21:
184–93.
[12] Li C, Tokuda M, Furusho J, Koyanagi K, Morimoto S, Hashimoto Y, Nakagawa A,
Akazawa Y.In: Research and Development of the Intelligently-Controlled
Prosthetic Ankle Joint; 2006.p. 1114–9.
[13] Hansen AH, Gard SA, Childress DS, Ruhe B, Williams R. Automatically-Adapt-
ing Ankle-Foot Prosthesis Concept. In: 12th World Congress of the Interna-
tional Society for Prosthetics and Orthotics; 2007.
[14] Braatz F, Alimusaj M, Wolf S, Doederlein L. Ein mikroprozessor-gesteuertes
Sprunggelenk: Verbessert ein aktiver Prothesenfuß die Gehfunktion? Orth
Technol 2006;8:608–11.
[15] Blumentritt S. A new biomechanical method for determination of static
prosthetic alignment. Prosthet Orthot Int 1997;21:107–13.
[16] Kadaba MP, Ramakrishnan HK, Wootten ME, Gainey J, Gorton G, Cochran GV.
Repeatability of kinematic, kinetic, and electromyographic data in normal
adult gait. J Orthop Res 1989;7:849–60.
[17] Nietert M. The human knee-joint as a biomechanical problem. Biomed Technol
(Berl) 1977;22:13–21.
[18] Simon JR, Alimusaj M, Wolf SI. Kinetics on ramps and stairs using floor level
mounted force plates. Gait Posture 2007;26:11–2.
[19] Wolf S, Loose T, Schablowski M, Doderlein L, Rupp R, Gerner HJ, et al. Auto-
mated feature assessment in instrumented gait analysis. Gait Posture
2006;23:331–8.
[20] Sanderson DJ, Martin PE. Gait Posture 1997;6:126–36.
[21] Selles RW, Korteland S, Van Soest AJ, Bussmann JB, Stam HJ. Lower-leg
inertial properties in transtibial amputees and control subjects and their
influence on the swing phase during gait. Arc Phys Med Rehabil 2003;84:
569–77.
[22] Browning RC, Modica JR, Kram R, Goswami A. The effects of adding mass to the
legs on the energetics and biomechanics of walking. Med Sci Sports Exerc
2007;39:515–25.
[23] Au S, Berniker M, Herr H. Powered ankle-foot prosthesis to assist level-ground
and stair-descent gaits. Neural Netw 2008;21:654–66.