Investigating the Gait of Lower Limb Amputees Regarding the

Present Classification of Mobility Grades

Katja Orlowski

, Kai-Uwe Mrkor

, Harald Loose

, Stephanie John

and Kerstin Witte

Department of Computer Science and Media, University of Applied Sciences Brandenburg,

Magdeburger Str. 50, 14770 Brandenburg/Havel, Germany

Institute III: Philology, Philosophy and Sport Science, Otto-von-Guericke University Magdeburg,

Zschokkestr. 32, 39104 Magdeburg, Germany

Keywords: Gait Parameters, Transfemoral Amputees, Mobility Grade, Objectivity.

Abstract: The mobility grade determined for German patients with a lower limb amputation based on the profile survey,

which is a subjective classification in one of the five mobility grades (0 to 4). It is recommendable to establish

objective examinations to determine the mobility grade of lower limb amputees. Gait parameters captured

with mobile sensors could be suitable for the distinction between amputees of the different groups (grade G2,

G3 or G4). Within a study, standard gait parameters were determined with the InvestiGAIT system based on

inertial sensors. A descriptive analysis of the data of the twenty-one subjects (G2: 4, G3: 6, G4: 11) indicates

that there are gait parameter (especially gait velocity, step and stride length) which can be used to make the

classification of the three mobility grades. The temporal gait parameters (stride duration, swing and stance

phase, one-leg-stance and double-leg-support) as well as angles during heel strike and toe off can be addition-

ally used for the classification. Nevertheless, further investigations have to done to get a larger database in

order to confirm the presented results regarding generalization and to check, whether the found classification

can be implemented as a kind of decision support system.

1 INTRODUCTION

The Medicare’s Functional Classification Level

(MFCL) system is used to distinguish persons with

lower limb loss in five functional levels (K-Level-0

to K-Level-4) (Balk et al., 2018). This classification

is based on the level of the amputee to walk with his

prosthesis (Agrawal et al. 2013a, 2013b, Gailey et al.

2002, 2006, Theeven et al. 2013). The established

MFCL system is a kind of an activity level categori-

zation (Dudek et al., 2008), which is questioned by

different researchers and research groups due to its

subjectivity. In Germany, a similar system is used for

the classification of lower limb amputees in order to

specify which prosthesis components should be fi-

nanced by the health insurances or other funding

agencies (e.g. trade association). The system is called

mobility grades and has the same categories as the

U.S. MFCL system. This classification is based on the

profile survey sheet of the German Medical Service

of Leading Associations of Healthcare Insurance Pro-

viders, which is also a very subjective assessment

(MDS, 2008).

Addressing the lack of objectiveness, Gailey et al.

(2002) developed the Amputee Mobility Predictor

(AMP), which is a clinical tool existing in two ver-

sions: AMPPRO (assessment with prosthesis) and

AMPnoPRO (assessment without prosthesis). The

AMP system is a 21-item measuring instrument con-

sidering different functional abilities such as sitting,

standing, walking, balancing, etc. As described in the

AMP instrument instructions, the average time for an

experienced examiner is less than ten minutes. The

low examination time is a big advantage of the AMP

tool. The result depends on the experience of the ex-

aminer. It is a “point system” with a maximum score

of 47 points, where the examiner assigns a number of

points (0 to 1, 0 to 2 or 0 to 5) per item (Gailey et al.,

2002). However, objective measures based on a stop-

watch are included in the AMP tool.

Nevertheless, the classification should be based

on an objective assessment. Agrawal et al. (2013a,

2013b) showed that there are different parameters

(such as external work and symmetry of work) which

are changing wearing different feet classified for the

functional level (tested K1-, K2- and K3-foot and gait

training with the different feet). That is why objective

336

Orlowski, K., Mrkor, K., Loose, H., John, S. and Witte, K.

Investigating the Gait of Lower Limb Amputees Regarding the Present Classiﬁcation of Mobility Grades.

DOI: 10.5220/0009175703360341

In Proceedings of the 13th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2020) - Volume 4: BIOSIGNALS, pages 336-341

ISBN: 978-989-758-398-8; ISSN: 2184-4305

methods should be used to find the best setting of

prosthesis components for each affected person.

A current research project deals with the develop-

ment of a diagnosis system based on different sensors

measuring and quantifying the abilities of the lower

limb amputees:

(1) the power of the hip muscles (maximum power,

endurance),

(2) the balance (static and dynamic), and

(3) the flexibility of the hip joint.

The developed diagnosis system addresses the

lack of objectiveness of the German profile survey ac-

tually used for the classification. The aim of the pro-

ject is the examination of different parameters

(power, balance, flexibility) regarding their potential

to be used for the determination of the mobility grade

(G0 to G4). Besides static and dynamic balance tests

conducted on a force plate, the gait of amputees is an

additional indicator of the dynamic balance, which is

also examined in the current research project. It could

be conducted without the diagnosis system and used

as an additional objective indicator.

The amputee gait is captured using the Investi-

GAIT system (Orlowski et al., 2015, 2016, 2017).

The results of the gait analysis are presented regard-

ing the classification of the transfemoral amputees

into one of the three mobility grades (G2 to G4),

which was performed during their individual clinical

or rehabilitation process after the amputation.

This paper attempts to answer the following ques-

tion: Are there gait parameters, which are different for

amputees of different mobility grades?

2 MATERIALS & METHODS

The InvestiGAIT system consisting of four inertial

sensors (6-DOF) is used to capture the gait of the

sub-

jects on a 9 m or 13 m long straight walkway, respec-

tively

. Two sensors are attached to the distal part of

the lower leg, slightly above the ankle, and two are

fixed to the upper body (in the middle of the posterior

superior iliac spine and at cervical vertebra II) using

elastic straps (Orlowski et al., 2017). The subjects were

asked to walk the walkway at least 12 times with their

self-selected gait velocity.

In-/Exclusion Criteria:

Subjects with a unilateral transfemoral amputation of

the lower limb were included in the study. The age cri-

teria was determined to 18 to 65 years, but due to dif-

ficulties to find enough subjects fulfill this criteria also

subjects older than 65 years were approved to the

study, when they feel healthy enough to perform the

given tests of the study (examination of power, bal-

ance, mobility and gait)

Participants:

Overall, twenty-one subjects with unilateral transfem-

oral amputation of the lower limb took part in the ex-

amination. Table 1 shows the distribution of subjects

in the mobility grades. Additionally, table 1 presents

the number of subjects, their mean age, body height,

stump length and time since amputation for each grade.

Informed written consent was obtained from all

subjects prior to study participation. The study was ap-

proved by the local ethics committee of the Otto-von-

Guericke university Magdeburg (no. of vote: 31/18)

and carried out in line with the Declaration of Helsinki

Data Analysis:

The gait parameters were calculated from the captured

acceleration and angular velocity based on the detec-

tion of gait events (initial contact (IC), midswing point

and terminal contact (TC)). The first two trials were

omitted from further statistical analysis due to their

training effect. The calculations were performed using

the in-house software InvestiGAIT developed in

MAT-LAB™ (TheMathworks Inc., Natick, MA,

USA).

Table 1: Anthropometric data (age body height, stump length and time since amputation) of the subjects with transfemoral

amputation. The number of subject per group (mobility grade) is given.

Mobility

Grade

Number of subjects

(male, female)

Age (yrs)

mean±sd

Height (cm)

mean±sd

Stump length (cm)

mean±sd

Time since Amputation (yrs)

mean±sd

(4 m, 0 f)

69.8

(±13.1)

177.5

(±14.1)

26.0

(±10.2)

33.0

(±30.1)

(4 m, 2 f)

63.5

(±11.6)

174.8

(±7.8)

30.8

(±4.5)

20.5

(±13.8)

(11 m, 0 f)

49.3

(±12.3)

182.9

(±7.0)

37.4

(±12.5)

13.1

(±13.0)

The investigations were conducted in both institutions

(Brandenburg and Magdeburg).

Investigating the Gait of Lower Limb Amputees Regarding the Present Classiﬁcation of Mobility Grades

337

Table 2: Gait parameters as mean value and standard deviation.

Parameters

Affected Leg Sound Leg

G2 G3 G4 G2 G3 G4

Stride (s)

1.35

(± 0.20)

1.12

(± 0.09)

1.13

(± 0.09)

1.34

(± 0.20)

1.12

(± 0.09)

1.13

(± 0.08)

Swing (%)

40.60

(± 5.38)

43.97

(± 9.56)

47.53

(± 5.23)

38.63

(± 1.82)

43.12

(± 2.51)

42.08

(± 3.27)

Stance (%)

59.40

(± 5.38)

56.03

(± 9.56)

52.47

(± 5.23)

61.37

(± 1.82)

56.88

(± 2.51)

57.92

(± 3.27)

One-Leg-Stance (%)

38.50

(± 1.90)

43.09

(± 2.56)

42.04

(± 3.42)

40.64

(± 5.39)

43.93

(± 9.51)

47.52

(± 5.37)

Double-Leg-Support (%)

18.86

(± 6.86)

12.69

(± 9.04)

10.49

(± 5.36)

16.92

(± 7.65)

12.87

(± 9.16)

10.49

(± 5.28)

Put-on-angle (°)

15.61

(± 2.61)

17.85

(± 5.84)

23.33

(± 3.57)

18.64

(± 4.75)

17.81

(± 2.02)

19.05

(± 4.09)

Take-off-angle (°)

-38.40

(± 2.69)

-38.97

(± 6.05)

-42.38

(± 7.24)

-40.23

(± 6.87)

-40.84

(± 5.18)

-44.40

(± 7.92)

Gait velocity (m/s)

0.77

(± 0.15)

1.16

(± 0.14)

1.31

(± 0.17)

Same Values as affected Leg, these pa-

rameters are not determined separately for

each leg

Stride length (m)

1.01

(± 0.08)

1.29

(± 0.10)

1.47

(± 0.14)

Step length (m)

0.60

(± 0.06)

0.71

(± 0.09)

0.77

(± 0.11)

0.62

(± 0.06)

0.73

(± 0.07)

0.85

(± 0.06)

Due to the small sample size a descriptive data

analysis was conducted. The mean values and stand-

ard deviation of standard gait parameters were con-

sidered to give an overview of the difference between

the groups (e.g. G2, G3, G4).

3 RESULTS

Table 2 contains the mean value and the standard de-

viation of the gait parameter displaying an overview

of the characteristics of the found differences of the

three considered groups (G2, G3 and G4). Further-

more, a comparison of both legs of each group and

compared to the other groups is possible.

The parameters gait velocity, stride and step

length show a clear increase from G2 over G3 to G4.

This increase can be registered for the affected and

the sound leg. While the gait velocity and the stride

length are the same for both legs, a difference for the

step length of the affected and the sound leg is clearly

visible. On average the step length of G2 (0.60 and

0.62 m) and G3 (0.71 and 0.73 m) is shorter than the

step length of G4 (0.77 and 0.85 m). Moreover, the

difference between the affected and the sound legs is

smaller in amputees of G2 and G3 (0.02 m) compared

to amputees of G4 (0.08 m).

Regarding the take-off-angle, it can be noted that

the mean values of the amputees G2 and G3 (affected:

-38.40 and -38.97°; sound: -40.23 and -40.84°) are

similar having a larger difference to the mean of the

G4 (affected: -42.38°; sound: -44.40°). The parame-

ters stride duration (s), stance phase (%), and double-

leg-support (%) show a decrease with the higher mo-

bility grades (G2 to G4), while the parameter swing

phase (%) and one-leg-stance (%) have an increase

for higher mobility grades with one exception of the

G3 for the affected leg.

4 DISCUSSION

The results suggest that a distinction between ampu-

tees of the mobility grades G2, G3 and G4 is possible

using the presented gait parameters. The gait param-

eters gait velocity, stride length and step length seem

to be characteristic and suitable for describing the

three considered mobility grades. All the other pa-

rameter are as well characteristic and can be addition-

ally used for the classification of transfemoral ampu-

tees based on gait analysis.

Batten et al. (2019) examined the gait speed of

amputees (transtibial 78, transfemoral 30, and knee

disarticulation 2 with mean age 63 (±13), range 24-85

years) based on the 10-m walk test, whereby the pa-

tients walked 12 m indoors on even floor. The deter-

mined gait speed found by Batten et al. (2019) was

slower for all groups (K2: 0.38 (0.25-0.54); K3: 0.63

(0.50-0.71); K4: 1.06 (0.95-1.18)) as well as for the

BIOSIGNALS 2020 - 13th International Conference on Bio-inspired Systems and Signal Processing

338

whole group (median (IQR): 0.52 (0.37-0.67), with a

range from 0 to 1.43 m/s).

Lemaire et al. (1993) published the average gait

speed for elderly transtibial amputees (8 subjects,

mean age of 68.75 (66 to 72) years) which is compa-

rable (mean: 1.20 (0.95 - 1.46) m/s) to the gait speed

determined in our examination. Unfortunately, no in-

formation about the mobility grade of the subjects is

given. Lemaire et al. (1993) compared the determined

gait speed with those of other investigations of ampu-

tees of little lower age groups (55-67 years, 43-77

years, 21-73 years, 39-57 years, 36 -76 years) detect-

ing an average speed ranging from 0.75 to 1.22 m/s

(1.22 m/s, 1.07 m/s, 1.07 m/s, 1.17 m/s, 0.75 m/s).

Lemaire et al. (1993) also investigated the aver-

age stride length of elderly amputees. A comparison

with data from other studies is presented. Lemaire et

al. (1993) determined an average stride length of

1.41 m of the subjects (mean age of 68.75 years)

which is little higher than the stride length found in

the other studies (range 1.10 to 1.40 m). Based on the

distinction in the mobility grades in our investigation

the determined stride length is comparable and com-

prehensible (G2: 1.01 m, G3: 1.29 m and G4: 1.47 m).

The main advantage of the gait analysis using the

InvestiGAIT system is the simplicity of usage. Gait

captures and analysis can be done within a few

minutes, are feasible almost anywhere and no large

spaces are needed. These facts support the usage in

the daily clinical routine. In contrast to the Investi-

GAIT system or other mobile gait analysis systems

(GAITRite walkway, MVN Biomech from Xsens),

3D motion capture systems (Vicon) or specialized

systems (GRAIL - Gait Real-time Analysis Interac-

tive Lab (Oude Lansink et al., 2017) or CAREN -

computer-assisted research environment (Darter and

Wilken, 2011)) are not very interesting for doctors

and clinicians working with patients in the clinical

routine due to time aspects. These systems have a

large overall complexity (required space, knowledge,

time) and do not appear to be appropriate for the clin-

ical routine.

Compared to 3D-systems the accuracy of inertial

gait systems is sometimes lower, especially to spatial

gait parameters, e.g. step and stride length. These pa-

rameters are affected by systematic errors. Due to

those facts, these parameters have small deviations.

Limitations of the Study:

Most of the subjects (38 % and 47 %) investigated in

this study were categorized with mobility grade 3 and

4. The mean age of these both groups are 59.1 (±10.9)

and 46.1 (±12.2) years. Considering the facts, it

should be noted that the more active and slightly

younger persons affected from lower limb loss took

part in the examination. While the group of subjects

with mobility grade 4 consists of 20 % persons older

than 62 years, 40 % older than 55 years and 40 %

younger than 40 years, the age distribution in G3 is

different. Within the group G3, there were 33.3 % of

subjects older than 63 years and 91.7 % older than 50

years. This does not seem to be representative of the

total amputee population.

Furthermore, the number of subjects (G2: 4, G3:

6, G4: 11) within each group has to be critically

viewed. Consequently, the results should be regarded

as tendency and a generalization is currently not pos-

sible.

There is another limitation in this study. The anal-

ysis strongly relied on the assessment of the grade of

mobility done by the subject's orthopedic techni-

cian/physician. Wurdeman et al. (2014) stated some

reasons why this is problematic: “[…] the ambiguity

under which patients are classified (Gailey et al.,

2002) and the undeniable fact that the activity level

for individuals ambulating with a prosthesis is not

possibly four distinct categories but rather repre-

sented as a continuum across a spectrum. The only

clinical tool available currently to help with patient

classification in the Amputee Mobility Predictor

(Gailey et al., 2002), but even this tool is known to

have large standard deviations making it difficult on

the individual level to objectively categorize pa-

tients.”. Gailey (2006) indicated that professionals are

able to determine the needs of the patients based on

their long-lasting experience, but there were exami-

nations from Stephen and Aitken (1987) showing that

significant differences in classifications of amputees

(assessment of mobility and self-care) were found in

clinicians of the same rehabilitation team. For that

reason, the validity of the subjective classification has

to be questioned. It is, therefore, particularly im-

portant to develop objective methods to classify the

patients. Additionally, it is necessary to implement

objective methods to find out which prosthesis com-

ponents are the best for each individual.

5 CONCLUSION

The results of the study indicate that selected gait pa-

rameters have the potential to be used to distinguish

amputees of different mobility grades. Considering

the three groups (G2, G3 and G4), the gait parameters

gait velocity, stride and step length show clear differ-

ences for all three comparisons (G2 vs G3, G2 vs G4,

G3 vs G4). These parameters are characteristic for the

gait of amputees of different mobility grades investi-

gated in the presented study. All the other presented

Investigating the Gait of Lower Limb Amputees Regarding the Present Classiﬁcation of Mobility Grades

339

gait parameters seem to be additionally useful for an

objective classification of transfemoral amputees.

The potential of the gait parameters contributing

to objective assessment has to be confirmed with fur-

ther investigations. Based on a larger sample size a

statistical analysis have to be performed in order to

make generalized statements and to develop an algo-

rithm to establish a decision support system for the

classification of patients with a transfemoral amputa-

tion in one of the three considered mobility grades.

Furthermore, it could be recommendable for fur-

ther studies to use the AMP tool additionally to the

determined mobility grade based on the profile survey

in order to have supplementary information.

ACKNOWLEDGEMENT

The authors are grateful for all subjects volunteering

to participate in this study. We also thank the German

Central Innovation Program SME (Zentrales Innova-

tionsprogramm Mittelstand - ZIM) for supporting the

project ‘‘Multifunctional diagnostic machine for pa-

tients of lower limb amputations” (ZF4096303TS6:

Multifunktionales Diagnostikgerät für Amputa-

tionspatentien) in which the diagnostic machine was

developed in cooperation with Guenther Bionics

GmbH and Peuker GmbH.

The authors wants to thank Hagen Theuer who

wrote his Bachelor Thesis in the course of the ZIM

project.

REFERENCES

Agrawal V, Gailey RS, O’Toole C, Gaunaurd I, Finnieston

A (2013). Influence of gait training and pros-thetic foot

category on external work symmetry during unilateral

transtibial amputee gait. Prosthetics and Orthotics In-

ternational, 37(5), 396–403. DOI: https://jour-

nals.sagepub.com/doi/10.1177/0309364612473501

Agrawal V, Gailey RS, Gaunaurd IA, O’Toole C, Fin-

nieston AA (2013). Comparison between micro-proces-

sorcontrolled ankle/foot and conventional prosthetic

feet during stair negotiation in people with unilateral

transtibial amputation. J Rehabil Res Dev. 2013; 50(7):

941–50. http://dx.doi.org/10.1682/JRRD.2012.05.0093

Balk EM, Gazula A, Markozannes G, et al. (2018). Lower

Limb Prostheses: Measurement Instruments, Compari-

son of Component Effects by Subgroups, and Long-

Term Outcomes [Internet]. Rockville (MD): Agency

for Healthcare Research and Quality (US); 2018 Sep.

(Comparative Effectiveness Review, No. 213.) Table 1,

Lower limb extremity prosthesis Medicare Functional

Classification Levels (K levels) Available from:

https://www.ncbi.nlm.nih.gov/

books/NBK531517/table/ch2.tab1/, (Online: 6.11.

2019)

Batten, H. R., McPhail, S. M., Mandrusiak, A. M., Var-

ghese, P. N., & Kuys, S. S. (2019). Gait speed as an

indicator of prosthetic walking potential following

lower limb amputation. Prosthetics and Orthotics Inter-

national, 43(2), 196–203. DOI: https://doi.org/

10.1177/0309364618792723

Darter BJ, Wilken JM (2011). Gait training with virtual re-

ality-based real-time feedback: improving gait perfor-

mance following transfemoral amputation. Phys Ther.

2011; 91(9): 1385–94. DOI: https://doi.org/

10.2522/ptj.20100360 PMID: 21757579

Dudek NL, Khan OD, Lemaire ED, Marks MB, Saville L

(2008). Ambulation monitoring of transtibial amputa-

tion subjects with patient activity monitor versus pe-

dometer. Journal of Rehabilitation Research & Devel-

opment 2008; 45 (4): 577–586.

Gailey R, Roach K, Applegate B, Nash M (2002). The Am-

putee Mobility Predictor (AMP): an instrument to as-

sess determinants of the lower limb amputee’s ability to

ambulate. Arch Phys Med Rehab 2002; 83: 613–627.

MDS Essen (ed.) (2008). Profilerhebungsbogen für die

Versorgung mit Beinprothesen. URL: https://www.

gkv-spitzen-ver-band.de/media/dokumente/krankenver

sicherung_1/hilfsmittel/fortschreibungen_aktuell/aelter

e/HiMi_Fortschr_Profilerhebungsbogen_Produktgrup

pe_24_2008.pdf, (Online: 18.11.2019).

Gailey RS (2006). Predictive Outcome Measures Versus

Functional Outcome Measures in the Lower Limb Am-

putee. American Academy of Orthotists and Prosthe-

tists. 2006; (6): 51-60.

Orlowski K, Loose H, Eckardt F, Edelmann-Nusser J, Witte

K (2015). Analyzing the Transfemoral Am-putee Gait

Using Inertial Sensors - Identifying Gait Parameters for

Investigating the Symmetry of Gait - A Pilot Study. In-

ternational Conference on Bio-inspired Systems and

Signal Processing (BIOSIGNALS), Lisbon, Portugal,

12.-15. January 2015.

Orlowski K, Loose H, Eckardt F, Edelmann-Nusser J, Witte

K (2016). Evaluation of gait parameters determined by

InvestiGAIT against a reference system. International

Conference on Bio-inspired Systems and Signal Pro-

cessing (BIOSIGNALS), Rome, Italy, February 2016.

Orlowski K, Eckardt F, Herold F, Aye N, Edelmann-

Nusser, J, Witte K (2017). Examination of the reli-abil-

ity of an inertial sensor-based gait analysis system. Bi-

omed. Eng.-Biomed. Tech. 2017; DOI 10.1515/bmt-

2016-0067.

Oude Lansink ILB, van Kouwenhove L, Dijkstra PU,

Postema K, Hijmans JM (2017). Effects of inter-ven-

tions on normalizing step width during self-paced dual-

belt treadmill walking with virtual reality, a randomized

controlled trial. Gait & posture. 2017; 58: 121–5.

Stephen P, Aitken R (1987). Morbidity survey of lower

limb amputees. Clin Rehabil 1987;1:181–186.

Theeven PJR, Hemmen B, Brink PRG, Smeets RJEM,

Seelen HEM (2013). Measures and procedures utilized

BIOSIGNALS 2020 - 13th International Conference on Bio-inspired Systems and Signal Processing

340

to determine the added value of microprocessor-con-

trolled prosthetic knee joints: a systematic review. BMC

Musculoskeletal Disorders 2013 14:333.

Wurdeman SR, Myers SA, Jacobsen AL, Stergiou N (2014)

Adaptation and Prosthesis Effects on Stride-to-Stride

Fluctuations in Amputee Gait. PLoS ONE 9(6): e100125.

doi:10.1371/journal.pone.0100125

Investigating the Gait of Lower Limb Amputees Regarding the Present Classiﬁcation of Mobility Grades

341