ReviewBiomechanics and physiological parameters during gait in
lower-limb amputees:AsystematicreviewYoshimasaSagawaJr.a,b,c,*,
KatiaTurcotf,Ste
phaneArmandf,AndreThevenonb,d,e,NicolasVuillermeg,h,EricWatelaina,b,c,iaUVHC,LAMIH,F-59313Valenciennes,FrancebUnivLilleNorddeFrance,F-59000Lille,FrancecCNRS,FRE3304,F-59313Valenciennes,FrancedPhysicalMedicineandRehabilitationDepartment,LilleUniversityHospital,Lille,FranceeLaboratoryofHumanMovementStudies,FacultyofSportsSciencesandPhysicalEducation,Lille,FrancefWillyTaillardLaboratoryofKinesiology,
GenevaUniversityHospitalsandGenevaUniversity,Geneva,SwitzerlandgLaboratoireTIMC-IMAGUMRUJFCNRS5525.EquipesARFIRM/AGIM3,Faculte
deMedecine,LaTronche,
FrancehCIC-IT805,INSERM/AP-HP,HopitalRaymondPoincare,EA4497,Garches,FranceiHandiBio,EA4322,UnivduSudToulonVar,LaGarde,France1.
IntroductionIn2007,
approximately1.7millionpeopleexperiencedlimblossintheUnitedStates[1].
Inthiscountry,
morethan185,000newamputationsareperformedeachyear[1,2].
Theprevalenceratein1996was4.9per1000persons.
Theprevalenceratewashigher for people aged 65 years and older: 19.4
per 1000 [1]. Theincidence rate was 46.2per 100,000for people
withvasculardisease, 5.86 per 100,000 for people with secondary
trauma and .35per 100,000for peoplewithboneor joint malignancy.
Annualacute and post-acute medical care costs associated with
caring forvascularamputees exceed $4.3billioninthe
UnitedStates[3].After lower-limb amputation, a person is routinely
prescribed aprosthesisthat mayincludeaprostheticfoot, pylon,
kneeandsocket, depending on the level of amputation and the cause.
Thereareanumber of cost-effectivecomponents
presentlyavailable[4,5], but until now, there has beennoconsensus
amongthedifferent professionals (e.g., doctors, physiotherapists,
prosthe-tists)intermsofthemaincriteriausedtoselectanappropriateGait&
Posture33 (2011)511526ARTI CLE I NFOArticlehistory:Received13
July2010Receivedinrevisedform3February2011Accepted6February2011Keywords:AmputeeBiomechanicsPhysiologicalGaitLower
limbProsthesisABSTRACTObjective: : The purpose of this systematic
reviewwas to identify which biomechanical and
physiologicalparametersarethemostrelevant, commonlyused,
abletodiscriminateand/orhavespecicclinicalrelevance
forthegaitanalysisoflower-limbamputees(LLA).Methods:
:WeperformedanelectronicsearchviathePubMed, EMBASEandISI Webof
Knowledgedatabases from1979toMay2009. Twoindependent reviewers
assessedthetitle andabstractof eachidentied study. The quality
assessment of the full text was undertaken using a 13-itemchecklist
dividedintothreelevels:A,B, andC.Results:
:Theliteraturesearchidentied584abstractstobeconsidered.
Afterapplyingtheinclusioncriteria, we reviewed the full text of a
total of 89 articles. The mean article quality was 8 2. No
A-levelarticle was found; the primary reason was a negative score
in blinded outcome assessment. Sixty-six articles(74%) corresponded
to a B-level, and two articles (2%) corresponded to a C-level.
Twenty-one articles (24%) didnot
acquireenoughpointstobeassignedtoanylevel. Inthisstudy, wepresent
anddiscussthemostcommonlyusedandmostrelevant 32parameters. Manyof
theparametersfoundwerenotreportedinenoughstudiesorin
enoughdetailtoallowausefulevaluation.Conclusion: : This systematic
reviewcanhelpresearchers compare,
chooseanddevelopthemostappropriate gait evaluation protocol for
their eld of study, based on the articles with best scores on
thecriterialistandtherelevance
ofspecicbiomechanicalandphysiologicalparameters.2011Elsevier B.V.
Allrightsreserved.Abbreviations: AB, able-bodied; EMG,
electromyography; FS, fast speed; GRF,ground reaction force; LLA,
lower-limb amputees; NS, normal speed; ROM, range ofmotion; SACH,
solid ankle cushion heel; TF, transfemoral amputation;
TT,transtibialamputation.* Corresponding author at: Laboratoire
dAutomatique et de Me caniquedInformatiqueIndustrielles et Humaines
(FRECNRS3304) Universite leMontHouy,
BatimentMalvache,59313Valenciennescedex9,France.Tel.:
+33(0)665177936/327511349;fax:+33
(0)327511317.E-mailaddress:[email protected](Y.Sagawa
Jr.).ContentslistsavailableatScienceDirectGait& Posturej our
nal homepage: www. el sevi er . com/ l ocat e/ gai t
post0966-6362/$seefrontmatter 2011 ElsevierB.V. Allrights
reserved.doi:10.1016/j.gaitpost.2011.02.003lower-limb prosthesis
that corresponds to the patient abilities andneeds. Therefore,
lower-limbprostheses areusuallyprescribedbased on empirical
knowledge. On the other hand, the healthcarepolicies require
objective and reliable criteria for prostheticprescriptions.
Depending on the prosthesis (e.g., mechanical
kneesversusmicroprocessorknees), costsmayvarywidely[6].
Somecountries havealready begunto clarifythese
healthcarepolicies,for example, the Health Care Financing
Administrations CommonProcedure Coding Systemin the USA [7] and the
Dutch Health CareInsuranceBoard inNetherlands[8].In the current
evidence-based medicine context, many
studieshavebeenconductedtoassesstherelationshipbetweenpatientcharacteristics,
the prosthesis and the environment (see referencesbelow). Most of
these studies used gait analysis to
assessbiomechanicalandphysiologicalaspectsofgait. Infact, walkingis
considered to be one of the most important aspects ofindependence
[9]. Moreover, biomechanical and
physiologicalanalysesallowaprecisequanticationofbothbodymovementsandenergy
expenditure.This systematic review aimed at identifying which
biomechan-ical and physiological parameters are the most relevant,
commonlyused,
abletodiscriminateand/orhavespecicclinicalrelevanceforthegaitanalysis
oflower-limb amputees (LLA).2. Methods2.1.
MethodusedtoidentifythestudiestoincludeWeperformedanelectronicsearchviathePubMed,
EMBASEand ISI Web of Knowledge databases from 1979 to May 2009.
Thesearchstrategywasbasedonacombinationofthefollowingsixkeywords:
amputee*, lower limb*, lower extremity*, gait*,locomotion* and
walking*. We further narrowed the eld toinclude only published
articles (i.e., conventional articles, compar-ativestudies,
evaluations) writteninEnglishorFrenchwithanentirely adultstudy
population (i.e.,+18years).2.2.
ApreliminaryselectionbasedontheabstractTwoindependentreviewersassessedthetitleandabstractofeach
study identied. Based on the abstract, studies were includedin the
full text reviewwhen they satised the following three set
ofcriteria: (1) patient characteristics, (2) descriptive or
interventionstudiesand(3) outcomemeasures.
Thesecriteriaaredescribedbelow.Participant characteristics: The
studies had to include apopulation with lower-limb amputation(s)
(1) with a traumat-ic, vascular or other cause; (2) with hip or
knee disarticulation,transfemoral (TF) or transtibial (TT)
amputation level; (3) withunilateral or bilateral involvement; (4)
with different types offoot, kneeor socket prosthesis; and/or (5)
walkingwithorwithoutassistivedevices.Descriptive or intervention
studies: The studies had to involve aparticular subpopulation
(e.g., above-knee amputations, vascu-lar causes, athletes),
aneffect of theprosthesiscomponents(e.g., sockets, knees, feet), or
an effect of tness, rehabilitationorother methodological aspects
(e.g., inuence of walking speed).Outcome measures: The studies had
to report gait-relatedbiomechanical parameters (e.g.,
spatio-temporal, kinematicand/or kinetic; transducers
basedmeasurements
andaccel-erometry)orphysiologicalparameters(e.g.,energyconsump-tion,energy
cost, heart rate,electromyography(EMG)).The exclusioncriteria of
this abstract-basedselectionweretheoretical studies, studies
validating a model/protocol,
andstudiesaboutosseointegratedprostheses, gaitunderconditionsother
than on level surface (e.g., uphill, downhill, stair
ambulation,obstacle crossing), and the effect of prosthesis mass or
prosthesissettings.In addition, the references of the full texts
selected wereexaminedtoextend our search.2.3.
MethodusedtoassessthequalityoftheselectedarticlesThe quality of the
articles selected was assessed using the 13-item checklist
developed by van der Linde et al. [8]. This
checklistwasadaptedtoevaluatenon-randomizedcontrolledtrialsusingtwoother
randomizedcontrolledtrial checklists [10,11]. Eachcriterion
wasscored 0if it wasinvalid or theanswer was noand 1 if it was
valid or the answer was yes. If a criterion was notapplicable, it
was scored 0. The focus of the article
reviewprocesswasnottheinterventionapproachesperse,
butratherthemainparametersusedduring these interventions.Four
independent reviewers piloted the adapted qualitychecklist
onthreerandomlychosenarticles inorder toassessthe content and to
certify reliable data extraction. The reviewersresults were
compared and the differences were resolved throughdiscussion. After
completing this pilot session, we standardized theitemdescriptions
toguaranteegoodinter-rater reliability. Thenal quality checklist
involved 13 items with a theoreticalmaximum score of 13 points. The
checklist covered three differentdomains: (a) adequacy of the
description of inclusion andexclusion criteria (four items, maximum
four points), (b)interventionandassessment (veitems,
maximumvepoints)and (c) statistical validity (four items, maximum
four points).
Thechecklistwasconvertedintoanelectronicdataextractionsheet,andthentwoindependentreviewersperformedthedataextrac-tion.2.4.
AnalysisInordertoensureagreementofthequalityassessments,
weperformed Kappa statistics and bootstrap condence intervals.
Allstudies included in this systematic reviewwere required
toappropriately control for selection and measurement
bias,similarly to in the review by van der Linde et al. [8].
Studies
wereclassiedas:A-levelstudiesStudieswithatotalscoreofatleast11of13points,
including 6 points in criteria sets a and b described
above;apositivescoreforblindedoutcomeassessment(criterionb7)andformeasurementtiming(criterionb8).
Thislastcriterionmeasured the time that the subjects were given to
adapt to thechangeinprosthesis. Infact,
anadequateadaptationperiodisrequired. According to Englishet al.
[12], transfemoral (TF)amputeesneedatleast3weeksof
walkingwithanewkneemechanism to ensure that their gait parameters
are stable, LLAneed a period of at least 1 week to adapt to a new
prosthetic footor to achange inprosthetic mass[12].B-level studies
Studies with a score between 6 and 12 points,including a positive
score for measurement timing (criterion b8).C-level studies Studies
with a score of at least six points out forthe criteria sets a and
b, with an invalid score on criteria b7 andb8.3. Results3.1.
ApreliminaryselectionbasedontheabstractThe preliminary literature
search identied 584 abstracts. Noneof the articles was excluded on
the basis of language. After applyingY.Sagawa Jr.etal. /
Gait&Posture33(2011)511526 512the inclusion criteria, we
included a total of 89 articles (i.e., 15%) inthefulltext review
(Fig.1).3.2. DataqualityThe agreement on data quality between the
two reviewers washigh [13]. The estimated mean Kappa value was
.92(SD .17) andthe95%condenceintervals rangedfrom.82to1.
Themeanquality score of studies was 8 (SD2) and these scores ranged
from2to 12. However, no articles corresponded to the A-level, due
mainlyto a negative score inblinded outcome assessment.
Sixty-sixarticles(74%)correspondedtotheB-level,
andtwoarticles(2%)correspondedtotheC-level. Twenty-onearticles(24%)
didnothaveenoughpoints to beassignedalevel.3.3.
ParticipantcharacteristicsThenumber of participants
rangedfrom2(intra-individualanalyses) [14] to94[15] (Table1).
Theparticipants fromthestudiesreviewedwereheterogeneous,
andablendof differentamputationlevels andcauses of amputationwas
oftenfound(Table 1).3.4. ParametersusedforgaitanalysisThe frequency
distribution of biomechanical and physiologicalparameters
usedingait analysis is illustratedinFig. 2. Afewstudies also
associated the psychological and cost parameters
withthebiomechanical andphysiological parameters(Fig. 2,
others).The main biomechanical parameters used were walking speed
(43times), knee angles (31 times), vertical ground reaction force
(30times), knee moments (27 times), hip power (26 times) and
ankleangles(22times).
ThemainphysiologicalparametersusedwereVO2(ml/min/kg) (30 times),
EMGof lower-limb muscle activity (17times)and VO2 cost(ml/m/kg)
(13times).InTable2,
thereferenceparametersaredistributedinfourareas: foot (26studies:
29%), knee(13studies: 15%), socket (6studies: 7%) and other (i.e.,
descriptive, rehabilitation, tnessstudies) (44 studies: 49%) are
represented. This list was based ontheoneproposedbyBenedetti et al.
[16], whichinvolves 122parameters related to spatio-temporal,
kinematic and
kineticparameters.Additionally,98specicparameterswerefoundandaddedtocreateanexhaustivelistof220parameters.Ofthe122parameters
proposed by Benedetti et al. [16], 78 (64%) have neverbeen
usedtoanalyze thegaitof LLA.Eight of the studies examined (9%)
[1724] had a combinationof biomechanical, physiological and
EMGoutcomes; 12 (13%)[15,2535] had a combination of biomechanical
and clinical/functional outcomes; 2 (2%) [36,37] had a combination
ofphysiological and clinical/functional outcomes; and 17
(19%)[26,28,31,3335,3747] hadacombinationof biomechanical
orphysiologicaloutcomeswithvariousquestionnairesaboutlevelsof
activity,prosthesis comfort and/orfunctionality.Table3shows
thefrequencydistributionof
the89articlesselectedfrom23biomedicaljournals.Themeanimpactfactorofthese
journals was 1.67 0.77 and ranged from .34 to 2.78. Thirty-three
percent of these studies were published in a specialized
journalconcerning LLA (impact factor .37).4. Discussion4.1.
ThemainobjectiveofthisstudyanditsrelevanceThe objective of this
systematic reviewwas to identify the mostrelevant biomechanical and
physiological parameters used toanalyze the gait of LLA. To our
knowledge, there are currently nostudies with this objective. This
reviewcan help researcherscompare, choose and develop the most
appropriate gait evaluationprotocol for their eld of study (Table
1), based on the articles withbest scores on the criteria list
(Tables 1 and 3) and the relevance ofspecicbiomechanical
andphysiological parameters(Fig. 2andTable 2). The review also
offers important information on researchelds and gait analysis
parameters used for LLA (Fig. 2 and Tables 1and 2).4.2.
ArticlequalityTwoindependent reviewersusedachecklist
adaptedtoLLAstudies to score the quality of the articles that
satised the criteria[8]. Similarly to the systematic reviewby van
der Linde et al. [8]
onthecontributionofdifferentprosthesiscomponents,
ourinvesti-gationobtainedlimitedunbiasedinformation. Inour study,
noarticlereceivedanA-levelscore,
andonlyonestudyhadablindassessor[48].In the study with a blind
assessor (10 points, B-level), Postemaet al. [48] compared four
different prosthetic feet and performed adouble-blind randomized
trial. This was possible because theprosthetic feet were covered by
a cosmetic overlay that mimickeda normal foot. In addition, the
prostheses were aligned by the sameorthopedic technician, who was
not involved in the trials. Adouble-blind experimental design is
more difcult when compar-ing other prosthetic components, such as
the knees or the socket,becausethesecomponentsareoftenapparent.
Thislimitationinperformingdouble-blindtrials,
whichpreventsmoreevidence-based results in non-pharmacological
experiments, has beenalready discussedintheliterature [4951].For
the sample size, van der Linde et al. [8] suggested that thenumber
of independent variables (K) was adequate if the ratio K:nexceeded
1:10. According to this ratio, 74% of the articles that
werevieweddidnothaveanadequatesamplesize(n = 17.2 14.2subjects,
range294subjects).
Areducedsamplesizereducestheabilitytovalidatethehypothesisandincreasestheriskof
type-IIerror. It also makes it difcult to illustrate the
discriminating capacityof a parameter. On the whole, due to the
methodological limitationsof thestudiesevaluated, it
isrecommendedtousecautionwheninterpretingthe parameters and
determining their relevance.Fig. 1. Procedurefor thestudyselections
withutilizeddatabases andfor
theliteraturesearch-and-selectioncriteria.Y.SagawaJr.etal. /
Gait&Posture33(2011) 511526
513Table1Methodologicalaspectsofreviewedarticles: authors,
participants,mainobjective(s)andlevelof evidence.Ref Articles
Participants (N) Sexandage Mainobjective(s) L[75] Aryaet al.(1995)
3Uni,Tt, ? 3M, 45? To assess the performance characteristics of the
Jaipur foot bycomparing its shock absorption capacity and inuence
on gaitstylewiththatof SACHandSeattlefeet, usingtheGRF5-U[76] Baeet
al.(2007) 8Uni,Tf, ?10Ab?M,? F,408?M,? F,242To evaluate the muscle
condition by acquiring the root meansquareelectromyogram7-B[29]
Bakerand Hewison(1990)20Uni, Tf,Tt,Tr,Va, Ca15M, 5F,61 Todetermine
therateat whichgait
recoversasmeasuredbytemporaldistancefactors(velocity
andsymmetry)6-B[77] Barnettetal.(2009) 8Uni,Tt, Va, ?7Uni,Tt, Va,
?7M, 1F,50165M, 2F,5811To investigate the gait patterns of
amputees, using eithertheamputeemobilityaid
orpneumaticpost-amputationaid9-B[33] Bergeetal.(2005) 15 Uni,Tt,
Tr,Va,In 15M, 519 Todetermine ifa
shock-absorbingpylonsaffectsgaitmechanics9-B[30] Boardetal.(2001)
11 Uni,Tt, Tr ?M, ? F,45? Tocompare thevolume changesassociated
withnormal andvacuumconditions using a total surface-bearing
suction socket8-B[45] Boonstraetal.(1993) 9Uni,Kd, Tr,Va,Ca 6M,
3F,4118 To investigate the gait patterns when wearing prostheses
ttedwitheitherthe MultiexorQuantumfoot8-B[18] Boonstraetal.(1994)
29 Uni,Tf, ? 24M, 5F,4113 Todescribe the gaitquantitatively 8-U[78]
Buckleyet al.(1997) 3Uni,Tf, Tr 3M, 4810 To quantify the
physiological energy cost of using the so-calledIntelligent
Prosthesiscompared to the cost of using
amoreconventionalpneumaticswing-phasecontrolled device8-U[37]
Casillasetal.(1995) 12 Uni,Tt, Va12 Uni,Tt, Tr10M, 2F,73712M,5014To
use bioenergetic parameters to assess a new energy-storingfoot
prothesis(Proteorfoot)bycomparingitwiththe SACHfoot
indifferentwalking situations9-B[36] Chin etal.(2002) 9Uni,Tf,
Va8Uni,Tt, Va?M,? F,632?M,?F, 722Toinvestigate whetherornot
%VO2maxasan indicatorofphysicaltnessis usefulinpredictingthe
outcome
foraprostheticrehabilitationafterdysvascularamputation7-B[79] Chin
etal.(2006) 49 Uni,Hd,Tf, Va,Tr,Ca, In34M, 15F, 676 To evaluate
physical tness and prosthetic ambulatory abilityandtoinvestigate
the leveloftnessrequiredforsuccessfulprostheticambulation8-B[80]
Chin etal.(2006) 4Uni,Tf, Tr,Ca14 Ab4M, 24810M,4F,254Toexamine the
impactofthe characteristicdifferencesbetween the Intelligent Knee
Prosthesis and C-Leg on walkingspeedandenergyexpenditureduring
walking6-B[81] Chin etal.(2009) 7Uni,Hd, Ca,In 6M, 1F,684 To
investigate the differences in energy consumption
betweenprostheticwalking andwheelchairlocomotion9-B[82]
Cortesetal.(1997) 8Uni,Tt, Tr7Ab8M, 35127M, 3210Topresent
anobjective quantitativemethod forstudyingprostheticgait,which
allowsgait patternstobecompared10-B[83] Culhamet al.(1986) 10Uni,
Tt,Tr,Va 8M, 2F,61? To evaluate the effect of the terminal
prosthetic component onthe electromyography activity of the
quadriceps and hamstringmusclegroupsduring gait7-B[39] Dattaet
al.(2004) 21 Uni,Tt, Tr,Va 19M, 2F,51.715 Toevaluate
gaitcharacteristics,costand
timeanalysis,andsubjectiveopinionwhensubjects changedfrom
aPTBtoanICEX1ttingtechnique8-B[43] Dattaet al.(2005) 10Uni, Tf ?M,
? F,38? Totestthe effect ofswitching toan
intelligentprosthesisonoxygenconsumption andgaitforusersof
pneumaticswing-phasecontrolkneejoints3-U[24]
Detrembleuretal.(2005)6Uni,Tf, Tr6Uni,Tt, Va?M, ? F,3812?M,?
F,5011To assess the inuence of self-selected gait speed, efciency
ofthe pendulummechanism andsmoothnessofcenter
ofbodymass(CMb)displacementonmetabolic energycosts8-B[84]
DoaneandHolt(1983) 8Uni,Tt, ? 8M, ?? Tocompare theSACHand
uni-axisfoot duringthe gait 7-U[85] Gaileyetal.(1994) 39 Uni,Tt,
?21 Ab39M, 461621M,316Tocompare themetabolic cost,heartrate,and
self-selectedspeedofambulation8-B[42] GardandKonz (2003) 10UniTt,
Tr,Vc 9M, 1F,5417 Toinvestigate the
effectthatshock-absorbingpylonhasonwalking10-B[86] Genin
etal.(2008) 10Uni, Tf,Tr9Uni,Tt, Tr13 Ab10M, 3559M,
35710M,3F,285Toinvestigate the effectofspeedon the
energyexpenditurerate12-B[73] Gitteretal.(1991) 5Uni,Tt, ?5Ab5M,
2050??M,? F,??To determine the biomechanical adaptations necessary
to walkwhilewearing aconventionalprostheticfoot-ankleassemblyand
subsequently to evaluate the effects of energy-storing feetinthe
restorationof normalgait characteristics8-B[19] Gitteretal.(1995)
8Uni,Tf, Tr,Ca8Ab7M, 1F,3778M, 32?To dene the relationships between
mechanical and metabolicfactorsinpathologicalgait8-B[59]
Gohetal.(1984) 11 Uni,Tf, Tt,? 11M, 4811 Toevaluate SACHanduniaxial
feetbiomechanically 6-U[87] Gohetal.(2004) 4Uni,Tt, Tr,Va 4M, 4010
Tocompare thepressuredistributionofthe pressurecast(PCast)socket
tothatofthe patellar-tendon-bearing(PTB)socket8-U[63]
Goujon-Pilletetal.(2008)27 Uni,Tf, Tr,Co,Ca33 Ab?M, ? F,5114?M,?F,
44.3?Toidentify
specic3-Dmotionpatternsforthepelvicandscapulargirdlesduringgait8-B[41]
Goujounet al.(2006) 4Uni,Tf, Tr6Uni,Tt, Tr35 Ab4M, 49135M,
1F,4513?M,? F,33?Toevaluate prostheticfeet withan
originalprotocolthatrecords fore-foot and ankle kinematics together
with the globalbodykinematicsandGRF duringgait3-U[46]
Grahametal.(2007) 6Uni,Tf, Tr 6M, 406 Toexplorethe
differencesofusingan energy-storingfootthrough
gaitanalysisandatimedwalking testtoproduceobjective measurements
and a comfort score for the patientssubjectiveopinion9-BY.Sagawa
Jr.etal. / Gait&Posture33(2011)511526 514Table 1(Continued )Ref
Articles Participants(N) Sexandage Main objective(s) L[65]
Hanetal.(2003) 6Uni,Tt, Tr,Va,Ca 4M,2F, 4110 To evaluatethe gait
patternsduring walkingwithandwithout shoesandto identifythe
differencesinbarefoot gaitpatternswhenusingdifferent
prostheticfeet9-B[71] Hansenet al.(2006) 14Uni, Tt 9M,5F, 4611 To
examinethe effect ofroll-overshapearc lengthon gait 5-U[22] Hoffman
etal.(1997) 5Bi,Tf,Tr,Co5Ab4M,1F, 2234M,1F,226To examinethe aerobic
demandsandcardio-respiratoryresponses during walking for a range of
speeds in addition tothe subjects chosenwalking speeds10-B[20]
Houdijk etal.(2009) 11Uni, Tt,Tr,Va11Ab? M,? F, 469? M,?F, 4711To
investigate whether the increasedenergy cost ofamputeegait
couldaccountforan increaseinthe mechanicalworkdissipatedduringthe
step-to-step transitioninwalking9-B[88] Hsu etal.(1999) 5Uni,Tt, ?
5M,324 To investigate and compare the differences in energy cost,
gaitefciency,andrelativeexerciseintensity ofmultiple-speedwalking
andrunningwiththree differenttypesof prostheticfeet:the
SACHfoot,the Flexfoot,andthe Re-Flex VerticalShock Pylon11-B[89]
Hurley etal.(1990) 7Uni,Tt, Tr,Va,Co4Ab? M,? F, 356? M,?F, 252To
investigatethe role ofthe contralaterallimb ingait
bydetermininglower limbs jointreactionforces andsymmetry6-B[90]
Isacov etal.(1885) 3Uni,Tf,Tr14Uni, Tf,Va3M,351113M,1F607To compare
a prosthesis with an open knee mechanism versusalocked
kneemechanismintermsof performanceandphysiologicalresponses7-U[56]
Isacov etal.(1996) 14Uni, Tt,Tr,Va 11 M,3F, 4013 To investigate
gait characteristics at two different speeds andthe inuence of
speed on symmetry of selected gait parametersobtained10-B[66]
Isakovetal.(2000) 14Uni, Tt,Tr 14 M,457 To outline
differencesbetweenbothlegsinterms ofthekinematic parametersandthe
activityofthe musclescontrolling the knees11-B[91]
Isakovetal.(2001) 11Uni, Tt,Tr 11 M,378 To investigatethe
activityofthe vastusmedialisandbicepsfemorismusclesduring
ambulation11-B[92] Jaegeret al.(1996) 11Uni, Tf,Tr,Ca3Ab11
M,36?3M,3812To study the electromyographic activity of the
supercial hipmusclesof bothlegs duringwalking8-B[25] Joneset
al.(1997) 10Uni,Tt, Va 10M, 676 To comparestanding
prostheticweight-bearingtolerance tothe forces experiencedduring
walking11-B[26] Kahleetal.(2008) 21Uni, Tf,Tr,Va,Ca, Co? M,? F,
5119 To compare subject performance using a
non-microprocessorkneemechanism versusaC-Leg9-B[93] Lacroix
etal.(1992) 5Uni,Tf,Tr,Ca3Uni,Tf,Tr4M,1F, 3943M,26?To describethe
energycostof gaitinyoung traumatictransfemoral
amputeesusingacontactsocket fordifferentkneeprosthesis3-U[94]
LeeandHong (2009) 5Uni,Tf,Tr 5M,482 To investigatethe effectof
articialanklemobility inthesaggitalplane onthe gait of
amputeeswearing astance-andswing-controlled kneeprosthesis6-U[34]
Leeetal.(2006) 4Uni,Tt, Tr,Co 4M,4116 To evaluate the gait
performance and perception of amputeeswhile using a exible
elliptican-shank monolimb as comparedto athickercircular-shank
monolimbandaconventionalmodular prosthesis9-C[17] Lehmann
etal.(1993) 10Uni,Tt, ? ? M,? F To
quantifymetabolicrateandefciency,biomechanicalgaitparameters,andprosthesiscomfort
ofthe SeattleAnkle/LiteFootcomparedto theSACHfoot8-B[27]
LemaireandFisher(1994)12Uni, Tt,Tr12Ab12 M,72412M,703To assess the
incidence of OA and relate these ndings to thekinematic
walkinggaitcharacteristics11-B[95] Lemaireetal.(1993) 8Uni,Tt, Tr
8M,692 To examinethe kinematicandkinetic gait parameters 8-U[14]
Linden etal.(1999) 2Uni,Tf,? 2M,341 To describeamethodology
forinvestigatingthe effectsofvarious prostheticfeet
onamputeegait7-U[40] McNealy andGard(2008)
4Bi,Tf,Tr,Co9Ab4M,41238M,1F,284To determine if adding prosthetic
ankle motion would improvegait7-B[96] Michaud etal.(2000)
9Uni,Tf,Tt, Tr 9M,4516 To assess the qualitative and quantitative
differences in pelvicobliquity6-B[32] Mizuno etal.(1992) 10Uni,Tt,
?5Ab10M, 51145M,41?To investigate the functional features of
various prostheses tofacilitatethe taskof prescribingthem3-U[97]
Murray (1980) 10Uni,Tf,Tr30Ab10M, 41?? M,?F, ? ?To document several
previously unreported motion patterns inthe lower limbs and the
trunk using prostheses with constant-friction kneecomponents8-B[98]
Murray etal.(1983) 10Uni,Tf,Tr? Ab10M, 41?? M,?F, ? ?To measurethe
multipledisplacementpatternsin ordertocompare
thestridedimensionsandtemporalcomponentsduring
slow,free-speed,andfastwalking, usinga constant-friction
kneecomponent toahydraulicswing-control kneecomponent8-B[31]
Nadollek etal.(2002) 22Uni, Tt,Va ? M,? F, 7210 To establish the
relationship between weight distribution, theanteriorposterior
andmedio-lateralcenterof exertedpressure,the strengthofthe
hipabductormuscle andgaitparameters10-B[68] NolanandLee(2000)
4Uni,Tf,Tr4Uni,Tt, Tr10Ab? M,? F, 288? M,?F, 416? M,?F, 2910To
quantify the sagittal plane kinematic characteristics and thejoint
momentandpowerdemands placedonthe intactlimbduring walking9-B[99]
Nyska etal.(2002) 3Uni,Tt, Tr 3M,50? To comparethreeprosthesesthe
SACH,the energy-storingSeattle prosthesisand theIndian
Jaipurprosthesis,whichismore
prevalentineasterncountries6-UY.SagawaJr.etal. /
Gait&Posture33(2011) 511526 515Table1(Continued )Ref Articles
Participants (N) Sexandage Mainobjective(s) L[100] Pagliaruloet
al.(1979) 15 Uni,Tt, Tr,Co 12M, 3F,2911 Todetermine themetabolic
energycostof walkingwithcrutchesandwith prostheses8-U[21]
Paysantetal.(2006) 10Uni, Tt,Tr10Ab10M, 391410M,39?Toinvestigate
the inuenceofgroundsurfaceonwalkinginreal-worldsituations12-B[101]
Pinzur etal.(1991) 7Uni,Tt, Va7Ab?M, ? F,64??M,? F,??To evaluate
the pressures applied at several at-risk locationson theplantar
surfaceofthe intactfoot6-B[58] Pinzur etal.(1991) 7Uni,Tt,
Va5Uni,Tt, Va?M, ? F,???M,? F,??Toevaluate thephasic myoelectric
activityof thequadricepsand the hamstring muscles in both the sound
and amputatedlimbs of active amputees that do a limited amount of
walking5-U[102] Pinzur etal.(1992) 25 Uni,Tf, Tt,Va5Ab, Va?M, ?
F,58??M,? F,54?Toevaluate themetabolic demandsduringgait 5-U[48]
Postemaetal.(1997) 10Uni, Tt,Tr,Va, Co9M, 1F,4911 Toobtain
abetterunderstandingof theuser benetsof theenergystoringand
releasebehaviorof someprosthetic feetthatareregularly usedinpatient
care10-B[67] Powersetal.(1994) 10Uni, Tt,Tr 10M, 4515 Toexamine the
jointmotionandGRF characteristicsofvedifferent
prostheticfeet10-B[15] Powersetal.(1996) 22 Uni,Tt, Va72 Ab15M,
7F,6011?M,? F,??To establish a relationship between muscular torque
capabilityandstridecharacteristics9-B[54] Powersetal.(1998) 10Uni,
Tt,Va10Ab10M, 6275M, 5F,519Touse
anintegratedapproach(i.e.EMG,kinematics,kinetics) to evaluate the
knee mechanics and to identify factorscontributingtoabnormal
knees11-B[60] Princeetal.(1998) 5Uni,Tt, Tr,Va ?M, ? F,4214
Toevaluate thenet energystored ordissipatedandthenrecovered, as
well as spring efciency, in order to
distinguishamongthreeprosthetic feet8-U[61] Rabuffetti etal.(2005)
11 Uni,Tf7Ab10M, 1F,36257M, 3819Todetermine theeffects
ofthebody/socket interfaceonamputeemotor strategies7-B[103]
RoyerandWasilewski(2006)10Uni, Tt,Tr,Va, Co9M, 1F,4110 Toexamine
frontalplanemoments 10-B[72] Sadeghietal.(2001) 5Uni,Tt, Tr,Va 3M,
2713 Togaininsight intohow hip musclepowers
cangenerateorabsorbmusclepower activityon
theamputatedsidetocompensateforthelack ofnormal anklemuscle
powerfunction,andhow these compensatorymechanismscaninuencemuscle
poweractivitiesinthe soundlimb9-B[55]
SandersonandMartin(1997)6Uni,Tt, Tr6Ab6M, 4076M, 337Toquantifythe
adaptationof thejointkinetics intheankle,kneeandhip ofboth
prostheticandintact limbs10-B[28] Sapinetal.(2008) 5Uni,Tf,
Tr,Co6Uni,Tf, Tr23 Ab5M, 53146M, 5811?M,? F,51?To describe amputee
gait patterns using two different
uni-axiskneejointswithahydraulicswing-phasecontrolandinparticulartostudy
the effectof amechanicalknee/ankle link8-B[23] Schmalzet al.(2002)
7Uni,Tt, Tr8Uni,Tt, Tr6Uni,Tf, Tr6Uni,Tf, Tr?M, ? F,4917?M,?
F,4417?M,? F,336?M,? F,369To dene more clearly the inuence of
prosthetic alignment onmetabolicenergy
consumptionduringwalking8-U[64] Segaletal.(2006) 8Uni,Tf, Tr9Ab7M,
1F,47136M, 3F,298Tocompare thedifferences ingaitbiomechanics
ofsubjectswearing theC-Legversusa
non-computerizedprosthesis(MauchSNS) usingaintra-subject
design6-B[104] Segaletal.(2009) 10Uni, Tt,Tr,Va, In,Ca10Ab9M,
1F,56126M, 4F,4414Todetermine ifa
commerciallyavailabletorsionadapter
canreducetranstibialamputeejointtorquescomparedtoarigidadapterduringstraight-linewalkingandturning
gait7-B[70] Seroussiet al.(1996) 8Uni,Tf, ?8Ab?M, ? F,37??M,?
F,32?To estimate the compensatory strategies of ankle, knee, and
hipmusclesinamputatedandintact limbs7-B[74] Silvermanetal.(2008) 14
Uni,Tt, Tr,Va10Ab13M, 1F,4597M,
3F,3312Tobetterunderstandcompensatorymechanismsbyexaminingthe
anterior/posteriorGRFimpulses andjointkinetics,acrossa
widerangeofsteady-statewalkingspeeds7-B[62] Sjodahletal.(2002)
9Uni,Tf, Tr,Ca18 Ab5M, 4F,33339M, 9F,368Todescribe the
effectofatraining programon temporalparametersand
onmovements,momentsandpower inthesagittalplaneinthe
pelvis,hip,kneeand anklejointssimultaneously8-B[105]
Snyderetal.(1995) 7Uni,Tf, Va 7M, 62, 8 To study the loading
patterns for ve different prosthetic feet 10-B[57] Suet al.(2007)
19 Bi,Tt14 Ab?M, ? F,5318?M,? F,26?Tocharacterize walkingpatterns
8-B[106] Torburnetal.(1995) 7Uni,Tt, Va7Uni,Tt, Tr7M, 6287M, 5116To
compare the effects of ve different
commercially-availableprostheticfeet onenergy expenditure9-B[107]
Traballesietal.(2008) 16 Uni,Tf, Va8Uni,Tt, Va11M, 5F,61116M,
2F,5617To verify whether or not the energy cost of treadmill
walkingtests and during free walking are really equivalent, and if
not,seewheretherearemeasurementdifferences9-B[44]
Underwoodetal.(2004)11 Uni,Tt, Tr 8M, 3F,4212 To examine the
effects of two prosthetic feet (the conventionalsemi-rigid foot
versus the dynamic elastic response Flex foot)on the3-D
kineticpatternsduringsteady-stategait9-C[108] VanJaarsveldet
al.(1990)5Uni,Tt, ? 5M, 3915 Toevaluate thedifferences
inabsorptionofhigh-levelaccelerationsamongcommercially-availableprosthetic
feetandthe inuenceofthe shoetype onthese differences9-B[35] Wirta
etal.(1991) 19 Uni,Tt, ? 15M, 4F,4816 To analyze the effect of ve
commonly prescribed devicesongait
andtoproposeadditionalguidelinesforselecting andprescribing
them6-B[38] Wright etal.(2008) 10Bi, Tt,Tf, Tr,Co10Ab10M, 401210M,?
F ?Todetermine thephysiological costofwalking 6-BY.Sagawa Jr.etal.
/ Gait&Posture33(2011)511526 5164.3.
ThemostcommonparametersforLLAgaitanalysisThearticleresultsoftenpresent
spatio-temporal parametersrst. We found 17 parameters (Table 2),
which were mostlycompared among AB groups or among other LLA
studies.Unsurprisingly, self-selectedgait speedwas themost
commonparameter(cited43timesin39.5%of thearticles),
followedbycadence, step length (19% of the articles) and stride
length (16% ofthe articles). These parameters can be obtained
withrelativeprecisionusing a variety of instruments (e.g.,
accelerometers,optoelectronic cameras, time-sensor cells,
footswitches) and theyrepresent a global gait predictor [52,53].
The most commonly usedkinematics and kinetic parameters are
discussed in Sections 4.4.24.4.4
respectively.Thirty-threeparameterswerebasedongroundreactionforce(GRF)
and impulse provided by force plates, mostly on the
verticalandanteroposterioraxes.
Thearticlesonprostheticfootcompo-nents employedthese parameters
most often(42.3%)
becausechangesintheabsorptionandpropulsivecharacteristicsof
theprostheticfeetcanbedirectlyreectedintheGRFandimpulses(Table
2).There were seven physiological parameters found in thearticles,
with oxygen consumption and stump EMG signals beingthe most
frequent. Oxygen consumption was expressed
inmillilitersperminuteperkilogramof bodyweight(ml/min/kg)or in
milliliters per meter per kilogram of body weight
(ml/m/kg).Theseparametersweremostoftenemployedtodescribegaitinfunction
of speed and to assess prostheses aiming to reduce
energyexpenditure or to achieve energy expenditure close to that of
able-bodied (AB) subjects (e.g., mechanical versus
microprocessorknees, SACH versusdynamic feet)(Table 2).The
EMGintensity and duration were employed to describe
andquantifystumpmuscleactivity.
Thiscannotbequantiedusingbiomechanical parameters since muscle
contractions do not occurto produce movement distally but to
compensate for the absenceof adjacentlower-limb structures and
tomaintain theprosthesisstability [54].4.4.
RelevantgaitparametersforLLA4.4.1.
Spatio-temporalparametersAdditional spatio-temporal parameters that
may be relevant forLLAgaitanalysisincludestancetime,
stancetimeratioandsteptimeratio. Duringthegaitcycle,
thestancephaseonthesoundlimb is slightly longer than on the
prosthetic side. This contributesto a more asymmetrical gait
[14,30,40,55,56]. Subjects withTable 1(Continued )Ref Articles
Participants(N) Sexandage Main objective(s) L[109]
ZhangandLee(2006) 12Uni, Tt 12 M,509 To evaluateandcompare
theload-tolerance ofdifferentregions ofthe stump9-B[47] Zmitrewicz
etal.(2006) 15Uni, Tt,Tr 14 M,1F, 587 To examine the inuence of
energy storage and return (ESAR)feet andmulti-axisankleson the
abilityto generateGRFs andthecorrespondingimpulsesduring
walkingatthesubjects self-selectedspeeds10-BAbbreviations: Ref,
references; L, level of evidence [8]; Ab, abled-body; Uni,
unilateral; Bi, bilateral; Hd, Hip disarticulation; Tf,
transfemoral; Kd, knee disarticulation; Tt,transtibial;
Tr,traumatic;Va,vascular;Ca, cancer; Co,congenital;In,
infection;?,not given;U,unassigned.Fig. 2.Thelistandthe frequencyof
mainparametersusedinthe
gaitanalysisoflower-limbamputees.Abbreviations: t, time; sup,
support; CV, coefcient of variation; A, articular; C, center; GRF,
ground reaction force; GRI, ground reaction impulse; COP, center of
pressure; Acc,acceleration;M,manual;EMG,electromyography;
RPE,ratingof perceivedexertion;n,
numbers;diag,diagnosis.Y.SagawaJr.etal. / Gait&Posture33(2011)
511526 517Table 2The list of references for all parameters in
biomechanics, physiology and other domains used in the gait
analysis of lower-limb amputees for different themes. The number
with a star means parameters with a signicant
difference.Therepeatedreferencenumber indicatesthe number
oftimesthataparameterwasused.Parameters NArt. Feet Knee Socket
OthersSpatio-temporal17parametersStancetime (s) 14 47*/83 98*/43
63*,63*,57*,72*,66*,73*,76*, 55*,77*/103,56Swingtime(s) 5 45*/45,83
66*/103,56Stridelength(m) 15 14*,54*,105*,67*/106, 83,100 61,39
102*,76*,15*,27*/31,23Cycletime (s) 3 47,83 76*Cadence(step/min) 17
40,47,54,106,106,83,100 61,39 57*,62*,72*,102*,76*,21*,15*,
77*/103,21Velocity(m/s) 35
46*,71*,41*,14*,105*/40,47,106,106,45,83,10064*,19*,37*,98*,26*,90*/4361,39
63*,62*,72*,102*,102*,20*,
81*,107*,76*,21*,15*,27*,29*/23,85Steptime(s) 3 45
66*/56Singlesupporttime (s) 2 83 66*Doublesupporttime (s) 4 41*/83
31,56Foot-attime (s) 3 54*,60*,59*Steplength(m) 17
47*,41*,17*,32*/40, 106,83 61* 63*,57*,62*,72*,66*,21*/57,
103,31,56,21Stepwidth(m) 1 57*Stancetime ratio (%) 4 14* 64*/64 30*
72*Swingtimeratio (%) 1 31Steplengthratio (%) 6 46*,32*/71,71 43
30* 21Timinggait events(%) 1 56CV(%) 3
72*,68*/95Groundreactionforces andimpulses23 parametersVert.max.
F.LR (N/kg) 9 46*,14*,105*,105*,67*,67*,17*/4664*/64
34*/57,55Vert.max. F.TS(N/kg) 3 46* 34*,55*Fore-aft.max.F.LR (N/kg)
4 47*,75* 55*/57Fore-aft.max.F.TS (N/kg) 5 47*/46 28* 55*/57Vert.F.
IC(ImpactF.peak)(N/kg) 2 75*,75* 42*Vert.F.
rateIC(ImpactF.rate)(N/s/kg) 1 75*,75*Vert.F.
diff.sound/prosthetic(%) 3 71*,71*,71*,71*/71 39 31*Vert.F.
excursion(N/kg) 2 41*,32*Vert.Imp. (Ns/kg) 2 75*
27*,27*Fore-aft.F.excursion(N/kg) 1 32*Fore-aft. +Imp. (Ns/kg) 4
47*,75*/47,48,75 74*,74*/74,74Fore-aft.Imp. (Ns/kg) 5
47*,47*,75*,17*/48 74*,74*Fore-aft.Imp.Ratio (%) 2 47*,47*
74*Fore-aft.F.pattern 1 32*Med-lat.F.pattern 1
82*Groundreactionmoment(Nm/kg) 1 63*COPexcursion (m) 1
31,31Effectivefootlength COP(m) 1 71*,71*,71*,71*Vertmin. F.MS
(N/kg),Fore-aft.Min. F.MS(N/kg), Med-lat.min. F.LR (N/kg),Med-lat
max.F.MS (N/kg),Med-lat.max.
F.TS(N/kg)Groundreactionforceandimpulsetimes(%stride)11
parametersTimeat Fore-aft.max.F.LR 2 47*,67*Timeat
Fore-aft.max.F.TS 2 47*,67*,67*Timeat Vert.max. F.LR,Time
atVertmin.F. MS,Timeat Vert.max. F.TS, Timeat Fore-aft.min.F.
MS,Time atMed-lat.min. F.LR,Time atMed-lat max.F.MS,Time atMed-lat.
max.F.TSTrunkangles(Deg)3parametersTotalsagittalplaneexcursion 1
63*Totalcoronalplaneexcursion 1 63*Totaltransversalplaneexcursion 1
63*Pelvisangles(Deg)
9parametersY.SagawaJr.etal./Gait&Posture33(2011)511526518Totalsagittal
planeexcursion 4 46 61* 63*,57*Totalcoronal planeexcursion 2
63*,96*Totaltranversalplaneexcursion 1
63*Pelvis/scapulargirdlesr.phase 1 63*Pelvissaggitalpattern 1
62Min.rot.sagittal plane,max.rot.coronalplane,max.rot.coronalplane,
max.rot.transverseplanePelvisangletimes(%stride)4parametersTime
atMin.rot.sagittal plane,timeat max.rot.coronalplane,time
atmax.rot.coronalplane,time atmax.
rot.transverseplaneHipangles(Deg)13 parametersMax.ex.atLR 2
62*/62,56Max.ext.instancephase 2 46 77*Totalsagittal planeexcursion
5 48*/40 61* 57*,92*Totalcoronal planeexcursion 1 45*Hipsagittal
pattern 1 95FlexionatIC,Flexionattoe off,max. ex.in
swingphase,max.add. instancephase,max. abd.inswing
phase,totaltransverseplaneexcursion, max.int.rot.instance
phase,max.exte.rot.inswingphaseHipangletimes
(%stride)7parametersTime at max. ex. at LR, time at max. ext. in
stance phase, time at max. ex. in swing phase, time at max. add. in
stance phase, time at max. abd. in swing phase, time at max. int.
rot. in stance phase, time at max. exte. rot. in
swingphaseKneeangles(Deg)13parametersMax.ex.atLR 10 40*,54* 64,64
57*,24*,65*,62*,66*, 56*,56*,33*Max.ext.instancephase 2
70*,62Flexionattoe off 2 64,64 56*Max.ex.inswingphase 4 46*/40
64*,64*/64 56*Totalsagittal planeexcursion 3 17*,17* 39
68*FlexionatIC,total coronalplaneexcursion,max.
add.instancephase,max. add.inswing
phase,totaltransverseplaneexcursion, max.int.rot.instancephase,max.
exte.rot.inswingphaseKneesagittalpattern 3 94
55*/95Kneeangletimes(%stride)7parametersTime at max. ex. at LR,
time at max. ext. in stance phase, time at max. ex. in swing phase,
time at max. add. in stance phase, time at max. add. in swing
phase, time at max. int. rot. in stance phase, time at max. exte.
rot. in swingphaseAnkleangles(Deg) 10parametersMax.plant.ex.atLR 5
14*,48*,59*, 84* 77*Max.dorsiex.instance phase 4 46*,105*,67*, 67*
57*Flexionattoe off 1 70*Max.dorsiex.inswing phase 1
28*Totalsagittal planeexcursion 4 48*,17* 68*,82*/68,68Totalcoronal
planeexcursion 1 102*Max.eversioninstance phase 1
44Anklesaggitalpattern 2 95*,55*FlexionatIC,max.inversion
inswingphaseAnkleangletimes(%stride)5parametersTime
atMax.plant.ex.atLR,timeat max.dorsiex.instance phase,time
atmax.dorsiex. inswingphase,time atmax.eversion instancephase,time
max.inversion inswingphaseLower-limbmoments(Nm/kg)
2parametersMax.lower-limbmoment 1 70*Lower-limbmomentpattern 1
55Hipjoint moments(Nm/kg) 8parametersMax.ex.momentatIC 5 40*,14*
68*,70*/15Max.ext.momentatTS 3 40 68*,70*Firstmax.add. moment 2
103*/15Max.exte.rot.moment 1 104*Max.ex.momentatMS 1 40Hipsagittal
pattern 1 55Max.int.rot.moment,secondmax. add.momentHipjoint
momenttimes(%stride)6parametersTime atmax.ex.momentatIC,time atmax.
ext.momentat TS,time atrst max.add.moment,time atsecondmax.
add.moment,time atmax.exte.rot.moment, timeat
max.int.rot.momentY.SagawaJr.etal./Gait&Posture33(2011)511526519Table2(Continued
)Parameters N Art. Feet Knee Socket OthersKneejointmoments(Nm/kg)
10parametersFirstmax.ext.moment 5 40*,44*, 14* 39,39
55*/55Firstmax.ex.moment 5 54* 64*,64*,64* 62*,70*,55*Secondmax.
ext.MomentTS 2 40 55Firstmax.add. moment 3 44*,44* 64
103*Max.int.rot.Moment 1 104*Secondmax. ex.moment 2 44*
68*,68*Kneesagittalmomentpattern 2 54 94Max.abd.Moment,
secondmax.add.Moment,Max.exte.rot.momentKneejointmomenttimes
(%stride)8parametersTime atrstmax.ex.moment 1 54*Time at rst max.
ext. moment, time at second max. ext. moment at TS, time at Max.
abd. moment, time at rst max. add. moment, time at second max. add.
moment, time at max. exte. rot. moment, time at max. int. rot.
momentAnklejointmoments(Nm/kg) 5parametersMax.plant.ex.moment 4
44*,14* 68*/62Max.dorsiex.Momentat TS 8 71*,71*,14*, 17*/40
57*,62*,70*, 70*,55*Max.dosiex.momentdiff.sound/prosthetic(%) 1
71*,71*Max.eversionmoment,
max.inversionmomentAnklejointmomenttimes(% stride)4parametersTime
atmax.plant.ex.moment 2 40* 55*Time atmax.dorsiex. momentatTS,
timeat max.eversionmoment, timeatmax. inversionmomentHipjoint
powers(W/kg) 8parametersMax.ext.generateLR (H1S) 9 40*,14*, 73*/44
64 74*,74*,57*,62*,70*Max.ex.absorb(H2S) 6 40,44 64
57*,62*/72Max.ex.generatePS(H3S) 9 40*,46*/44 64
57*,62*/72,68,70Max.abd.absorb (H1F) 1 72*Max.abd.generate (H3F) 1
72*Max.ext.rot.generate(H2T) 1
72*Kneejointpowers(W/kg)6parametersMax.ext.absorbLR(K1S) 7 40*,14*,
73*,73* 64 62*,72*,27*Max.ext.generate(K2S) 7 54*,73* 64
74*,74*,62*,72*,27*Max.ext.absorbTS(K3S) 5 44*/40 64
72*,68*,68*Max.ext.absorb(K4S) 1 64Max.add,abd. generate(K2F) 2 44*
72*Anklejointpowers(W/kg)3parametersMax.dorsiex.absorb(A1S) 3
44*,44* 64 74*Max.plant.ex.generateatPS(A2S) 12 40*,46*,
74*,73*/44,48 64
74*,57*,62*,72*,70*Recoveryenergy(plantarexgenerate/dosiexabsorb(%))
1 60*Physiological5parametersHeartrate(bpm) 7 100*, 88* 90*, 90*
22*,85*/102,107Bloodpressure(mmHg) 1 100VO2(ml/min/kg) 15 17,106
79*,43*,78*, 37*/43,78,37,37,8086*,23*,23*,23*,23*,
22*,85*,107*,21*,21*,36*/86, 23,23,85,20,107,21VO2cost(ml/m/kg) 12
100*, 88* 19*,37*,80*/37 86*,22*,102*,81*,
107*/24,20Respiratoryquotient(%) 1
86*EMG2parametersStumpmuscleactivityintensity 7 54*,83*,83*/95
66*,92*,76*/66, 66,58Stumpmuscleactivitytime 3 54*
91*,91*,91*,91*,92*,
92*,92*,92*/91,91,91,91,91,92,92Psychological4parametersProsthesiscomfort
1 46Prosthesispreference 2 45*
64*Y.SagawaJr.etal./Gait&Posture33(2011)511526520Satisfactionrate
5 44 43*,37*,26* 39*Rateof perceivedexertion 3 42*,22*/33Others26
parametersHipmuscle force 1 31Weight-bearingdistalendability 1
109Socketpressurepeak 3 99* 109*, 109* 101*Socketpressurecontact
time 1 99*Socketpressurecontact area 1 99*Footstore/dissipateenergy
1 60*Shockfactor(accelerometer) 1
17*Trunkcenterofmassverticalexcursion 1 19COMverticalexcursion 1
24*COMext.mechanicalenergy 2 24*,20*COMenergyrecovery 1
24COMnegativeexternalmechanicalwork ofthe leadinglimb 1
20*COMpositiveexternal mechanicalwork ofthe trailinglimb 1
20*COMnegativeexternalmechanicalwork tostep-to-step transition 1
20*Costof components(total) 1 39*Costnumber ofvisits 1
39Costprosthesisprocesstime 1 39*Timedwalkingtest 1 46Steps/week 1
33Falls 2 26*,26* 36*PEQscore 1 26Comorbiditynumber 1
36*Socketvolume 1 30*, 30*Socketdisplacement 1 30*,
30*Osteoarthritisclinicaldiagnosis 1
27*,27*Number:reference.Number*:referenceandsignicant
differenceusingsuchparameter.Others:
descriptive,rehabilitation,tnessandpylonstudies.Abbreviations:
NArt, number of articles, which utilized such parameter; F, force;
Imp, impulse; Vert, vertical; Med-lat, mediolateral; IC, initial
contact; LR, loading response; MS, mid stance; TS, terminal stance;
diff, difference;
ex,exion;ext,extension;add,adduction;abd,abduction;int.internal;
rot,rotation;exte.external;plant,plantar; dorsiex,dorsiexion;
r,relative;COMcenterofbody
mass.Y.SagawaJr.etal./Gait&Posture33(2011)511526521unilateral
amputations rely more on their sound leg to compensatefor some of
the deciencies associated with prostheses
[57].However,theincreasedloading periodmayexplainthedevelop-mentof
complications intheremaining limb [58].Some studies have
demonstrated positive effects from
specicprostheticcomponents.Subjectsfeltmorecondentwithsuchaprosthetic
foot and therefore compensated less on the sound side[14].
Theuseanappropriatedsocket, andthroughappropriatetting, the degree
of gait asymmetry can be reduced
[30].AppropriatedttingmaybeofferingLLAbettercontrolovertheprosthesis
position, possibly by improving proprioceptionandforce transfer to
the prosthesis. This probably improves thesymmetryof their
gait[30].Foot at time is another spatio-temporal parameter that
seemsappropriatefor LLAgait analysis. ABsubjectsreachedfoot
atphaseat1217%of
thegaitcycle[54,59]comparedtosubjectsusingasolidanklecushionheel(SACH)foot,whowerefoundtomakeheel
contactonlyfortherst20%[60]or44.5%[59]. Theinability to achieve
foot at contact during loading phase could beattributed to the
compromised plantar exion of some prostheticfeet. For example,
Seattle Light foot has been reported to provideonly 238 of motion
[54]. Reduced motion during weightacceptance mayresult ina periodof
instabilityduringwhichbalance relies onthe rear foot. It is
therefore likely that thecontractionof theadjacent muscles(e.g.,
quadricepsandham-strings) could represent an attempt to provide
joint stability duringthisphaseof thegaitcycle.Powers et al. [54]
suggest that individuals with transtibial
(TT)amputationneedtostabilize the knee duringweight acceptance
duetotheprolongedheel-onlycontactcausedbyreducedprostheticfoot
mobility. Nevertheless, other prosthetic feet, such as
theGolden-Ankle one, attainedthe foot at phase signicantlyearlier
inthegaitcycle(14%) comparedwiththosettedwitheithertheSeattle Light
foot (21%), or the SACH foot (20%) [60]. Golden-Anklefoot users
showed a more natural gait pattern, probably because itsankle
articulation approached natural ankle function [60].4.4.2.
JointangleparametersReducedhiprangeof motion(ROM) inthesagittal
planeisassociated with ischiumsocket interference in TF amputees
[61].ReducedhipROM, mainlyduringhipextension,
alsoemployscompensatory mechanisms on the pelvis and on the sound
side inorder to maintain adequate speed. Rabuffetti et al. [61]
suggestedthat, when the hip on the prosthetic side is extended the
ischiumsocket interference limits the physiological ROM. In order
tomaintain functional step length, the pelvic range of motion
(ROM)in the sagittal plane and the hip exion of the sound side
increasedcomparedtoABsubjects. Inaccordancewithprevious
studies[61,62], Goujon-Pillet et al. [63] found that, for TF
subjects (88 58),the pelvis ROMis twice that of AB subjects (48
18). In the long-term,this motor strategy may cause lower back
pain, whichis oftenreportedbyTF amputees. The literature
reviewedsuggests, thatprosthetic feet have no inuence [40] or a
minor inuence [48] on thehip motion inthe sagittal plane inTT and
TF amputees.Pelvic ROMin the frontal plane is increased, in LLA
compared toABsubjects. Suet al. [57] foundasignicant
differencewhencomparing LLA at self-selected speeds (8.48 2.88) to
AB subjects atslowspeeds (6.38 2.18). LLA lifted the pelvis on the
swing side whilewalking. Thiscompensatorymotion, knownashiphiking,
isoftenobserved inindividualswith unilateral TTor TF amputationsand
isbelievedtocompensatefortheinabilitytodorsiextheprostheticankle.
Hip hiking increases the prosthetic foot clearance [57];however, it
mayalsorequire additional metabolic energytoliftbody mass against
gravity, thus reducing gait efciency. Individualswithbilateral
TTamputations displaybilateral hiphiking, whichprobably requires
increased energy expenditure during walkingcompared
tounilateralamputees[57].Knee exion during the loading phase has a
shock-absorbingeffectwhichisimportantinthepreventionofwearandtearofweight-bearing
joints [56]. For AB subjects or the LLAs sound side,this parameter
value is about 15188 [54,57,64]. For TT
amputees,itislimitedto9128[54,56,57];however, forTFamputees,
itisoftenabsent or negative[64]. Duringgait speedchanges,
thisparameter is less variable on the prosthetic side than on the
soundside [56,57]. It can be inuenced by many conditions, such
shoed orbarefootwalking[65],
thetypeofsocket[66]andrehabilitation[62]. However, Segal et al.
[64] demonstrated that TF patients whoused microprocessor
controlled knee (e.g., C-Leg) designed toallow controlled stance
phase knee exion, this did not normalize.They suggested that,
although stance phase knee exion ispossible, it is difcult for TF
amputees toachieve it, possiblybecausethey associatethis action
with buckling andfalling.Table3Themainscientic journalspertaining
togait analysesandthe distributionofarticles.Journals IF2008 N
Articles L(meanSD,minmax)JBiomech 2.784 1 14 7GaitPosture 2.743 8
20,23,24,54,55,74,103,107 91.3, 711IEEETransNeuralSyst RehabilEng
2.7 1 34 9Phys Ther 2.190 2 15,100 8.50.7,89ArchPhys MedRehabil
2.159 11 17,22,27,37,46,47,63,67,70,78,98 8.91.2,711ScandJ
RehabilMed 215 2 18,90 7.50.7,78ClinBiomech 2 2 44,76 81.4,
79JRehabilMed 1.983 1 91 10EurJ ApplPhysiol 1.931 1 86
12ClinOrthopRelatRes 1.893 1 92 8ClinRehabil 1.840 1 43 3AmJPhys
MedRehabil 1.695 6 19,65,72,73,80,81 8.30.5,89JRehabilResDev 1.446
13 21,26,33,35,42,57,60,64,87,96,104106 8.31.8,612BullProsthet Res
1.29 1 97 8FootAnkle 1.061 1 58 6Orthopedics 0.588 2 58,102
5PhysiotherResInt 0.561 1 31 10ProsthetOrthot Int 0.377 30
25,2830,32,36,3841,45,48,56,59,61,62,66,68,71,75,77,80,8285,89,95,108,1097.52,311IntJRehabilRes
0.343 1 99 6AnnPhys RehabilMed 1 93 3JOrthopSports PhysTher 1 88
11ProcInst ofMechEngPartH, JEng Med 1 94 6Abbreviations:IF,impact
factor;N,number;L, levelofevidenceobtainedfromthe vanderLinde
etal.[8] criterialist withamaximum scoreof13 points.Y.Sagawa
Jr.etal. / Gait&Posture33(2011)511526 522Plantar exion in the
early stance phase is an importantparameter, which quanties the
prosthetic foots capabilities to beat ontheground,
allowingmoreoorcontact,
whichpermitsbetterstabilityduringthestancephase.
Thisparameterchangesconsiderably with the prosthetic foot design.
Most of the
dynamicprostheticfeetarecomposedbyabladewithoutanarticulatedanklejoint,
sothat themeasuredankleplantar exionmotionduring the early stance
phase is primarily due to heel compression.This feature has been
reported by Postema et al. [48], who foundthat
theMultiexarticulatedfoot hasbetter plantar
exionatnormalspeeds(NS)(8.38 .48)andfastspeeds(FS)(6.58 .78)than
other so-called exible feet: Springlite II (NS: 6.18 3.28; FS:4.98
.28); CarboncopyII (NS: 4.58 .98; FS: 4.68 .78); andSeattle light
foot (NS: 4.68 .68; FS: 4.88 .78).The next important parameter is
the dorsiexion motion duringthemid-to-latestancephase.
Prostheticfeet providelessanklemotion during the stance phase than
the natural ankle motion inAB subjects: 12.58 3.18 at self-selected
speeds for LLAversus20.28 3.58atslowspeedsforABsubjects[48,67].
Thisparameteralsoseemstobegreatlyinuencedbyprostheticfootdesigns.
Asexplained above, for dynamic prosthetic feet the dorsiexion
motionalsodependsofthebladecapacitytobend.Poweretal.[67]foundbetter
results for Flex foot (23.28) and Quantum (19.58) feet than
forSeattle (15.18),Carbon copy II(12.18)orSACH (128)feet.The last
parameter is the total ankle ROM in the sagittal plane.Nolan and
Lee [68] reported a 218 ROM for AB subjects, a 208 ROMfor the LLAs
prosthetic side and a 268 ROMfor their sound side.
TheincreasedROMforthesoundlimbwasattributedtothelimitedanklemovement
ontheprostheticlimb: LLAneedtoincreasesound limb length in order to
clear the prosthetic limb during theswingphase,
acompensationmethodthat hasbeenpreviouslyreported forTF amputee
gait.Nolan andLee [68] also foundthatthe type of prosthesis used
can affect lower-limb kinematics. Theyobserved in their study that
TT amputees using a SACHfoot had 108ROM less in the prosthetic
ankle and 158 ROM greater in the soundanklethanwhenthey
usedamulti-axis foot.4.4.3. JointmomentparametersMcNealy and Gard
[40] found that the mean hip joint momentin the sagittal plane at
initial contact in LLA was +.8 Nm/kg, whichwas more than twice the
one observed in AB subjects (+.3 Nm/kg).It appearedthat thehipjoint
intheLLAgroupwas critical ingeneratingpower duringthe earlystance
phase. Since the TFamputees investigated did not have any active
control of the ankleand knee joints, the moment generated at the
hip probably assistedforwardprogression[40].
NolanandLee[68]alsofoundsimilarresults for the same moment on the
sound hip side. Seroussi et al.[70] found that the hip moment
became a exor moment(external) substantially earlier for the sound
limb than theprosthetic limbof LLAand both limbsofAB subjects.The
greatest difference between LLAand ABsubjects for the
kneemomentinthesagittal planeoccursduringtheexternal exionmoment in
the loading-response phase. According to Powers et al.[54],
ABsubjectsrelyonthekneeextensors, whichareactingtocontrol knee
exion during weight acceptance. This was conrmedthroughanEMGof
thevastus lateralis, whichdemonstratedanaverage intensity of 29% of
a maximal muscle contraction test andlasted until 22% of the gait
cycle [54]. In contrast, the moment dataobtained fromthe TT
amputees suggested negligible extensordemandas the knee exionmoment
was signicantlysmallerduringthestancephase. However,
theEMGresultsshowedthatactivityof thevastus lateralis
intheTTgroupwas signicantlygreaterthanthatoftheABgroup. Onaverage,
theTTamputeesdemonstrateda25%increaseinthemeanintensityofthevastuslateralisandsignicantlylongerdurationofactivity(lastinguntil33%
of the gait cycle) compared to normal values
[54].Theinconsistencybetweenthemoment
dataandtheEMGresultsforthispopulationindicatesadiscrepancybetweenthemechanical
measurements of knee extensor demandandthephysiological response.
Hence, cautionis neededwhileinter-preting the results [54]. For TF
amputees, the knee exionmomentis even smaller because the
prosthetic knee cannot replacequadricepsmuscleactivity. However,
Segal etal. [64] foundasmall but signicant exor knee moment when TF
subjects usedthemicroprocessor kneeC-Leg(.14 .05 N m/kg),
comparedtosubjectsusingamechanicalknee(.06 .07 N m/kg).
Thesevaluesare still modest when compared with those of AB
subjects(.47 .1 N
m/kg).LLAdevelopanexternalplantaranklemomentovertherst20%ofthegaitcycle,
whereasABsubjectsdisplaythismomentonly for the rst 9% of the gait
cycle [40]. AB subjects are able torapidly achieve the foot at
phase and to transfer the load onto
theleadinglegduringtheloading-responsephaseinpreparationforsinglesupport.
Indoingso,
theyadvancethecenterofpressureunderthefoottoapositionanteriortotheanklejointaxis[40].Additionally,
the stance phase knee exion in AB subjects
probablyhelpstoreducethedurationofthenegativeanklemoment[40].LLAspendmoretime,
comparedtoABsubjects, rotatingtheirprosthetic legs forward until
the foot at phase is achieved. This isdue to reduced movement of
the prosthetic ankle, combined withthe absenceofstance phaseknee
exion[40].Sjodahl et al. [62] found that, both before and after
arehabilitationprogram, therewasabumpintheexternal
ankledorsiexionmomentat26%ofthegaitcycleonthesoundside,indicating
that LLA were using a vaulting movement.
Thisadaptationwasprobablyusedtofacilitatetoeclearanceontheprosthetic
side[55,68].Suet al. [57] reportedthat the external ankle
dorsiexionmoment at the end of the stance phase of unilateral TT
amputeeswas only 6070%of the moment foundinABsubjects.
Theysuggested that this reduction was due to the absence of the
ankleplantar exors. Another possibility is that the keel of the
prostheticfoot was functionally shorter thanthat of the biological
foot,reducingthemomentarmbetweentheGRFsandtheanklejointcenter
during stance [71]. However, the ankle moment on both theprosthetic
andsoundsidesmayvaryaccordingtotherehabilita-tion program [62] or
prosthetic foot characteristics [14,17,44]. AsUndewood et al. [44]
have shown, LLA using a dynamic Flex foothave 15% greater ankle
dorsiexion moment when compared to anon-dynamic
footprosthesis.4.4.4.
JointpowerparametersOnboththeprostheticandsoundsides,
LLAhaveagreateramplitude and duration of hip joint power lasting
throughout therst half of the stance phase (5560% of gait cycle)
than AB subjects(20%of gaitcycle) [72]. Sadeghi etal.
[72]calledthishipjointpowertherstlimbpropellerparameter.
Theincreaseinhipextensor power output in TT amputees corresponds to
an increasein the gluteus maximus and hamstring activity, as shown
by EMG.Increased hip extensor use in the early stance phase appears
bothto control knee exion during limb loading and to pull the
trunkforward after heel strike as the foot makes contact with the
oor.Using hip extensors as a source of power represents a
compensa-tion for the lack of appropriate ankle function during
push-off [72].The excessive work produced at the hip may further
contribute tothe increasedenergy expenditure inLLAwalking[57].The
increased hip power has been conrmed by other
studies[44,57,62,73,74]. Seroussi et al. [70] found an increase of
270% inthe concentric hip extensor work in the early stance phase
on thesound side of LLA when the prosthetic limb was pushing off.
Gitteret al. [73] have shown an increased use of concentric hip
extensorenergy generation during the early and mid stance (H1S)
inY.SagawaJr.etal. / Gait&Posture33(2011) 511526 523ambulation
with a non-dynamic foot, such as the SACH. Inaddition, Underwood et
al. [44] conrmed the inuence of the typeofprostheticfoot.
Theyshowed, thattheH1Spowergenerationtended to be smaller on both
the prosthetic limb (a 66% decrease)and the sound limb (a 75%
decrease) when LLA were wearing thedynamic Flex foot, compared to a
SACH foot. This nding suggeststhat the dynamic Flex foot requires
less passive and active stability,compared toaconventional
prostheticfoot.ComparedtoABsubjects,
LLAexhibitanotherlargesourceofpower generation at the hip joint
prior to toe off (H3S) [40,70,72].The H3S power burst represents
theactionof the hipexorspulling the lower-limbupwardandforward. As
showninABsubjects, the ankle power generation and the H3S hip power
burstsoccur simultaneouslyduringthegait
cycleastheleadinglimbprepares for the swing phase. Thus, LLA
apparently compensate forthe lackof energyproducedinthe prosthetic
ankle (A2S) byincreasing power generation at the hip (H3S)
immediately prior totoeoff [40,70,72].Sadeghi et al. [72] reported
a reductionof 63%in the knee powerabsorption during the
loading-response phase (K1S) on
theprostheticsideofTTamputees,comparedtothesoundside.Theproblemwas
worse for TF amputees because the negative externalknee moment did
not allow prosthetic knee motion in the stancephase.
Thustherewasnoenergystorage(K1S)orenergyreturn(K2S) in the knee
power curves during this phase of the gait cycle[40]. Consequently,
thekneejoint didnot
contributetoshockabsorption,energystorageorenergy
returnsformostofthegaitcycle[40].Both theTTandTFamputees appearedto
usethehipjoint as an alternative means of shock absorption,
powerabsorption and power generation [40,72], or they
overloadedtheirsound knees[62].There are twoimportant parameters
for ankle power: the
anklepowerabsorption(A1S)attheloading-responsephaseandthemid stance
phase and the ankle power generation (A2S) at the endof the stance
phase. The deformation properties of dynamic feetallowfor a greater
power absorption (A1S) during weightacceptance and, consequently, a
trend towards a greater externaldorsiexionmoment; power
generation(A2S) takes place atpush-off. Althoughtheuseof
thedynamicfoot increasedthedorsiexion moment and push-off power of
the prosthetic ankle,thesevalues werestill well belowthosefor
theanklesof ABsubjects [44]. Seroussi et al. [70] found that the
LLAs prostheticankle at push-off reachedonlyabout 20%of the ankle
workgenerated by the AB subjects ankle.Suetal.
[57]foundthattheprostheticsideof
TTamputees,walkingataself-selectedspeed(.38 .18 W/kg),
generatedfourtimes less ankle power than the AB subjects, walking
at a slow speed(1.26 .38 W/kg). For the same ankle parameter,
Sadeghi et al. [72]found a reduction of 76% for the TT prosthetic
side compared to thesound side, which was expected because of the
limited deformationcapability of prosthetic feet [72]. Ankle power
parameters seem to begreatly inuencedby the type of prosthetic foot
[14,44,46,48,70] or bytherehabilitationprogram [62].It has been
shown that the ankle plantarexors at push-off are amajor source of
energy generation when walking [70]. Therefore,the decrease in
prosthetic ankle push-off represents a substantiallossof
themechanical workgeneratedbythelower extremityduringwalking,
resultinginanumberofpossiblecompensatorymechanisms[70,72].Seroussietal.[70]describedthreepossiblecompensatory
mechanisms in their study. First, the sound ankle ofLLA generates
approximately one third more work than the anklesof
theABsubjectsduringpush-off. Second, thedecreaseintheprosthetic
ankle push-off creates an increase in the concentric hipextensor
work in the early stance phase on the sound side of LLA.Third,
there is a relative increase in the concentric hip pull-off
intheprosthetic limb(H3S) [70].4.5.
ThemainlimitationsofstudiesandparametersdiscussedThereisalackof
studiesintheliteratureonbiomechanicalmodels adapted for LLA, which
would take into account thecharacteristics of prosthetic components
(i.e., mass, center ofarticulation, center of mass, moment of
inertia). All resultsforankle motion in the sagittal plane
described above were obtainedusing a method in which the joint
position of the prosthetic anklewas assumed to be in the same
position as that of an intact ankle.However,
arecentstudy[69]hasshownthatthemotionoftheprosthetic feet is
different from that of the intact ankle. Thus,
theuseofthismethodissubjecttosystematicerrorsasitcouldnotreect the
real motion of the prosthetic foot. The same is probablythe case
for other prosthetic components, such as the
polycentrickneemechanisms, andthe sameerrors
shouldbeexpected.Moreover, manyof thestudies investigatedstatedthat
theprostheticcomponentswereadjustedandalignedbyanexperi-enced
prosthetist, without providing any more information.However,
individual tting characteristics may affect
motionanalysisandleadtoerror. Themanufacturersof somedynamicfeet,
for example, recommend that for an optimal function, thesefeet
require to be set in slight plantar exion. Another example
isthesocketof TFsubjects, whichisoftenmanufacturedwithaninitial
exion to facilitate support during gait. Thus, if thisinformationis
not explicit, it couldbias interpretations of
LLAgaitanalysis.Indeed, in our opinion, further work is required to
establish amethodological consensus or a guideline before
clinicallymeaningful measurements canbe condently based on LLA
gaitanalysis.5. ConclusionIn this study, we presented and discussed
the 32 most commonparameterspublisheduntilnow.
Manyoftheparametersfoundwere not reported in enough studies or in
enough detail to allow auseful discussion. The diversity in the
outcomes selected todescribe the LLA gait cannot be explained by
differences inresearchobjectivesonly.
Thisparameterdiversitysuggeststhatthere is a lack of consensus
among researchers about the aspects ofgaitthat areimportantwhen
assessingLLAoutcomes.Finally, due to the methodological
inconsistency of the
studiesandtheparameterdiversity,ithasbeendifculttoidentifythemainparametersthat
shouldbeusedingait
analysisforLLA.Althoughthissystematicreviewcannotcorrectthebiasesandmethodological
awsobservedintheoriginal studies, itcouldhelp guide future studies
for choosing parameters, thus bringingabout a more evidence-based
compromise. Further researchemphasizing the clinical usefulness of
LLA gait analysis may helpdetermine which gait parameters provides
the most usefulinformation.AcknowledgmentsThe present research work
has been supported by OSEO projectnumberA0607009N, International
CampusonSafetyandInter-modalityinTransportation,
theNord-Pas-de-Calais Region, theEuropeanCommunity,
theRegionalDelegationforResearchandTechnology,
theMinistryofHigherEducationandResearch, andthe National Center for
Scientic Research. The authors gratefullyacknowledgethe
supportofthese institutions.Conict of interestThe authors state
that no conicts of interest are present in thisresearch.Y.Sagawa
Jr.etal. / Gait&Posture33(2011)511526 524Appendix A.
Supplementary dataSupplementary data associated with this article
can be found, intheonline version,at
doi:10.1016/j.gaitpost.2011.02.003.References[1]
Amputationcoalitionof America[cited2010March, 23];
Availablefrom:www.amputee-coalition.org.[2] Dillingham TR, Pezzin
LE, MacKenzie EJ. Limb amputation and limb decien-cy: epidemiology
and recent trends inthe United States. SouthMed J2002;95:87583.[3]
DillinghamTR, Pezzin LE, Shore AD. Reamputation, mortality, and
health carecostsamongpersonswithdysvascularlower-limbamputations.
ArchPhysMedRehabil2005;86:4806.[4] HafnerBJ, SandersJE, Czerniecki
J, FergasonJ. Energystorageandreturnprostheses: does patient
perception correlate with biomechanical analysis?ClinBiomech
2002;17(5):32544.[5] Michael JW. Modernprosthetickneemechanisms.
ClinOrthopRelat Res1999;361:3947.[6]
GerzeliS,TorbicaA,FattoreG.Costutilityanalysisofkneeprosthesiswithcompletemicroprocessorcontrol
(C-leg)comparedwithmechanical tech-nology
intrans-femoralamputees.Eur JHealthEcon2009;10:4755.[7] GaileyRS,
RoachKE, ApplegateEB, ChoB, CunniffeB, LichtS, etal.
Theamputeemobilitypredictor:aninstrumenttoassessdetermiantsof
thelower-limb amputees ability to ambulate. Arch Phys Med
Rehabil2002;83:61327.[8] vander LindenML, HofstadCJ, Geurts AC,
PostemaK, GeertzenJH, vanLimbeek J. A systematic literature review
of the effect of different prostheticcomponets on human functioning
with a lower-limb prosthesis. J Rehabil ResDev2004;41:55570.[9]
SansamK, NeumannV, OCornnorR, BhaktaB.
Predictingwalkingabilityfollowinglowerlimbamputation:
asystematicreviewof theliterature. JRehabilMed2009;41:593603.[10]
van Tulder MW, Assendelft WJ, Koes BW, Bouter LM. Method guidelines
forsystematicreviewsintheCochraneCollaborationBackReviewGroupforspinaldisorders.Spine1997;22:232330.[11]
Verhagen AP, de Vet HC, de Bie RA, Kessels AG, Boers M, Bouter LM,
et al. TheDelphi list: a criteria list for quality assessment of
randomized clinical trialsfor condctingsystematic reviews
developedbyDelphi consensus. J Clinepidemiol 1998;51:123541.[12]
English RD, Hubbard WA, McElroy GK. Establishment of consistent
gait aftertting ofnewcomponents. JRehabilResDev1995;32:325.[13]
Landis JR, Koch GG. An application of hierarchical kappa-type
statistics in theassessment of majorityagreement
amongmultipleobservers. Biometrics1977;33:36374.[14] van der Linden
ML, Solomonidis SE, Spence WD, Li N, Paul JP. A methodologyfor
studying the effects of various types of prosthetic feet on the
biomechan-ics oftrans-femoralamputeegait. JBiomech
1999;32:87789.[15] Powers CM, Boyd LA, Fontaine CA, Perry J. The
inuence of lower-extremitymuscle force on gait characteristics in
individuals with below-knee amputa-tions
secondarytovasculardisease.Phys Ther1996;76:36977.[16] Benedetti
MG, Catani F, Leardini A, Pignotti E, Giannini S. Data managementin
gaitanalysisforclinicalapplications.ClinBiomech 1998;13:20415.[17]
Lehmann JF, Price R, Boswell-Bessette S, Dralle A, Questad K.
Comprehensiveanalysis of dynamic elastic response feet: Seattle
Ankle/Lite Foot versus SACHfoot. ArchPhys
MedRehabil1993;74:85361.[18] Boonstra AM, Schrama J, Fidler V,
Eisma WH. The gait of unilateral transfe-moral amputees.Scand
JRehabilMed1994;26:21723.[19] Gitter A, Czerniecki J, Weaver K. A
reassessment of center-of-mass dynamicsas a determinate of the
metabolic inefciency of above-knee amputeeambulation. AmJPhys
MedRehabil1995;74:3328.[20] Houdijk H, Pollmann E, Groenewold M,
Wiggerts H, Polomski W. The energycost for the step-to-step
transition in amputee walking. Gait Posture2009;30:3540.[21]
Paysant J, Beyaert C, Datie AM, Martinet N, Andre JM. Inuence of
terrain
onmetabolicandtemporalgaitcharacteristicsofunilateraltranstibialampu-tees.
JRehabilResDev2006;43:15360.[22] Hoffman MD, Sheldahl LM, Buley KJ,
Sandford PR. Physiological
comparisonofwalkingamongbilateralabove-kneeamputeeandable-bodiedsubjects,and
a model to account for the differences in metabolic cost. Arch Phys
MedRehabil1997;78:38592.[23] Schmalz T, Blumentritt S, Jarasch R.
Energy expenditure and biomechanicalcharacteristics of lower
limbamputeegait: theinuenceof prostheticalignment anddifferent
prosthetic components. Gait Posture2002;16:25563.[24] Detrembleur
C, Vanmarsenille JM, De Cuyper F, Dierick F.
Relationshipbetweenenergycost, gait speed, vertical displacement of
centreof bodymass and efciency of pendulum-like mechanism in
unilateral amputee gait.Gait Posture2005;21:33340.[25] JonesME,
SteelJR, BashfordGM, DavidsonIR.
Staticversusdynamicpros-theticweightbearinginelderlytrans-tibial
amputees. ProsthetOrthotInt1997;21:1006.[26] Kahle JT, Highsmith
MJ, Hubbard SL. Comparison of nonmicroprocessor kneemechanism
versusC-Legonprosthesisevaluationquestionnaire,stumbles,falls,
walkingtests, stair descent andkneepreference. J Rehabil
ResDev2008;45:113.[27] LemaireED, FisherFR.
Osteoarthritisandelderlyamputeegait.
ArchPhysMedRehabil1994;75:10949.[28] Sapin E, Goujon H, de Almeida
F, Fode P, Lavaste F. Functional gait analysis oftrans-femoral
amputees usingtwodifferent single-axis prosthetic
kneeswithhydraulicswing-phasecontrol:kinematicandkineticcomparisonoftwoprosthetic
knees.ProsthetOrthot Int2008;32:20118.[29] Baker PA, Hewison SR.
Gait recovery pattern of unilateral lower-limb
ampu-teesduringrehabilitation.ProsthetOrthot Int1990;14:804.[30]
Board WJ, Street GM, Caspers C. A comparison of trans-tibial
amputee suctionandvacuumsocket conditions.ProsthetOrthot
Int2001;25:2029.[31] NadollekH, BrauerS, IslesR.
Outcomesaftertrans-tibial amputation:therelationship between quiet
stance ability, strength of hip abductor musclesandgait.
PhysiotherResInt2002;7:20314.[32] Mizuno N, Aoyama T, Nakajima A,
Kasahara T, Takami K. Functional
evalua-tionbygaitanalysisofvariousankle-footassembliesusedbybelow-kneeamputees.ProsthetOrthotInt
1992;16:17482.[33] BergeJS, CzernieckiJM, KluteGK.
Efcacyofshock-absorbingversusrigidpylonsforimpact
reductionintranstibial amputeesbasedonlaboratory,eld,andoutcome
metrics.JRehabilResDev2005;42:795807.[34] LeeWCC, ZhangM, ChanPPY,
BooneDA. Gaitanalysisoflow-costexible-shank transtibial prostheses.
IEEE Trans Neural Syst Rehabil Eng 2006;14:3707.[35] WirtaRW,
MasonR, CalvoK, GolbransonFL. Effect ongait
usingvariousprostheticankle-footdevices.J
RehabilResDev1991;28:1324.[36] Chin T, Sawamura S, Fujita H, Ojima
I, Oyabu H, Nagakura Y, et al. %VO2max asan indicator of prosthetic
rehabilitation outcome after dysvascular amputa-tion.Prosthet
OrthotInt2002;26:449.[37] Casillas JM, Dulieu V, Cohen M, Marcer I,
Didier JP. Bioenergetic comparisonofa new energy-storing foot and
SACH foot in traumatic below-knee vascularamputations.ArchPhys
MedRehabil1995;76:3944.[38] Wright DA, Marks L, Payne RC. A
comparative study of the physiological costsofwalkinginten
bilateralamputees.ProsthetOrthot Int2008;32:5767.[39] Datta D,
Harris I, Heller B, Howitt J, Martin R. Gait, cost and
timeimplicationsforchangingfromPTBtoICEXsockets.
ProsthetOrthotInt2004;28:11520.[40] McNealy LL, Gard SA. Effect of
prosthetic ankle units onthe gait of persons withbilateral
trans-femoral amputations. Prosthet Orthot Int 2008;32:11126.[41]
GoujonH, BonnetX, Sautreuil P, MaurissetM, DarmonL, FodeP, etal.
Afunctionalevaluationofprostheticfootkinematicsduringlower-limbam-puteegait.
ProsthetOrthot Int2006;30:21323.[42] Gard SA, Konz RJ. The effect
of a shock-absorbing pylon on the gait of
personswithunilateraltranstibialamputation.
JRehabilResDev2003;40:10924.[43] Datta D, Heller B, Howitt J. A
comparative evaluation of oxygen consumptionand gait pattern in
amputees using Intelligent Prostheses and conventionallydampedknee
swing-phasecontrol. ClinRehabil2005;19:398403.[44] Underwood HA,
Tokuno CD, Eng JJ. A comparison of two prosthetic feet on
themulti-joint and multi-plane kinetic gait compensations in
individuals with aunilateraltrans-tibialamputation.
ClinBiomech2004;19:60916.[45] BoonstraAM,FidlerV,Spits GM,TuilP,
Hof AL.Comparisonof gait usingaMultiexfoot versus a Quantumfoot
inknee disarticulationamputees.ProsthetOrthot Int1993;17:904.[46]
GrahamLE, DattaD, Heller B, Howitt J, Pros D.
Acomparativestudyofconventionalandenergy-storingprostheticfeetinhigh-functioningtrans-femoralamputees.ArchPhys
MedRehabil2007;88:8016.[47] Zmitrewicz RJ, Neptune RR, Walden JG,
Rogers WE, Bosker GW. The effect
offootandankleprostheticcomponentsonbrakingandpropulsiveimpulsesduringtranstibialamputeegait.
ArchPhys MedRehabil2006;87:13349.[48] Postema K, Hermens HJ, de
Vries J, Koopman HF, Eisma WH. Energy storageand release of
prosthetic feet. Part 1: biomechanical analysis related to
userbenets.ProsthetOrthot Int1997;21:1727.[49] BoutronI, TubachF,
GiraudeauB, RavaudP. Methodologicaldifferencesinclinical trials
evaluatingnonpharmacological andpharmacological treat-mentsofhip
andkneeosteoarthritis.JAMA 2003;290:106270.[50] BoutronI, TubachF,
GiraudeauB, RavaudP. Blindingwas
judgedmoredifculttoachieveandmaintaininnonpharmacologicthanpharmacologictrials.JClinEpidemiol
2004;57:54350.[51] Boutron I, Guittet L, Estellat C, Moher D,
Hrobjartsson A, Ravaud P. Reportingmethods of
blindinginrandomizedtrials assessing
nonpharmacologicaltreatments.PLoSMed2007;4:e61.[52] Montero-Odasso
M, Schapira M, Varela C, Pitteri C, Soriano ER, Kaplan R, et
al.Gait velocity in senior people. An easy test for detecting
mobility impairmentincommunityelderly. JNutr
HealthAging2004;8:3403.[53] Montero-Odasso M, Schapira M, Soriano
ER, Varela M, Kaplan R, Camera LA,et al. Gait velocity as a single
predictor of adverse events in healthy seniorsaged75
yearsandolder.J GerontolABiolSciMedSci2005;60:13049.[54] PowersCM,
RaoS, PerryJ. Kneekineticsintrans-tibialamputeegait.
GaitPosture1998;8:17.[55] Sanderson DJ, Martin PE. Lower extremity
kinematic and kinetic adaptationsinunilateral
below-kneeamputeesduringwalking. Gait Posture1997;6:12636.[56]
Isakov E, Burger H, Krajnik J, Gregoric M, Marincek C. Inuence of
speed ongait parameters and on symmetry in trans-tibial amputees.
Prosthet OrthotInt1996;20:1538.Y.SagawaJr.etal. /
Gait&Posture33(2011) 511526 525[57] Su PF, Gard SA, Lipschutz
RD, Kuiken TA. Gait characteristics of persons
withbilateraltranstibialamputations.J
RehabilResDev2007;44:491501.[58] Pinzur MS, Asselmeier M, Smith D.
Dynamic electromyography in active andlimited
walkingbelow-kneeamputees.Orthopedics1991;14:5357.[59]
GohJCH,SolomonidisSE,SpenceWD,PaulJP.BiomechanicalevaluationofSACHanduniaxial
feet.Prosthet OrthotInt1984;8:14754.[60] PrinceF, Winter DA,
SjonnensenG, Powell C, WheeldonRK. Mechanicalefciencyduringgaitof
adultswithtranstibial amputation: apilotstudycomparing the SACH,
Seattle, and Golden-Ankle prosthetic feet. J Rehabil
ResDev1998;35:17785.[61]
RabuffettiM,RecalcatiM,FerrarinM.Trans-femoralamputeegait:socket-pelvis
constraints and compensation strategies. Prosthet Orthot
Int2005;29:18392.[62] Sjodahl C, JarnloGB, SoderbergB, PerssonBM.
Kinematicandkineticgaitanalysisinthesagittal planeof trans-femoral
amputeesbeforeandafterspecialgait re-education.Prosthet
OrthotInt2002;26:10112.[63] Goujon-PilletH, SapinE, FodeP,
LavasteF. Three-dimensional motionsoftrunk and pelvis during
transfemoral amputee gait. Arch Phys Med Rehabil2008;89:8794.[64]
Segal AD, Orendurff MS, Klute GK, McDowell ML, Pecoraro JA, Shofer
J, et al.Kinematic and kinetic comparisons of transfemoral amputee
gait using C-LegandMauchSNSprosthetic knees.J
RehabilResDev2006;43:85770.[65] HanTR, ChungSG, ShinHI.
Gaitpatternsof transtibial amputeepatientswalkingindoors
barefoot.AmJPhys MedRehabil2003;82:96100.[66] Isakov E, KerenO,
Benjuya N. Trans-tibial amputee gait:
time-distanceparametersandEMGactivity.Prosthet
OrthotInt2000;24:21620.[67] Powers CM, Torburn L, Perry J, Ayyappa
E. Inuence of prosthetic foot
designonsoundlimbloadinginadultswithunilateral
below-kneeamputations.ArchPhys MedRehabil1994;75:8259.[68] Nolan L,
Lees A. The functional demands on the intact limb during walking
foractivetrans-femoral andtrans-tibial amputees. Prosthet
OrthotInt2000;24:11725.[69] RusawD, Ramstrand N. Sagittal plane
position of the functional joint centre ofprosthetic
foot/anklemechanisms.ClinBiomech2010;25:71320.[70] Seroussi RE,
Gitter A, Czerniecki JM, Weaver K. Mechanical work adaptations
ofabove-knee amputee ambulation. Arch Phys Med Rehabil
1996;77:120914.[71] Hansen AH, Meier MR, Sessoms PH, Childress DS.
The effects of prosthetic footroll-over shape arc length on the
gait of trans-tibial prosthesis users.
ProsthetOrthotInt2006;30:28699.[72] Sadeghi H, Allard P, Duhaime
PM. Muscle power compensatory mechanismsinbelow-knee
amputeegait.AmJ Phys MedRehabil2001;80:2532.[73] Gitter A,
Czerniecki JM, DeGroot DM. Biomechanical analysis of the inuenceof
prosthetic feet on below-knee amputee walking. Am J Phys Med
Rehabil1991;70:1428.[74] SilvermanAK, FeyNP, PortilloA, WaldenJG,
BoskerG, NeptuneRR. Com-pensatory mechanisms in below-knee amputee
gait in response to
increasingsteady-statewalkingspeeds.GaitPosture2008;28:6029.[75]
Arya AP, Lees A, Nirula HC, Klenerman L. A biomechanical comparison
of theSACH, Seattle and Jaipur feet using ground reaction forces.
Prosthet Orthot Int1995;19:3745.[76] BaeTS, ChoiK, HongD, MunM.
Dynamicanalysisofabove-kneeamputeegait.
ClinBiomech2007;22:55766.[77] Barnett C, Vanicek N, Polman R,
Hancock A, Brown B, Smith L, et al. Kinematicgait adaptations
inunilateral transtibial amputees
duringrehabilitation.ProsthetOrthot Int2009;33:13547.[78] Buckley
JG, Spence WD, Solomonidis SE. Energy cost of walking: comparisonof
intelligent prosthesis withconventional mechanism.
ArchPhysMedRehabil1997;78:3303.[79] Chin T, Sawamura S, Shiba R.
Effect of physical tness on prosthetic ambula-tion
inelderlyamputees.AmJ Phys MedRehabil2006;85:9926.[80] ChinT,
Machida K, Sawamura S, Shiba R, OyabuH, Nagakura Y, et
al.Comparison of different microprocessor controlled knee joints on
the energyconsumptionduringwalkingintrans-femoral amputees:
intelligent kneeprosthesis(IP)versusC-Leg.Prosthet
OrthotInt2006;30:7380.[81] Chin T, Oyabu H, Maeda Y, Takase I,
Machida K. Energy consumption duringprosthetic walking and
wheelchair locomotion by elderly hip
disarticulationamputees.AmJPhys MedRehabil2009;88:399403.[82]
CortesA, VioscaE, HoyosJV, PratJ, Sanchez-LacuestaJ.
Optimisationoftheprescriptionfor trans-tibial (TT) amputees.
Prosthet Orthot Int 1997;21:16874.[83] Culham EG, Peat M, Newell E.
Below-knee amputation: a comparison of
theeffectoftheSACHfootandsingleaxisfootonelectromyographicpatternsduringlocomotion.
ProsthetOrthot Int1986;10:1522.[84] Doane NE, Holt LE. Acomparison
of the SACHand single axis foot in the gait
ofunilateralbelow-kneeamputees.Prosthet OrthotInt1983;7:336.[85]
Gailey RS, Wenger MA, Raya M, Kirk N, Erbs K, Spyropoulos P, et al.
Energyexpenditure of trans-tibial amputees during ambulationat
self-selectedpace.Prosthet OrthotInt1994;18:8491.[86] Genin JJ,
Bastien GJ, Franck B, Detrembleur C, Willems PA. Effect of speed
onthe energy cost of walking in unilateral traumatic lower limb
amputees. Eur JApplPhysiol 2008;103:65563.[87] GohJC, Lee PV, Chong
SY. Comparative study between patellar-tendon-bearingand pressure
cast prosthetic sockets. J Rehabil Res Dev 2004;41:491501.[88]
HsuMJ, NielsenDH, YackHJ, Shurr DG. Physiological measurements
ofwalking and running in people with transtibial amputations with 3
differentprostheses.J OrthopSportsPhys Ther1999;29:52633.[89]
HurleyGRB, McKenneyR, RobinsonM, ZadravecM, PierrynowskiMR. Therole
of the contralateral limb in below-knee amputee gait. Prosthet
Orthot Int1990;14:3342.[90] IsakovE, SusakZ, Becker E.
Energyexpenditureandcardiacresponseinabove-kneeamputeeswhileusingprostheseswithopenandlockedkneemechanisms.ScandJRehabilMed
1985;17:10811.[91] IsakovE,BurgerH,KrajnikJ,GregoricM,Marincek
C.Kneemuscle activityduring ambulation of trans-tibial amputees. J
Rehabil Med 2001;33:1969.[92] Jaegers SM, Arendzen JH, de Jongh HJ.
An electromyographic study of the hipmuscles of transfemoral
amputees in walking. Clin Orthop Relat Res1996;328:11928.[93]
LacroixB, JaninF, FouquetB, GrandmontagneM, PasteneJ, Jehl JL,
etal.Energy consumption of walking in young traumatic leg amputated
patientsupperthekneeaccordingtodifferent typesof kneeprotheses.
AnnPhysRehabilMed1992;35:2731.[94] Lee S, Hong J. The effect of
prosthetic ankle mobility in the sagittal plane
onthegaitoftransfemoralamputeeswearingastancephasecontrolledkneeprosthesis.
In: Proceedings of the Institution of Mechanical Engineers Part H.
JEng Med2009;223:263271.[95] Lemaire ED, Fisher FR, Robertson DG.
Gait patterns of elderly men with trans-tibialamputations.
ProsthetOrthot Int1993;17:2737.[96] MichaudSB, GardSA, ChildressDS.
Apreliminaryinvestigationof
pelvicobliquitypatternsduringgaitinpersonswithtranstibialandtransfemoralamputation.
JRehabilResDev2000;37:110.[97] Murray MP, Sepic SB, Gardner GM,
Mollinger LA. Gait patterns of above-kneeamputees using
constant-friction knee components. Bull Prosthet Res
1980;10:3545.[98] Murray MP, Mollinger LA, Sepic SB, Gardner GM,
Linder MT. Gait patterns inabove-kneeamputeepatients:hydraulicswing
controlvsconstant-frictionkneecomponents. ArchPhys
MedRehabil1983;64:33945.[99] Nyska M, Shabat S, Arya A, McCabe C,
Linge K, Klenerman L. A comparativestudy of different below-knee
prostheses by dynamic foot pressure analysis.IntJ
RehabilRes2002;25:3414.[100] PagliaruloMA, WatersR, HislopHJ.
Energycostofwalkingofbelow-kneeamputeeshavingnovasculardisease.Phys
Ther1979;59:53843.[101] Pinzur MS, Perona P, Patwardhan A, Havey R.
Loading of the contralateral footin peripheral vascular insufciency
below-knee amputees. Foot Ankle1991;11:36871.[102] Pinzur MS,
GoldJ, SchwartzD, GrossN. Energydemandsfor walkingindysvascular
amputeesasrelatedtothelevel of amputation.
Orthopedics1992;15:10336.[103] RoyerTD, WasilewskiCA.
Hipandkneefrontalplanemomentsinpersonswithunilateral,trans-tibial
amputation.GaitPosture 2006;23:3036.[104] Segal AD, Orendurff MS,
Czerniecki JM, Shofer JB, Klute GK. Transtibialamputee joint
rotation moments during straight-line walking and a commonturning
task with and without a torsion adapter. J Rehabil Res
Dev2009;46:37583.[105] Snyder RD, Powers CM, Fontaine C, Perry J.
The effect of ve prosthetic feet onthe gait and loading of the
sound limb in dysvascular below-knee amputees.
JRehabilResDev1995;32:30915.[106] TorburnL, Powers CM, Guiterrez R,
PerryJ. Energyexpenditure duringambulation in dysvascular and
traumatic below-knee amputees: a compari-sonofveprosthetic feet.J
RehabilResDev1995;32:1119.[107] Traballesi M, PorcacchiaP, AvernaT,
Brunelli S. Energycost of walkingmeasurements in subjects with
lower limb amputations: a comparison
studybetweenoorandtreadmilltest.GaitPosture 2008;27:705.[108] Van
Jaarsveld HWL, Grootenboer HJ, De Vries J. Accelerations due to
impact atheel strike using below-knee prosthesis. Prosthet Orthot
Int 1990;14:636.[109] Zhang M, Lee WC. Quantifying the regional
load-bearing ability of trans-tibialstumps.ProsthetOrthot
Int2006;30:2534.Y.Sagawa Jr.etal. / Gait&Posture33(2011)511526
526