original article upper-limb joint power and its distribution in...

O

UIVRA

MdsA

ams

u

t

pi2cv.M.sv

owwppt

SVDsiPUNC

(

so

vp

456

A

RIGINAL ARTICLE

pper-Limb Joint Power and Its Distribution in Spinal Cordnjured Wheelchair Users: Steady-State Self-Selected Speedersus Maximal Acceleration Trials

obert Price, MSME, Zachary R. Ashwell, MSME, Michael W. Chang, MD, PhD, Michael L. Boninger, MD,
licia M. Koontz, PhD, RET, Sue Ann Sisto, PT, MA, PhD
i

cR

Swaltj

baiatgmpvs

lmccmgdstpcteteawipfi

desskt

ABSTRACT. Price R, Ashwell ZR, Chang MW, BoningerL, Koontz AM, Sisto SA. Upper-limb joint power and its

istribution in spinal cord injured wheelchair users: steady-tate self-selected speed versus maximal acceleration trials.rch Phys Med Rehabil 2007;88:456-63.

Objective: To compare upper-limb joint power magnitudend distribution between the shoulder, elbow, and wrist duringaximal acceleration (MAC) versus steady-state, self-selected

peed (SSS) manual wheelchair propulsion.Design: Cross-sectional biomechanic study.Setting: Research university and teaching hospital.Participants: Volunteer sample of 13 manual wheelchair

sers with spinal cord injury below T1.Interventions: Not applicable.Main Outcome Measures: Propulsive joint power magni-

ude and fractional distribution among upper-limb joints.Results: Wilcoxon signed-rank testing revealed shoulder

ower was larger for MAC versus SSS (median peak, 101.5W;nterquartile range [IQR], 74.6; median peak, 37.7W; IQR,2.9; respectively) (P�.01). Elbow and wrist power were un-hanged. Peak shoulder power fraction was larger for MACersus SSS (median peak, 1.055; IQR, .110 vs peak, .870; IQR,252) (P�.01). Peak elbow power fraction was smaller for

AC versus SSS (median peak, �.012; IQR, .144 vs peak,146; IQR, .206) (P�.05). Peak wrist power fraction wasmaller for MAC versus SSS (median peak, �.058; IQR, .057s peak, �.010; IQR, .150) (P�.05).Conclusions: Power at the shoulder was larger than at

ther joints. Peak shoulder joint power and power fractionas larger during MAC versus SSS propulsion. Elbow andrist power fractions were smaller for MAC versus SSSropulsion. Higher joint power, present under MAC, mayredispose manual wheelchair users to injury, particularly athe shoulder.

From the Department of Rehabilitation Medicine, University of Washington,eattle, WA (Price, Ashwell, Chang); Human Engineering Research Laboratories,eterans Affairs Pittsburgh HealthCare System, Pittsburgh, PA (Boninger, Koontz);epartments of Rehabilitation Science and Technology and Bioengineering, Univer-

ity of Pittsburgh, Pittsburgh, PA (Boninger, Koontz); Department of Physical Med-cine and Rehabilitation, University of Pittsburgh Medical Center Health System,ittsburgh, PA (Boninger); Department of Physical Medicine and Rehabilitation,niversity of Medicine and Dentistry of New Jersey, New Jersey Medical School,ewark, NJ (Sisto); and Kessler Medical Rehabilitation Research and Educationorp, West Orange, NJ (Sisto).Supported by the National Institute on Disability and Rehabilitation Research

grant no. H133A011107).No commercial party having a direct financial interest in the results of the research

upporting this article has or will confer a benefit upon the author(s) or upon anyrganization with which the author(s) is/are associated.Reprint requests to Robert Price, MSME, Dept of Rehabilitation Medicine, Uni-

ersity of Washington, Box 356490, 1959 NE Pacific St, Seattle, WA 98053, e-mail:[email protected].

q0003-9993/07/8804-11385$32.00/0doi:10.1016/j.apmr.2007.01.016

rch Phys Med Rehabil Vol 88, April 2007

Key Words: Biomechanics; Kinetics; Movement; Rehabil-tation; Wheelchairs.

© 2007 by the American Congress of Rehabilitation Medi-ine and the American Academy of Physical Medicine andehabilitation

HOULDER PAIN, CARPAL TUNNEL syndrome, and otherupper-limb pathologies are common problems in manual

heelchair users.1,2 They result from high mechanical stressccompanying the propulsion of manual wheelchairs. Over-oading injuries of this nature require a biomechanic analysis ofhe propulsion technique to understand the relationship to in-ury.3

In the analysis of locomotion, joint power offers insight into theiomechanics of the activity. Joint power has been used in gaitnalysis to study biomechanics,4,5 as an outcome to evaluatenterventions,6,7 and as a reference against which to measure theccuracy of clinical evaluations.8,9 Mechanical power describeshe time rate of (net) work performed by the muscles. Eitherenerative or absorptive, it is the result of concentric or eccentricuscle contractions, respectively. Applied to joint movement,

ower is the product of net joint moment and joint angularelocity and thus captures the effects of both net muscle effort andpeed of joint rotation.10

Guo et al11 investigated the general characteristics of upper-imb segmental mechanical energy and power flow in healthyale adults propelling a manual wheelchair under steady-state

onditions over a laboratory floor. Using a kinematic dataollection system and an instrumented pushrim, they deter-ined that at the beginning of the propulsion stroke, power is

enerated by the upper arm, or transferred from the trunkownward to the forearm and hand. At the end of the propul-ion stroke, joint power is transferred upward to the trunk fromhe forearm and upper arm. They found that work computed viaower flow analysis was greater than work computed by thehange in mechanical energy, implying that the power suppliedo the upper-limb segments is greater than is required. Withoutlaboration, they speculated that guidelines for the configura-ion of wheelchairs could be developed to improve propulsionfficiency and reduce the risk of upper-limb injury. Guo et al12

gain used power analysis to determine the mechanical cost ofheelchair propulsion and found pushrim diameter was an

mportant factor in wheelchair propulsion mechanics. As such,ower analysis may contribute to optimizing wheelchair con-guration and design.Rodgers et al13 examined joint power shifting under fatigue

uring wheelchair propulsion. Using a mix of nonusers andxperienced manual wheelchair users propelling against a re-isting ergometer, they showed that with fatigue, joint powershift from the shoulder to the elbow and wrist. For someinematic measures, experienced users compensated for fa-igue differently than nonusers, for example, in stroke fre-
uency, contact time, and pushrim and joint forces. They

cmippi

taalftbrjsi

lwtpImcssbc

ujlpjioatujoeam

pdrwlGrpftpcm

wsp

ptos

vhsoDsGsp

fbmbtihpcwsipv

armemecotbalpss

n2obediupfuns

ar

457JOINT POWER IN WHEELCHAIR USERS, Price

oncluded that power shifting may be a necessary strategy foraintaining constant power output under fatigue and may

ncrease the risk of upper-limb injury. Their study showed thatower analysis may contribute to understanding and optimizingropulsion strategies employed by wheelchair users and maynform us on the potential for injury.

In bringing a manual wheelchair up to speed, 2 stagesypically exist, an acceleration stage, as the speed is increasing,nd a steady-state stage, when the desired speed has beenchieved. The relative contribution of either stage to upper-imb injury and pain is unknown. Koontz et al14 showed thatorces and moments are higher when starting from a dead stophan when propelling at a steady-state speed, other factorseing equal. This would support the view that although ofelatively short duration, a rapid acceleration results in higheroint moments, velocities, and powers than does comfortableteady-state propulsion. High joint loading may result in acutenjury.

Conversely, during comfortable steady-state propulsion overevel smooth surfaces, the joint moments, velocities, and powersould tend to be smaller. Although acute injury may not occur,

he long term, repetitive nature of the comfortable steady-stateropulsion mode may invite chronic upper-limb injury and pain.15

n actuality, it may be the combined effects of both propulsiveodes that lead to upper-limb injury and pain. Should this be the

ase, it is important to understand the factors involved in bothtages of wheelchair propulsion. This should include an under-tanding of how joint power is redistributed under accelerationecause shifting power from large to small joints may in-rease the risk of injury.13

The literature reveals that joint power analysis enhances ournderstanding of wheelchair propulsion. In the same way thatoint reaction forces are likely responsible, in part, for upper-imb injuries,16 it is reasonable to expect that excessive jointower generation and absorption may be injurious. Excessiveoint power may represent an essential, but in and of itselfnsufficient, contributor to injury. Other factors such as anat-my, limb orientation, and tissue tolerance, may play a medi-ting role. Whether power analysis proves a clinically usefulechnique to understand the role propulsion mechanics plays inpper-limb pain and injury remains in question. Nevertheless,oint power analysis may benefit the mechanistic investigationf the origin of upper-limb pain and injury, inform us onrgonomic treatment methods for wheelchair users, and gener-lly lead to improved clinical insight. We hope the novelethods and findings described herein contribute to that goal.This study’s objective was to compare the magnitude of joint

ower in the upper limb and its distribution between the shoul-er, elbow, and wrist during maximal acceleration (MAC)elative to steady-state, self-selected speed (SSS) manualheelchair propulsion. Power distribution among the upper-

imb joints can be defined either in absolute or relative terms.iven the disparity between the 2 propulsion conditions, a

elative measure of power distribution was deemed most ap-ropriate. As used herein, power distribution refers to theraction of power exhibited at each joint, normalized to theotal upper-limb power, that is, the joint power fraction. A jointower fraction of .65 would equate to a concentric powerontribution of 65% to the total power, for example. Statedathematically:

Joint power fraction � Pj ⁄ PT (1)

here Pj is power at a specific upper-limb joint and PT is theum of power for all upper-limb joints (ie, total upper-limb
ower). s
The following hypotheses were tested: (1) upper-limb jointower is higher during MAC trials relative to steady-state, SSSrials; and (2) the distribution of joint powers, when expressedn a fractional basis is unchanged when comparing MAC andteady-state, SSS propulsion trials.

Hypothesis 1, included for the sake of completeness and facealidity, relies on the finding that forces and moments areigher during start-up.14 Also, from a mechanical energy per-pective, power input is required to increase the kinetic energyf the system from rest to achieve steady-state propulsion.uring steady-state propulsion, the power requirement is pre-

umably less because it serves only to offset frictional losses.iven the lack of information on power distribution, hypothe-

is 2 was stated in the form of a null hypothesis withoutresuming a difference.

METHODSThe following comprised the inclusion and exclusion criteria

or the study: complete or incomplete spinal cord injury (SCI)elow T1, injury occurred at age 18 years or older, using aanual wheelchair as a primary means of mobility (�40h/wk),

etween 18 and 65 years of age at the time of the study, greaterhan 1 year since injury, no history of fractures or dislocationsn upper limbs from which user has not fully recovered, noistory of arm or shoulder pain as a result of a syrinx, notregnant at the time of the study, and currently using a wheel-hair equipped with quick release axles and 61-cm (24-in)heels. The last criterion ensured compatibility with the in-

trumented pushrim (see below). All participants provided theirnformed consent for the study, which was reviewed and ap-roved by the University of Washington Human Subjects Re-iew Committee.We collected 3-dimensional upper-limb kinematic data usingQualisysa passive-marker motion capture system. Spherical

etro-reflective markers were placed on the following land-arks and captured at 120Hz: temporomandibular joint (bilat-

ral), C7 and T2 vertebrae, sternal notch, along with unilateralarkers on the nondominant side: acromion, elbow lateral

picondyle, olecranon, radial, and ulnar styloids and third meta-arpal joint. In addition, the hub and a spoke of the SmartWheelb

n the nondominant side were marked. The protocol followed inhis study is part of a larger study investigating the relationshipetween wheelchair propulsion biomechanics and upper-limb painnd injury. We studied the nondominant side because it is lessikely to be injured from repetitive use other than wheelchairropulsion. Wheelchair propulsion biomechanics are generallyimilar between sides; Boninger et al17,18 found that left and rightide kinetic and kinematic data are correlated.

The triaxial pushrim forces and moments applied on theondominant side were captured using a SmartWheel at40Hz (fig 1). This is the refined and commercialized versionf the instrumented wheelchair pushrim developed and testedy Asato19 and VanSickle20 and colleagues. The pushrim wasither anodized metal or vinyl-coated, according to the style inaily use by the participant. Only 1 SmartWheel was available;t was installed on the participant’s own wheelchair, on theser’s nondominant side. A standard wheel with a matchingushrim style was installed on the dominant side. Both wheelseatured solid tire inserts irrespective of the tire style in dailyse by the participant. Kinematic data capture was synchro-ized using the SmartWheel data capture system as a triggerource.

Participants propelled with their wheelchair secured, via andjustable strap and rail, to the platform of a passive, dual-oller system (see fig 1); each roller acted independently and
upported 1 rear wheelchair wheel. A visual feedback system
Arch Phys Med Rehabil Vol 88, April 2007

plaan

tpestdhessgs

p2Stjbwia

teiuatie

munult

sbsbwcsntmbCiefflotca

ptubittijpdpuutw

dosipaSTto

Fp

458 JOINT POWER IN WHEELCHAIR USERS, Price

A

rovided speed and directional information (to ensure straight-ine propulsion), to the participant and investigators. This wasccomplished by using a tachometer-generator for each rollernd a laptop computer with analog-to-digital capability, run-ing custom LabViewc software.

We asked participants to propel under 2 conditions: (1) atheir SSS and (2) under an MAC propulsion condition. Thisrotocol has the effect of varying the level of effort between 2xtremes with no attempt to prescribe a fixed speed betweenubjects. We reasoned that holding speed constant at 2 ex-remes may not be truly reflective of the level of effort forifferent people; for some people the highest speed may notave been achievable because size, strength, experience level,tc, vary widely among users. In other words, a high fixedpeed requiring great effort for one person may not represent aignificant effort for another person and any conclusions re-arding the distribution of joint power under these circum-tances may be clouded.

Participants were permitted to acclimate to the roller systemrior to data collection. Data were captured for a minimum of0 and a maximum of 40 seconds for each condition. For theSS trial, data capture was performed such that at a minimum,

he last 20 seconds of the trial were at a steady-state pace in theudgment of an investigator monitoring the visual speed feed-ack system. For the maximal acceleration trial, the participantas instructed, via the computer’s display, to accelerate max-

mally (from a stop) for 5 seconds, brake and rest for 5 seconds,

ig 1. Reflective markers applied to volunteer, SmartWheel, andassive dual-roller system.

ccelerate maximally again for 5 seconds, and then coast for t


he remainder of the trial. The MAC trial was an attempt tolicit maximum propulsive power output over a short timenterval. The wheel speed during the MAC trial varied contin-ously over the trial and featured both the start-up speed fromdead stop and the participant’s maximum limiting speed over

he acceleration period. Although the speed was variable, thentent of the protocol was that the participant’s propulsionffort was maximal.

After collection, we adapted the 240-Hz kinetic data toatch the 120-Hz sampling frequency of the kinematic data by

sing every other frame. The synchronized 3-dimensional ki-ematic and kinetic data were forward and reverse filteredsing a second-order digital Butterworth filter with a 5-Hzow-pass cutoff for each stage. Filter parameters are similar tohose described by Winter10 and Cooper et al.21

A link-segment model of the upper-limb on the nondominantide was generated using anthropometric parameters publishedy Winter10 (based on subjects without physical disabilities). Aeparate anthropometric model was created for each participantased on their weight and measured anatomic landmarks alongith the published parameters to compute the limbs’ inertial

haracteristics. Joint moments were calculated based on theegmental accelerations from kinematic measurements, exter-al loads measured using the SmartWheel and inertial charac-eristics and segmental gravitational forces based on anthropo-etric data. The point of force application was approximated

y the location of the third metacarpal joint marker. Althoughooper et al22 observed that the point of force application was

n general not collocated with any anatomic landmark, Sabickt al23 concluded that kinematic methods of estimating point oforce application were the most stable. Joint centers of rotationor the shoulder, elbow, and wrist were approximated by theocation of the acromion, lateral epicondyle, and the mid-pointf the radial- and ulnar-styloid markers, respectively. Segmen-al angular velocities were calculated and joint powers wereomputed as the scalar product of the joint moment and jointngular velocity vectors.

Because the contribution of each joint to the total upper-limbower output likely varied over the propulsion cycle, we usedhe time of peak exertion, given by the time of maximum totalpper-limb power output. The algebraic sum of shoulder, el-ow, and wrist powers gives the total net upper-limb power; itncludes any negative power that may be present. In addition,he timing of the half-power points on either side of the peakotal upper-limb power (ie, the points at which the total powers 50% of the peak) would reveal if the contribution of eachoint changes prior, and subsequent to, the peak. Also, these 3oints (pre-peak, peak, post-peak) provided a convenientlyefined boundary encompassing the time of greatest total pro-ulsive effort. Note that the peak powers of each joint individ-ally may occur at different times from the peak of the totalpper-limb power. Investigating the joint power fraction underhese circumstances leads to difficulties in interpretation andas therefore not performed.To focus the analysis on the start-up acceleration strokes

uring the MAC trials, we evaluated the first 4 strokes of eachf the 2 acceleration sets. Koontz et al14 showed that thetart-up moments after the first 4 strokes are statistically sim-lar. Because they are likely dissimilar, the joint power andower fraction for the first 4 strokes, for each of 2 successivettempts, were averaged to characterize the MAC trials. For theSS trials, the last 8 full propulsion strokes were evaluated.hus, for each condition, 8 strokes were averaged to estimate

he power outputs for the shoulder, elbow, and wrist at the timef peak total and half-power points. Also, the power outputs for
he shoulder, elbow, and wrist at the time of peak and half-

p(j

aSmPs.1aem

9Irwhps

(ficTacitpa

sa2oarnMtecj3�a

a

4FuSStsp(siS


ower points were used to determine the joint power fractionsee equation 1). This served as a relative measure of eachoint’s contribution to the total power output.

To evaluate the statistical significance of any data trends, wepplied the Wilcoxon signed-rank test for paired data24 usingPSS.d The modest sample size dictated the use of a nonpara-etric test, because power data normality could not be ensured.airing the data effectively compared changes within eachubject. Significance levels of P less than .01 and P less than05 were applied. A 1-tailed test was applied to test hypothesis

and a 2-tailed test was applied to test hypothesis 2. Inddition, to test the significance of power differences withinach condition and between the upper-limb joints, the Fried-an test25 was used because the data across joints are related.

RESULTSThe volunteer participant sample consisted of 4 women and

men, ranging from 32 to 61 years of age (mean, 46.5�10.3y).njury levels ranged from T3 to L1, with time since injuryanging from 2 to 33 years (mean, 15�9.4y). Participanteight ranged from 49 to 88kg (mean, 65.4�13.1kg) andeight ranged from 1.5 to 1.93m (mean, 1.73�0.12m). All 13articipants were right-hand dominant and were therefore in-trumented on the left side.

For reference purposes, sample power versus time curvesfor a person) for each joint during an SSS trial are given ingures 2A through 2C. The corresponding power versus timeurves for an MAC trial are shown in figures 3A through 3C.hese graphs display several cycles of propulsion and includestep function that indicates when a propulsive moment ex-

eeding .25Nm was applied to the SmartWheel, effectivelyndicating when contact on the pushrim occurred. In addition,he total upper-limb power and the individual joint power arelotted. Positive power values denote concentric contractionnd negative values denote eccentric contraction.

Wheelchair speed during the SSS condition averaged �tandard deviation 1.29�0.41m/s. The peak wheelchair speedttained on the fourth stroke of the MAC condition averaged.07�0.43m/s. Descriptive statistics for the joint-specific powerutput at the time of peak total upper-limb power under SSSnd MAC conditions are given in table 1. Wilcoxon signed-ank statistical testing revealed that, at a P less than .01 sig-ificance level, median shoulder power was greater underAC conditions (102W) than under SSS conditions (38W). At

he time of peak total upper-extremity power, power at thelbow and wrist did not differ significantly for the propulsiononditions. At the time of peak total upper-limb power, medianoint power was greater (P�.001) for the shoulder (SSS,7.7W; MAC, 101.5W) than for the elbow (SSS, 3.2W; MAC,1.8W) and wrist (SSS, �0.3W; MAC, �5.0W), for both SSS

nd MAC conditions, according to the Friedman test.The descriptive statistics for joint power fractions under SSS

nd MAC conditions are displayed in table 2. Wilcoxon

™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™ig 2. (A) Power versus time for the shoulder joint (solid) and totalpper-limb power (dashed) at steady-state, self-selected speed.tep function indicates when a propulsion moment is applied to themartWheel. (B) Power versus time for the elbow joint (solid) andotal upper-limb power (dashed) at steady-state, self-selectedpeed. Step function indicates when a propulsion moment is ap-lied to the SmartWheel. (C) Power versus time for the wrist jointsolid) and total upper-limb power (dashed) at steady-state, self-elected speed. Step function indicates when a propulsion moment
s applied to the SmartWheel. Abbreviations: Pwr, power; SW,martWheel; UE, upper extremity.

sd(aMMp((pnPtt

tefii

pscjoptefiero

ptclo

4FucPpptum

Na*



igned-rank statistical testing revealed that median peak shoul-er power fraction was higher (P�.01) for MAC versus SSS1.055 vs .870). Median pre-peak shoulder power fraction waslso higher (P�.01) for MAC versus SSS (.722 vs .489).edian peak elbow power fraction was lower (P�.05) forAC versus SSS (�.012 vs .146). Median pre-peak elbow

ower fraction was also lower (P�.05) for MAC versus SSS.369 vs .495). Median peak wrist power fraction was lowerP�.05) for MAC versus SSS (�.058 vs �.010). Medianre-peak wrist, and post-peak comparisons for all 3 joints, didot differ significantly between SSS and MAC conditions.eak power fraction was greater (P�.001) for the shoulder

han for the elbow and wrist, for both SSS and MAC condi-ions, according to the Friedman test.

Graphical comparisons of the joint power fractions at theime of peak net upper-limb power output for the shoulder,lbow, and wrist for SSS and MAC conditions are shown ingure 4. Likewise, pre-peak graphical comparisons are shown

n figure 5.

DISCUSSIONOnly at the shoulder did the data support hypothesis 1 that

eak joint power is higher during acceleration than duringteady-state propulsion; the elbow and wrist joints showed noonsistent change (see table 1). Examination of the sampleoint power data (see figs 2A–C, 3A–C) reveals an asymmetryf the total upper-limb joint power, with greater concentricower (positive) relative to eccentric power (negative). Fur-hermore, the concentric power during the propulsion phasexceeds the concentric power during the recovery phase. Bothndings are consistent with the participant working against thexternal load imposed by the passive dual-roller system due toolling friction, similar to what might be expected duringverground propulsion.During the propulsion phase, the shoulder dominates the

ower output behavior between the half-power points, relativeo the other joints, for both propulsion conditions. This wasonfirmed statistically by the Friedman test to compare powerevels between joints at the time of peak total upper-limb powerutput. The finding that the shoulder power was greater during

™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™ig 3. (A) Power versus time for the shoulder joint (solid) and totalpper-limb power (dashed) under acceleration. Step function indi-ates when a propulsion moment is applied to the SmartWheel. (B)ower versus time for the elbow joint (solid) and total upper-limbower (dashed) under acceleration. Step function indicates when aropulsion moment is applied to the SmartWheel. (C) Power versusime for the wrist joint (solid) and total upper-limb power (dashed)

Table 1: Joint Powers at Time of Peak Upper-Extremity PowerDuring SSS and MAC Conditions

Shoulder Elbow Wrist

Value SSS MAC SSS MAC SSS MAC

Median 37.7* 101.5* 3.2 �1.8 �0.3 �5.0IQR 22.9 74.6 6.3 8.7 8.0 4.6Min 11.0 26.9 �3.7 �13.6 �8.1 �33.2Max 104.7 178.8 10.9 23.8 6.7 23.9

OTE. Values are median, interquartile range (IQR), minimum (Min),nd maximum (Max). Values in watts. N�13.P�.01 (Wilcoxon signed-rank test).

nder acceleration. Step function indicates when a propulsion mo-ent is applied to the SmartWheel.

tgpmp

wcc

p

N*

Fl


he MAC condition reflects the dominant role of the shoulder inenerating concentric power under circumstances of maximalropulsive effort. By contrast, the elbow and wrist joints areinor contributors to concentric power output under both pro-

ulsive conditions and they show no significant difference

Table 2: Joint Power Fractions at Time of Peak, Pre-Peak, and Po

Peak Shoulder PowerFraction

Value SSS MAC S

Median 0.870* 1.055* 0.IQR 0.252 0.110 0.Min 0.68 0.64 �0.Max 1.13 1.15 0.

Pre-Peak Shoulder PowerFraction

Median 0.489* 0.722* 0.IQR 0.442 0.351 0.Min 0.05 0.42 0.Max 0.93 1.06 0.

Post-Peak Shoulder PowerFraction

Median 1.158 1.013 �0IQR 0.713 0.562 0Min 0.26 0.24 �0Max 1.63 1.88 0

OTE. Values are median, IQR, minimum, and maximum. N�13.P�.01; †P�.05 (Wilcoxon signed-rank test).

ig 4. Joint power fractions (median and quartiles) at time of upper-imb peak power for SSS and MAC trials.

Fl

hen comparing SSS to MAC propulsion. The elbow powerurves indicate a tendency for the elbow to act in an eccentricapacity during the later half of the propulsion stroke.

During the recovery phase, the shoulder again dominatesower generation and absorption. This is likely the result of the

ak Upper-Extremity Power During the SSS and MAC Conditions

ak Elbow PowerFraction

Peak Wrist PowerFraction

MAC SSS MAC

�0.012† �0.010† �0.058†

0.144 0.150 0.057�0.10 �0.17 �0.22

0.18 0.26 0.21

eak Elbow PowerFraction

Pre-Peak Wrist PowerFraction

0.369† �0.044 �0.0600.446 0.099 0.066

�0.11 �0.14 �0.170.74 0.16 0.26

Peak Elbow PowerFraction

Post-Peak Wrist PowerFraction

0.053 �0.062 �0.0450.410 0.145 0.061

�0.72 �0.22 �0.190.78 0.22 0.24

st-Pe

Pe

SS

146†

2060429

Pre-P

495†

5010393

Post-

.078

.636

.51

.77

ig 5. Joint power fractions (median and quartiles) at time of upper-imb pre-peak power for SSS and MAC trials.


ldwtmbm

2o2(psgoiftdccpbmdtw

cwpRtbg1p1Tpoae

emjaatrpcntstp

lsppTe

pc

itjiscopuenIwTpi

aeacfslsdMmdeerstiwcfda

tppidbiwa

ciapieap(


A

arger inertial load applied to the shoulder in accelerating andecelerating the upper limb in comparison to the elbow andrist, as well as a larger angular velocity likely to be present at

he shoulder. In the MAC trial example, the elbow power isore pronounced during recovery than in the SSS trial, possi-

ly a reflection of a more vigorous retraction of the arm in theaximal acceleration trial.Comparisons of the SSS joint power data (typified by figs

A–C) with power flow data of Guo et al11 reveals similarutcomes, despite differences in the methodology between thestudies. The Guo study computed segmental power flows

upper arm, forearm, hand), whereas the current study com-uted joint powers (shoulder, elbow, wrist). Also, the Guotudy relied on nonwheelchair users propelling over levelround, rather than on experienced users with SCI, propellingn a passive dual-roller system. Nevertheless, it may be reveal-ng to compare the peak upper-limb power during propulsionor each method. The mean total upper-limb power flow, ob-ained by Guo, revealed an initial positive peak of about 22Wuring early propulsion followed by a nearly continuous de-rease in power flow until the onset of the recovery phase. Theurrent study found a somewhat larger mean peak upper-limbower (40.6W) during the SSS propulsion phase, as determinedy the sum of individual mean joint powers in table 1. Thisay be reflective of differences due to propelling on a passive,

ual-roller system by experienced wheelchair users in contrasto nonusers rolling over ground. The type of tire used by Guoas not specified and differences in tires may also play a role.Shoulder joint power fraction was greater for the MAC

ondition relative to the SSS condition, whereas elbow andrist power fractions were less, contrary to hypothesis 2, whichosited no change in power fraction (see table 2, figs 4, 5).odgers et al13 observed a shift in power from the shoulder to

he elbow and wrist under conditions of fatigue. In their mix ofoth wheelchair users and nonusers working against an er-ometer, they found that the shoulder provided 61%, the elbow5%, and the wrist 24% of the total power during nonfatiguingropulsion. Under fatigue, the joint power fraction was 50%,9%, and 31% for the shoulder, elbow, and wrist, respectively.he wrist power fraction observed in the present study ap-eared to be considerably lower (see table 2, figs 4, 5) than thatbserved by Rodgers. It is possible that differences in the loadpplied by their ergometer relative to our roller system mayxplain the wrist power difference.

The power fraction at the shoulder was greater than at thelbow and wrist for both SSS and MAC conditions, as deter-ined by the Friedman test comparing power fractions across

oints. Comparison of joint power fractions between the SSSnd MAC conditions revealed that the shoulder joint generatedsignificantly larger fraction of the power during the MAC

rials. This behavior parallels that of the shoulder power andeinforces the finding that the shoulder dominates during theropulsion phase, especially during acceleration. This was inontradistinction to the elbow and wrist, which exhibited aumerically lower peak power fraction under the MAC condi-ion. Contrary to the shoulder and elbow, the wrist tended tohow power absorption under both conditions, more so underhe MAC condition, as reflected by a negative value of theower fraction.Pre-peak power fractions for the shoulder and elbow fol-

owed the same pattern as the peak power fractions, but weretatistically weaker. No significant difference in pre-peakower fraction was observed for the wrist. The post-peakower fraction did not show a systematic change for any joint.his supports the rationale that the pre- and post-peak regions
ffectively bracket the behavior of interest, namely, the time of f

eak propulsive effort when systematic differences betweenonditions and joints may be at their greatest.

The shoulder power fraction data (see figs 4, 5) revealed thatn some cases the power fraction exceeded 1.0. This impliedhat a power absorbing, eccentric contraction occurred at otheroints. This was confirmed because the elbow and wrist exhib-ted a concurrent negative power fraction in those cases. Atudy by van der Linden et al26 showed that negative powerould be generated under certain conditions of force and jointrientations in conjunction with the kinematic constraints im-osed by the trajectory of the hand. At the time of peak totalpper-limb power output, the shoulder joint power fractionxceeded 1.0, and the elbow and wrist power fractions wereegative, in more cases under MAC than under SSS conditions.n effect it appears that the shoulder concentric contraction wasorking against the wrist and elbow eccentric contractions.hus, it can be argued that for many of the participants,ropulsion was less effective under maximal acceleration thant could have been.

There appeared to be a shift in power toward the shoulder andway from the elbow during the MAC trials. This power shift mayxpose the shoulder to greater risk of injury and/or pain duringcceleration. Three factors present under MAC conditions mayontribute to greater injury risk, namely, higher power and powerraction at the shoulder and the higher likelihood of exhibitingimultaneous eccentric and concentric contractions over the upperimb. The higher level, and fraction, of power being shifted to thehoulder during acceleration implies that the muscles at the shoul-er have the excess power generating capacity to perform theAC task that the muscles of the more distal joints may lack. Itay also imply that the shoulder is at greater risk for acute injury

uring maximal acceleration than the other joints. Exhibitingxcess concentric contraction at the shoulder while simultaneouslyxhibiting eccentric contractions at the distal joints may extend theisk of injury and and/or pain in those persons using this propul-ion strategy. Correlations of a user’s injury and pain status withheir joint power fraction pattern under acceleration may facilitatenvestigation of shoulder pain and may provide direction forheelchair user training to minimize pain and injury. In relation to

hronic injury, it may be revealing to examine the joint powerraction patterns during SSS conditions, for persons who have noteveloped chronic pain and injury, to serve as a training model ors an indication of when an intervention may be warranted.

The study methodology could not account for co-contrac-ions of the upper-limb joints and thus permitted only netower calculations. The alternatives invoking electromyogra-hy-based muscle modeling to estimate co-contraction forcesntroduce other assumptions and limitations, however. In ad-ition, the anthropometric parameters used in modeling wereased on published values for persons without physical disabil-ty; however, the statistical testing relied on pairing of dataithin a subject (ie, differential changes); this would presum-

bly minimize the associated error.To leverage these findings and methods to improve the

linical care of manual wheelchair users, future research shouldnvestigate: (1) the association between upper-limb joint powernd the development of upper-extremity pain or injury (eg, Areersons with greater shoulder power fraction at a higher risk fornjury?) and (2) the sensitivity of this metric to detect differ-nces in wheelchair modifications (eg, pushrim type, seat/backdjustments, postural supports), training interventions (eg, Doower fractions change with training?), and user characteristicseg, How does shoulder strength influence the joint’s power
raction?).

pdsApbgsswsfmtaBpmFapfi

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

ab

c


CONCLUSIONSThis study’s objective was to compare the magnitude of joint

ower in the upper limb and its distribution between the shoul-er, elbow, and wrist during maximal acceleration relative toteady-state, self-selected speed manual wheelchair propulsion.t the time of peak total upper-limb power output, the shoulderlayed the dominant role, relative to the elbow and wrist, foroth propulsion conditions. Also, shoulder joint power wasreater during maximum acceleration, relative to self-selectedteady-state propulsion, whereas the wrist and elbow jointshowed no consistent change. Shoulder joint power fractionas greater during maximum acceleration relative to self-

elected steady-state propulsion, but elbow and wrist powerractions tended to be smaller. Some participants showed si-ultaneous concentric contractions at the shoulder and eccen-

ric contractions at the elbow and wrist during the maximalcceleration trials, indicating reduced propulsion effectiveness.ecause they are higher, joint powers and power fractionsresent under maximal acceleration conditions may predisposeanual wheelchair users to injury, especially at the shoulder.urther study is required to better understand how joint powersnd power fractions relate to the development of upper-limbain or injury in manual wheelchair users and how thesendings may be applied clinically.

References1. Boninger ML, Cooper RA, Baldwin MA, Shimada SD, Koontz

AL. Wheelchair pushrim kinetics: body weight and median nerveconduction. Arch Phys Med Rehabil 1999;80:910-5.

2. Boninger ML, Cooper RA, Robertson RN, Rudy TE. Wrist bio-mechanics during two speeds of wheelchair propulsion: an anal-ysis using local coordinate system. Arch Phys Med Rehabil 1997;78:364-72.

3. van der Woude LH, Veeger HE, Dallmeijer AJ, Janssen TW,Rozendaal LA. Biomechanics and physiology in active manualwheelchair propulsion. Med Eng Phys 2001;10:713-33.

4. Quanbury AO, Winter DA, Reimer GD. Instantaneous power andpower flow in body segments during walking. J Hum Mov Stud1975;1:59-67.

5. Williams KR, Cavanagh PR. A model for the calculation ofmechanical power during distance running. J Biomech 1983;16:115-28.

6. Brostrom E, Hagelberg S, Haglund-Akerlind Y. Effect of jointinjections in children with juvenile idiopathic arthritis: evaluationby 3D-gait analysis. Acta Paediatr 2004;93:906-10.

7. Cooney KM, Sanders JO, Concha MC, Buczek FL. Novel biome-chanics demonstrate gait dysfunction due to hamstring tightness.Clin Biomech (Bristol, Avon) 2006;21:59-66.

8. McGinley JL, Goldie PA, Greenwood KM, Olney SJ. Accuracyand reliability of observational gait analysis data: judgments ofpush-off in gait after stroke. Phys Ther 2003;83:146-60.

9. McGinley JL, Morris ME, Greenwood KM, Goldie PA, Olney SJ.Accuracy of clinical observations of push-off during gait afterstroke. Arch Phys Med Rehabil 2006;87:779-85.

0. Winter DA. Biomechanics and motor control of human move-ment. 3rd ed. New York: John Wiley & Sons; 2005.

1. Guo LY, Su FC, Wu HW, An KN. Mechanical energy and powerflow of the upper extremity in manual wheelchair propulsion. Clin
Biomech (Bristol, Avon) 2003;18:106-14. d
2. Guo LY, Su FC, Wu HW, An KN. Effect of handrim diameter onmanual wheelchair propulsion: mechanical energy and power flowanalysis. Clin Biomech (Bristol, Avon) 2006;21:107-15.

3. Rodgers MM, McQuade KJ, Rasch EK, Keyser RE, Finley MA.Upper-limb fatigue-related joint power shifts in experiencedwheelchair users and nonwheelchair users. J Rehabil Res Dev2003;40:27-38.

4. Koontz AM, Cooper RA, Boninger ML, Yang Y, Impink BG, vander Woude LH. A kinetic analysis of manual wheelchair propul-sion during start-up on select indoor and outdoor surfaces. JRehabil Res Dev 2005;42:447-58.

5. Hagberg M, Morganstern H, Kelsh M. Impact of occupations andjob tasks on the prevalence of carpal tunnel syndrome. Scand JWork Environ Health 1992;18:337-45.

6. Boninger ML, Cooper RA, Robertson RN, Shimada SD. Three-dimensional pushrim forces during two speeds of wheelchairpropulsion. Am J Phys Med Rehabil 1997;76:420-6.

7. Boninger ML, Baldwin MA, Cooper RA, Koontz AM, Chan L.Manual wheelchair pushrim biomechanics and axle position. ArchPhys Med Rehabil 2000;81:608-13.

8. Boninger ML, Cooper RA, Baldwin MA, Shimada SD, Koontz A.Wheelchair pushrim kinetics: body weight and median nervefunction. Arch Phys Med Rehabil 1999;80:910-5.

9. Asato KT, Cooper RA, Robertson RN, Ster JF. SMART-Wheels:development and testing of a system for measuring manual wheel-chair propulsion dynamics. IEEE Trans Biomed Eng 1993;40:1320-4.

0. VanSickle DP, Cooper RA, Robertson RN. SmartWheels: devel-opment of a digital force and moment sensing pushrim. In: Pro-ceedings of the 18th International RESNA conference; 1995 June9–14; Vancouver (BC). Washington (DC): RESNA Pr; 1995. p352-4.

1. Cooper RA, DiGiovine CP, Boninger ML, Shimada SD, KoontzA, Baldwin MA. Filter frequency selection for manual wheelchairbiomechanics. J Rehabil Res Dev 2002;39:323-6.

2. Cooper RA, VanSickle DP, Robertson RN, Boninger ML, Ens-inger GJ. A method for analyzing center of pressure duringmanual wheelchair propulsion. IEEE Trans Rehabil Eng 1995;3:289-98.

3. Sabick MB, Zhao KD, An KN. A comparison of methods tocompute the point of force application in handrim wheelchairpropulsion: a technical note. J Rehabil Res Dev 2001;38:57-68.

4. Walpole RE, Myers RH. Probability and statistics for engineersand scientists. 2nd ed. New York: McMillan; 1978.

5. Hollander M, Wolfe DA. Nonparametric statistical methods. NewYork: John Wiley & Sons; 1973.

6. van der Linden ML, Valent L, Veeger HE, van der Woude LH.The effect of wheelchair handrim tube diameter on propulsionefficiency and force application. IEEE Trans Rehabil Eng 1996;4:123-32.

Suppliers. Qualisys AB, Packhusgatan 6, S-411 13 Gothenburg, Sweden.. Three Rivers Holdings LLC, 1826 W Broadway Rd, Ste 43, Mesa,

AZ 85202.. National Instruments Corp, 11500 N Mopac Expwy, Austin, TX

78759-3504.
. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606-6307.

original article upper-limb joint power and its distribution in...

Documents