surplus value of hip adduction in leg press exercise in patients with patellofemoral pain syndrome

Surplus Value of Hip Adduction inLeg-Press Exercise in Patients WithPatellofemoral Pain Syndrome:A Randomized Controlled TrialChen-Yi Song, Yeong-Fwu Lin, Tung-Ching Wei, Da-Hon Lin, Tzu-Yu Yen,Mei-Hwa Jan

Background. A common treatment for patients with patellofemoral pain syn-drome (PFPS) is strength (force-generating capacity) training of the vastus medialisoblique muscle (VMO). Hip adduction in conjunction with knee extension is com-monly used in clinical practice; however, evidence supporting the efficacy of thisexercise is lacking.

Objective. The objective of this study was to determine the surplus effect of hipadduction on the VMO.

Design. This study was a randomized controlled trial.

Setting. The study was conducted in a kinesiology laboratory.

Participants. Eighty-nine patients with PFPS participated.

Intervention. Participants were randomly assigned to 1 of 3 groups: hip adduc-tion combined with leg-press exercise (LPHA group), leg-press exercise only (LPgroup), or no exercise (control group). Training consisted of 3 weekly sessions for 8weeks.

Measurements. Ratings of worst pain as measured with a 100-mm visual analogscale (VAS-W), Lysholm scale scores, and measurements of VMO morphology (in-cluding cross-sectional area [CSA] and volume) were obtained before and after theintervention.

Results. Significant improvements in VAS-W ratings, Lysholm scale scores, andVMO CSA and volume were observed after the intervention in both exercise groups,but not in the control group. Significantly greater improvements for VAS-W ratings,Lysholm scale scores, and VMO volume were apparent in the LP group comparedwith the control group. There were no differences between the LP and LPHA groupsfor any measures.

Limitations. Only the VMO was examined by ultrasonography. The resistancelevel for hip adduction and the length of intervention period may have been inade-quate to induce a training effect.

Conclusions. Similar changes in pain reduction, functional improvement, andVMO hypertrophy were observed in both exercise groups. Incorporating hip adduc-tion with leg-press exercise had no impact on outcome in patients with PFPS.

CY Song, PT, MS, is a PhD student,School and Graduate Institute ofPhysical Therapy, College of Med-icine, National Taiwan University,Taipei, Taiwan.

YF Lin, MD, PhD, is OrthopedicSurgeon, Department of Orthope-dics, West Garden Hospital, Tai-pei, Taiwan.

TC Wei, PT, MS, is Physical Thera-pist, Yeong-An Clinic, Taipei,Taiwan.

DH Lin, MD, is Orthopedic Sur-geon, Department of Orthope-dics, En Chu Kong Hospital, Tai-pei, Taiwan.

TY Yen, PT, MS, is Physical Thera-pist, Yeong-An Clinic.

MH Jan, PT, MS, is Associate Pro-fessor, School and Graduate Insti-tute of Physical Therapy, Collegeof Medicine, National Taiwan Uni-versity, 3F, No. 17, Xuzhou Rd,Zhongzheng District, Taipei 100,Taiwan, Republic of China. Ad-dress all correspondence to Ms Janat: [email protected].

[Song CY, Lin YF, Wei TC, et al.Surplus value of hip adduction inleg-press exercise in patients withpatellofemoral pain syndrome: arandomized controlled trial. PhysTher. 2009;89:409–418.]

© 2009 American Physical TherapyAssociation

Research Report

Post a Rapid Response orfind The Bottom Line:www.ptjournal.org

May 2009 Volume 89 Number 5 Physical Therapy f 409

Patellofemoral pain syndrome(PFPS) is a common musculo-skeletal problem of the knee.

Patients with PFPS often have ret-ropatellar or peripatellar pain.1 Thispain is aggravated by activities thatincrease patellofemoral compressiveforce, such as walking up and downstairs, squatting, running, and pro-longed sitting with bent knees.2–4

Patellofemoral pain syndrome isthought to be associated with lateralmalalignment of the patella.5 Pos-sible causes of this malalignmentinclude hypotrophy or atrophy ofthe vastus medialis oblique muscle(VMO), changes that are commonlyseen in patients with PFPS, and animbalance or delay in the activationof the VMO relative to the vastuslateralis muscle (VL).5,6 Several ca-daver studies have shown that theVMO fiber angle was 50 to 57 de-grees off the long axis of the femurin the frontal plane, and the proxi-mal vastus medialis muscle fiberorientation was approximately 15degrees.7–10 Anatomically, the direc-tion of VMO muscle pull was morehorizontal. It also has been demon-strated in a cadaver study that all ofthe quadriceps femoris muscles, ex-cept the VMO, can extend the knee(regardless of the force applied) andthat the role of the VMO is to main-tain medial tracking of the patelladuring knee extension.11 An in vivostudy12 showed that electrical stimu-lation of the VMO resulted in pre-

dominantly medial patellar pull. Be-cause the VMO plays an importantrole in medial stabilization of the pa-tella,11–13 any dysfunction may leadto reduced medial stabilization ofthe patella against the counterforceof lateral pull exerted by the VL andother quadriceps femoris musclecomponents.14,15

Numerous rehabilitation protocolshave been described for treating peo-ple with patellofemoral problems.Quadriceps femoris muscle strength-ening (increased force-generating ca-pacity), especially that emphasizingthe VMO, is generally considered tobe a conservative treatment.14,16–19

Incorporating hip adduction withknee extension is a popular strategyfor strengthening the VMO.14,16 Thisintervention takes into considerationthe fact that the VMO is connectedto the adductor magnus and longusmuscles.20 Training of the adductormuscles uses this anatomical link toprovide a more-stable proximal at-tachment and transfers physiologicalstretch to the VMO, thereby enhanc-ing the contraction force.15,21

Because hip adduction exercise wasfound to selectively activate theVMO,22 numerous studies have ex-amined the electrical activity ofthe VMO and VL with hip adduc-tion during various knee extensionexercises.15,21,23–29 Findings havebeen disparate. Selective recruit-ment of the VMO was reported in 2studies where VMO activity washigher than VL activity following hipadduction with knee extension froma semi-squatting position in subjectswho were healthy.21,28 In patientswith PFPS, however, incorporationof hip adduction has been found topromote a more-balanced VMO/VLratio29 or increased whole quadri-ceps femoris muscle activity.24,25

Hodges and Richardson21 and Grel-samer and McConnell30 have advo-cated incorporating hip adduction

with prone-lying or squatting posi-tions. These positions, however, arenonfunctional or weight bearing,where the tendency of increasingthe dynamic Q-angle may exacerbatestress on the patellofemoral joint.31

Furthermore, some patients may nottolerate this position because thegradually increasing joint stress mayaggravate the pain. Leg-press (LP)exercise, therefore, be used as asubstitute to train patients in a func-tional position without aggravatingsymptoms.

To date, it remains unclear as towhether the addition of hip adduc-tion to LP exercise would facilitateVMO hypertrophy and result in a bet-ter treatment outcome. In additionto the pain and functional scales thattypically are used, ultrasonographyprovides a further quantitative mea-sure for assessing muscle morphol-ogy in clinical trials involving pa-tients with PFPS. Ultrasonography isa noninvasive and low-cost tech-nique that is now extensively usedfor morphological investigations inthe field of rehabilitation.32 The mea-surement validity of ultrasound com-pared with the magnetic resonanceimaging or computed tomographygold standards has been reported tobe acceptable.33,34 Indeed, althoughVMO strength cannot be directlyassessed in vivo, morphologicalchanges following exercise trainingcan be both observed and quantifiedvia ultrasonography and, thus, mus-cle force and excursion capabilitycan be determined.35

The purpose of this study was toinvestigate the surplus effect of hipadduction to seated LP exercise onVMO morphology, pain, and func-tion in patients with PFPS. We hy-pothesized that incorporation ofhip adduction with leg-press train-ing (LPHA) would result in more-beneficial effects on VMO hypertro-phy, pain, and functional improve-

Available WithThis Article atwww.ptjournal.org

• The Bottom Line clinicalsummary

• The Bottom Line Podcast

• Audio Abstracts Podcast

This article was published ahead ofprint on March 19, 2009, atwww.ptjournal.org.

Surplus Value of Hip Adduction in Leg-Press Exercise in PFPS

410 f Physical Therapy Volume 89 Number 5 May 2009

http://www.ptjournal.org/cgi/content/full/89/5/409/DC1

http://www.ptjournal.org/cgi/content/full/89/5/DC1

http://www.ptjournal.org/cgi/content/full/89/5/DC1

ment compared with LP trainingalone.

MethodSetting and ParticipantsA total of 123 patients with a diagno-sis of PFPS were referred to our ki-nesiology laboratory by an orthope-dic surgeon (YFL). The inclusioncriteria were: (1) experience of an-terior or retropatellar knee pain afterperforming at least 2 of the followingactivities: prolonged sitting, stairclimbing, squatting, running, kneel-ing, hopping and jumping, and deepknee flexing; (2) insidious onset ofsymptoms unrelated to traumatic ac-cident; (3) presence of pain for morethan 1 month; and (4) age of 50 yearsand under (to eliminate the possibil-ity of osteoarthritis). In addition, par-ticipants had to exhibit at least 2of the following positive signs ofanterior knee pain during the initialphysical examination: (1) patellarcrepitus, (2) pain following isomet-ric quadriceps femoris muscle con-traction against suprapatellar resis-tance with the knee in slight flexion(Clarke’s sign), (3) pain followingcompression of the patella againstthe femoral condyles with the kneein full extension (patellar grind test),(4) tenderness upon palpation ofthe posterior surface of the patella orsurrounding structures, and (5) painfollowing resisted knee extension.

Participants were excluded if theyhad: (1) self-reported clinical evi-dence of other knee pathology; (2)patellar tendinitis or knee plica; (3) ahistory of knee surgery; (4) centralor peripheral neurological patholo-gy; (5) knee radiographic abnormal-ities (eg, knee osteoarthritis) orlower-extremity malalignment (eg,foot pronation); (6) severe knee pain(visual analog scale [VAS] score of�8); or (7) received nonsteroidalanti-inflammatory drugs, injections,or physical therapy intervention inpreceding 3 months.

Of the 123 patients initially screened,98 met the study inclusion criteria.Nine of the 98 recruited participantsdeclined to participate before ran-domization. Therefore, a total of 89participants were enrolled in thisstudy. Sample size was calculatedusing a predetermined difference be-tween treatment groups of 1.5 cmfor worst pain on a 10-cm VAS. As-suming a standard deviation of 2 cm,at least 29 participants per treatmentgroup were required to attain 80%power.

Participants were randomly allo-cated to 1 of 3 groups: a group thatreceived hip adduction combinedwith leg-press exercise (LPHA), agroup that received LP exercise only,or a group that received no exercise(control group). Ten participantslater dropped out of the study due topersonal factors (not knee pain) orwork commitment. Seventy-nine par-ticipants completed the trial (27 ineach exercise group and 25 in thecontrol group, Fig. 1). The demo-graphic data for the 3 study groupsare presented in Table 1. There wereno significant between-group differ-ences for any of the demographicvariables. None of the participantswere engaged in regular sportingactivities.

RandomizationAll volunteers were enrolled afterproviding written informed consent.The study was performed in a blind(nondiscriminatory) manner. A sin-gle physical therapist, unaware ofthe purpose of the study, was re-sponsible for randomization and in-terventions. Stratified allocation wascarried out with regard to the num-ber of affected sides (unilateral orbilateral) and symptom severity(Lysholm scale scores �65 or �65).Participants were randomly assignedto the LP group, LPHA group, or con-trol group in blocks of 9 (chosenthrough numbered opaque enve-lopes) and participated in 3 weekly

exercise sessions for 8 weeks. Twoassessment sessions were performedby another physical therapist (blindedto each patient’s grouping) beforeand after the 8-week intervention.

InterventionsSimple LP exercise. Leg-press ex-ercise was performed unilaterallystarting from 45 degrees of knee flex-ion to full extension using an EN-Dynamic Track machine.* Exercisewithin the functional range was con-sidered safe for patients with PFPS.36

A blue Thera-Band† was tied to eachpatient’s thigh (without resistance)to maintain consistent tactile stimu-lation among groups. Prior to the be-ginning of exercise training, the uni-lateral 1-repetition maximum (RM)strength of the lower extremity wasdetermined by Odvar Holten Pyra-mid diagram37 with repetition-to-fatigue testing. Patients were unilat-erally trained at 60% of 1 RM for 5sets of 10 repetitions. The 1 RM wasre-measured every 2 weeks, and theexercise intensity was adjusted ac-cordingly. A 60-Hz metronome wasused to control the exercise paceat 2-second concentric and eccentriccontractions from 45 degrees ofknee flexion to full extension. Therewere 2-second breaks between rep-etitions and 2-minute breaks be-tween sets. Limbs were alternativelytrained between exercise sets.

LPHA. This exercise was per-formed as per the LP, except that a50-N hip abduction force was ap-plied to the distal one third of thethigh. This force was achieved bytying a blue Thera-Band to an armof the EN-Dynamic Track machine(Fig. 2). Therefore, this exercise wasa combination of LP and 50-N isomet-ric hip adduction.

* Enraf-Nonius BV, Vareseweg 127, 3047 AT,Rotterdam, the Netherlands.† The Hygenic Corp, 1245 Home Ave, Akron,OH 44310-2575.



Table 1.Demographic Data for Study Participantsa

VariableLPHA Group

(n�29)LP Group(n�30)

Control Group(n�30) P

Sex (male:female) 8:21 8:22 4:26 .337

Age (y) 38.6�10.8 40.2�9.9 43.9�9.8 .129

Height (cm) 162.3�7.2 161.3�8.4 159.7�5.2 .370

Weight (kg) 58.3�9.0 60.1�11.2 57.4�6.9 .505

Body mass index (kg/m2) 22.2�3.2 23.0�3.0 22.5�2.1 .498

Involved side (bilateral:unilateral) 19:10 18:12 18:12 .882

Duration of symptoms (mo) 41.8�36.1 38.3�34.2 27.7�41.0 .056

a Data are presented as mean�SD. LPHA�hip adduction with leg-press exercise, LP�leg-press exercise.

Figure 1.Flow chart demonstrating the progression of participants through the trial. LPHA�hip adduction with leg-press exercise, LP�leg-press exercise.



A hot pack was applied to the quad-riceps femoris muscle for 15 minutesbefore exercise was commenced. Af-ter exercise completion, participantswere asked to stretch the quadri-ceps, hamstring, iliotibial band, andcalf muscle groups and were given acold pack to apply for 10 minutes.Stretches were maintained for 30seconds and were repeated 3 timesfor each muscle group. All study par-ticipants were asked not participatein any form of sport or exercise dur-ing the intervention period.

Control group. Control groupparticipants did not receive any ex-ercise intervention, but were pro-vided with health educational mate-rial regarding patellofemoral pain.They were advised not to perform orreceive any exercise program or in-tervention. Neither tape nor bracewas used. Exercise training was im-plemented after the 8-week controlperiod.

Outcome MeasurementsVAS pain assessment. The VAS isa reliable, well-validated, and widelyused tool for assessing knee pain.38–41

A 100-mm VAS was used to measurethe worst pain (VAS-W) experiencedin the previous week.

Functional evaluation. The Lys-holm scale was used to measurefunctional ability. The scale rangesfrom 0 to 100 points (with a score of100 indicating maximal function)and was originally designed to eval-uate symptoms and functions per-taining to knee injury.42 There are8 components to this assessment:stair climbing (10 points), squatting(5 points), pain (25 points), pres-ence of a limp (5 points), locking(15 points), instability (25 points),swelling (10 points), and the require-ment of support when walking (5points). The reliability, validity, androbustness of the Lysholm scale havebeen well documented.42–48 TheLysholm scale was found to correlate

highly with the Kujala anterior kneepain scale (r�.86)45 and has beenused as a knee function evaluationtool in patients with patellofemoraldisorders.45,49–51

Measurement of VMO morphol-ogy. Vastus medialis oblique mus-cle morphology was assessed by ul-

trasonography (HDI 5000‡) with a5- to 12-MHz broadband linear-arraytransducer (38 mm). The ultrasono-graphic measurements included VMOcross-sectional area (CSA) on thepatellar-base level and VMO volume

‡ Advanced Technology Laboratories, 22100Bothell Everett Hwy, Bothell, WA 98041-3003.

Figure 2.Hip adduction with leg-press exercise (a), with a close lateral view of the setup ofresisted hip adduction via Thera-Band (b).



under the patellar-base level.52 Allmeasurements were obtained whileparticipants were lying on a bed,with both legs relaxed (feet werepositioned in a frame to prevent legrotation) and a thick padded towelplaced underneath the knee to main-tain resting position.

The longitudinal length of the patella(in millimeters) was determinedfrom the upper border to the lowerborder with calipers. The VMO vol-ume under the patellar base was ap-proximated from a series of VMOCSAs using the trapezoidal rule.52 Toobtain a valid calculation of VMO vol-ume from the sonographic image, acustom-made holder was used to fixthe probe.52 The holder was cali-brated to quantify movement of thetransducer by synchronizing with ascaled hub, which was turned in afull circle to mobilize the transducerby 1 mm from the proximal patellarbase toward the distal patellar apexalong a line perpendicular to thehorizontal representing the upperborder of the patella. The first VMO

CSA was taken from the line pass-ing through the patellar-base level(Fig. 3). Serial VMO CSAs were ob-tained every 2 mm until the VMOimage on the visual display faded. Tocontrol for any potential confound-ing pressure exerted by the probeholder, gel was applied to the skinsuch that there was no direct contactbetween the probe and the skin.53

The image was carefully monitoredby the examiner to ensure that theVMO was not being compressed(Fig. 3).

The intraclass correlation coeffi-cients for between-day test-retest re-liability of VMO CSA and volumemeasurements were .96 and .94, re-spectively. The actual day-to-day dif-ferences (X�SD) were 0.02�0.30cm2 and 0.06�0.68 cm3, respec-tively. The standard errors of mea-surement were 0.29 and 0.52,respectively.

Data AnalysisData obtained from the most symp-tomatic knee were analyzed using

SPSS version 11.0.§ Data were sub-jected to an intention-to-treat analy-sis and included all dropouts. De-scriptive statistics (X�SD) wereused to determine participant char-acteristics. Prior to statistical analy-sis, the Kolmogorov-Smirnov testwas performed to assess the normal-ity of continuous data. Normallydistributed baseline demographicvariables (age, body height, bodyweight, and body mass index) werecompared by 1-way analysis of vari-ance (ANOVA). Non-normally dis-tributed variables (symptom dura-tion) were compared by Kruskal-Wallis test with an alpha of .05. Sexand numbers of afflicted sides (bilat-eral versus unilateral) were com-pared by chi-square test with analpha of .05. For each outcome vari-able measured, a 2 (preinterventionand postintervention) � 3 (treat-ment groups: LPHA, LP, and control)2-way mixed ANOVA was per-formed. When a significant 2-way in-teraction was detected, post hocanalysis was performed using Bonfer-roni adjustment (P�.008).

ResultsAll exercise intervention participantsexcept one attended all scheduledexercise sessions. One participant inthe LP group completed only half ofthe intervention and subsequentlydropped out of the study due towork commitments (Fig. 1). The re-maining study participant dropoutsin both exercise groups completedthe exercise programs but did notattend postintervention evaluations.

Results pertaining to VAS-W, Lysh-olm scale scores, VMO CSA, andVMO volume for the 3 groups beforeand after the 8-week interventionperiod are shown in Table 2. Therewere no significant baseline differ-ences among the groups. The 2-wayANOVA for repeated measures re-

§ SPSS Inc, 233 S Wacker Dr, Chicago, IL60606.

Figure 3.Measurement of vastus medialis oblique muscle cross-sectional area on patellar-baselevel.



vealed significant interactions forVAS-W, Lysholm scale scores, VMOCSA, and VMO volume. Post hocanalyses indicated that VAS-W, Lysh-olm scale scores, VMO CSA, andVMO volume significantly increasedfollowing intervention in the LP andLPHA groups (P�.008), but not thecontrol group (Tab. 2). Furthermore,values pertaining to the VAS-W andLysholm scale were significantly bet-ter in both exercise groups com-pared with the control group afterintervention (P�.008) (Tab. 3). Thevalues for VMO volume were signif-icantly higher in the LP group com-pared with the control group afterintervention (P�.008), whereas thebetween-group difference in VMOCSA did not reach the level of ad-justed significance (P�.012). Thevalues for VMO CSA and VMO vol-ume were not different between theLPHA group and the control groupafter intervention (P�.046 andP�.02, respectively). No differenceswere detected between the LP andLPHA groups (Tab. 3).

The pretest-posttest effect size in theLPHA group ranged from 0.48 to0.77 for VMO morphology and 0.76to 1.10 for VAS-W and Lysholm scalescores, and the corresponding valuesfor the LP group were 0.71 to 0.75and 0.89 to 0.95, respectively. Whencomparing the effect between theLPHA and control groups, effect sizevalues were 0.78 for VAS-W, 1.12 forthe Lysholm scale scores, and 0.56 to0.77 for VMO CSA and VMO volume.These effect sizes were consistentlysmaller than those associated withthe LP group and control group com-parison, where effect size valueswere 0.92 for VAS-W and 0.75 to0.77 for VMO CSA and VMO volume.An exception was Lysholm scalescores (effect size�1.01). However,there was no significant difference inimprovement for any variable be-tween the LP and LPHA groups.

Tab

le2.

Com

par

ison

Betw

een

Prei

nter

vent

ion

and

Post

inte

rven

tion

Cha

nges

ofPa

in,

Func

tion,

and

Vast

usM

edia

lisO

bliq

ueM

uscl

e(V

MO

)M

orp

holo

gyfo

rth

eH

ipA

dduc

tion

With

Leg-

Pres

sEx

erci

se(L

PHA

)G

roup

,th

eLe

g-Pr

ess

Exer

cise

(LP)

Gro

up,

and

the

Con

trol

Gro

upa

Var

iab

le

LPH

AG

rou

p(n

�2

9)

LPG

rou

p(n

�3

0)

Co

ntr

ol

Gro

up

(n�

30)

Pre

-in

terv

enti

on

Po

st-

inte

rven

tio

n

Mea

nD

iffe

ren

ce(9

5%

CI)

PP

re-

inte

rven

tio

nP

ost

-in

terv

enti

on

Mea

nD

iffe

ren

ce(9

5%

CI)

PP

re-

inte

rven

tio

nP

ost

-in

terv

enti

on

Mea

nD

iffe

ren

ce(9

5%

CI)

P

VAS-

W4.

80�

2.26

2.62

�2.

51�

2.18

(�3.

17to

�1.

19)

�.0

05*

4.85

�2.

492.

26�

2.20

�2.

58(�

3.56

to�

1.61

)�

.005

*4.

99�

2.18

4.81

�2.

55�

0.18

(�1.

16to

0.80

).7

15

Lysh

olm

scal

e74

.8�

12.1

85.7

�8.

510

.93

(7.2

7to

14.5

9)�

.005

*75

.7�

12.8

86.5

�10

.410

.73

(7.1

3to

14.3

3)�

.005

*75

.1�

9.3

75.7

�10

.90.

67(�

2.93

to4.

27)

.714

VMO

CSA

(cm

2)

3.67

�1.

454.

24�

1.43

0.57

(0.2

2to

0.92

).0

04*

3.75

�1.

594.

46�

1.90

0.71

(0.3

4to

1.08

)�

.005

*3.

39�

1.47

3.38

�1.

52�

0.01

(�0.

38to

0.36

).9

62

VMO

volu

me

(cm

3)

3.04

�2.

184.

12�

1.83

1.08

(0.5

8to

1.59

)�

.005

*3.

38�

2.37

4.45

�2.

521.

06(0

.57

to1.

56)

�.0

05*

2.76

�2.

012.

82�

1.91

0.05

(�0.

45to

0.55

).8

38

aD

ata

are

pre

sent

edas

mea

n�SD

.VA

S-W

�w

orst

pai

nas

mea

sure

dby

the

100-

mm

visu

alan

alog

scal

e,C

SA�

cros

s-se

ctio

nala

rea,

CI�

confi

denc

ein

terv

al.

Ast

eris

kde

note

ssi

gnifi

cant

diffe

renc

ebe

twee

np

rein

terv

entio

nan

dp

ostin

terv

entio

nva

lues

(P�

.008

).



DiscussionStrengthening of the knee extensorvia hip adduction is a very commontherapeutic approach for treatingpeople with patellofemoral pain.14,16

The effects of 8-week LP and LPHAexercise interventions on pain re-duction, functional improvement,and VMO hypertrophy were compa-rable, indicating that there is no ad-ditive beneficial effect of incorporat-ing hip adduction with LP exercise.This finding may be attributed to thefact that during simple LP exercise,the hip adductor magnus and longusmuscles are simultaneously activatedto stabilize hip movement.54

This is the first study, to our knowl-edge, that investigated the clinicaleffects of adding hip adduction toLP exercise for the management ofPFPS. Limited knowledge might begained if change of the training po-sition would result in different out-comes. However, the effects of exer-cise training on pain reduction in thecurrent study can be considered clin-ically significant based on a previousreport that a VAS change of 1.5 mmin patients with PFPS is the minimaldifference to be considered clinicallyimportant.38 These findings are in ac-cordance with the results of previ-ous studies regarding pain reductionwhere various quadriceps femorismuscle retraining exercise protocolswere used.49,55–59 Functional perfor-mance (as indicated by Lysholmscale scores) following our exercise

intervention improved by approxi-mately 11 points. This finding is sim-ilar to those of a previous study inwhich Lysholm scale scores in-creased from 67.6�6.4 points to81.1�9.4 points after 6 weeks of iso-kinetic training.49 Unlike the VAS,the minimal change in the Lysholmscale scores representing a clinicallyrelevant improvement in functionalstatus is yet to be determined. Out of100 points, a score of 95 to 100 in-dicates excellent function, a score of84 to 94 indicates good function, ascore of 65 to 83 indicates fair func-tion, and a score of �65 indicatespoor function.42 Overall, the patientsin our study exhibited significantfunctional improvement from a levelof fair to good following LP or LPHAtraining. Functional improvementsin our study were significantly corre-lated with reductions in VAS-W (r��.451, P�.005). This finding is com-parable to that from a previous re-port (r��.424, P�.009).49 It mustbe borne in mind that the resultsof intention-to-treat may even beunderestimated.

Because the Lysholm scale is not aPFPS-specific scale, we made a fur-ther subanalysis of the stair-climbingand squatting items. Initially, 93%and 82% of the exercise interventionparticipants had difficulty perform-ing the stair-climbing and squattingtasks, respectively. After exercise in-tervention, 50% to 52% of the pa-tients achieved maximal scores for

stair climbing and squatting. Ourfindings of decreased pain and in-creased functional capacity agreewith previous research demonstrat-ing that patients with PFPS were ableto perform significantly more step-ups, step-downs, and squats beforepain onset after quadriceps femorismuscle training.59 The relationshipbetween quadriceps femoris musclestrength and locomotor function inpatients with PFPS has been docu-mented previously by Powers andPerry.60 Leg-press exercise training,especially in the eccentric contrac-tion mode, is better suited for indi-viduals with PFPS who demonstrateweaker eccentric than concentricquadriceps femoris muscle strength.61

Given the improvements in 1 RM(from 90�30 kg to 145�50 kg inthe LPHA group and from 89�33 kgto 138�51 kg in the LP group), it isnot surprising that both exercisegroups exhibited significant im-provements in the stair-climbing andsquatting scores on the Lysholmscale (r�.347, P�.05).

Folland and Williams62 concludedthat the primary morphological ad-aptation after resistance exercise isrelated to an increase in the CSA ofthe whole muscle and individualmuscle fibers (caused by an increasein myofibril size and number). Afterboth LPHA and LP exercise interven-tion in our study, VMO hypertrophywas observed as the result of trainingbecause the amount of improvement

Table 3.Comparison Among Group Changes of Pain, Function, and Vastus Medialis Oblique Muscle (VMO) Morphology for the HipAdduction With Leg-Press Exercise (LPHA) Group, the Leg-Press Exercise (LP) Group, and the Control Groupa

Variable

LPHA Group vs Control Group LP Group vs Control Group LPHA Group vs LP Group

Mean Difference(95% CI) P



VAS-W �2.19 (�3.44 to �0.93) .001* �2.54 (�3.79 to �1.30) �.005* 0.35 (�0.90 to 1.61) .577

Lysholm scale 9.99 (4.81 to 15.17) �.005* 10.73 (5.60 to 15.87) �.005* �0.74 (�5.92 to 4.44) .776

VMO CSA (cm2) 0.86 (0.02 to 1.70) .046 1.09 (0.25 to 1.92) .012 �0.23 (�1.07 to 0.62) .598

VMO volume (cm3) 1.30 (0.21 to 2.40) .020 1.63 (0.55 to 2.71) .004* �0.33 (�1.42 to 0.77) .556

a CSA�cross-sectional area, CI�confidence interval. Asterisk denotes significant difference among groups (P�.008).



was greater than the measurementerror. This is the first study usingnoninvasive ultrasonography to in-vestigate changes in VMO morphol-ogy in patients with PFPS followingLPHA or LP training. We speculatethat this positive outcome may bepartially a consequence of VMO hy-pertrophy. As all patients were givenhot packs before exercise and wereinstructed to stretch and apply coldpacks after exercise, it is possiblethat the positive outcomes may nothave been due to LP exercise alone.A previous study63 demonstratedthat centralization of the patella canresult from VMO strengthening andstretching procedures. Further re-search is needed to determinewhether patellar alignment or track-ing is altered with VMOhypertrophy.

We found that LPHA did not result infurther beneficial effects comparedwith LP exercise alone. It is possiblethat hip adductor activation duringLP training54 subtracted from theeffect of additional isometric hipadduction. However, it also is possi-ble that isometric hip adductiondoes not preferentially recruit theVMO.15,23–25,27 Further study is war-ranted to clarify this issue.

This study had several limitations.Only the VMO was examined by ul-trasonography after exercise inter-vention, and the VL and other com-ponents of the quadriceps femorismuscle were not assessed. Addi-tional investigation is warranted toexamine morphological changes inindividual quadriceps femoris mus-cles. A second potential limitationwas the use of 50 N as the force levelfor hip adduction. This force levelmay have been inadequate to inducea training effect. We also note thatoutcomes were assessed followingan 8-week exercise intervention pe-riod. Determining how these out-comes change with a longer-term

intervention period would be ofinterest.

ConclusionThe findings suggest that an 8-weekexercise program involving simpleLP training (from 45° of knee flexionto full extension) and stretching caninduce significant VMO hypertro-phy, improve knee function, and re-duce pain in patients with PFPS. Wefound that adding 50 N of hip adduc-tion to LP exercise had no furtherbeneficial effects on outcome com-pared with LP exercise alone afteran 8-week intervention in patientswith PFPS.

Ms Song, Dr YF Lin, Dr DH Lin, and Ms Janprovided concept/idea/research design. MsSong and Ms Jan provided writing. Ms Songand Mr Wei provided data collection. MsSong and Dr DH Lin provided data analysis.Ms Jan provided project management, facil-ities/equipment, and institutional liaisons. DrYF Lin and Dr DH Lin provided participants.Mr Wei and Ms Yen provided clerical sup-port. Dr YF Lin provided consultation (in-cluding review of manuscript before submis-sion).

The study protocol was approved by the Re-search Ethics Committee of National TaiwanUniversity Hospital.

An oral presentation of the results of thisstudy was given at the International Con-gress of the World Confederation for PhysicalTherapy; June 5, 2007; Vancouver, BritishColumbia, Canada.

This article was received June 23, 2008, andwas accepted February 3, 2009.

DOI: 10.2522/ptj.20080195

References1 Gerbino PG, Griffin ED, d’Hemecourt PA,

et al. Patellofemoral pain syndrome: eval-uation of location and intensity of pain.Clin J Pain. 2006;22:154–159.

2 Zappala FG, Taffel CB, Scuderi GR. Reha-bilitation of patellofemoral joint disorders.Orthop Clin North Am. 1992;23:555–566.

3 Tria AJ, Palumbo RC, Alicea JA. Conserva-tive care for patellofemoral pain. OrthopClin North Am. 1992;23:545–553.

4 Dixit S, Difiori JP, Burton M, Mines B. Man-agement of patellofemoral pain syndrome.Am Fam Physician. 2007;75:194–202.

5 Witvrouw E, Werner S, Mikkelsen C, et al.Clinical classification of patellofemoralpain syndrome: guidelines for non-operative treatment. Knee Surg SportsTraumatol Arthrosc. 2005;13:122–130.

6 Cowan SM, Bennell KL, Hodges, et al. De-layed onset of electromyographic activityof VMO relative to VL in subjects withpatellofemoral pain syndrome. Arch PhysMed Rehabil. 2001;82:183–189.

7 Hubbard JK, Sampson HW, Elledge JR.Prevalence and morphology of the media-lis oblique muscle in human cadavers.Anat Rec. 1997;249:135–142.

8 Hubbard JK, Sampson HW, Elledge JR. Thevastus medialis oblique muscle and its re-lationship to patellofemoral joint deterio-ration in human cadavers. J Orthop SportsPhys Ther. 1998;28:384–391.

9 Peeler J, Cooper J, Porter MM, et al. Struc-tural parameters of the vastus medialismuscle. Clin Anat. 2005;18:281–289.

10 Raimondo RA, Ahmad CS, Blankevoort L,et al. Patellar stabilization: a quantitativeevaluation of the vastus medialis obliquemuscle. Orthopedics. 1998;21:791–795.

11 Lieb FJ, Perry J. Quadriceps function: ananatomical and mechanical study usingamputated limbs. J Bone Joint Surg Am.1968;50:1535–1548.

12 Lin F, Wang G, Koh JL, et al. In vivo andnoninvasive three-dimensional patellartracking induced by individual heads ofquadriceps. Med Sci Sports Exerc. 2004;36:93–101.

13 Toumi H, Poumarat G, Benjamin M, et al.New insights into the function of the vas-tus medialis with clinical implications.Med Sci Sports Exerc. 2007;39:1153–1159.

14 Malone T, Davies G, Walsh WM. Muscularcontrol of the patella. Clin Sports Med.2002;21:349–362.

15 Karst GM, Jewett PD. Electromyographicanalysis of exercises proposed for differ-ential activation of medial and lateral quad-riceps femoris muscle components. PhysTher. 1993;73:286–299.

16 Powers CM. Rehabilitation of patellofemo-ral joint disorders. J Orthop Sports PhysTher. 1998;28:345–354.

17 Syme G, Rowe P, Martin D, Daly G. Dis-ability in patients with chronic patel-lofemoral pain syndrome: a randomisedcontrolled trial of VMO selective trainingversus general quadriceps strengthening.Man Ther. 2008 Apr 22 [Epub ahead ofprint].

18 Cowan SM, Bennell KL, Crossley KM, et al.Physical therapy alters recruitment ofthe vasti in patellofemoral pain syn-drome. Med Sci Sports Exerc. 2002;34:1879–1885.

19 Fagan V, Delahunt E. Patellofemoral painsyndrome: a review on associated neuro-muscular deficits and current treatmentoptions. Br J Sports Med. 2008;42:489–495.

20 Bose K, Kanagasuntheram R, Osman MBH.Vastus medialis oblique: an anatomic andphysiologic study. Orthopedics. 1980;3:880–883.



21 Hodges PW, Richardson CA. The influenceof isomeric hip adduction on quadricepsfemoris activity. Scand J Rehabil Med.1993;25:57–62.

22 Hanten WP, Schulthies SS. Exercise effecton electromyographic activity of the vas-tus medialis oblique and vastus lateralismuscles. Phys Ther. 1990;70:561–565.

23 Laprade J, Culham E, Brouwer B. Compar-ison of five isometric exercises in the re-cruitment of the vastus medialis obliquein persons with and without patellofemo-ral pain syndrome. J Orthop Sports PhysTher. 1998;27:197–204.

24 Cerny K. Vastus medialis oblique/vastuslateralis muscle activity ratios for selectedexercises in persons with and withoutpatellofemoral pain syndrome. Phys Ther.1995;75:672–683.

25 Earl JE, Schmitz RJ, Arnold BL. Activationof the VMO and VL during dynamic mini-squat exercises with and without isomet-ric hip adduction. J Electromyogr Kine-siol. 2001;11:381–386.

26 Zakaria D, Harburn KL, Kramer JF. Prefer-ential activation of the vastus medialisoblique, vastus lateralis, and hip adductormuscles during isometric exercises in fe-males. J Orthop Sports Phys Ther. 1997;26:23–28.

27 Hertel J, Earl JE, Tsang KW, Miller SJ. Com-bining isometric knee extension exerciseswith hip adduction or abduction does notincrease quadriceps EMG activity. Br JSports Med. 2004;38:210–213.

28 Miller JP, Sedory D, Croce RV. Vastus me-dialis oblique and vastus lateralis activityin patients with and without patellofemo-ral pain syndrome. J Sports Rehab. 1997;6:1–10.

29 Coqueiro KR, Bevilaqua-Grossi D, BerzinF, et al. Analysis on the activation of theVMO and VLL muscles during semisquatexercises with and without hip adductionin individuals with patellofemoral painsyndrome. J Electromyogr Kinesiol. 2005;15:596–603.

30 Grelsamer RP, McConnell J. The Patella: ATeam Approach. Gaithersburg, MD: As-pen Publishers Inc; 1998:127–129.

31 Powers CM. The influence of alteredlower-extremity kinematics on patel-lofemoral joint dysfunction: a theoreticalperspective. J Orthop Sports Phys Ther.2003;33:639–646.

32 Whittaker JL, Teyhen DS, Elliott JM, et al.Rehabilitative ultrasound imaging: under-standing the technology and its applica-tions. J Orthop Sports Phys Ther. 2007;37:434–449.

33 Dupont AC, Sauerbrei EE, Fenton PV, et al.Real-time sonography to estimate musclethickness: comparison with MRI and CT.J Clin Ultrasound. 2001;29:230–236.

34 Hides JA, Richardson CA, Jull GA. Mag-netic resonance imaging and ultrasonogra-phy of the lumbar multifidus muscle: com-parison of two different modalities. Spine.1995;20:54–58.

35 Lieber RL, Friden J. Clinical significance ofskeletal muscle architecture. Clin Orthop.2001;383:140–151.

36 Grelsamer RP, Klein J. The biomechanicsof the patellofemoral joint. J Orthop SportsPhys Ther. 1998;28:286–297.

37 Gustavsen R, Streek R. Training Therapy:Prophylaxis and Rehabilitation. Stutt-gart, Germany: Georg Thieme Verlag:1993.

38 Crossley KM, Bennell KL, Cowan SM,Green S. Analysis of outcome measures forpersons with patellofemoral pain: whichare reliable and valid? Arch Phys Med Re-habil. 2004;85:815–822.

39 Bennell KL, Bartram S, Crossley KM, GreenS. Outcome measures in patellofemoralpain syndrome: test-retest reliability andinterrelationships. Phys Ther Sport. 2000;1:32–34.

40 Eng JJ, Pierrynowski MR. Evaluation of softfoot orthotics in the treatment of patel-lofemoral pain syndrome. Phys Ther.1993;73:62–70.

41 Chesworth BM, Culham EG, Tata GE, et al.Validation of outcome measures in pa-tients with patellofemoral pain syndrome.J Orthop Sports Phys Ther. 1989;17:302–308.

42 Lysholm J, Gillquist J. Evaluation of kneeligament surgery results with special em-phasis on use of a scoring scale. Am JSports Med. 1982;10:150–154.

43 Marx RG, Jones EC, Allen AA, et al. Reli-ability, validity, and responsiveness offour knee outcome scales for athletic pa-tients. J Bone Joint Surg Am. 2001;83:1459–1469.

44 Kocher MS, Steadman JR, Briggs KK, et al.Reliability, validity, and responsiveness ofthe Lysholm knee scale for various chon-dral disorders of the knee. J Bone JointSurg Am. 2004;86:1139–1146.

45 Paxton EW, Fithian DC, Stone ML, Silva P.The reliability and validity of knee- specificand general health instruments in assess-ing acute patellar dislocation outcomes.Am J Sports Med. 2003;31:487–492.

46 Irrgang JJ, Snyder-Mackler L, Wainner RS,et al. Development of a patient-reportedmeasure of function of the knee. J BoneJoint Surg Am. 1998;80:1132–1145.

47 Bollen S, Seedhom BB. A comparison ofthe Lysholm and Cincinnati knee scoringquestionnaires. Am J Sports Med 1991;19:189–190.

48 Risberg MA, Holm I, Steen H, Beynnon BD.Sensitivity to changes over time for theIKDC form, the Lysholm score, and theCincinnati knee score. Knee Surg SportsTraumatol, Arthrosc. 1999;7:152–159.

49 Alaca R, Yilmaz B, Goktepe AS, et al. Effi-cacy of isokinetic exercise on functionalcapacity and pain in patellofemoral painsyndrome. Am J Phys Med Rehabil. 2002;81:807–813.

50 Lin YF, Lin JJ, Jan MH, et al. Role of thevastus medialis obliquus in repositioningthe patella: a dynamic computed tomogra-phy study. Am J Sports Med. 2008;36:741–746.

51 Natri A, Kannus P, Jarvinen M. Which fac-tors predict the long-term outcome inchronic patellofemoral pain syndrome? A7-year prospective follow-up study. MedSci Sports Exerc. 1998;30:1572–1577.

52 Lin YF, Lin JJ, Cheng CK, et al. Associationbetween sonographic morphology of vas-tus medialis obliquus and patellar align-ment in patients with patellofemoral painsyndrome. J Orthop Sports Phys Ther.2008;38:196–202.

53 Jan MH, Chai HM, Wang CL, et al. Effectsof repetitive shortwave diathermy forreducing synovitis in patients with kneeosteoarthritis: an ultrasonographic study.Phys Ther. 2006;86:236–244.

54 Enocson AG, Berg HE, Vargas R, et al. Sig-nal intensity of MR-images of thigh mus-cles following acute open- and closed-chain kinetic knee extensor exercise:index of muscle use. Eur J Appl Physiol.2005;94:357–363.

55 Witvrouw E, Lysens R, Bellemans J, et al.Open versus closed kinetic chain exer-cises for patellofemoral pain. Am J SportsMed. 2000;28:687–694.

56 Herrington L, Al-Sherhi A. A controlledtrial of weight-bearing versus non-weight-bearing exercises for patellofemoral pain.J Orthop Sports Phys Ther. 2007;37:155–160.

57 Hazneci B, Yildiz Y, Sekir U, et al. Efficacyof isokinetic exercise on joint positionsense and muscle strength in patellofemo-ral pain syndrome. Am J Phys Med Reha-bil. 2005;84:521–527.

58 Clark DI, Downing N, Mitchell J, et al.Physiotherapy for anterior knee pain: arandomized controlled trial. Ann RheumDis. 2000;59:700–704.

59 Crossley KM, Bennell KL, Green S, et al.Physical therapy for patellofemoral pain: arandomized, double-blinded, placebo-controlled trial. Am J Sports Med. 2002;30:857–865.

60 Powers CM, Perry J. Are patellofemoralpain and quadriceps strength associatedwith locomotor function? Phys Ther.1997;77:1063–1074.

61 Werner S. An evaluation of knee extensorand flexor torques and EMGs in patientswith PFPS in comparison with matchedcontrols. Knee Surg Sports Traumatol Ar-throsc. 1995;3:89–94.

62 Folland JP, Williams AG. The adaptationsto strength training: morphological andneurological contributions to increasedstrength. Sports Med. 2007;37:145–168.

63 Doucette SA, Goble EM. The effect of ex-ercise on patellar tracking in lateral patel-lar compression syndrome. Am J SportsMed. 1992;20:434–440.



surplus value of hip adduction in leg press exercise in patients with patellofemoral pain syndrome

Education