VASTUS MEDIALIS: ANATOMICAL AND FUNCTIONAL

CONSIDERATIONS AND IMPLICATIONS BASED UPON

HUMAN AND CADAVERIC STUDIES

Richard Lefebvre, PhD,a Alain Leroux, PhD,b Georges Poumarat, PhD,c Bruno Galtier, MD,d

Michel Guillot, MD,e Guy Vanneuville, PhD,f and Jean P. Boucher, PhDg

ABSTRACT

a Adjunct Profesdu Quebec a MontrLaboratoire Orthop

b Assistant ProfeUniversity MontreReadaptation dubridge Rehabilitati

c Full Professor,siques et Sportives

d Medical Doctotionnelles, Hopital

e Medical Doctotionnelles, HopitaAssociate Researccine, Place Henri

Objective: To provide an electrophysiological and functional description of the vastus medialis (VM) and contrast it to

an anatomical description.

Methods: Motor points of all superficial portions of the quadriceps were identified on the dominant side of 8 human

subjects and electrically stimulated to achieve a light contraction to trace and measure the orientation of the fibers.

Electromyography of theVMwas then recorded over 2motor points during isometric and isokineticmaximumknee extensions.

An independent laboratory dissected 39 cadaveric specimens focusing on fiber orientations and distal insertions of the VM.

Results: Results revealed 5 motor points for the quadriceps: 1 point for the vastus lateralis, 1 point for the rectus femoris

(RF), and 3 points for the VM. The 3 VM motor points suggest 3 separate groups of fibers: proximal (pf), median (mf),

distal (df). Fiber orientations ranged from 458 for VMpfs to 558 for VMdfs. Motor point stimulation and anatomical

dissection clearly showed that the VMpfs and VMmfs were inserted on a tendon common to the RF, whereas VMdfs were

attached directly to the medial aspect of the patella. Furthermore, the VMpfs were more active (P b .05) than VMdfs

during maximum knee extensions.

Conclusion: The anatomy, motor points, and electromyography clearly support an important distinction between the

VMpfs and VMdfs. The role of the VMpfs would be one of assisting the RF in knee extension, whereas the VMdfs would

track the patella medially without participating in knee extension. Because of these anatomical and functional differences,

the VMpfs and VMdfs should be addressed very differently during quadriceps rehabilitation in patellofemoral

dysfunctions. (J Manipulative Physiol Ther 2006;29:139-144)

Key Indexing Terms: Patellofemoral Pain Syndrome; Knee; Anatomy; Vastus Medialis

Overuse and misuse pertaining to joint dysfunc-

tions have recently been the focus of many

research endeavors.1-6 Because of its prevalence

in active adolescents and young adults,4 patellofemoral

dysfunctions (PFDs) captivate clinicians and researchers

alike.4-12 According to Smilie,13 PFDs were associated

with specific quadriceps muscle atrophy, especially in the

vastus medialis (VM) muscle, and limitations in terminal

sor, Departement deKinanthropologie, Universiteeal, Montreal, Quebec, Canada; Director of R&D,edique Medicus, Montreal, Quebec, Canada.ssor, Department of Exercise Science, Concordiaal and Centre de Recherche Interdisciplinaire enMontreal Metropolitain-Site, Constance-Leth-on Centre, Montreal, Quebec, Canada.U.F.R. Sciences et Techniques des Activites Phy-, Universite Blaise Pascal, Aubiere Cedex, France.r, Service de Reeducation Readaptation Fonc-Nord, Clermont-Ferrand Cedex 2, France.r, Service de Reeducation Readaptation Fonc-l Nord, Clermont-Ferrand Cedex 2, France;her, Laboratoire d’Anatomie, Faculte de Mede-Dunant, F-63000 Clermont-Ferrand, France.

knee extension. Clinical observations led Smilie13 to

conclude that the VM was responsible for terminal knee

extension (last 158). In a succession of 2 studies inves-

tigating the VM anatomy and function, Lieb and Perry14,15

showed that this muscle was divided into 2 independent

portions: the VM oblique (VMO) and the VM longus

(VML). Although Lieb and Perry15 showed that the VMO

was active throughout the range of knee extension, clinical

139

f Chairman of the Laboratory, Laboratoire d’Anatomie, Faculte deMedecine, Place Henri Dunant, F-63000 Clermont-Ferrand, France.

g Full Professor, Departement de Kinanthropologie, Universitedu Quebec a Montreal, Montreal, Quebec, Canada.

Sources of support: This study was partially supported by theNatural Sciences and Engineering Research Council of CanadaPostgraduate Scholarships Program.Submit requests for reprints to: Jean P. Boucher, PhD, Departe-

ment de Kinanthropologie, Universite du Quebec a Montreal,Montreal (Quebec), Canada H3C 3P8(e-mail: boucher.jean_p@uqam.ca).Paper submitted March 3, 2005; in revised form July 30, 2005.0161-4754/$32.00Copyright D 2006 by National University of Health Sciences.doi:10.1016/j.jmpt.2005.12.006

Fig 1. Measurement of motor point location and fiber orientation.X’s indicate the motor point locations, the vertical line represents

140 Journal of Manipulative and Physiological TherapeuticsLefebvre et al

February 2006Investigation of the Vastus Medialis

belief stemming from Smilie’s observations persisted and

their results were interpreted nonetheless as a confirmation

that the VM was more active in the last degrees of

extension.16-18 However, recent work in electrophysiology

and kinesiology failed again to verify that assumption, and

it was shown that the VM, and especially the VMO, was

in fact more active at around 908 of knee flexion.8,12,19,20

Adding to the anatomical and functional debate concern-

ing the VM, and fueling the clinical discussions, recent

anatomical findings2,21-23 support differences in the inner-

vation and architecture of the VM. Thiranagama22 found that

the VM is divided into 3 portions (upper, middle, and lower)

and is innervated by 2 nerves (branches) arising from the

femoral nerve: the lateral nerve supplies the upper fibers of

the VM, whereas the medial nerve supplies the middle and

lower fibers. Furthermore, these 3 portions revealed different

origins and insertions. Gqnal et al23 also reported additional

VM innervation via the saphenous nerve. This nerve would

arise from the medial portion of the femoral nerve and

supply only the lower fibers. Finally, very recently, Lin

et al,24 also combining anatomical and electrophysiological

approaches, reported significant mechanical and functional

differences between the components of the VM.

Even if the anatomical and functional pictures appear

clearer, the controversy persists. The fact that very little

efforts have been vested in conjugating anatomical and

functional aspects of the VM certainly contributed to the

clinical dilemma of understanding and treating PFDs. The

purpose of this study was to perform an anatomical,

electrophysiological, and functional investigation of the VM.

the thigh axis, and the oblique lines depict the orientation ofthe fibers.

METHODS

Experiment 1Eight volunteers aged between 20 and 28 years, free of

musculoskeletal disorders, participated in this experiment to

investigate the quadriceps muscles through electrophysio-

logical procedures. Before beginning the study, every

subject signed an informed consent document. Ethical

approval was obtained by the ethics committee of the

Universite du Quebec a Montreal (Quebec, Canada).

The experiment was conducted in 2 stages. The compo-

sition of each superficial quadriceps compartment was first

investigated by locating the motor points through standard

electrophysiological techniques. Second, motor point stim-

ulation was conducted to determine fiber orientations.

A motor point is defined as the point where fiber

recruitment is obtained with the least amount of current. To

locate the motor points of the quadriceps, each muscle was

investigated by moving a monopolar probe over its belly.

The anode was placed over the proximal aspect of the rectus

femoris (RF). The electrical stimulation was delivered by a

Biostim stimulator (model 6050; Mazet Electronique, Le

Mazet Saint Voy, France). The current was composed of

symmetrical biphasic square wave pulses (ie, duration of

200 milliseconds) delivered percutaneously to the muscle at

a frequency of 80 Hz. Stimulation train duration was set at

3 seconds. The location of each motor point was measured

in relation with the center of the patella (Fig 1).

The muscle fibers were stimulated at the motor points to

obtain a slight contraction, which made it possible to trace

the orientation of a fascicle. The electrode positions and the

electrical stimulation parameters were the same as presented

in the motor point procedure section except for the train

duration, which was set at 5 seconds. Fiber angles were

measured relative to a reference line that joined the anterior-

superior iliac spine (ASIS) to the center of the patella (Fig 1).

Experiment 2Experiment 2 was performed immediately after experi-

ment 1 on the same group of subjects and investigated VM

function through isokinetic measurements. This investiga-

tion measured isokinetic concentric and eccentric extension

torque output. These measurements, done on a KinCom

dynamometer (Chattecx Corp, Chattanooga, Tenn), were

Table 1. Mean values and standards deviations of muscle fiberorientations and motor point locations in the frontal plane

Angles (degree)

Motor points

X location (cm) Y location (cm)

VMpfs 45.17 F 8.18 2.40 F 0.64 6.40 F 1.26

VMmfs 56.00 F 10.53 2.77 F 0.70 3.50 F 0.35

VMdfs 55.83 F 9.24 2.83 F 0.63 1.93 F 0.20

RF 17.67 F 8.02 �1.23 F 0.45 12.72 F 0.92

VL 41.33 F 1.97 �2.88 F 0.45 7.10 F 1.10

Fig 2. Integrated EMG activity during maximum knee extensions(isokinetic contractions). Each bar represents the mean valuesand standard deviations for 8 subjects. (A) Comparison betweenthe activation level of the distal fibers and proximal fibers ofthe VM. (B) Activation level of the VMdfs as a function of 2selected angles.

Lefebvre et alJournal of Manipulative and Physiological Therapeutics

Investigation of the Vastus MedialisVolume 29, Number 2141

taken on 3 trials on the dominant side. During each

contraction, electromyographic (EMG) activity of 2 portions

of the VM corresponding to the proximal and distal motor

points was monitored. The maximum isometric contractions

with the knee at 908 of flexion were measured for

normalization purposes. The isokinetic portion involved a

range of motion from 1008 to 108 of knee flexion at a speed

of 308 per second. The EMG signal was recorded using

bipolar surface electrodes (standard Beckman Ag/AgCl;

Beckman Instruments Inc, Fullerton, Calif) and amplified

through preamplifiers (Chattecx). The active electrode was

located over the motor point of the muscle, the reference

electrode was secured 25 mm distal along the fiber

orientation, and the ground electrode was placed on the

lateral condyle of the femur. The surface electrodes were left

in place only after the electrode impedance was reduced by

standard techniques to 5 kV or less. The analog signal was

full wave rectified and integrated (16.7 milliseconds time

window averaging). The integrated EMG signal was then

acquired online at a rate of 60 Hz per channel using an

analog-to-digital interface (Chattecx).

The EMG quantification involved a technique developed

in our laboratories and presented elsewhere.8 Briefly, the

EMG signal preceding the peak torque (ie, 100 milliseconds

time window) reached over the range of motion (from 1008to 108 of knee flexion) during isokinetic contractions was

averaged for each muscle. The averaged EMG activity from

the 3 trials was used to evaluate the level of muscle

activation. In addition, the EMG activity amplitude of the

VMdf only was compared at 158 and 908 of knee flexion

during isokinetic contractions. The statistical analysis was

done with a muscle � contraction type � knee angle 3-way

factorial design with repeated measures on all factors.

Experiment 3Thirty-nine anterior regions of the thigh were taken from

22 human cadavers (3 fresh, 18 injected, and 1 fetus) to

investigate the VM through anatomical procedures. After a

visual inspection, none of the specimens showed the

presence of knee surgery or other knee problems. Dissection

of the VM was classically carried out by longitudinal

section from the ASIS to the patella and completed with

upper and lower transverse sections. The upper transverse

section was performed from the ASIS to the pubic

symphysis, whereas the lower section was tangential to

the apex of the patella. Muscle dissection was executed level

by level from superficial to deep layers. Dissection of the

femoral nerve and its branches to the VM was performed in

the cranio-caudal direction.

RESULTS

Experiment 1The electrophysiology results revealed that, contrary to

the RF and vastus lateralis (VL), where only 1 motor point

was found, the VM presented 3 motor points. The position

of these motor points suggests that the VM is divided into

3 sets of fibers: proximal (VMpf), medium (VMmf), and

distal (VMdf). The standard deviations of the motor point

positions showed very small values (Table 1). The small

standard deviations suggest that the positions of the motor

points are consistent across subjects and can be useful for

selecting EMG electrode positions. Moreover, a wide range

of muscle fiber angles were found. Mean muscle fiber

angles varied from 17.678 for the RF to 56.008 for the VMpf

(Table 1). Muscle fiber orientations, obtained from the

motor point stimulation technique, clearly showed that

VMpfs and VMmfs were inserted on a tendon common to

the RF, whereas VMdfs were attached directly to the medial

aspect of the patella (Fig 1).

Experiment 2Functional EMG procedures, the muscle main effect

specifically, revealed that the VMpfs were significantly

Fig 3. A medial view of the VM specimen. This view shows theVMpfs (a) and VMmfs (b) attaching directly to the RF distaltendon, and VMdfs (c) attaching directly to the medial aspect of thepatella.

142 Journal of Manipulative and Physiological TherapeuticsLefebvre et al

February 2006Investigation of the Vastus Medialis

more active (P b .05) than the VMdfs during knee extension

exercises (Fig 2A). When comparing the VMdf angle

effects, no significant differences were found (P N .05);

however, it could be seen that the VMdf had a tendency to

be more active at 908 of knee flexion (Fig 2B).

Experiment 3The morphology of the VM was similar in all specimens.

As shown in Fig 3, no fascial plane or definite separations

were found. Hence, none of the specimens presented

separated anatomical compartments of the muscle body.

The fiber alignment did not reveal any abrupt changes.

Proximally to distally, all the fibers showed an oblique

orientation. This transverse obliquity increased gradually

from the proximal to the distal portions. Furthermore, there

was clear evidence that the distal fibers attach directly to the

patella, whereas the medium and proximal fibers attach to

the RF tendon.

Two modes of innervation were found for the VM. In

22 cases, including the fetus, 2 distinct nerve pedicles

coming from the terminal branch of the femoral nerve were

present: (1) a short deep pedicle, rapidly perforating and

ramifying in the deep layers of the proximal portion of the

muscle; (2) a long superficial pedicle, located closely to the

first in the proximal portion of the muscle, traveling along

the medial and the distal portions and perforating the distal

portion and ramifying. In 17 cases, a unique nerve pedicle

coming from the terminal branch of the femoral nerve was

found. This pedicle travels along the muscle body and

perforates all portions of the VM.

DISCUSSION

The results of this study showed that the VM consistently

presented 3 motor points, each associated with a set of

fibers. The fiber orientation was 458, 568, and 568 for the

proximal, medium, and distal fibers, respectively. These data

differ from Lieb and Perry14 who found 188 for the VML

referred to as proximal fibers in this study. This difference is

associated with the way in which the fiber angle was

measured: in relationship with the center of the patella and

not the fiber insertion and orientation as conducted in the

present study. In fact, in our study, if a line is drawn from the

VMpf to the center of the patella, similar results are

obtained. Weinstabl et al21 also found a 158 to 188orientation for the proximal fibers. However, they presented

a picture showing a VM preparation (Weinstabl et al,21 Fig 1)

where all fibers revealed an oblique orientation. Being

mainly interested in understanding the VM function and,

more specifically, the effect of its architecture and how it

pertains to rehabilitation, we traced the orientations and

insertions and not the projections of the proximal portion on

the center of the patella. Hence, we believed our results to be

more relevant and functionally valid.

Anatomically, there was no evidence of a fascial plane

separating the proximal and distal portions of the VM as

reported by Peeler et al.25 Nonetheless, results obtained

from electrophysiological procedures were supported by

other more meaningful anatomical findings. In this study, all

specimens showed an oblique orientation of the VM fibers

throughout the muscle body, and this transverse obliquity

increased gradually from the proximal to the distal portions.

Galtier et al2 also reported a constant fiber obliquity in all

portions of the VM. Visual inspection of the anatomical

specimens reveals that the VMdfs attach directly to the

patella, whereas the VMmfs and VMpfs attach to the RF

tendon. These observations corroborate results obtained

from electrical stimulation of the VM motor points and are

similar to those presented by Lin et al.24 Because of the

different insertions, the VM portions would produce differ-

ent actions at the knee joint. In a study investigating VM

function, Leroux et al26 showed that electrical stimulation of

the motor points of VMpfs leads to knee extension, whereas

the VMdfs acted in tracking the patella medially without

Lefebvre et alJournal of Manipulative and Physiological Therapeutics

Investigation of the Vastus MedialisVolume 29, Number 2143

participating in knee extension. Lin et al24 also showed that

the proximal portion of the VM, or VML, and the distal

portion, or VMO, produce significantly different contribu-

tion to patella tracking. These results, along with ours, are in

agreement with Lieb and Perry14 and Goh et al3 who

showed that the distal fibers are responsible for patellar

medial alignment during knee extension.

Interpretation of the VM functional aspects is consistent

with results obtained through kinesiological EMG inves-

tigation. It was found that the VMpfs were 20% more active

than the VMdfs during maximum knee extension. Hence, as

expected, higher levels of activation would be required from

the VMpfs to assist the RF in knee extension as compared to

the VMdf activity needed to perform patellar alignment

during knee extension. Discrepancies found in the innerva-

tion of the VM portions could also contribute to the differ-

ences observed between VMpf and VMdf activation levels.

Peeler et al25 observed that bThe nerve never entered the

VM muscle belly at more than one site.Q They reported no

evidence of separate innervation of the VMO.25 In contrast,

the present study shows through deep muscle dissection

that, in a majority of the anatomical specimens, the VM is

innervated by 2 distinct nerve pedicles arising from the

terminal branch of the femoral nerve: a short deep pedicle

supplies the VMpfs, whereas a long superficial pedicle

supplies the VMdfs. This multiple innervation of the VM

was also well described by Galtier et al2 and Thiranagama.22

Gqnal et al23 also reported an additional innervation of the

VMdfs via the saphenous nerve. Furthermore, Lin et al24

showed that the different portions of the VM and the VL

revealed no coactivation (or insignificant levels) when

independently recruited through electrical stimulation.

Again, these results support the finding of independent

innervation within the VM.

Concerning the knee angle effect, it was shown that the

VMdfs have a tendency to be more active at 908 of knee

flexion during concentric and eccentric contractions. These

findings corroborate earlier results8,12,19,20 and further con-

tradict the clinical belief according to which the VMdfs

would be more active in the last degrees of knee extension.

When elaborating treatment protocols in view of restoring

VM function, it is then important to consider its architecture

and insertions, which make the VMdf compartment

responsible for medial patella tracking and not knee

extension, and that during knee extension it reaches its

highest level of activation at around 908 of flexion.

CONCLUSION

The results presented show that the VM should not be

considered a single muscle. The anatomy, electrophysiology,

and functional electromyography suggest an important

distinction between the VMpfs and VMdfs. The noninvasive

techniques used in the present study yielded results similar

to those found with the anatomical preparations. Further-

more, our data support the fact that the VM should not be

regarded as a muscle simply separated into a long and

oblique portion as suggested in other works.14,21,24,25

Instead, we propose that the VM should be perceived as a

continuum of oblique fibers divided into 3 functional

compartments spanning a major portion of the medial

aspect of the thigh: proximal fibers and medium fibers

attaching to the distal tendon of the RF, thus assisting the RF

in knee extension; and the distal fibers attaching directly to

the patella, having then a negligible role in knee extension.

The role of distal fibers would be one of tracking the patella

medially, against the external traction of the VL and RF

exerted at angles of 418 and 188, respectively, during knee

extension. Therefore, if the distal fibers of the VM are not

involved in knee extension, it could explain the great

difficulty in reversing their atrophy in PFDs through

traditional rehabilitation involving mostly knee extension

exercises. Our results, taken with others,8,14,21,24 strongly

support that the proximal and distal portions of the VM

should be addressed very differently during quadriceps

rehabilitation because of their important anatomical and

functional differences.

REFERENCES

1. Young A, Stokes M, Iles JF. Effects of joint pathology onmuscle. Clin Orthop Relat Res 1987;219:21-7.

2. Galtier B, Guillot M, Vanneuville G. Anatomical basis of therole of vastus medialis muscle in femoro-patellar degenerativearthropathy. Surg Radiol Anat 1995;17:7-11.

3. Goh JCH, Lee PYC, Bose K. A cadaver study of the functionof the oblique part of vastus medialis. J Bone Joint Surg 1995;77:225-31.

4. Avramidis K, Strike PW, Taylor PN, Swain ID. Effectivenessof electric stimulation of the vastus medialis muscle in therehabilitation of patients after total knee arthroplasty. ArchPhys Med Rehabil 2003;84:1850-3.

5. Palmieri RM, Ingersoll CD, Edwards JE, Hoffman MA, StoneMB, Babington JP, et al. Arthrogenic muscle inhibition is notpresent in the limb contralateral to a simulated knee jointeffusion. Am J Phys Med Rehabil 2003;82:910-6.

6. Hopkins J, Ingersoll CD, Edwards J, Klootwyk TE. Cryother-apy and transcutaneous electric neuromuscular stimulationdecrease arthrogenic muscle inhibition of the vastus medialisafter knee joint effusion. J Athl Train 2002;37:25-31.

7. Schmidt DR, Henry JH. Stress injuries of the adolescentextensor mechanism. Clin Sports Med 1989;8:343-55.

8. Boucher JP, King MA, Lefebvre R, Pepin A. Quadricepsfemoris muscle activity in patellofemoral pain syndrome. Am JSports Med 1992;20:527-32.

9. Karst GM, Jewett PD. Electromyographic analysis of exercisesproposed for differential activation of medial and lateralquadriceps femoris muscle components. Phys Ther 1993;73:286-99.

10. Cerny K. Vastus medialis oblique/vastus lateralis muscleactivity ratios for selected exercises in persons with andwithout patellofemoral pain syndrome. Phys Ther 1995;75:672-83.

11. Leroux A, Belanger M, Boucher JP. Pain effect on mono-synaptic and polysynaptic reflex inhibition. Arch Phys MedRehabil 1995;76:576-82.

144 Journal of Manipulative and Physiological TherapeuticsLefebvre et al

February 2006Investigation of the Vastus Medialis

12. Sczepanski TL, Gross MT, Duncan PW, Chandler JM. Effect ofcontraction type, angular velocity, and arc of motion on VMO:VL EMG ratio. J Orthop Sports Phys Ther 1991;14:256-62.

13. Smilie IS. The quadriceps in relation to recovery from injuriesof the knee. Physiotherapy 1949;35:53-7.

14. Lieb FJ, Perry J. Quadriceps function: an anatomical andmechanical study using amputated limbs. J Bone Joint Surg1968;50:1535-48.

15. Lieb FJ, Perry J. Quadriceps function: an electromyographicstudy under isometric conditions. J Bone Joint Surg 1971;53:749-58.

16. Brunet ME, Stewart GW. Patellofemoral rehabilitation. ClinSports Med 1989;8:319-29.

17. Henry JH. Conservative treatment of patellofemoral sublux-ation. Clin Sports Med 1989;8:261-77.

18. Kramer PG. Patellar malalignment syndrome: rationale toreduce excessive lateral pressure. J Orthop Sports Phys Ther1986;8:301-9.

19. Brownstein BA, Lamb RL, Mangine RE. Quadriceps torqueand integrated electromyography. J Orthop Sports Phys Ther1985;6:309-14.

20. Signorile JF, Kacsik D, Perry A, Robertson B, Williams R,Lowensteyn I, et al. The effect of knee and foot position on theelectromyographical activity of the superficial quadriceps.J Orthop Sports Phys Ther 1995;22:2-9.

21. Weinstabl R, Scharf W, Firbas W. The extensor apparatus ofthe knee joint and its peripheral vasti: anatomic investigationand clinical relevance. Surg Radiol Anat 1989;11:17-22.

22. Thiranagama R. Nerve supply of the VM muscle. J Anat1990;170:193-8.

23. Gqnal I, Arac S, Sahinoglu K, Birvar K. The innervation ofvastus medialis obliquus. J Bone Joint Surg 1992;74:624.

24. Lin F, Wang G, Koh JL, Hendrix RW, Zhang LQ. In vivo andnoninvasive three-dimensional patellar tracking induced byindividual heads of quadriceps. Med Sci Sports Exerc2004;36:93-101.

25. Peeler J, Cooper J, Porter MM, Thliveris JA, Anderson JE.Structural parameters of the vastus medialis muscle. Clin Anat2005;18:281-9.

26. Leroux A, Poumarat G, Boucher JP. Investigation of thequadriceps femoris function through electrical stimulation.J Athl Train 1997;32:115-8.