ZMPCZM016000.11.04 A clinical trail of NMES in improving quadraceps muscle strength and activation...

A Clinical Trial of NeuromuscularElectrical Stimulation in ImprovingQuadriceps Muscle Strength andActivation Among Women With Mildand Moderate OsteoarthritisRiann M. Palmieri-Smith, Abbey C. Thomas, Carrie Karvonen-Gutierrez,MaryFran Sowers

Background. Neuromuscular electrical stimulation (NMES) has demonstratedefficacy in improving quadriceps muscle strength (force-generating capacity) andactivation following knee replacement and ligamentous reconstruction. Yet, data arelacking to establish the efficacy of NMES in people with evidence of early radio-graphic osteoarthritis.

Objective. The purpose of this study was to determine whether NMES is capableof improving quadriceps muscle strength and activation in women with mild andmoderate knee osteoarthritis.

Design. This study was a randomized controlled trial.

Methods. Thirty women with radiographic evidence of mild or moderate kneeosteoarthritis were randomly assigned to receive either no treatment (standard ofcare) or NMES treatments 3 times per week for 4 weeks. The effects of NMES onquadriceps muscle strength and activation were evaluated upon study enrollment, aswell as at 5 and 16 weeks after study enrollment, which represent 1 and 12 weeksafter cessation of NMES among the treated participants. The Western Ontario andMcMaster Universities Osteoarthritis Index and a 40-foot (12.19-m) walk test wereused at each testing session.

Results. Improvements in quadriceps muscle strength or activation were notrealized for the women in the intervention group. Quadriceps muscle strength andactivation were similar across testing sessions for both groups.

Limitations. Women were enrolled based on radiographic evidence of osteo-arthritis, not symptomatic osteoarthritis, which could have contributed to our nullfinding. A type II statistical error may have been committed despite an a priori powercalculation. The assessor and the patients were not blinded to group assignment,which may have introduced bias into the study.

Conclusions. Four weeks of NMES delivered to women with mild and moderateosteoarthritis and mild strength deficits was insufficient to induce gains in quadricepsmuscle strength or activation. Future research is needed to examine the dose-response relationship for NMES in people with early radiographic evidence of osteo-arthritis.

R.M. Palmieri-Smith, PhD, ATC, isAssociate Professor, School of Ki-nesiology, University of Michigan,4745G CCRB, 401 WashtenawAve, Ann Arbor, MI 48109 (USA),and Bone & Joint Injury Preven-tion & Rehabilitation Center, Uni-versity of Michigan. Address allcorrespondence to Dr Palmieri-Smith at: [email protected].

A.C. Thomas, MEd, ATC, is a doc-toral candidate in the School of Ki-nesiology, University of Michigan.

C. Karvonen-Gutierrez, MPH, isResearch Analyst/Epidemiologistand doctoral student, Departmentof Epidemiology, School of PublicHealth, University of Michigan,and Center for Integrated Ap-proaches to Complex Diseases,School of Public Health, Universityof Michigan.

M.F. Sowers, PhD, is John G.Searle Professor of Public Healthand Professor of Epidemiology,Department of Epidemiology,School of Public Health, Universityof Michigan, and Center for Inte-grated Approaches to ComplexDiseases, School of Public Health,University of Michigan.

[Palmieri-Smith RM, Thomas AC,Karvonen-Gutierrez C, Sowers MF.A clinical trial of neuromuscularelectrical stimulation in improvingquadriceps muscle strength andactivation among women withmild and moderate osteoarthritis.Phys Ther. 2010;90:1441–1452.]

© 2010 American Physical TherapyAssociation

Research Report

Post a Rapid Response tothis article at:ptjournal.apta.org

October 2010 Volume 90 Number 10 Physical Therapy f 1441

Tibiofemoral osteoarthritis (OA)is the most common cause ofchronic disability in the United

States1 and is the most frequent indi-cation for total knee arthroplasty.2

No cure exists for OA, and currenttreatment approaches are quite lim-ited, with surgery becoming an in-creasingly frequent endpoint. In re-cent years, OA has been identified inadults at younger ages,3 accentuatingthe need to develop interventions ef-fective in restoring normal functionand capable of retaining or slowingthe process of joint degeneration. Ifeffective interventions are not iden-tified, joint replacement surgery willbe needed at earlier ages to maintainmobility and quality of life.

Quadriceps muscle weakness is com-monly associated with tibiofemoralOA,4–7 is linked with physical disabil-ity,5 and may play a role in diseasepathogenesis.8 Quadriceps musclestrength (force-generating capacity)appears to be highly related to func-tional performance,9 and minimiz-ing weakness has been shown toresult in clinical or mechanical im-provements in a variety of popula-tions.10–12 Enhancing quadricepsmuscle strength, therefore, is consid-ered to be of benefit, as it may im-prove quality of life. Research is lim-ited on the benefits of quadricepsmuscle strengthening early in the OAdisease process, as most investiga-tions have targeted people in the endstages of the disease or following to-tal knee arthroplasty.11 Improvingquadriceps muscle strength during

the early stages of the disease pro-cess may prove beneficial, not onlyfor maximizing function and mini-mizing pain but also for delaying therate of disease progression. Further-more, enhancing quadriceps musclestrength in people with radiographicevidence of the disease who arewithout symptoms may contributeto preventing the onset of symptom-atic OA.13

Neuromuscular electrical stimula-tion (NMES) delivered at high inten-sities to the quadriceps muscle hasbeen successful at improving quadri-ceps muscle strength and activationin patients who have undergone an-terior cruciate ligament reconstruc-tion and total knee arthroplas-ties.14–16 However, the efficacy ofNMES to improve quadriceps musclefunction in people with early-stageOA is lacking.

In this study, we evaluated whetherNMES was capable of improvingquadriceps muscle strength and acti-vation in people with mild to mod-erate OA. Our hypothesis was thatparticipants who received the NMESintervention would demonstrate im-provements in both quadriceps mus-cle strength and activation comparedwith participants who received notreatment.

MethodParticipantsParticipants were recruited from anOA registry at the University of Mich-igan. All volunteers in this registryhave radiographic evidence of OA,defined as a score of �2 on the Kell-gren and Lawrence (K-L) scale.17

One hundred twenty-three womenlisted in the registry had a K-L scalescore of 2 or 3 and were screened foreligibility via a telephone interview

Available WithThis Article atptjournal.apta.org

• The Bottom Line Podcast

• Audio Abstracts Podcast

This article was published ahead ofprint on July 29, 2010, atptjournal.apta.org.

The Bottom Line

What do we already know about this topic?

Neuromuscular electrical stimulation (NMES) has been shown to beeffective in improving quadriceps femoris muscle strength and activationin patients who have undergone knee replacement and ligament recon-struction. It is unknown whether NMES is capable of improving quadri-ceps function in a nonsurgical population, such as people with mild tomoderate knee osteoarthritis.

What new information does this study offer?

Neuromuscular electrical stimulation, as delivered in this study, wasincapable of improving quadriceps strength and activation. Increasing thedosage of the intervention, either by increasing the intensity of thestimulation or prolonging the length of delivery, may result in morefavorable outcomes; however, further research is necessary to confirm orrefute this premise.

If you’re a patient, what might these findings meanfor you?

Women with radiographic evidence of mild or moderate knee osteo-arthritis are not likely to improve thigh muscle (quadriceps) strength witha short course of NMES.

Neuromuscular Electrical Stimulation in Women With Osteoarthritis

1442 f Physical Therapy Volume 90 Number 10 October 2010

http://ptjournal.apta.org/content/90/10/suppl/DC1

http://ptjournal.apta.org/content/90/10/suppl/DC1

(Fig. 1). Women were excluded ifthey: (1) had previously undergone atotal knee arthroplasty or tibial os-teotomy; (2) had diagnosed arthritisof the hip, ankle, or foot; (3) had abody mass index of �40; (4) used anassistive device while ambulating;(5) had a disease of the central orperipheral nervous system; (6) hadany cardiac pathology; (7) reported aprevious ligamentous knee injury; or(8) had previously undergone NMEStherapy for OA. Women also wereexcluded if they were currently un-dergoing physical therapy for anylower-extremity orthopedic condi-tion, taking COX-2 inhibitors, or re-

ceiving corticosteroid or hyaluronicacid injections. All participants pro-vided written informed consentprior to enrollment.

Radiographic EvaluationAll women in the OA registry hadanterior-posterior knee radiographs(model X-GE MPX-80*) within 12months (mean�9.2 months) of studyinitiation. Bilateral, weight-bearing,semi-flexed radiographs were assessedfor evidence of OA according to theK-L scale criteria, a standardized ap-

proach in which higher scores indi-cate greater disease severity, by one ofthe authors (M.F.S.) and a musculo-skeletal radiologist.17 Radiographs ofknee joints where K-L scale scoreagreement was not achieved be-tween the 2 readers were re-readand, if required, subjected to a con-sensus reading performed by a sepa-rate musculoskeletal radiologist. Theinter-reader reliability between thestudy investigator and musculoskele-tal radiologist was high (kappa�.92).The methods and procedures forreading and standardizing the radio-graphs are described elsewhere.18

* General Electric Co, 5730 N Glen Park Rd,Milwaukee, WI 53209-4403.

▼

▼

▼

▼

▼

▼ ▼

▼

▼

▼

Assessed for eligibility (n=123)

Randomized (n=30)

Intervention group (n=16) Control group (n=14)

Lost to follow-up (n=1)Unable to report (n=1)

Lost to follow-up (n=4)

Analyzed (n=14)Excluded (n=0)

Analyzed (n=16)Excluded (n=0)

Discontinued NMES (n=1)Lost to follow-up (n=3)

Discontinued NMES (n=1)

Unable to contact (n=14)Refused to participate (n=21)Did not meet inclusion criteria• BMI ≥40 (n=18)• Other injury/disease (n=21)• Pain medication (n=9)• Other (n=10)

Enro

llmen

tA

lloca

tio

n16

-Wee

kFo

llow

-up

5-W

eek

Follo

w-u

pA

nal

ysis

Figure 1.Consolidated Standards of Reporting Trials diagram. BMI�body mass index, NMES�neuromuscular electrical stimulation.



Each knee joint was classified as:0�normal knee (no OA), 1�doubtfulOA, 2�mild OA, 3�moderate OA, or4�severe OA. Women eligible for in-clusion had a K-L scale score of 2 or 3.Women with bilateral knee OA wereincluded; however, these participantshad to have a K-L scale score nogreater than 3 in both knees.

Randomization and InterventionRandomization and assignment wereblind and conducted independentlyby one of the authors (R.P.S.) inde-pendently of intervention implemen-tation and data analysis. Eligible par-ticipants were randomly assigned to1 of 2 groups using a computer-generated random-number sequencefrom a randomization table and as-signed into blocks of 4 participants.Allocation to groups included usingsealed envelopes with slips of paperlabeled with an I or C, signifying as-

signment to the intervention groupor the control group, respectively.Envelopes were opened, and ran-domization occurred only after a par-ticipant met all inclusion criteria andsuccessfully completed the baselinestudy evaluation.

Sixteen women were assigned to theintervention group, and 14 womenwere assigned to the control group.Women assigned to the interventiongroup received NMES 3 times perweek over 4 weeks, for a total of 12NMES treatment sessions. Womenassigned to the control group re-ceived no intervention during thecourse of the study, as this is consid-ered the standard of care for peoplenot currently seeking physician in-volvement or treatment of any kind.

Each of the 12 NMES treatment ses-sions consisted of 10 electrically in-

duced contractions of the quadricepsmusculature. This course of interven-tion has been reported to improvequadriceps muscle strength.10,15 TheNMES was delivered to one limbonly, either the osteoarthritic limb incases of unilateral OA or the weakerlimb identified during the baselinestudy evaluation in cases of bilateralOA. During each treatment session,each woman was seated in a chairwith her leg positioned in 90 degreesof flexion and fixed to a pad that wasattached to a load cell. Self-adhesiveelectrodes (2.75 � 5 in [6.98 � 12.7cm], Dura Stick II†) were positionedproximally over the rectus femorismuscle and distally over the vastusmedialis muscle. Quadriceps musclecontractions were elicited using acommercial electrical stimulating

† Chattanooga Group Inc, 4717 Adams Rd,Hixson, TX 37343.

Figure 2.Study testing timeline. All participants completed 3 testing sessions. Testing sessions took place at the time of study enrollment andat 5 and 16 weeks after study enrollment. The 5- and 16-week time points represent 1 week and 12 weeks after neuromuscularelectrical stimulation (NMES), respectively, for the intervention group. Dotted line indicates randomization to group; dashed-dottedline indicates the end of the intervention period.



unit (Vectra Genisys†) delivering a2,500-Hz alternating current, modu-lated at 50 bursts per second, with aramp-up time of 2 seconds.16 Theelectrical current was set for a se-quence of 10 seconds on (which in-cludes the 2-second ramp-up time)16

and 50 seconds off. Current intensitywas set at each woman’s maximumtolerance, although a target intensityof at least 35% of the participant’sdaily knee extension maximum vol-untary isometric contraction (MVIC)was encouraged. Throughout thetreatment sessions, the women wereinstructed to not contract their quad-riceps muscles during the applica-tion of the stimulus. The same inves-tigator delivered the intervention toall of the women assigned to receiveNMES.

The percentage of the quadricepsmuscle MVIC generated by the NMESintervention was determined duringeach treatment session for each par-ticipant to ascertain how closely thetolerated stimulus intensity chosenby each woman came to matchingthe target intensity set by the inves-tigators (ie, a stimulus intensity thatwould produce 35% of their dailyMVIC). After the baseline evaluationsession, women in the control groupdid not receive any intervention, butwere instructed to maintain theircurrent activities throughout thecourse of the study. Participants inthe intervention group were givensimilar instructions regarding themaintenance of activity.

Testing ProceduresEvery participant, regardless ofgroup assignment, was assessed atthe time of study enrollment (base-line), as well as at 5 (primary timeendpoint) and 16 weeks followingstudy entry (Fig. 2). Follow-up timepoints were selected to assess boththe immediate and sustained effectsof the NMES therapy on the pri-mary outcomes (quadriceps musclestrength and activation). Each of

the 3 testing sessions included: (1)quadriceps muscle strength and ac-tivation assessment, (2) administra-tion of the Western Ontario andMcMaster Universities Osteoarthri-tis Index (WOMAC),19 and (3) atimed walk. Testing order was ran-domized for each participant priorto enrollment. All data collectionstook place between June 2008 andAugust 2009.

Data were collected in a single re-search laboratory by one of the au-thors (A.C.T.), a certified athletictrainer, who was trained to assess alloutcome measures, but who was notblinded to group assignment. At thestart of the trial, the assessor wasable to conduct testing for quadri-ceps muscle strength, quadricepsmuscle activation, and the timedwalk reliably within and betweensessions (intraclass correlation coef-ficient �.92).

Quadriceps Muscle Strength andActivation AssessmentParticipants were positioned in anisokinetic dynamometer (System 3‡)with their hips flexed at 85 degrees,their knees flexed to 90 degrees, andtheir backs supported. The test limbwas secured to the dynamometerarm, and a stabilization strap was se-cured over the pelvis. For women inthe intervention group, the test limbwas the limb subjected to the NMESintervention. The test limb for par-ticipants in the control group waseither the osteoarthritic limb (incases of unilateral OA) or the weakerlimb established at the time of base-line testing when bilateral OA waspresent. Two self-adhesive elec-trodes of the same size as those usedto deliver the NMES interventionwere applied to the quadriceps mus-cle. Participants were asked to com-plete 3 submaximal isometric kneeextension contractions to familiarize

themselves with the testing proce-dures. Following the practice trials, 3quadriceps muscle MVICs were per-formed, with 2 minutes of rest pro-vided between contractions. Whenvolunteers reached their peak torqueduring each of the MVIC trials, a su-pramaximal burst of electrical stim-uli (100 pps, 600-�s pulse duration,100-millisecond train duration, and130 V) was delivered (Grass S88§ andSIU8T§) in accordance with the burstsuperimposition technique.20

Torque data were exported in realtime during the collections to a sep-arate data acquisition unit (MP100�).Quadriceps muscle activation wascalculated using the central activa-tion ratio (CAR), where the peaktorque generated during the MVIC isdivided by the peak torque gener-ated from the electrical stimulation.A CAR of 1.00 represents completequadriceps muscle activation. Thetrial with the largest quadriceps mus-cle MVIC torque was normalized tobody mass (N�m/kg) and submittedto statistical analysis. The CAR calcu-lated from the trial with the largestMVIC also was used in the dataanalysis.

Pain and Function AssessmentsWomen rated their knee pain, stiff-ness, and disability on a scale from 1to 5 or none to extreme using the5-point Likert version of theWOMAC.19 Both the pain and disabil-ity components were scored toquantify patient-reported knee painand function. Lower scores are asso-ciated with less pain and betterfunctioning.

Participants also completed a timed40-foot (12.19-m) walk,21,22 whichserved as a quantitative measure offunctional performance. Participants

‡ Biodex Medical Systems Inc, 20 Ramsey Rd,Shirley, NY 11967-4704.

§ Astro-Med Inc, 600 E Greenwich Ave, WestWarwick, RI 02893.� Biopac Systems Inc, 42 Aero Camino, Goleta,CA 93117.



were instructed to walk down a tiledhallway at a purposeful (ie, brisk)pace. Participants completed 3 trials,and the trial with the fastest speedwas used in the data analysis.

Data AnalysisAn a priori power analysis indicatedthat 12 participants per group wereneeded to achieve 80% statisticalpower based on previous work ex-amining the effects of NMES onquadriceps muscle strength and CARafter total knee arthroplasty.16 Pro-jected effect sizes (Cohen d) of 3.14(strength) and 1.19 (CAR) were cal-culated using data that comparedthe CAR and MVIC prior to and 3months following NMES in a treat-ment group.16

Therapeutic efficacy was assessedusing an intent-to-treat analysis. Theprimary outcome measures for this

study were the quadriceps muscleMVIC and CAR, and the secondaryoutcome measures were the WOMACpain and disability subscales and thetimed walk. To ensure that covari-ates were successfully randomizedbetween the intervention and con-trol groups, demographic and clini-cal characteristics were comparedfrom the trial baseline assessment us-ing chi-square tests, independent ttests, and Fisher exact tests, asappropriate.

Linear mixed modeling with re-peated measures was used to assesswhether there was a change in thedependent variables. The modelspecifications included 2 indepen-dent variables (ie, group and time),as well as a group � time interactionterm, to assess whether there weredifferent group effects at the 3 timepoints. Time was treated as a class

variable, which allowed the compar-ison of week 5 versus baseline andweek 16 versus baseline. An unstruc-tured covariance matrix was se-lected for use in the mixed model,which allows for potential missingdata assumed to be missing at ran-dom. Missing data were accountedfor using case-wise deletion. The al-pha level for all tests was set atP�.05. We used SAS version 9.2# forstatistical analyses.

Change scores (follow-up � base-line) were calculated for both pri-mary and secondary outcome mea-sures. Effect sizes were quantifiedfor each group between the base-line and follow-up time points us-ing the Cohen d method (groupmean baseline � group mean atfollow-up divided by the pooledstandard deviation).

ResultsThere were no statistically signifi-cant differences in baseline demo-graphic or clinical characteristics(Tab. 1), suggesting successful ran-domization. Data for all 30 partici-pants who were randomly assignedto a group, including those who ei-ther violated protocol or were lost tofollow-up, were included in the anal-yses (Fig. 1).

Adverse EffectsNo adverse events related to theNMES treatment were reported.One participant in the interventiongroup had reported having heartpalpitations, which she attributedto the burst superimposition test-ing. The heart palpitations did notoccur during any of the testing ortreatment sessions, but rather oc-curred at home and did not beginon days in which the woman par-ticipated instudy-related activities.At the time the complaint was re-ported, the woman had finished re-

# SAS Institute Inc, PO Box 8000, Cary, NC27513.

Table 1.Baseline Demographic and Clinical Characteristics for the Intervention and ControlGroupsa

Variable

InterventionGroup(n�16)

ControlGroup(n�14) P

Age (y) 58 (2.7) 56.8 (2.9) .25

Height (cm) 162.3 (7.4) 163.3 (4.5) .61

Mass (kg) 86.3 (13.6) 87.8 (12.5) .76

BMI 32.7 (4.1) 32.1 (5.1) .76

Quadriceps muscle CAR 0.90 (0.08) 0.83 (0.15) .10

Quadriceps muscle MVIC(N�m/kg)

1.24 (0.35) 1.29 (0.46) .74

WOMAC pain subscalescore (range�5–25)

7.6 (2.7) 8.9 (4.3) .35

WOMAC disability subscalescore (range�17–85)

27.1 (12.1) 28.8 (15.3) .75

Timed walking speed (m/s) 1.43 (0.19) 1.40 (0.13) .67

Kellgren and Lawrence scalescore (% of grade 2)

93.8 92.9 1.00

Bilateral knee OA (%) 56 71 .47

Symptomatic OA (%) 50 50 1.00

a Values are means (SD) unless otherwise noted. No significant differences between groups were notedfor characteristics at baseline. For the purpose of defining a woman as symptomatic or notsymptomatic, we used the Likert scale included within the Western Ontario and McMaster UniversitiesOsteoarthritis Index (WOMAC). This scale asks participants to rate the level of pain in their kneesduring the 48 hours prior to testing. Women rating their knee pain as 2 or greater were considered tobe symptomatic. BMI�body mass index, CAR�central activation ratio, MVIC�maximum voluntaryisometric contraction, OA�osteoarthritis.



ceiving the intervention. Her phy-sician cleared her to participate inthe remaining portions of the study(ie, the 5- and 16-week follow-upsessions), which she did but optednot to undergo the burst superim-position testing.

Protocol AdherenceOf the 16 women assigned to theintervention group, 14 completed all12 treatment sessions, as per thetreatment protocol. The women inthe intervention group were able totolerate a stimulus intensity toachieve the target MVIC percentage

of 35% or greater in 93% of the treat-ment sessions.

Primary Outcome Measures:Quadriceps Muscle Strengthand CARNo significant main effects wereidentified for the intervention group

Figure 3.A graphical representation of the means (SD) for all primary and secondary outcome measures over the 3 time points (baseline, 5weeks, and 16 weeks) for both groups (intervention and control): (A) quadriceps muscle maximum voluntary isometric contraction(MVIC), (B) central activation ratio, (C) Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscalescore, (D) WOMAC disability subscale score, and (E) timed walking speed. No significant differences were noted between the groupsat any of the 3 time points (P�.05).



versus the control group (MVIC:P�.74; CAR: P�.08) or for week 5(MVIC: P�.60; CAR: P�.85) or week16 (MVIC: P�.42; CAR: P�.23) ver-sus baseline for either of the primaryoutcomes. The group � time inter-action terms in the modeling werenot statistically significant, indicatingthat the intervention and controlgroups did not differ at week 5 orweek 16 on either the MVIC (P�.59and P�.15, respectively) or the CAR(P�.43 and P�.09, respectively)(Fig. 3). Table 2 presents the changescores and 95% confidence intervalsfor the primary outcome measures.Table 3 provides the within-groupeffect sizes and 95% confidence in-tervals for the primary outcomemeasures.

Secondary Outcome Measures:WOMAC and Timed WalkNo significant main effects wereidentified for the intervention group

versus the control groups (WOMACpain subscale: P�.90; WOMAC dis-ability subscale: P�.66; timed walk:P�.89) or for week 5 versus baseline(WOMAC pain subscale: P�.47;WOMAC disability subscale: P�.84;timed walk: P�.34). However, whencomparing week 16 versus baseline,the average WOMAC pain andWOMAC disability scores were sig-nificantly greater (P�.05 and P�.03,respectively), but there was no sig-nificant effect for timed walk(P�.39). The group � time interac-tion term for week 16 was significantfor WOMAC pain and disability sub-scale scores, indicating that therewere differences between the inter-vention and control groups at thistime point (P�.04 and P�.003, re-spectively); with women assigned tothe control group having greaterWOMAC pain and disability subscalescores than those assigned to the in-tervention group. Neither of the

group � time interaction terms weresignificant for the timed walk, indi-cating there were no differences inwalking speed between the interven-tion and control groups at week 5(P�.73) or week 16 (P�.96). Table 2presents the change scores and 95%confidence intervals from baselinefor the secondary outcome mea-sures. Table 3 provides the within-group effect sizes and 95% confi-dence intervals for the secondaryoutcome measures.

DiscussionQuadriceps muscle weakness im-pairs functional ability and may berelated to the onset of OA. Thus, iden-tifying interventions capable of im-proving quadriceps muscle strengthmay improve quality of life in patientswith this disease. This clinical trial ofNMES was designed to evaluate thepotential to improve quadriceps mus-cle strength and activation in a group

Table 2.Mean Change Scores (95% Confidence Intervals) for the Primary and Secondary Outcome Measures at the 5- and 16-WeekFollow-up Time Pointsa

Measure

Intervention Group (n�16) Control Group (n�14)

5 Weeks 16 Weeks 5 Weeks 16 Weeks

CAR �0.02 (�0.07 to 0.02) �0.04 (�0.11 to 0.03) 0.01 (�0.09 to 0.10) 0.02 (�0.06 to 0.10)

MVIC �0.03 (�0.21 to 0.14) �0.21 (�0.50 to 0.08) 0.05 (�0.22 to 0.31) 0.14 (�0.20 to 0.49)

WOMAC pain subscale score �0.53 (�2.37 to 1.30) �0.54 (�1.83 to 0.75) 0.00 (�1.16 to 1.16) 1.4 (�0.26 to 3.06)

WOMAC disability subscale score �4.86 (�11.29 to 1.56) �4.92 (�10.89 to 1.05) 0.00 (�5.00 to 5.00) 5.0 (0.59 to 9.41)

Timed walking speed �0.04 (�0.09 to 0.01) �0.05 (�0.13 to 0.03) 0.00 (�0.02 to 0.20) 0.04 (�0.09 to 0.17)

a Negative values reflect a decrease from baseline; positive values reflect an increase from baseline. Baseline data were subtracted from follow-up data tocalculate the scores. CAR�central activation ratio, MVIC�maximum voluntary isometric contraction, WOMAC�Western Ontario and McMaster UniversitiesOsteoarthritis Index.

Table 3.Within-Group Effect Sizes (95% Confidence Intervals) at the 5- and 16-Week Follow-up Pointsa

Measure

Intervention Group (n�16) Control Group (n�14)

5 Weeks 16 Weeks 5 Weeks 16 Weeks

CAR 0.20 (�0.53 to 0.91) 0.42 (�0.36 to 1.18) 0.00 (�0.78 to 0.78) �0.33 (�1.15 to 0.51)

MVIC 0.12 (�0.60 to 0.84) 0.46 (�0.33 to 1.21) �0.06 (�0.85 to 0.72) �0.29 (�1.11 to 0.55)

WOMAC pain subscale score 0.26 (�0.45 to 0.96) 0.30 (�0.44 to 1.03) 0.07 (�0.70 to 0.84) �0.45 (�1.25 to 0.39)

WOMAC disability subscale score 0.55 (�0.18 to 1.25) 0.55 (�0.21 to 1.27) �0.08 (�0.85 to 0.69) �0.55 (�1.36 to 0.29)

Timed walking speed 0.06 (�0.67 to 0.79) 0.31 (�0.47 to 1.06) �0.45 (�1.21 to 0.35) �0.18 (�0.99 to 0.54)

a Effect sizes were computed using the Cohen d method (mean difference between baseline and follow-up/pooled standard deviation). CAR�centralactivation ratio, MVIC�maximum voluntary isometric contraction, WOMAC�Western Ontario and McMaster Universities Osteoarthritis Index.



of women with mild or moderate kneeOA. Contrary to our hypothesis, thewomen who received NMES did notdemonstrate significantly increasedstrength or activation following com-pletion of the intervention protocol.

The current clinical trial used anNMES intervention protocol similarto those demonstrated to be effectivein people following anterior cruciateligament reconstruction10,14,15,23 andtotal knee arthroplasty.16 Protocolsdeemed efficacious do vary slightly,but most include a 2,500-Hz, alternat-ing current, a current intensity elicit-ing at least 10% of the individual’sMVIC, and a duration of 4 to 10 weeks.As the total duration of the interven-tion and the intensity at which it wasdelivered were similar to those ofthese other protocols, we believe theprimary reason for the absence of aneffect is related to the magnitude ofquadriceps muscle strength and acti-vation deficits demonstrated by ourparticipants at the time of enrollment.Previously published data demon-strated that quadriceps muscle weak-ness was present in people with mildto moderate OA (K-L scale scores of 2and 3), but that the magnitude of thisweakness was not large, with appar-ent strength deficits equaling approx-imately 10% to 14% compared toadults who were healthy and of a sim-ilar age.13,24 On the contrary, the mag-nitude of weakness evident initially af-ter total knee arthroplasty can be atleast 40% when compared with theunaffected limb,25 and the weaknesspresent initially after anterior cruciateligament reconstruction can be up-wards of 70% compared with the un-injured limb.26 Similarly, the magni-tude of activation failure can beupwards of 30% following total kneearthroplasy27 and 25% after anteriorcruciate ligament reconstruction.28 Asthese magnitudes of strength and acti-vation deficits are substantially largerthan those found in the current study,it could be argued that effectiveness ofNMES may be dependent on the mag-

nitude of activation or strength defi-cits. Although there are no data to di-rectly support this contention, ininstances where NMES has beenproven to be effective, the activa-tion deficits demonstrated weremarkedly larger (Mintken et al29:CAR�0.729; Stevens et al16:CAR�0.594) compared with thosefound in the current study.

Another consideration relates to themagnitude of quadriceps muscleweakness and activation failure andthe NMES dosage. In instanceswhere quadriceps muscle strengthand activation deficits are great, as isoften the case following knee re-placement and ligament reconstruc-tive surgery, NMES may be effectiveat lower doses (eg, over shorter timeperiods and at lower intensities),whereas people with less weaknessand activation failure may requiregreater doses to realize gains in quad-riceps muscle strength and activa-tion. This observation is supportedby the work of Talbot et al,30 whonoted increased quadriceps musclestrength following 12 weeks ofNMES (completed at home using abattery-operated stimulator, 3 timesper week for 12 weeks, 15 contrac-tions per session, with progressiveincreases in intensity every 4 weeks[weeks 1–4: 10%–20% of MVIC;weeks 5–8: 20%–30% of MVIC; andweeks 9–12: 30%–40% of MVIC]) inelderly people with symptomaticOA. Additionally, recent work sug-gests a dose-response relationshipfor NMES, with higher doses ofNMES associated with less quadri-ceps muscle strength loss followingtotal knee arthroplasty.31 The idealdose of NMES for people with mildto moderate OA is unknown andwarrants further study; however, it isapparent that 4 weeks of NMES atintensities sufficient to generate atleast 35% of the MVIC are inadequatein this population.

Quadriceps muscle weakness relatedto OA may stem from voluntary acti-vation failure (ie, an inability to fullyactivate the quadriceps muscle) ormuscle atrophy.27,32,33 Neuromuscu-lar electrical stimulation is thoughtto improve activation and induce hy-pertrophy, resulting in gains instrength.16,29,34–36 However, basedon its duration and intensity, ourNMES protocol would be more likelyto improve activation rather than in-duce muscle hypertrophy (reducingactivation deficits may be more crit-ical than reducing atrophy, as physi-cal function is more severely affectedin patients with OA who havegreater activation deficits4).

Muscle hypertrophy is typicallythought not to be discernible until8 to 12 weeks after initiating resis-tance training,37– 40 although someresearch suggests hypertrophy canoccur earlier.41 Therefore, strengthgains achieved before 8 weeks gen-erally are thought to result fromneural improvements, such as a de-cline in activation failure.37– 40 Fur-thermore, research completed inelderly people who were healthysuggests training intensities up-ward of 50% to 60% of the MVIC areneeded to induce hypertrophy.42

When considering these findings,our NMES protocol probably did notalter the structural characteristics ofthe muscle. This distinction is impor-tant because we are aware of no datathat highlight the contributions ofboth activation failure and atrophy toquadriceps muscle strength in peo-ple with early to moderate OA. Ifatrophy is the primary contributor toquadriceps muscle weakness, thenthe intervention, as delivered, maybe unlikely to improve musclestrength, and this fact could havecontributed to our null results. How-ever, if activation failure predomi-nantly causes the muscle weakness,then we would have expected to seesome benefit from the intervention,



if the NMES intensity was sufficientfor women with mild to moderateOA. As the women participating inthis trial had an average CAR of 0.873(coefficient of variation�7%) and el-derly people who were healthy dis-played an average activation ratio of0.955,43 we would surmise that vol-untary activation failure was likelypresent in many of women in ourstudy groups and at least partiallycontributed to the existing quadri-ceps muscle weakness. Therefore,we again suggest that the dosage pro-vided was not of a sufficient intensityto improve muscle strength or acti-vation in our study population. Fu-ture research should focus on exam-ining the dose-response relationshipin people at various stages of the OAdisease process.

If the dosage of NMES is critical inimproving quadriceps muscle strengthand activation in people with earlyOA, altering the protocol for futurestudies is reasonable. Such protocolchanges may include increasing thequantity and duration of NMES treat-ments; however, increasing the inten-sity of the stimulation may not be re-alistic. To achieve quadriceps musclecontractions of a sufficient intensityfor the current trial, some participantswere quite uncomfortable. Further-more, in some instances, the stimuluswas delivered near the maximum al-lowable by the electrical stimulationunit. Therefore, in instances whereminimal to moderate muscle weak-ness is present and where higher dos-ages of NMES treatment may be war-ranted, increasing the duration of theintervention may be necessary, but fu-ture studies are needed to evaluatethis. An additional consideration mer-iting discussion is using NMES alongwith quadriceps muscle resistance ex-ercises, which yielded favorable re-sults in a study by Petterson et al.11

The intervention for our study wasdelivered to a group of women withradiographic evidence of OA who re-

ported mild symptoms and minimaldisability. It is possible, therefore, thatthe use of NMES in women with fewsymptoms may not be effective de-spite the dosage utilized. Interveningin people with minimal or no symp-toms and radiographic evidence of OAappeared justified given available datasuggesting that quadriceps musclestrength in women may be involved inincident symptomatic OA.13

Although NMES appeared to have lit-tle effect on quadriceps musclestrength and activation, it did appearto influence WOMAC scores. Womenin the intervention group displayedlower WOMAC pain and disability sub-scale scores compared with women inthe control group at the 16-week timepoint, suggesting less pain and greaterfunction in the women who under-went NMES therapy. When examiningthe data closely (see Fig. 3, graphs Cand D), it becomes apparent that painand function worsened by week 16 inthe control group, whereas women inthe intervention group had similar(pain) or improved (function) scoresat week 16. This finding may suggestthat NMES could be effective at delay-ing or preventing the progression oronset of OA symptoms. As neithergroup displayed changes in quadri-ceps muscle strength or activation dur-ing the study period, the symptom-preventing effect that the interventiongroup displayed is likely not the resultof improved quadriceps muscle func-tion, nor is the worsening of OA symp-toms in the control group likely re-lated to decrements in quadricepsmuscle function. The reasons why self-reported pain and function wereworse in the control group comparedwith the intervention group at week16 are unknown and warrant furtherstudy.

Although statistically significant dif-ferences between groups at 5 weekswere not apparent for the WOMACdisability subscale score, it should bepointed out that a moderate effect

(Cohen d�0.55) was noted betweenbaseline and the 5-week follow-upfor the intervention group. Womenundergoing NMES therapy did seemto realize clinically relevant improve-ments in physical functioning, as de-fined by the WOMAC, after receivingtreatment. These clinically meaning-ful self-reported improvements alsomay have manifested during the timedwalk, although the effect size for thisvariable was small. At week 5, thewomen in the control group slowedtheir walking speed by �0.08 m/s,which in older adults is considered tobe a meaningful change in physicalfunction.44,45 Whether this rathersmall decline in speed is clinicallymeaningful in people with OA remainsunknown and should be examined.

This study is not without limitations.First, the women in our clinical trialwere enrolled based on radiographicevidence of OA and did not neces-sarily report symptoms of the dis-ease. Only half of the present studysample reported symptoms. Peoplewith radiographic OA and withoutsymptoms may not respond to NMEStherapy, and this fact could havecontributed to our null finding. Fu-ture studies should consider examin-ing the effects of NMES on indi-viduals who are symptomatic andasymptomatic separately, especiallyconsidering that people with symp-tomatic OA may have fewer pain-freeexercise alternatives, and thus it isimportant to ascertain its efficacy inthis population. Second, though weconducted an a priori power evalu-ation to determine the number ofparticipants to enroll in the trial,these calculations were completedusing data where NMES had a largeeffect (ie, the values were taken be-fore and after NMES therapy in peo-ple following total knee arthroplas-ty16). Therefore, the study may sufferfrom a type II statistical error, as theeffect of NMES on people with mildto moderate OA is not as strong.Third, we did not blind the assessor



or the patients to group assignment,which may have introduced bias intothe study. Finally, participants weretested and treated in the same posi-tions, which may have affected ourstrength and activation outcomes.Given that no effect was demon-strated for the treatment, this con-cern is almost negated.

ConclusionsIn this clinical trial, women with radio-graphic evidence of mild or moderateknee OA who were randomly assignedto receive an NMES intervention didnot realize gains in quadriceps musclestrength or activation compared withthe untreated group. The inability ofNMES to improve quadriceps musclefunction may be the result of the lowmagnitude of quadriceps muscle dys-function present in the women in ourstudy or the dosage of the interven-tion. The dose-response relationshipfor NMES should be determined inpeople with evidence of mild to mod-erate OA.

Dr Palmieri-Smith and Dr Sowers providedconcept/idea/research design and writing.Dr Palmieri-Smith, Ms Thomas, and MsKarvonen-Gutierrez provided data collec-tion. Dr Palmieri-Smith, Ms Karvonen-Gutierrez, and Dr Sowers provided dataanalysis. Dr Palmieri-Smith provided projectmanagement, fund procurement, and facil-ities/equipment. Ms Thomas and MsKarvonen-Gutierrez provided consultation(including review of manuscript beforesubmission).

This study was approved by the InstitutionalReview Board of the University of Michigan.

This study was supported by a grant fromthe Michigan Chapter of the Arthritis Foun-dation to Dr Palmieri-Smith.

Trial registration: ClinicalTrials.gov Identifier:NCT00500448.

This article was submitted October 6, 2009,and was accepted May 23, 2010.

DOI: 10.2522/ptj.20090330

References1 Guccione AA, Felson DT, Anderson JJ,

et al. The effects of specific medical con-ditions on the functional limitations of el-ders in the Framingham Study. Am J Pub-lic Health. 1994;84:351–358.

2 Praemer A, Furner S, Rice DP. Musculo-skeletal Conditions in the United States.2nd ed. Park Ridge, IL: American Academyof Orthopaedic Surgeons; 1992:145–170.

3 Sowers M, Lachance L, Hochberg M,Jamadar D. Radiographically defined os-teoarthritis of the hand and knee in youngand middle-aged African American andCaucasian women. Osteoarthritis Carti-lage. 2000;8:69–77.

4 Fitzgerald GK, Piva SR, Irrgang JJ, et al.Quadriceps activation failure as a modera-tor of the relationship between quadri-ceps strength and physical function in in-dividuals with knee osteoarthritis.Arthritis Rheum. 2004;51:40–48.

5 Hurley MV, Scott DL, Rees J, Newham DJ.Sensorimotor changes and functionalperformance in patients with knee os-teoarthritis. Ann Rheum Dis. 1997;56:641– 648.

6 Lewek MD, Rudolph KS, Snyder-MacklerL. Quadriceps femoris muscle weaknessand activation failure in patients withsymptomatic knee osteoarthritis. J OrthopRes. 2004;22:110–115.

7 Machner A, Pap G, Awiszus F. Evaluationof quadriceps strength and voluntary acti-vation after unicompartmental arthro-plasty for medial osteoarthritis of theknee. J Orthop Res. 2002;20:108–111.

8 Slemenda C, Heilman DK, Brandt KD,et al. Reduced quadriceps strength relativeto body weight: a risk factor for knee os-teoarthritis in women? Arthritis Rheum.1998;41:1951–1959.

9 O’Reilly SC, Jones A, Muir KR, Doherty M.Quadriceps weakness in knee osteoarthri-tis: the effect on pain and disability. AnnRheum Dis. 1998;57:588–594.

10 Snyder-Mackler L, Ladin L, Schepsis AA,Young JC. Electrical stimulation of thethigh muscles after reconstruction of theanterior cruciate ligament: effects ofelectrically elicited contraction of thequadriceps femoris and hamstring mus-cles on gait and on strength of the thighmuscles. J Bone Joint Surg Am. 1991;73:1025–1036.

11 Petterson SC, Mizner RL, Stevens JE, et al.Improved function from progressivestrengthening interventions after totalknee arthroplasty: a randomized clinicaltrial with an imbedded prospective co-hort. Arthritis Rheum. 2009;61:174–183.

12 Jenkinson CM, Doherty M, Avery AJ, et al.Effects of dietary intervention and quadri-ceps strengthening exercises on pain andfunction in overweight people with kneepain: randomised controlled trial. BMJ.2009;339:b3170.

13 Segal NA, Torner JC, Felson D, et al. Effectof thigh strength on incident radiographicand symptomatic knee osteoarthritis in alongitudinal cohort. Arthritis Rheum.2009;61:1210–1217.

14 Snyder-Mackler L, Delitto A, Stralka S, Bal-ley SL. Use of electrical stimulation to en-hance recovery of quadriceps femorismuscle force production in patients fol-lowing anterior cruciate ligament recon-struction. Phys Ther. 1994;74:901–907.

15 Snyder-Mackler L, Delitto A, Bailey SL,Stralka S. Strength of the quadriceps fem-oris muscle and functional recovery afterreconstruction of the anterior cruciate lig-ament: a prospective, randomized clinicaltrial of electrical stimulation. J Bone JointSurg Am. 1995;77:1166–1173.

16 Stevens JE, Mizner RL, Snyder-Mackler L.Neuromuscular electrical stimulation forquadriceps muscle strengthening after bi-lateral total knee arthroplasty: a case se-ries. J Orthop Sports Phys Ther. 2004;34:21–29.

17 Kellgren JH, Lawrence JS. The Epidemiol-ogy of Chronic Rheumatism. Atlas ofStandard Radiographs of Arthritis. Vol II.Philadelphia, PA: FA Davis Co; 1963.

18 Lachance L, Sowers MF, Jamadar D, Hoch-berg M. The natural history of emergentosteoarthritis of the knee in women. Os-teoarthritis Cartilage. 2002;10:849–854.

19 Bellamy N, Buchanan WW, Goldsmith CH,et al. Validation study of WOMAC: a healthstatus instrument for measuring clinicallyimportant patient relevant outcomes toantirheumatic drug therapy in patientswith osteoarthritis of the hip or knee.J Rheumatol. 1988;15:1833–1840.

20 Snyder-Mackler L, De Luca PF, WilliamsPR, et al. Reflex inhibition of the quadri-ceps femoris muscle after injury or recon-struction of the anterior cruciate ligament.J Bone Joint Surg Am. 1994;76:555–560.

21 Sowers M, Jannausch ML, Gross M, et al.Performance-based physical functioningin African-American and Caucasianwomen at midlife: considering body com-position, quadriceps strength, and kneeosteoarthritis. Am J Epidemiol. 2006;163:950–958.

22 Sowers M, Tomey K, Jannausch M, et al.Physical functioning and menopause states.Obstet Gynecol. 2007;110:1290–1296.

23 Fitzgerald GK, Piva SR, Irrgang JJ. A mod-ified neuromuscular electrical stimulationprotocol for quadriceps strength trainingfollowing anterior cruciate ligament re-construction. J Orthop Sports Phys Ther.2003;33:492–501.

24 Palmieri-Smith RM, Thomas AC, Karvonen-Gutierrez CA, Sowers MF. Isometric quad-riceps strength in women with mild, mod-erate, and severe knee osteoarthritis. Am JPhys Med Rehabil. In press.

25 Zeni JA Jr, Snyder-Mackler L. Early postop-erative measures predict 1- and 2-year out-comes after unilateral total knee arthro-plasty: importance of contralateral limbstrength. Phys Ther. 2010;90:43–54.

26 Kobayashi A, Higuchi H, Terauchi M, et al.Muscle performance after anterior cruci-ate ligament reconstruction. Int Orthop.2004;28:48–51.

27 Stevens JE, Mizner RL, Snyder-Mackler L.Quadriceps strength and volitional activa-tion before and after total knee arthro-plasty for osteoarthritis. J Orthop Res.2003;21:775–779.



28 Hart JM, Pietrosimone B, Hertel J, IngersollCD. Quadriceps activation following kneeinjuries: a systematic review. J Athl Train.2010;45:87–97.

29 Mintken PE, Carpenter KJ, Eckhoff D,et al. Early neuromuscular electrical stim-ulation to optimize quadriceps musclefunction following total knee arthroplasty:a case report. J Orthop Sports Phys Ther.2007;37:364–371.

30 Talbot LA, Gaines JM, Ling SM, Metter EJ.A home-based protocol of electrical mus-cle stimulation for quadriceps musclestrength in older adults with osteoarthritisof the knee. J Rheumatol. 2003;30:1571–1578.

31 Stevens-Lapsley JE, Lbalter JE, Eckhoff DG,Kohrt WM. Dose-response relationship forneuromuscular electrical stimulation tothe quadriceps after total knee arthro-plasty. Paper presented at: 57th AnnualMeeting of the American College ofSports Medicine; May 27–30, 2009; Seat-tle, Washington.

32 Mizner RL, Petterson SC, Stevens JE, et al.Early quadriceps strength loss after totalknee arthroplasty: the contributions ofmuscle atrophy and failure of voluntarymuscle activation. J Bone Joint Surg Am.2005;87:1047–1053.

33 Fink B, Egl M, Singer J, et al. Morphologicchanges in the vastus medialis muscle inpatients with osteoarthritis of the knee.Arthritis Rheum. 2007;56:3626–3633.

34 Wigerstad-Lossing I, Grimby G, Jonsson T,et al. Effects of electrical muscle stimula-tion combined with voluntary contrac-tions after knee ligament surgery. Med SciSports Exerc. 1988;20:93–98.

35 Cabric M, Appell HJ, Resic A. Fine struc-tural changes in electrostimulated humanskeletal muscle: evidence for predominanteffects on fast muscle fibres. Eur J ApplPhysiol Occup Physiol. 1988;57:1–5.

36 Pinczewski LA, Deehan DJ, Salmon LJ,et al. A five-year comparison of patellartendon versus four-strand hamstring ten-don autograft for arthroscopic reconstruc-tion of the anterior cruciate ligament [er-ratum in Am J Sports Med. 2005;33:927].Am J Sports Med. 2002;30:523–536.

37 Hakkinen K, Newton RU, Gordon SE,et al. Changes in muscle morphology,electromyographic activity, and forceproduction characteristics during pro-gressive strength training in young andolder men. J Gerontol A Biol Sci Med Sci.1998;53:B415–B423.

38 Higbie EJ, Cureton KJ, Warren GL III, PriorBM. Effects of concentric and eccentrictraining on muscle strength, cross-sectionalarea, and neural activation. J Appl Physiol.1996;81:2173–2181.

39 Narici MV, Roi GS, Landoni L, et al.Changes in force, cross-sectional area andneural activation during strength trainingand detraining of the human quadriceps.Eur J Appl Physiol Occup Physiol. 1989;59:310–319.

40 Young A, Hughes I, Round JM, EdwardsRHT. The effect of knee injury on the num-ber of muscle fibres in the human quadri-ceps femoris. Clin Sci (Lond). 1982;62:227–234.

41 Stevens JE, Walter GA, Okereke E, et al.Muscle adaptations with immobilizationand rehabilitation after ankle fracture. MedSci Sports Exerc. 2004;36:1695–1701.

42 Kalapotharakos VI, Michalopoulou M,Godolias G, et al. The effects of high- andmoderate-resistance training on musclefunction in the elderly. J Aging Phys Act.2004;12:131–143.

43 Stackhouse SK, Stevens JE, Lee SC, et al.Maximum voluntary activation in nonfa-tigued and fatigued muscle of young andelderly individuals. Phys Ther. 2001;81:1102–1109.

44 Kwon S, Perera S, Pahor M, et al. What is ameaningful change in physical perfor-mance: findings from a clinical trial inolder adults (the LIFE-P study). J NutrHealth Aging. 2009;13:538–544.

45 Perera S, Mody SH, Woodman RC, Studen-ski SA. Meaningful change and responsive-ness in common physical performancemeasures in older adults. J Am GeriatrSoc. 2006;54:743–749.



ZMPCZM016000.11.04 A clinical trail of NMES in improving quadraceps muscle strength and activation...

Health & Medicine

Transcript of ZMPCZM016000.11.04 A clinical trail of NMES in improving quadraceps muscle strength and activation...