O

AMJJC

OdwDSPrQcsIewrMtmRohgCtrs

I

Tayadds5mDdpts

P1

P

riginal Research

Randomized Controlled Trial of Exercise to Improveood After Traumatic Brain Injury

eanne M. Hoffman, PhD, Kathleen R. Bell, MD, Janet M. Powell, PhD,ames Behr, MD, Erin C. Dunn, PhD, Sureyya Dikmen, PhD,

harles H. Bombardier, PhD

Jc1AjDt

KUD

JcD

JbD

EPD

SaiD

CcD

DC

bjective: To test the hypothesis that a structured aerobic exercise regimen wouldecrease the severity of depressive symptoms in people with traumatic brain injury (TBI)ho reported at least mild depression severity at baseline.esign: Prospective, randomized, controlled trial.etting: Community gymnasium.articipants: Subjects with a history of a prior TBI (6 months to 5 years post-injury),ecruited from the community. Inclusion criteria included scoring �5 on the Patient Healthuestionnaire-9. Subjects were excluded if they were non-English speakers, had a medical

ondition precluding exercise, had suicidal ideation, regularly exercised, or could not usetandard aerobic exercise equipment.ntervention: Weekly supervised exercise sessions over a 10-week period consisted ofducation, warm-up, 30 minutes of aerobic exercise, and cool down. The exercise intensityas adjusted to reach a heart rate goal of 60% of the participant’s estimated maximal heart

ate.ain Outcome Measurement: Beck Depression Inventory (BDI) comparing exercise

o control groups. Post hoc analyses compared groups exercising �90 minutes or �90inutes per week.esults: Between-group comparisons at 10 weeks revealed no difference between groupsn the BDI (P � .250). For the groups divided by minutes exercised per week, theigh-activity group had significantly better depression scores than those in the low-activityroup (P � .033).onclusions: Although there was no statistically significant difference between the

reated and the control group on mood after intervention, those persons with TBI whoecounted higher levels of exercise per week also reported less depression and improvedleep, community participation, and overall quality of life.

PM R 2010;2:911-919

NTRODUCTION

raumatic brain injury (TBI) is a major cause of disability in the United States and representsserious public health issue with almost 1.5 million people sustaining nonfatal injuries eachear [1, 2]. This figure may underestimate the true incidence because many cases of mild TBIre never directed to hospitals or even evaluated in emergency departments. Emotionalistress and depression commonly occur after a brain injury [3]. Estimates of the rate ofepression after TBI vary substantially ranging from 6% to 77% [4-6]. In a recent rigoroustudy of depression after TBI, Bombardier and colleagues [7] examined mood disorders in59 TBI patients admitted to a trauma hospital. Subjects were assessed monthly for 6onths, then at 8, 10, and 12 months after TBI. Fifty-three percent of the sample metiagnostic and Statistical Manual of Mental Disorders (4th edition) criteria for majorepression at some point during the first year after TBI. Depression is not a transienthenomenon for those with TBI. The risk for depression remains elevated for decades afterhe TBI [8, 9]. Mild TBI is associated with depression at comparably high levels to more

evere TBI [10].

Sa

M&R © 2010 by the American Academy of Physical Me934-1482/10/$36.00

rinted in U.S.A. D

.M.H. Department of Rehabilitation Medi-ine, University of Washington, Box 356490,959 NE Pacific Street, Seattle, WA 98195.ddress correspondence to: J.M.H.; e-mail:

eanneh@uw.eduisclosure: 8B, NIDRR, Department of Educa-

ion grant #H133A070032

.R.B. Department of Rehabilitation Medicine,niversity of Washington, Seattle, WAisclosure: 8B, NIDRR grant

.M.P. Department of Rehabilitation Medi-ine, University of Washington, Seattle, WAisclosure: 8A

.B. Orthopedic Associates, P.A., Spartan-urg, SCisclosure: nothing to disclose

.C.D. St. Paul’s Hospital Eating Disordersrogram, Vancouver, BC, Canadaisclosure: nothing to disclose

.D. Departments of Rehabilitation Medicinend Neurological Surgery, University of Wash-

ngton, Seattle, WAisclosure: nothing to disclose

.H.B. Department of Rehabilitation Medi-ine, University of Washington, Seattle, WAisclosure: 8B, NIDRR Aging RRTC

isclosure Key can be found on the Table ofontents and at www.pmrjournal.org

ubmitted for publication March 5, 2010;ccepted June 12, 2010.

dicine and RehabilitationVol. 2, 911-919, October 2010

OI: 10.1016/j.pmrj.2010.06.008911

ipuatamassttc

lemfpnamrsaa[swtdslpwm

ptsedgwosrlcesve

Gett8ihtddsgeacn

alaaradapoa

apmtsa

M

Ttic

S

Ptwfitbws

912 Hoffman et al EXERCISE TO IMPROVE MOOD AFTER BRAIN INJURY

Identification and treatment of depression after TBI aremportant because untreated depression may negatively im-act functional recovery from TBI and can exacerbate resid-al cognitive and behavioral difficulties. People with TBI whore also depressed demonstrate worse subjective and objec-ive cognitive functioning [11], increased aggressive behaviornd anxiety [6,12], greater functional disability [12-14], andore somatic symptoms compared to people with TBI who

re not depressed [12, 15]. Additionally, depression in per-ons after TBI is associated with significantly higher rates ofuicidal plans [16], attempts [17], and completed suicidehan in the general population and non–brain-injured con-rols [18]. This may be attributed in part to executive dys-ontrol and impulsive decision-making in this population.

The treatment of depression in non–brain-injured popu-ations has been studied extensively and there are numerousffective treatment modalities available, with antidepressantedications and cognitive behavioral therapy used most

requently [19-21]. Unfortunately, to date, there are fewublished studies on the treatment of depression in TBI ando adequately powered, randomized, placebo-controlled tri-ls [22]. Open label trials of sertraline and citalopram forajor depression after TBI showed that 28%-87% of subjects

esponded positively to treatment [23, 24]. Treatment re-ulted in significant improvement in psychological distress,nger, general functioning and postconcussive symptoms,nd performance on standardized neuropsychological tests23, 25]. A small study comparing methylphenidate andertraline reported improvement in symptoms of depressionith both drugs to a comparable degree [26]. A recent study

hat combined 52 people with mild to severe TBI and severeepression found no differences between groups treated withertraline and placebo for depression, anxiety, or quality ofife [27]. Therefore, to date, no clear therapeutic effect can beredicted for specific antidepressant treatment for personsith TBI. There are no studies using psychotherapy or otherodalities of treatment.Although more studies of antidepressant medications and

sychotherapy are needed, we chose to study exercise as areatment for depression in people with TBI for several rea-ons. First, 2 well-controlled studies have demonstrated thatxercise is an effective treatment for major depression inepressed adults. In one study, 45 minutes of supervisedroup aerobic conditioning 3 times per week for 16 weeksas found to be as effective as sertraline among depressedlder adults [28]. At 10 months after the completion of thetudy, exercisers were significantly less likely to report aelapse of major depression compared to those taking sertra-ine. In a second randomized controlled trial, Dunn andolleagues demonstrated that individually supervised mod-rate intensity aerobic exercise resulted in lower depressioneverity scores compared to a group that received no inter-ention and a group that engaged in only low-intensity

xercise [29]. (For a complete review of this literature, see i

reer [30] and Mead [31].) Second, we anticipated thatxercise would be a highly acceptable form of depressionreatment for TBI survivors. Recent research on depressionreatment preferences among those with TBI indicates that9% of the depressed subgroup would be likely to participate

n exercise treatment [32]. Exercise as a treatment modalityas already been demonstrated to improve associated condi-ions including anxiety and somatic complaints such as sleepifficulties, fatigue, and pain [33, 34]. A study of poststrokeepressive symptoms and exercise resulted in fewer depres-ive symptoms and improved quality of life in the exerciseroup [35]. The only study examining the effects of aerobicxercise on persons with TBI found no differences in anxietynd depression despite significant improvement in exerciseapacity for the exercise group. However, these subjects didot have depression on entry to the study [36].

Additionally, evidence is accumulating that exercise isssociated with improved cognitive function in adult popu-ations [37-39], in community-dwelling elderly people [40],nd adults with Alzheimer disease [41]. Furthermore, innimals, aerobic exercise is associated with changes in neu-otrophic factors such as brain-derived neurotrophic factornd intracellular signaling proteins such as kinases that un-erlie cognitive processes [37, 42, 43]. Although we are notware of any studies comparing pharmacotherapy alone withharmacotherapy combined with exercise for the treatmentf depression, it seems reasonable to expect there might be andditional benefit with no added risk.

This study evaluated our hypothesis that a structurederobic exercise regimen would decrease the severity of de-ressive symptoms in people with TBI who reported at leastild depression severity at baseline. We also hypothesized

hat exercise would have the effect of reducing other negativeymptoms associated with TBI such as anxiety, insomnia,nd pain.

ETHODS

he study was a 2-group randomized, prospective clinicalrial. The experimental group received a 10-week exercisentervention, and both the experimental and control groupsompleted baseline and 10-week assessments.

ubjects

articipants were recruited through posted and online adver-isements in local rehabilitation clinics, newspapers, andebsites. Local rehabilitation physician and psychologist of-ces were given information and flyers for the study. Presen-ations were made to local support groups for people withrain injuries. In addition, participants in previous studiesho had completed those studies and met criteria for this

tudy were given information at their final contact if they had

ndicated interest in future studies at the time of initial

cpiri

tibHdmsEwwepmsodmhairstt

R

Pctccettrrmt

P

Epwteo

stsmomstwBlhhtrTTpoI(lsan1n

iiepatHe

Hpestaiatclt

tihb

913PM&R Vol. 2, Iss. 10, 2010

onsent. Initial telephone contact with interested partici-ants included a description of the study and a review of

nclusion and exclusion criteria. After participants were en-olled and randomized, transportation support was providedf necessary.

Inclusion criteria were: 1) self-reported TBI severe enougho have required medical evaluation or hospital admissionmmediately after injury, 2) TBI from 6 months to 5 yearsefore enrollment, 3) score of 5 or greater on the Patientealth Questionnaire-9 [44] indicating at least a mild level ofepressive symptoms, and 4) sufficient cognitive ability toaintain an exercise log and independently participate in the

tudy or have an involved support person for assistance.xclusion criteria were: 1) having a medical condition thatould limit or preclude exercise, 2) current suicidal ideationith intent or plan, 3) current pregnancy, 4) current regular

xercise program 3 times per week or more, and 5) anyhysical barrier to the use of standard aerobic exercise equip-ent. An additional exclusion was being a non-English

peaker. Most of the measures used in this study do not havefficial translations. Also, because this geographic regionoes not have a single predominant ethnic minority, havingultiple interpreters present for exercise sessions wouldave been prohibitively costly. If the study coordinator hadny concerns about the potential participant’s physical abil-ties, the study physician (K.B.) reviewed available pertinentecords and examined the patient. After this screening, theubject’s informed consent was obtained with the approval ofhe Human Subjects Division of the University of Washing-on.

andomization

articipants were randomized into the exercise or controlondition using sealed envelopes prepared before the start ofhe study. Those individuals who were randomized into theontrol group were given instruction that they would beontacted for assessment in 10 weeks and that they would beligible to participate in the exercise program. Individuals inhe control group were offered this delayed participation inhe supervised exercise program to encourage maximal en-ollment by offering them the same intervention as thoseandomized into the treatment group. No specific recom-endations regarding abstention from exercise were given to

hose individuals in the control group.

rocedures

xercise Intervention. Supervised exercise training tooklace once per week at a local community college gymnasiumith a research educational trainer and a certified athletic

rainer. At each session, participants received 15 minutes ofducation on an exercise-related topic. This included a hand-

ut on relevant topics and a review of their home exercise log b

heets with attention to overcoming barriers to exercise. Theyhen participated in 15 minutes of warm-up exercises con-isting of stretching and walking. This was followed by 30inutes of aerobic exercise. Subjects could choose the mode

f aerobic exercise for each session from the available equip-ent including treadmill, stair-stepper, rowing machine,

tationary bicycle, or an indoor track. Participants were trainedo monitor their own heart rate during the exercise sessions, asell as during their home exercise program, and were taught theorg Scale of Perceived Exertion [45] using a target intensityevel of 12. The exercise intensity was adjusted to reach aeart rate goal of 60% of the participant’s estimated maximaleart rate (220 - age). This heart rate goal is designed to havehe participant exercise at about 50% of aerobic capacity andeach a rating of perceived exertion of 12 on a 20-point scale.his corresponds to moderate level of exercise intensity [46].his exercise level was chosen for several reasons. First,revious literature has found benefits of an exercise intensityf 60% of maximal heart rate in decreasing depression [47].n addition, the American College of Sports MedicineACSM) guidelines indicate that, because individuals with aow level of fitness will respond to a low-exercise intensityuch as 40% to 60% of aerobic capacity beginning exercise atlower aerobic level may improve exercise adherence. Fi-ally, each exercise session concluded with a cool-down of5 minutes of stretching, slow walking, and planning for theext week of home exercise activities.

Educational topics covered during the exercise sessionsncluded general benefits of exercise, clothing and footwear,njury prevention and safety, motivation and barriers toxercise, nutrition and hydration, common physical com-laints during exercise, exercising in the community, pacingnd cross-training, and relapse prevention. There was alsoime reserved for questions and planning for home exercise.andouts and information about community resources for

xercises were distributed at each session.

ome Program. As part of the exercise program, eacharticipant was asked to perform 30 minutes of aerobicxercise at least 4 times during the week in addition to theupervised exercise session. Participants could choose anyype of aerobic exercise as long as they met their heart ratend perceived exertion goal. Each participant was instructedn how to use a paper exercise log that was reviewed weeklyt the beginning of each supervised session with one of therainers to assess for compliance with recommended out-of-lass exercise. If participants did not meet the recommendedevel of exercise, the trainer would work with each individualo identify and resolve barriers to regular exercise.

We used principles of social cognitive theory [48] and theranstheoretical stages of change model [49] to guide ourntervention to generalize the exercise intervention to theome environment as well as strategies to maintain exerciseeyond the 10-week trial. These principles and models have

een used in widely cited successful studies of exercise pro-

mmeeet

M

Bmaesm

dcPI[

C

SLpwiafptm

D

Dwupawicu

R

Eagraat4gib(ose

Ta

Fd

T

AMWSHW

N

914 Hoffman et al EXERCISE TO IMPROVE MOOD AFTER BRAIN INJURY

otion in the general medical literature [50, 51]. Knownediators of successful behavior change such as self-efficacy,

njoyment of exercise, highlighting perceived benefits ofxercise, barrier reduction, and building social support forxercise were systematically addressed by the educationalrainer at the beginning and end of each exercise session.

easures

asic demographics were collected including age, gender,arital status, current living situation, and work status. We

ssessed each participant at baseline and at 10 weeks afternrollment into the study. Our primary outcome was thecore on the Beck Depression Inventory (BDI), a widely usedeasure of depression severity [52].We also assessed physical symptoms often associated with

epression and general health status. These measures in-luded an analog assessment of pain and fatigue, the Briefain Inventory [53], Pittsburgh Sleep Inventory [54], Headnjury Symptom checklist [55], and SF-12 Health Survey56].

Assessment of activity and participation included the

igure 1. Consolidated Standards of Reporting Trials (CONSORT)iagram of patient flow through the study.

raig Handicap Assessment and Reporting Technique – g

hort Form (CHART-SF) [57], and the Perceived Quality ofife (PQOL) [58]. The CHART-SF is a measure of the lack ofarticipation in societal roles (ie, limitations and restrictions,hich occur over time as a result of impairments and disabil-

ties). The PQOL measures satisfaction with 12 differentspects of life including an individual’s contact with family orriends, contribution to the community, meaning and pur-ose in life, and happiness. In addition, a 6-minute walkingest was also obtained at baseline and at 10 weeks as aeasure of physical conditioning.

ata Analysis

emographic data from the control and experimental groupsere compared to evaluate the adequacy of randomizationsing t-tests or �2 tests depending on the variable. Therimary analysis examined differences on all outcome vari-bles between the experimental and control groups at 10eeks and was evaluated on an intent-to-treat basis, adjust-

ng for baseline scores. Analysis of covariance was used for allontinuous outcome measures and Mann-Whitney tests weresed for nonparametric measures.

ESULTS

ighty-four participants were enrolled, completed baselinessessment, and were randomized to either the experimentalroup or wait-control group. Of the 84 participants who wereandomized, 4 (2 from each group) were excluded fromnalyses because they were found to have been exercising atlevel that exceeded our exclusion criteria before the start of

he study (Figure 1). We were unable to contact or schedulesubjects for the 10-week assessment (3 from the treatment

roup and one from the control group). A total of 76 partic-pants completed their 10-week assessment. The mean num-er of sessions attended by the exercise group was 5.88minimum � 0, maximum � 10, mode � 9). More than halff the exercise group attended 7 or more of the supervisedessions. Four individuals attended none of the sessions afternrollment.

Table 1 includes baseline demographics for the 2 groups.he majority of the sample was female, white, never married,nd had completed high school. Tests of balance between the

able 1. Characteristics of control and treatment groups

Control Group(n � 40)

Treatment Group(n � 40)

ge (y) 37.1 39.7ale 50% 38%hite 80% 68%

ingle, never married 68% 50%igh school or above 85% 95%orking 35% 38%

ote: No significant differences were found between characteristics of the 2

roups.

2rdct

s2idtw.

pbio(tcss

pcstmhv

pelamogetthepasasa

D

Wrdsgwcep

T

BEECWAPAFBC

TT

PPSSW

B

915PM&R Vol. 2, Iss. 10, 2010

groups on demographic and study variables suggested thatandomization was effective. Only one variable was found toiffer between groups at baseline. Individuals in the exerciseondition reported significantly higher PQOL compared tohose in the control group at baseline.

Both control and exercise condition groups achieved sub-tantial increases in total minutes per week of exercise (Table). There was no difference between groups in the total

ncreased minutes of exercise per week (P � .064). However,espite no difference in minutes exercised per week, those inhe exercise group did report exercising on more days pereek than the control group (3.68 days versus 2.05, P �

004).After adjusting for baseline values, between-group com-

arisons at the 10-week assessment revealed no differenceetween groups on the BDI, our primary outcome variable

ndicating severity of depression (P � .250). For secondaryutcomes, the exercise group reported less pain interferenceP � .021) and greater improvement on the Brief Pain Inven-ory (P � .030) compared to the control group. No signifi-ant between-groups differences were found for head injuryymptoms, perceived quality of life, sleep, general healthtatus, heart rate, or ability to walk.

Because the intervention did not achieve the goal of im-roving minutes of exercise more in the intervention versusontrol group, we wanted to examine the theoretical relation-hip between improved exercise and improved depressionhrough post-hoc analyses. Specifically, we wanted to deter-ine whether those who approached a recommended publicealth dose of exercise (at least 90 minutes of moderate or

able 2. Comparison of outcomes based on group assignme

Measure

Co

Baseline

eck Depression Inventory 24.7xercise recall Total minutes per week 58xercise recall No. days per week 1.47urrent pain 3.33orst pain 5.98verage pain 4.33ain interference 3.58verage fatigue 5.65atigue interference 5.45rief Pain Inventory 3.54raig Handicap Assessment and ReportingTechnique – Short Form

84.5

raumatic brain injury symptom checklist (severity) 2.55raumatic brain injury symptom checklist Total

number of symptoms11.4

erceived quality of Life 45ittsburgh Sleep 10.6hort Form-12 Physical 41.4hort Form-12 Mental 28.2alking test 1272

old values indicate significance.

igorous physical activity per week) improved in terms of our t

rimary and secondary outcomes. Therefore, we divided thentire sample into 2 groups, those who reported exercising ateast 90 minutes (“high active,” n � 23 at baseline and n � 43t outcome) and those who reported exercising less than 90inutes (“low active,” n � 57 at baseline and n � 33 at

utcome), regardless of initial group assignment. When theroups were compared at baseline, they were equivalentxcept that the SF-12 Mental subscore was significantly bet-er in the high active group (P � .029; Table 3). Not all ofhose who started in the active group at baseline remainedigh active at outcome. In fact, 7 of the 23 reduced theirxercise level below 90 minutes per week. Analysis of theost-treatment data comparing groups showed that the highctive group at outcome had significantly lower depressioncores than those in the low activity group (P � .033). Inddition, those subjects in the high active group reportedignificantly more community activity, better quality of life,nd better general mental health.

ISCUSSION

e hypothesized that an aerobic exercise program wouldesult in improved mood for those with TBI with moodisorder before the study intervention. We did not find aignificant difference between the exercise and controlroups on our primary outcome, the BDI, at the end of 10eeks even though the intervention group did report exer-

ising more days per week and displayed a trend towardxercising more minutes per week. The exercise group diderceive a decrease in fatigue and impact of pain as compared

Exercise

Statistical Test Significanceeek Baseline 10-Week

1.2 21.5 16.4 Regression .2503 66 252 Mann-Whitney .0642.05 1.28 3.68 Regression .0043.08 3.35 2.73 Mann-Whitney .7175.44 5.33 4.65 Mann-Whitney .9664.10 3.68 3.14 Mann-Whitney .1943.79 4.13 3.24 Mann-Whitney .0215.74 5.75 5.16 Mann-Whitney .3125.47 5.63 4.89 Mann-Whitney .1843.51 3.82 3.12 Mann-Whitney .0303.3 85.7 85 Mann-Whitney .489

2.41 2.44 2.41 Regression .4311.4 11.0 11.8 Mann-Whitney .682

9 54 58 Regression .3880.9 10.0 9 Regression .1069.5 41.6 42 Regression .2242.5 31.8 38.3 Regression .2386 1343 1491 Regression .566

nt

ntrol

10-W

214

8

1

4133

139

o the control group.

gsdbettongrgp

ntettriwDgaptosm

“tfas

wgwtrotiadta

ttTwmppimto

T

BCWAPAFBC

TT

PPSSW

*

916 Hoffman et al EXERCISE TO IMPROVE MOOD AFTER BRAIN INJURY

There were no significant differences between the 2roups at baseline except for reports of higher quality of lifecores for those randomized to the exercise group. It isoubtful that this affected the outcome of the study becauseoth groups improved equally on the PQOL measure. Inter-stingly, we had more female participants and higher educa-ional levels in this study than we would have predicted onhe basis of TBI statistics. A wide range of recruitment meth-ds was used for this study including advertisements inewsletters (paper and electronic), web postings, supportroup presentations, physician/clinic letters, and flyers toeduce the possibility of bias [59]. It is difficult to know howeneralizable our results are to more representative groups ofersons with TBI.

We speculate that at least 3 factors contributed to ouregative findings and should be carefully considered in fu-ure studies. First, despite not participating in the supervisedxercise program, many of the control subjects increasedheir exercise duration substantially during the 10 weeks ofhe study, matching the increase by the exercise group. Byandomizing only those who were interested in participatingn an exercise program, we may have recruited people whoere highly motivated to begin exercising on their own.espite the possibility of future participation in the exerciseroup in 10 weeks, it appears that many individuals werelready “ready for exercise.” High initial readiness to exerciseredicts greater increases in physical activity during interven-ion trials [60]. To enroll participants, we had to inform themf the exercise intervention so we were unable to blindubjects to their group assignment. Future studies should

able 3. Comparison of outcomes by groups sorted by minut

Measure

<90Minutesn � 57

eck Depression Inventory 24.3urrent pain 3.2orst pain 5.6verage pain 3.9ain interference 3.5verage fatigue 5.8atigue interference 5.5rief Pain Inventory 3.4raig Handicap Assessment and ReportingTechnique – Short Form

84.6

raumatic brain injury symptom checklist (severity) 2.5raumatic brain injury symptom checklist (total

number of symptoms)11.3

erceived quality of life 48.1ittsburgh Sleep 10.4hort Form-12 Physical 43.1hort Form-12 Mental 28.1alking test 1319

MWU � Mann-Whitney U test.

easure initial readiness to exercise and closely monitor the c

no exercise” condition. It is possible that having an alterna-ive “no exercise” intervention (eg, support group sessions)or the control group during the 10 weeks between baselinend outcome may have better controlled this aspect of thetudy.

Next, the intervention required participants to attendeekly supervised training sessions held at a communityymnasium on Sunday afternoons. Although taxi vouchersere given to subjects when needed to overcome transporta-

ion barriers, this aspect of the intervention may have limitedecruitment and participation by a more representative groupf people with TBI. Ideally, studies should provide alterna-ive places and times to engage in supervised exercise butn-person supervision in this scenario would be difficult toccomplish. Another option would be not to require atten-ance in a supervised exercise program, but instead, to useelephone or other electronic monitoring to facilitate exercisedoption in the person’s home environment.

Third, we included people with mild depressive symp-oms. Those with mild depressive symptoms have less roomo improve and may be more likely to remit spontaneously.rials of antidepressant medications routinely exclude thoseith less severe depression because less severe depression isore likely to remit spontaneously [21, 61]. In addition,rior research in non-TBI populations has suggested that theositive effect of single exercise sessions on mood was signif-

cantly greater in those with more symptoms of depressedood [62]. Future studies should consider including only

hose with more severe depression or major depressive dis-rder. The problem of high rates of spontaneous remission

rcised

aseline Outcome

0tes

t-testP

Value*

<90Minutes

>90Minutes

t-testP

Value*23 n � 33 n � 43

0.2 .089 .157 25.8 20.7 .023 .0333.7 .424 .382 3.7 3.0 .248 .2895.8 .836 .748 5.7 5.6 .964 .9414.2 .741 .760 4.3 3.6 .273 .2574.6 .192 .253 4.2 3.6 .399 .3655.4 .600 .720 5.8 5.5 .624 .6925.6 .959 .747 5.9 5.2 .303 .2714.3 .195 .203 3.9 3.5 .524 .5606.2 .592 .651 82.1 86.7 .100 .028

2.5 .834 .936 2.5 2.4 .747 .7101.0 .783 .666 11.2 11.0 .901 .992

2.5 .367 .341 43.5 53.4 .028 .0340.2 .858 .773 11.4 9.2 .030 .0237.5 .086 .119 42.8 41.6 .692 .6754.7 .022 .029 26.0 32.6 .015 .0247 .682 .885 1217 1376 .089 .036

es exe

B

>9Minun �

2

8

1

5133

127

ould also be countered by requiring potential participants to

dp

cmgpa9tc

iseSdsaDts

imtgtbstps3t

tnslta[witwpreca

a

wrtftsamiawa

crsacop

C

BmmtaFts

A

Tntaf

R

917PM&R Vol. 2, Iss. 10, 2010

emonstrate persistent depression, for example, using a re-eated screening evaluation 2 weeks apart.

Our post-hoc analyses suggested that mood was signifi-antly higher for study participants who were exercisingore than 90 minutes per week at outcome, regardless of

roup assignment. These exercising subjects reported im-roved community participation, improved sleep, and over-ll better quality of life than did subjects exercising less than0 minutes per week. However, because we did not designhe study to evaluate the effect of more or less exercise, weannot say that a cause-effect relationship was established.

These results offer support for an association betweenncreased physical activation and improved mood for per-ons with TBI. However, it appears that a certain threshold ofxercise, in terms of duration, is required for this effect.imply performing a low level of physical activity over moreays was not associated with improved mood. There is auggestion of a dose-response relationship between exercisend depression for non-TBI populations [63]. In the study byunn and colleagues, a dose-dependent relationship be-

ween exercise and improvement of depression was demon-trated [29].

The most recent guidelines from the ACSM for maintain-ng physical health include moderate level of exercise for 30

inutes 5 times weekly or 20 minutes of vigorous exercise 3imes weekly as a minimum [64]. We chose to separate theroups into those who exercised less than 90 minutes versushose who exercised 90 minutes or more weekly rather thany the ACSM criteria to allow for the fact that some of ourubjects had physical challenges and needed some modifica-ions to duration of exercise. However, because subjects whoerformed exercise above the 90-minute threshold demon-trated improvements with weight and rate of walking (Table), they likely had a physiological exercise effect in additiono an effect on mood.

Few prior studies on the effect of exercise on mood inhose with TBI exist. A retrospective study by Gordon et aloted that TBI exercisers were less depressed and had fewerymptoms than those who did not exercise, despite similarevels of disability and handicap [33]. Bateman et al reportedhat aerobic exercise after severe TBI resulted in improvederobic capacity, but did not alter fatigue and depression36]. Exercise has relatively few adverse effects comparedith pharmacological treatment of depression and can be

nexpensive. In addition, exercise offers a potential avenueoward increased social activity and leisure participation,hich may also alleviate symptoms of depression. Manyeople with significant TBI have physical challenges that mayequire a change in preferred exercise activity or adaptedquipment. However, we were able to adapt standard exer-ise equipment for use with our patients with physical dis-bilities in this study with simple alterations.

Given the potential benefits of exercise, fortunately, there

re no specific contraindications to aerobic exercise for those

ith TBI. Guidelines on return to play after concussionecommend that return to exercise/sport should occur afterhe athlete is asymptomatic at rest. Light aerobic exercise canollow (walking, swimming, or stationary cycling) with in-ensity at less than 70% MHPR with progression to sports-pecific activities, noncontact play, and eventually full sportsctivities [65]. Generally, for more severe injuries, recom-endations are made to avoid sports with the potential to

ncur another brain injury during the first year after injurynd to wear head protection whenever possible. For thoseith craniotomy, limitations on weight lifting are generally

dvised for 6-12 weeks postoperatively.Another barrier to exercise for those with TBI may be

ognitive impairments, negatively affecting adherence withecommended exercise regimens. In our study, those withignificant memory impairments sometimes chose to includefamily member or friend who acted as a cognitive “aide” forarrying out the home exercise program. Other potentialptions to increase participation might be group exerciserograms or electronic or telephone reminders.

ONCLUSION

ecause of its potential positive effect on cognition as well asood, as well as its attractiveness to people with TBI as aeans of treating depression, further efforts should be made

o investigate the efficacy of exercise and means of fosteringdherence with exercise prescriptions in those with TBI.uture research studies will need to carefully plan to monitorhe control condition to prevent unintended “positive” re-ults for participants in these types of studies.

CKNOWLEDGMENTS

he authors would like to thank our study educators/coordi-ators, Aaron Scrol and Denise Hansen, and the physicalrainers who made this study possible. The authors wouldlso like to acknowledge North Seattle Community Collegeor the use of their outstanding exercise facilities.

EFERENCES1. Langlois J, Rutland-Brown W, Thomas K. Traumatic brain injury in the

United States: emergency department visits, hospitalizations, anddeaths. Atlanta, GA; Centers for Disease Control and Prevention/Na-tional Center for Injury Prevention and Control 2004.

2. Thurman D, Alverson C, Dunn K, Guerrero J, Sniezek J. Traumaticbrain injury in the United States: a public health perspective. J HeadTrauma Rehabil 1999;14:602-615.

3. Rosenthal M, Christensen B, Ross T. Depression following traumaticbrain injury. Arch Phys Med Rehabil 1998;79:90-103.

4. Kreutzer J, Seel R, Gourley E. The prevalence and symptom rates ofdepression after traumatic brain injury: a comprehensive examination.Brain Inj 2001;15:563-576.

5. Jorge R, Robinson R, Arndt S, Starkstein S, Forrester A, Geisler F.Depression following traumatic brain injury: a 1 year longitudinal

study. J Affect Disord 1993;27:233-243.

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

4

4

4

4

4

918 Hoffman et al EXERCISE TO IMPROVE MOOD AFTER BRAIN INJURY

6. Jorge R, Robinson R, Moser D, Tateno A, Crespo-Facorro B, Arndt S.Major depression following traumatic brain injury. Arch Gen Psychiatry2004;61:42-50.

7. Bombardier C, Fann J, Temkin N, Esselman P, Barber J, Dikmen S.Rates of major depressive disorder and clinical outcomes followingtraumatic brain injury. JAMA 2010;303:1938-1945.

8. Koponen S, Taiminen T, Portin R, Himanen L, Isoniemi H, HeinonenH, et al. Axis I and II psychiatric disorders after traumatic brain injury:a 30-year follow-up study. Am J Psychiatry 2002;159:1315-1321.

9. Holsinger T, Steffens D, Phillips C, Helms M, Havlik R, Breitner J, et al.Head injury in early adulthood and the lifetime risk of depression. ArchGen Psychiatry 2002;59:17-22.

0. Dikmen S, Bombardier C, Machamer J, Fann J, Temkin N. Naturalhistory of depression in traumatic brain injury. Arch Phys Med Rehabil2004;85:1457-1464.

1. Chamelian L, Feinstein A. The effect of major depression on subjectiveand objective cognitive deficits in mild to moderate traumatic braininjury. J Neuropsychiatry Clin Neurosci 2006;18:33-38.

2. Fann J, Katon W, Uomoto J, Esselman P. Psychiatric disorders andfunctional disability in outpatients with traumatic brain injuries. Am JPsychiatry 1995;152:1493-1499.

3. Hibbard M, Ashman T, Spielman L, Chun D, Charatz H, Melvin S.Relationship between depression and psychosocial functioning aftertraumatic brain injury. Arch Phys Med Rehabil 2004;85(4 Suppl 2):S43-S53.

4. Rapoport M, McCullagh S, Streiner D, Feinstein A. The clinical signif-icance of major depression following mild traumatic brain injury.Psychosomatics 2003;44:31-37.

5. Rapoport M, McCullagh S, Shammi P, Feinstein A. Cognitive impair-ment associated with major depression following mild and moderatetraumatic brain injury. J Neuropsychiatry Clin Neurosci 2005;17:61-65.

6. Kishi Y, Robinson RG, Kosier JT. Suicidal ideation among patients withacute life-threatening physical illness: patients with stroke, traumaticbrain injury, myocardial infarction, and spinal cord injury. Psychoso-matics 2001;42:382-390.

7. Silver JM, Kramer R, Greenwald S, Weissman M. The associationbetween head injuries and psychiatric disorders: findings from the NewHaven NIMH Epidemiologic Catchment Area Study. Brain Inj 2001;15:935-945.

8. Teasdale TW, Engberg AW. Suicide after traumatic brain injury: apopulation study. J Neurol Neurosurg Psychiatry 2001;71:436-440.

9. Cuijpers P, van Straten A, Andersson G, van Oppen P. Psychotherapyfor depression in adults: a meta-analysis of comparative outcomestudies. J Consult Clin Psychol 2008;76:909-922.

0. Freeman M. Complementary and alternative medicine (CAM): consid-erations for the treatment of major depressive disorder. J Clin Psychi-atry 2009;70(Suppl 5):4-6.

1. Fournier J, DeRubeis R, Hollon S, Dimidjian S, Amsterdam J, Shelton R,et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA 2010;303:47-53.

2. Fann J, Hart T, Schomer K. Treatment for depression after traumaticbrain injury: a systematic review. J Neurotrauma 2009;26:2383-2402.

3. Fann J, Uomoto J, Katon W. Sertraline in the treatment of majordepression following mild traumatic brain injury. J NeuropsychiatryClin Neurosci 2000;12:226-232.

4. Rapoport M, Chan F, Lanctot K, Herrmann N, McCullagh S, FeinsteinA. An open-label study of citalopram for major depression followingtraumatic brain injury. J Psychopharmacol 2008;22:860-864.

5. Fann JR UJ, Katon WJ. Cognitive improvement with treatment ofdepression following mild traumatic brain injury. Psychosomatics2001;42:48-54.

6. Lee H, Kim S, Kim J, Shin I, Yang S, Yoon J. Comparing effects of

methylphenidate, sertraline and placebo on neuropsychiatric sequelae

in patients with traumatic brain injury. Hum Psychopharmacol2005;20:97-104.

7. Ashman T, Cantor J, Gordon W, Spielman L, Flanagan S, Ginsberg A, etal. A randomized controlled trial of sertraline for the treatment ofdepression in persons with traumatic brain injury. Arch Phys MedRehabil 2009;90:733-740.

8. Blumenthal J, Babyak M, Doraiswamy P, Watkins L, Hoffman B,Barbour K, et al. Exercise and pharmacotherapy in the treatment ofmajor depressive disorder. Psychosom Med 2007;69:587-596.

9. Dunn A, Trivedi M, Kampert J, Clark C, Chambliss H. Exercise treat-ment for depression: efficacy and dose response. Am J Prev Med2005;28:1-8.

0. Greer T, Trivedi M. Exercise in the treatment of depression. CurrPsychiatry Rep 2009;11:466-472.

1. Mead G, Morley W, Campbell P, Greig C, McMurdo M, Lawlor D.Exercise for depression. Cochrane Database Syst Rev 2009(3):CD004366.

2. Fann J, Jones A, Dikmen S, Temkin N, Esselman P, Bombardier C.Depression treatment preferences after traumatic brain injury. J HeadTrauma Rehabil 2009;24:272-278.

3. Gordon W, Sliwinski M, Echo J, McLoughlin M, Sheerer M, Meili T.The benefits of exercise in individuals with traumatic brain injury: aretrospective study. J Head Trauma Rehabil 1998;13:58-67.

4. Ströhle A. Physical activity, exercise, depression and anxiety disorders.J Neural Transm 2009;116:777-784.

5. Lai S, Studenski S, Richards L, Perera S, Reker D, Rigler S, et al.Therapeutic exercise and depressive symptoms after stroke. J AmGeriatr Soc 2006;54:240-247.

6. Bateman A, Culpan F, Pickering A, Powell J, Scott O, Greenwood R.The effect of aerobic training on rehabilitation outcomes after recentsevere brain injury: a randomized controlled evaluation. Arch PhysMed Rehabil 2001;82:174-182.

7. Grealy M, Johnson D, Rushton S. Improving cognitive function afterbrain injury: the use of exercise and virtual reality. Arch Phys MedRehabil 1999;80:661-667.

8. Kleim J, Jones T, Schallert T. Motor enrichment and the induction ofplasticity before or after brain injury. Neurochem Res 2003;28:1757-1769.

9. Pontifex M, Hillman C, Fernhall B, Thompson K, Valentini T. The effectof acute aerobic and resistance exercise on working memory. Med SciSports Exerc 2009;41:927-934.

0. Williamson J, Espeland M, Kritchevsky S, Newman A, King A, Pahor M,et al. Changes in cognitive function in a randomized trial of physicalactivity: results of the lifestyle interventions and independence forelders pilot study. J Gerontol A Biol Sci Med Sci 2009;64:688-694.

1. Lautenschlager N, Cox K, Flicker L, Foster J, van Bockxmeer F, Xiao J,et al. Effect of physical activity on cognitive function in older adults atrisk for Alzheimer disease: a randomized trial. JAMA 2008;300:1027-1037.

2. Ferris L, Williams J, Shen C. The effect of acute exercise on serumbrain-derived neurotrophic factor levels and cognitive function. MedSci Sports Exerc 2007;39:728-734.

3. Hu S, Ying Z, Gomez-Pinilla F, Frautschy S. Exercise can increase smallheat shock proteins (sHSP) and pre- and post-synaptic proteins in thehippocampus. Brain Res 2009;1249:191-201.

4. Spitzer RL, Kroenke K, Williams JB. Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study. PrimaryCare Evaluation of Mental Disorders. Patient Health Questionnaire.JAMA 1999;282:1737-1744.

5. Borg G. Borg’s Perceived Exertion and Pain Scales. Champaign, IL:Human Kinetics; 1998.

6. Franklin B, Whaley M, Howley E. ACSM’s Guidelines for ExerciseTesting and Prescription. Philadelphia: Lippincott, Williams &

Wilkins; 2000.

4

4

4

5

5

5

5

5

5

5

5

5

5

6

6

6

6

6

6

Tcgdn

919PM&R Vol. 2, Iss. 10, 2010

7. Martinsen E, Hoffart A, Solberg O. Comparing aerobic with nonaerobicforms of exercise in the treatment of clinical depression: a randomizedtrial. Compr Psychiatry 1989;30:324-331.

8. Bandura A. Social Foundations of Thought and Action. EnglewoodCliffs, NJ: Prentice-Hall; 1986.

9. Prochaska J, DiClemente C, Norcross J. In search of how peoplechange. Am Psychologist 1992;47:1102-1114.

0. King A, Sallis J, Dunn A, Simons-Morton D, Albright C, Cohen S, et al.Overview of the Activity Counseling Trial (ACT) intervention forpromoting physical activity in primary health care settings. ActivityCounseling Trial Research Group. Med Sci Sports Exerc 1998;30:1086-1096.

1. Dunn A, Marcus B, Kampert J, Garcia M, Kohl Hr, Blair S. Comparisonof lifestyle and structured interventions to increase physical activity andcardiorespiratory fitness: a randomized trial. JAMA 1999;281:327-334.

2. Beck A, Ward C, Mendelson M, Mock J, Erbaugh J. An inventory formeasuring depression. Arch Gen Psychiatry 1961;4:561-571.

3. Cleeland C, Ryan K. Pain assessment: global use of the Brief PainInventory. Ann Acad Med 1994;23:129-138.

4. Buysse D, Reynolds C, Monk T, Berman S, Kupfer D. The PittsburghSleep Quality Index: a new instrument for psychiatric practice andresearch. Psychiatry Res 1989;28:193-213.

5. Dikmen S, McLean A, Temkin N. Neuropsychological and psychoso-cial consequences of minor head injury. J Neurol Neurosurg Psychiatry1986;49:1227-1232.

6. Ware JE, Kosinski M, Keller SD. SF-12: How to score the SF-12Physical and Mental Health Summary Scales. 2nd ed. Boston: TheHealth Institute, New England Medical Center; 1995.

7. Hall K, Dijkers M, Whiteneck G, et al. The Craig Handicap Assessmentand Reporting Technique (CHART): metric properties and scoring.

Top Spinal Cord Inj Rehabil 1998;4:16-30.

8. Patrick DL, Danis M, Southerland LI, Hong G. Quality of life followingintensive care. J Gen Intern Med 1988;3:218-223.

9. Bell K, Hammond F, Bickett A, Temkin N, Dikmen S. Participantrecruitment and retention in rehabilitation research. Am J Phys MedRehabil 2008;87:330-338.

0. Steptoe A, Rink E, Kerry S. Psychosocial predictors of changes inphysical activity in overweight sedentary adults following counseling inprimary care. Prev Med 2000;31:183-194.

1. Forsell Y. A three-year follow-up of major depression, dysthymia,minor depression and subsyndromal depression: results from a popu-lation-based study. Depress Anxiety 2007;24:62-65.

2. Lane A, Lovejoy D. The effects of exercise on mood changes: themoderating effect of depressed mood. J Sports Med Phys Fitness2001;41:539-545.

3. Teychenne M, Ball K, Salmon J. Physical activity and likelihood ofdepression in adults: a review. Prev Med 2008;46:397-411.

4. Haskell W, Lee I, Pate R, Powell K, Blair S, Franklin B, et al. Physicalactivity and public health: updated recommendation for adults fromthe American College of Sports Medicine and the American HeartAssociation. Med Sci Sports Exerc 2007;39:1423-1434.

5. McCrory P, Meeuwisse W, Johnston K, Dvorak J, Aubry M, Molloy M,et al. Consensus statement on concussion in sport - the Third Interna-tional Conference on Concussion in Sport held in Zurich, November2008. Phys Sportsmed 2009;37:141-159.

his CME activity is designated for 1.0 AMA PRA Category 1 Credit™ andan be completed online at me.aapmr.org. Log on to www.me.aapmr.org,o to Lifelong Learning (CME) and select Journal-based CME from therop down menu. This activity is FREE to AAPM&R members and $25 foron-members.