depression in heart failure: a meta-analytic review …zunis.org/jacc sota 7-06 to 8-08/depression...

doi:10.1016/j.jacc.2006.06.055 2006;48;1527-1537; originally published online Sep 25, 2006; J. Am. Coll. Cardiol.

J. Mills Thomas Rutledge, Veronica A. Reis, Sarah E. Linke, Barry H. Greenberg, and Paul

Intervention Effects, and Associations With Clinical OutcomesDepression in Heart Failure: A Meta-Analytic Review of Prevalence,

This information is current as of August 9, 2008

http://content.onlinejacc.org/cgi/content/full/48/8/1527located on the World Wide Web at:

The online version of this article, along with updated information and services, is

by on August 9, 2008 content.onlinejacc.orgDownloaded from

http://content.onlinejacc.org/cgi/content/full/48/8/1527

http://content.onlinejacc.org

DAITBS

Odittptdmbwcvcr

cfircr

oLSD

a

Journal of the American College of Cardiology Vol. 48, No. 8, 2006© 2006 by the American College of Cardiology Foundation ISSN 0735-1097/06/$32.00P

epression in Heart FailureMeta-Analytic Review of Prevalence,

ntervention Effects, and Associations With Clinical Outcomeshomas Rutledge, PHD,*† Veronica A. Reis, BSC,*‡ Sarah E. Linke, BA,§arry H. Greenberg, MD, FACC,† Paul J. Mills, PHD†an Diego and Los Angeles, California

This article describes a meta-analysis of published associations between depression and heartfailure (HF) in regard to 3 questions: 1) What is the prevalence of depression among patientswith HF? 2) What is the magnitude of the relationship between depression and clinicaloutcomes in the HF population? 3) What is the evidence for treatment effectiveness inreducing depression in HF patients? Key word searches of the Medline and PsycInfodatabases, as well as reference searches in published HF and depression articles, identified 36publications meeting our criteria. Clinically significant depression was present in 21.5% of HFpatients, and varied by the use of questionnaires versus diagnostic interview (33.6% and19.3%, respectively) and New York Heart Association–defined HF severity (11% in class I vs.42% in class IV), among other factors. Combined results suggested higher rates of death andsecondary events (risk ratio � 2.1, 95% confidence interval 1.7 to 2.6), trends towardincreased health care use, and higher rates of hospitalization and emergency room visitsamong depressed patients. Treatment studies generally relied on small samples, but alsosuggested depression symptom reductions from a variety of interventions. In sum, clinicallysignificant depression is present in at least 1 in 5 patients with HF; however, depression ratescan be much higher among patients screened with questionnaires or with more advanced HF.The relationship between depression and poorer HF outcomes is consistent and strong acrossmultiple end points. These findings reinforce the importance of psychosocial research in HFpopulations and identify a number of areas for future study. (J Am Coll Cardiol 2006;48:

ublished by Elsevier Inc. doi:10.1016/j.jacc.2006.06.055

1527–37) © 2006 by the American College of Cardiology Foundation

et(r

inses1ipaodomWi

M

Aairw

ver the past 2 decades, associations between clinicalepression and cardiovascular disease risk have become an

ncreasingly common, if not always consistent, finding inhe cardiovascular literature (1–4). Meta-analytic reviews ofhe depression relationship with heart disease outcomesrovide strong evidence for prospective connections be-ween depression and the incidence of coronary arteryisease (CAD) (5,6), between depression and prematureortality among patients with documented CAD (7,8), and

etween depression and all-cause mortality in populationsith and without CAD (9). However, despite these asso-

iations, there has yet to be a successful randomized inter-ention trial for depression affecting objective clinical out-omes (10), raising important questions about the biologicalole of depression in CAD.

The surge of research interest in depression and CAD hasarried over into related cardiovascular diseases. Heartailure (HF), a condition affecting nearly 5 million patientsn the U.S. alone, has become a major focus of depressionesearch in recent years, with a growing number of publi-ations suggesting poorer clinical outcomes for HF patientseporting symptoms of depression (11–14). Research inter-

From the *VA San Diego Health Care System, San Diego, California; †Universityf California, San Diego, San Diego, California; ‡University of Southern California,os Angeles, California; and §San Diego State University/University of California,an Diego Joint Doctoral Program in Clinical Psychology, San Diego, California.avid S. Sheps, MD, MSPH, acted as the guest editor for this article.

dManuscript received March 6, 2006; revised manuscript received May 10, 2006,

ccepted June 16, 2006.

content.onlinejaDownloaded from

st in the psychosocial dimensions of HF is reinforced byhe enormous health care costs associated with the condition15,16), as well as by the high rates of clinical depressioneported among patients with HF in numerous studies (17,18).

Although recent literature reviews of depression and HFdentify important themes (19,20), quantitative methods areecessary to precisely define the magnitude of the relation-hip between clinical depression and HF. To our knowl-dge, this article is the first meta-analytic review of depres-ion and HF. The focus was guided by 3 primary questions:) What is the prevalence of clinically significant depressionn HF? To what extent is the heterogeneity in reportedrevalence estimates explained by differences in depressionssessment methods, HF severity, age, gender composition,r other demographic characteristics? 2) What is the evi-ence for longitudinal associations between depression andbjective clinical outcomes in HF, including HF incidence,ortality and cardiovascular events, and hospitalization? 3)hat are the effects of pharmacologic and nonpharmacologic

nterventional efforts on depression among patients with HF?

ETHODS

rticle selection and literature search. Using the Medlinend PsycInfo databases, two of the authors (V.R. and S.L.)ndependently identified relevant articles published in peer-eviewed journals by September 12, 2005. Primary keyords included depression and HF, congestive heart failure,
epressive disorder, depressive symptoms, quality of life,
by on August 9, 2008 cc.org


ptqatciu2aidtipspsapssrtaDitS(SNDfIMiSSsISHSCpDm

pasoDirttpc

dsaiisadqrptcvfidt(dsttvecracli

bseives

sm

1528 Rutledge et al. JACC Vol. 48, No. 8, 2006Depression and Heart Failure October 17, 2006:1527–37

sychosocial factors, stress, cardiomyopathy, emotional fac-ors, psychological distress, and mental health. Subse-uently, these abstracts were reviewed for inclusion based onssessments of the published article. We identified addi-ional eligible articles using references from the articlesollected through the database search. Criteria for selectionncluded the following: 1) reporting a rate of depressionsing either clinical interview or a validated questionnaire;) describing prospective relationships between depressionnd mortality, secondary cardiovascular events, rehospital-zation, or health care costs; and 3) documenting changes inepression, measured before and after treatment, attributedo an intervention. In cases in which articles containednsufficient statistics, we attempted to contact the study’srimary investigators to provide additional information. Alltudies required the use of a sample composed exclusively ofatients diagnosed with HF, or the reporting of statisticspecifically for a patient subgroup with HF. Using theforementioned methods, we identified a total of 36 inde-endent articles. These include 27 articles reporting depres-ion prevalence information, 14 describing prospective as-ociations between depression and HF outcomes, and 6eporting depression changes from treatment (some con-ained multiple association types, accounting for the non-dditive article total).

epression measures. Depression diagnostic methods fellnto 3 primary categories: 1) clinical interviews, includinghe Structured Clinical Interview for the Diagnostic andtatistical Manual of Mental Disorders, Fourth Edition21), Structured Clinical Interview for the Diagnostic andtatistical Manual of Mental Disorders, Fourth Edition–onpatient Edition (22), Diagnostic Interview Schedule forepression (23), Modified Diagnostic Interview Schedule

or Depression (24), Composite International Diagnosticnterview Short Form (25), and Primary Care Evaluation of

ental Disorders (26); 2) depression symptom inventories,ncluding the Beck Depression Inventory (BDI) (27), Zungelf-Rating Depression Scale (28), Geriatric Depressioncale (29), Center for Epidemiological Studies–Depressioncale (30), Hospital Anxiety and Depression Scale (31),nventory to Diagnose Depression (32), Hamilton Ratingcale for Depression (33), the depression scale from theopkins Symptom Checklist (34), Medical Outcomes

tudy–Depression (35), and the Multiple Affect Adjectivehecklist (36); and 3) a diagnosis of depression based on theatient’s medical record (e.g., International Classification ofiseases, 9th Edition code) or use of antidepressant

Abbreviations and AcronymsBDI � Beck Depression InventoryCAD � coronary artery diseaseHF � heart failureIL � interleukinNYHA � New York Heart Association

edication. Ccontent.onlinejaDownloaded from

In several cases, investigators used more than 1 of thereviously mentioned criteria. For example, a number ofrticles used questionnaires as screening tools, and higher-coring patients were subsequently interviewed for evidencef a major depressive disorder.atabase construction and coding. Because of the relative

ndependence of our review questions, we developed sepa-ate databases for the prevalence, clinical outcome, andreatment effects themes. Some studies contributed to morehan 1 database (e.g., by reporting baseline depressionrevalence in addition to a prospective or interventionomponent).

From studies describing depression prevalence, we codedepression rates by HF severity, gender, minority status,tudy location, depression assessment method, mean age,nd inpatient/outpatient status, as available. Several studiesncluded sample information regarding HF subtypes (e.g.,schemic vs. nonischemic HF), however, there were notufficient data from which to make depression comparisonscross different forms of HF. The coding scheme definedepression assessment methods according to the use ofuestionnaires, clinical interviews, or patient medicalecords/patient use of antidepressants. We also categorizedrevalence rates from individual studies as either “conserva-ive” or “liberal” depending on the rigor of their criteria forlassifying participants as depressed. To qualify as conser-ative, a prevalence rate had to meet one or more of theollowing criteria: included an interview component, exam-ned patients’ medical records for evidence of a formaliagnosis of depression, or used a questionnaire with a cutoffhat explicitly screened for moderate to severe depressionrepresentative of a DSM diagnosis) as opposed to mildepression or depressive symptoms. An example of a con-ervative versus a liberal cutoff is a score of 17 versus 10 onhe BDI. All other rates were classified as liberal: most ofhem were determined from using a lower cutoff on aalidated questionnaire, whereas 1 rate (15) was based onvidence of an antidepressant prescription without an ac-ompanying diagnosis of depression in the patients’ medicalecords, and therefore, could have included patients takingntidepressants for other reasons (e.g., sleep or smokingessation). Studies that used 2 cutoffs (1 conservative and 1iberal) and reported separate rates for each were representedn both categories.

The categorization of treatment studies included sortingy intervention type and study duration. Because of themall number of available treatment studies, we did notxclude studies based on qualitative factors. All of thedentified treatment studies assessed depressive symptomsia questionnaires before and after treatment; treatmentffects were determined, in part, by comparing changes incores between these 2 time periods.

From the longitudinal outcome studies, we categorizedtudies by duration, study method, level of covariate adjust-ent, and type of clinical outcome, among other variables.
linical outcomes included HF incidence, health care costs


aowaqrislQ(MNpFpi[

�NLURiroig

pisnmdi

mmrovsfpCS

tcwuvia

sidreawwael

R

Pcea(ata6

iagmar(

de48s59dgwAtfs

idf(aAtt

1529JACC Vol. 48, No. 8, 2006 Rutledge et al.October 17, 2006:1527–37 Depression and Heart Failure

ssociated with HF, hospitalization, and combined clinicalutcomes in the form of death and secondary events. Asith the prevalence articles, the sorting of depression

ssessment methods used categories reflecting the use ofuestionnaires, diagnostic interviews, or patient medicalecords. The coding of all articles included demographicnformation (i.e., gender, age, and minority composition),ample size, inpatient and outpatient sample type, and studyocation (e.g., U.S., Europe).

uantitative methods. The software SPSS, version 11.5SPSS Inc., Chicago, Illinois), and Comprehensive

etaAnalysis, Version 2 (BioStat Software, Engelwood,ew Jersey), served as the statistical platforms for com-

leting all statistical tests and associated graphic results.or the summation of the prevalence findings, we com-uted prevalence point estimates and 95% confidencentervals using the formulas Logit Event Rate � LogEvent Rate / (1 � Event Rate)], Logit Event Rate SE �

1⁄�Event Rate � Num Tot� � 1 / [(1 � Event Rate) �um Tot], and 95% confidence intervals as Lower Limit �ogit Event Rate � (1.96 � Logit Event Rate SE) andpper Limit � Logit Event Rate � (1.96 � Logit Eventate SE). To standardize results for meta-analytic summary

n the clinical outcomes articles, we converted reported oddsatios into risk ratio values using the formula: risk ratio �dds ratio/([1 � P0] � [1 � P0]), where P0 represents thencidence of the outcome in the nonexposed (nondepressed)roup (37).

We centered the display of prevalence estimates around aoint of 0.50 to provide a reference standard for purposes ofllustration. In addition to the overall random effects model,ensitivity analyses broke down prevalence results across aumber of methodologic factors to assess evidence ofoderation. These factors included gender, minority status,

epression assessment method, HF severity, and use ofnpatient versus outpatient HF patient samples.

The small number of overall treatment studies and theixed nature of the interventions prohibited the use ofeta-analytic tests for this section. Descriptive summa-

ies of the treatment articles divided studies on the basisf using pharmacologic versus nonpharmacologic inter-entions. We calculated effects sizes for the treatmenttudies in the form of Cohen d values according to theormula d � M1 � M2/spooled. To address the issue ofublication bias, a fail-safe N was calculated using theomprehensive Meta-Analysis software program (Bio-tat).Heterogeneity assessments preceded all meta-analytic

ests concerning depression prevalence and clinical out-omes. In each case regarding prevalence rate analysis, thereas significant heterogeneity, hence we calculated resultssing a random effects model and reported corresponding palues and I2 values. Clinical outcome article categoriesncluded HF incidence, health care use and hospitalization,
nd mortality and clinical events. The number of mortality- n

pecific findings was insufficient to assess this outcomendependently. Log-transformed risk ratios and 95% confi-ence intervals were calculated for each study using theeported effect size and estimates of the standard error ofach effect drawn from data reported in the article. Whenrticles reported multiple models, we selected the modelith the highest level of covariate adjustment. In cases inhich parallel models from different time points were

vailable, we used the model with the longest time estimate,xcept in cases in which we extracted both short-term andonger-term model results for comparisons as described.

ESULTS

revalence. Figure 1 shows the prevalence rates of clini-ally significant depression among HF patients reported forach of the 27 studies and an aggregated estimate ofpproximately 21.5%, determined using meta-analytic testspreviously described). The failsafe N statistic indicated thatn additional 8,680 missing studies were required to bringhe p value from 0.00 to 0.05. The prevalence rates reportedcross these studies varied widely, ranging from 9% (38) to0% (39).Sixteen of the 27 studies included sufficient data regard-

ng gender to analyze prevalence rates of depression in malend female HF patients (12,15,17,18,38–60). The aggre-ated prevalence rate for women was higher than that foren, with point estimates of 32.7% (p � 0.000; I2 � 89.2%)

nd 26.1% (p � 0.000; I2 � 94.4%), respectively. Reportedanges again varied widely, with a range of 11% (40) to 67%41) for women and 7% (15) to 63% (42) for men.

Ten of the 27 studies provided prevalence rates ofepression among HF patients broken down by race orthnicity. The range of minority prevalence rates was 7% to4% with an aggregated estimate of 18.7% (p � 0.000; I2 �5.2%). In comparison, Caucasian HF patients within theame 10 studies had a depression prevalence range of 9% to4% with an aggregated estimate of 25.3% (p � 0.000; I2 �1.2%). Thus, there was some evidence of depressionifferences between minority and nonminority ethnicroups. Minority groups included within these 9 studiesere African Americans, Afro-Caribbeans, Asians, Nativemericans, and Pacific Islanders. An attempt to break

hese data down further for analysis by ethnicity was noteasible because of the heterogeneity of groups and smallample sizes.

An analysis of the reported prevalence rates of depressionn HF patients based on 5 studies that presented sufficientata for analysis by New York Heart Association (NYHA)unctional class showed an aggregated estimate of 27.8%p � 0.001; I2 � 92.1%) and higher prevalence ratesssociated with higher NYHA functional class (Table 1).lthough the differences between prevalence rates for pa-

ients within classes I or II and classes III or IV are similar,he rate of depression in patients with class III HF was
early double that of patients with class II HF. We further


opHsy

rdC2qiawiittr

e pat

TC

*f


bserved a nearly 4-fold increase in depression rates betweenatients with class I and class IV HF. The depression andF class relationship may further be affected by age, as

hown by Freedland et al. (17), with younger patients (�60ears) being more vulnerable to depression.

Figure 1. Prevalence of depression in heart failur

able 1. Depression Prevalence Reported by NYHA Functionallass in Patients With HF

NYHA Functional Class n* Depression Rate

I 222 0.11II 774 0.20III 638 0.38IV 155 0.42

Estimates compiled from 5 studies reporting depression rates among specific NYHA

punctional classes of HF patients.

HF � heart failure; NYHA � New York Heart Association.


Prevalence rates by the type of assessment measure usedanged from 10% to 54%. The method of assessment was thenichotomized into “questionnaires” or “interview included.”ompared with the aggregate depression rate of approximately2%, prevalence estimates for patients assessed solely withuestionnaires versus those whose assessment included annterview were 33.6% (heterogeneity p � 0.001, I2 � 90.3%)nd 19.3% (p � 0.001, I2 � 94.1%), respectively. Depressionas assessed with both a self-report questionnaire and an

nterview component in 9 of the 26 studies, solely an interviewn 2 of the 26 studies; the combined prevalence rate reported inhese studies was approximately 17.7%, nearly 25% less thanhe overall aggregate rate. The remaining studies used self-eport questionnaires alone to assess depression and re-

ients, and 95% confidence intervals (27 studies).

orted prevalence rates ranging from 30% to 44%. When by on August 9, 2008 cc.org


tc

spo((

ptwobc2ct

ecoqcdc

dr

iH(Cpp6tacadoinHpri1de

T

* iagnosc

T

AWHSFKRDFFJJMSV

CMf


he studies that used the same depression measure wereombined and compared with each other, the rates differed.

The overall weighted conservative prevalence depres-ion rate mean (based on 9 studies with 5,053 partici-ants) was 20.3% (p � 0.000; I2 � 86.7%), whereas theverall weighted liberal prevalence depression rate meanbased on 13 studies with 5,376 participants) was 35.5%p � 0.000; I2 � 95.5%).

The data were also analyzed for differences in depressionrevalence rates among inpatient and outpatient HF pa-ients. Five of the 27 studies included for prevalence dataere designed to measure participants’ depression levelsver time and hence captured data while participants wereoth inpatients and outpatients; thus, these studies wereoded as “combined” for this portion of the analysis. Of the2 studies with participants who were not consideredombined, 10 fell into the inpatient category and 12 fell intohe outpatient category.

Although overall results suggested no important differ-nces between inpatient and outpatient samples, we wereoncerned about the risk of measurement bias such thatutpatients may have been more likely to be assessed withuestionnaires. As shown in Table 2, we addressed thisoncern formally by examining 14 studies that reportedepression rates using either a conservative cutoff, a liberalutoff (e.g., 17 vs. 10 on the BDI), or both. Even broken

able 2. Depression Rates Among Heart Failure Patients by Inp

Depression Prevalence Rates Among Patient ClassificationsUsing Conservative Cutoffs*

PatientClassification

PrevalenceRate (SD)

Number of Studies,(Patients)

Inpatient 0.16 (0.09) 8 (2,921)Outpatient 0.14 (0.02) 8 (1,095)Combined 0.14 (0.06) 3 (1,037)

Conservative depression definitions included protocols with a structured interview, dutoff in the case of measures with multiple cutoff points.

able 3. A Description of HF Studies Reporting Relationships B

Study Depression Measure Duration

bramson et al. (61) CES-D 4.5 yrs illiams et al. (44) CES-D 14 yrs imelhoch et al. (62) Medical records 1 yr

ullivan et al. (15) Medical records 3 yrs ulop et al. (58) SCID interview 6 months oenig et al. (48) DIS interview 1 yr umsfeld et al. (14) MOS-D 6 weeks e Denus et al. (49) Medical records 7.5 months aris et al. (50) Medical records 4 yrs reedland et al. (60) DIS interview 1 yr

iang et al. (12) DIS interview 1 yr unger et al. (13) HADS-D 24 months

urberg et al. (63) Zung 2 yrs ullivan et al. (11) PRIME-MD interview 3 yrs accarino et al. (38) Geriatric depression 6 months

ES-D � Center for Epidemiological Studies–Depression; DIS � Diagnostic Inter
OS-D � Medical Outcomes Study–Depression; NA � not available; PRIME-MD � P
or DSM-IV.


own by depression assessment method, rates of depressionemained similar for inpatient and outpatient samples.

Analyses based on geographic location showed nearlydentical point estimates of prevalence of depression rates in

F patients in the U.S. and Canada (20.3%) and Europe21%).

linical outcomes. Table 3 provides a description of the 14rospective investigations reporting associations between de-ression and HF-related events (11–15,38,44,48–50,58,60–3). The studies included a range of depression assessmentools, protocol lengths, and sample sizes. Women composed

significant percentage of most study samples. Outcomeategories included: 1) the relationship between depressionnd HF incidence in a healthy or at-risk sample; 2)epression effects among HF patient samples on measuresf health care cost or health care system use; and 3) thempact of depression on mortality and/or secondary eventsot including hospitalization.F incidence. Two studies reported data concerning de-

ression and HF development. Covariate adjusted hazardatios (including age and disease severity) were 1.5 and 2.6n these studies, which followed up participants for 4.5 and4 years, respectively. The stronger relationship betweenepression and HF incidence was noted in the protocolnrolling “high risk” patients with systolic hypertension (61).

t/Outpatient Status and Diagnostic Threshold

Depression Prevalence Rates Among Patient ClassificationsUsing Liberal Cutoffs*

PatientClassification

PrevalenceRate (SD)

Number of Studies,(Patients)

Inpatient 0.38 (0.13) 7 (2,667)Outpatient 0.38 (0.11) 11 (1,036)Combined 0.32 (0.06) 5 (1,673)

is by an MD or PhD mental health professional, or use of a stringent questionnaire

en Depression and Clinical Outcomes

Sample Size % Women Outcome(s)

4,538 57 Incident HF2,501 58 Incident HF

139,089 NA Health service use, hospitalization1,098 53 Health care costs, clinical events

203 53 Hospitalization107 52 Hospitalization466 24 Hospitalization171 36 Clinical events396 26 Hospitalization, clinical events60 57 Hospitalization, mortality

357 36 Hospitalization, clinical events209 28 Clinical events119 29 Clinical events142 23 Clinical events391 49 Clinical events

chedule; HADS-D � Hospital Anxiety and Depression Scale; HF � heart failure;

atien

etwe

view S
rimary Care Evaluation of Mental Disorders; SCID � Structured Clinical Interview


Hrudtifmptmmt(t

tPssae(wD4twMdmaci

dnanwso

aetcsncbft(6aeHiidrTcipdn

F(


ealth care use and hospitalization. A total of 7 studieseported information concerning relative rates of health carese among HF patients with lower versus higher levels ofepression, including 6 studies comparing rates of rehospi-alization. Despite the wide range of use variables, the datandicate a consistent pattern of increased health care usageor patients with depression. These end points included aore than 2-fold risk of emergency room visits for de-

ressed versus nondepressed patients (62); a 29% increase inotal health care costs (aggregated from separate measures ofental health visits, and inpatient and outpatient treat-ents) for patients with a depression diagnosis or receiving

reatment for depression versus those without depression11); and increases in both short-term (4-week) and longer-erm (6-month) medical encounters (58).

Among studies reporting hospitalization comparisons, pro-ocol lengths varied from 6 weeks to more than 4 years.rotocol length influenced the relationship between depressioneverity and rehospitalization rates. Briefer studies typicallyuggested no or small differences during the initial monthsfter discharge (60), compared with more substantial differ-nces appearing by 1 year of follow-up. The study by Koenig48) provides an illustration of this theme. Among 39 HF patientsith high scores on the Center for Epidemiologic Studies–epression scale, inpatient readmission rates were 31.6%, 54.8%,

0.7%, and 25% at 3, 6, 9, and 12 months, respectively. Rates overhe same intervals among 45 nondepressed patients in this sampleere 35.7%, 27.8%, 17.7%, and 16.1%.ortality and associated clinical events. Eight indepen-

ent cohort studies, ranging in length from 6 months toore than 4 years, tracked the incidence of mortality and

ssociated cardiac events (e.g., heart transplantation, newardiac events) in association with depression. Figure 2llustrates the range of effects from these studies.

igure 2. Effect sizes and 95% confidence intervals among studies reportin8 studies).


Study protocols included a variety of definitions ofepression, including elevated questionnaire scores, diag-oses based on information in the patients’ medical records,nd the administration of a diagnostic interview. Theumber of studies including an interview-based diagnosisas insufficient to make comparisons between depression

ymptoms and mood disorders in the prediction of clinicalutcomes.The aggregated risk estimate from the 8 studies suggestedgreater than 2-fold risk of death and associated clinical

vents for HF patients with heightened depressive symp-oms or a depressive disorder (relative risk � 2.1, 95%onfidence interval 1.7 to 2.6). Despite the variation acrosstudies noted above, this estimation is based on homoge-ous results (heterogeneity p � 0.87; I2 � 0%). Overall, theombined evidence did not indicate that the relationshipetween depression and HF outcomes differed by length ofollow-up, however, the results of individual studies suggesthat this is an area requiring additional study. Junger et al.13) reported depression and mortality associations by-month intervals across 30 months of follow-up, indicatingstrong, graded relationship over time. In contrast, differ-

nces in death rates between depressed and nondepressedF patients were stable across 1-, 2-, and 3-year estimates

n a separate protocol (50). Few if any of the studies to datencluded samples sizes that were adequately powered toetect differences in short-term mortality and clinical eventates.reatment effects. Table 4 summarizes the methodologic

haracteristics and intervention outcomes for the 6 articlesncluded as treatment studies (56,64–68). Each study com-rised a pre-post design using HF patients among whomepression severity was assessed using a validated question-aire to gauge treatment effectiveness. Two studies also

g associations between depression and mortality and secondary events



raaowilp(

mmcp4HtCsgduC

D

TtappiicoidswDombwscqdptDp(slle

4.A

Sum

mar

yof

HF

Inte

rven

tion

Stud

ies

Des

crib

ing

Cha

nges

inD

epre

ssio

nSe

veri

ty

Aut

hors

n(W

omen

)M

ean

Age

,yr

s(S

D)

Inte

rven

tion

Dur

atio

nD

epre

ssio

nM

easu

reO

utco

mes

rver

a-T

inde

let

al.

64)

39(N

A)

63.2

(10.

1)H

ome

wal

king

exer

cise

prog

ram

12w

eeks

Mul

tiple

Affe

ctA

djec

tive

Che

cklis

t(M

AA

CL

)M

AA

CL

com

plia

ntvs

.non

-com

plia

ntco

mpl

eter

s:d

��

0.21

MA

AC

Lco

mpl

iant

com

plet

ers

vs.d

rop-

outs

:d�

0.14

stis

et

al. (

65)

20 (6

) 65

.7 (6

.1)

Com

bine

d ex

erci

se, d

ieta

ryco

ntro

l,&

grou

pC

BT

vs.

digo

xin

vs.p

lace

bo

12w

eeks

Bec

kD

epre

ssio

nIn

vent

ory

(BD

I),

Ham

ilton

Dep

ress

ion

Rat

ing

Scal

e(H

DR

S)

Mea

nde

pres

sion

chan

ge(c

ombi

ned

BD

I&

HD

RS)

Non

phar

mac

olog

ictx

:d�

�0.

64D

igox

intx

:d�

0.35

Pla

cebo

:d�

0.35

er e

t al

. (66

) 58

9 (1

56)

64.6

(11.

7)

Dig

oxin

vs.

pla

cebo

4

and

12m

onth

sC

ente

rfo

rE

pide

mio

logi

cSt

udie

s–D

epre

ssio

nSc

ale

(CE

S-D

)

Dep

ress

ion

scor

ech

ange

nons

igni

fican

t,sa

me

inbo

thgr

oups

:8to

7D

igox

in6

MW

T:d

�0.

16P

lace

bo6

MW

T:d

�0.

14pe

ranc

e et

al.

(56)

28

(8)

59.6

(NA

) N

efaz

odon

e 12

wee

ks

BD

I or

PR

IME

-MD

&SC

ID&

HD

RS

HD

RS

chan

ge:d

��

2.20

BD

Ich

ange

:d�

�1.

54sk

in e

t al

. (67

) 33

(20)

66

(9)

Gro

up s

tres

s m

anag

emen

ttr

aini

ngvs

.con

trol

10w

eeks

Ger

iatr

icD

epre

ssio

nSc

ale

(GD

S)T

xG

DS

chan

ge:d

��

0.16

Con

trol

GD

Sch

ange

:d�

0.03

Tx

6M

WT

chan

ge:d

�0.

12C

ontr

ol6

MW

Tch

ange

:d�

�0.

02ze

witz

et

al. (

68)

88 (2

1)

65.8

(8.2

) M

uscl

e st

reng

th a

nd b

icyc

leer

gom

eter

trai

ning

;6-m

inw

alk

4w

eeks

Hos

pita

lAnx

iety

&D

epre

ssio

nSc

ale–

Dep

ress

ion

(HA

DS)

HA

DS-

Dch

ange

:d�

�0.

136

MW

Tch

ange

:d�

0.99

T �

cog

nitiv

e be

havi

or t

hera

py; 6

MW

T �

6 m

inut

e w

alk

test

; Tx

� t

reat

men

t gr

oup;

oth

er a

bbre

viat

ions

as

in T

able

3.


eported patients’ results on the 6-min walk test before andfter treatment to determine whether objective measures ofbility were affected by the intervention. In addition, somef the studies also reported results of other variables onhich HF patients experienced changes attributed to the

ntervention, such as physical abilities (e.g., VO2max, weightifted), physiological measurements (e.g., weight, bloodressure, QT intervals), and scores on other questionnairese.g., quality of life, anxiety).

The treatment studies generally suffered from 1 or moreethodologic flaws, including small samples, brief treat-ent durations, and a lack of a placebo or control group

omparison. The mean decrease in depression scores, ex-ressed in Cohen d units, was 0.42, translating to just over0% of a standard deviation difference between groups.owever, this result was skewed by large depression symp-

om reductions observed in 1 study (56). The medianohen d value across all studies was 0.16 (i.e., 16% of a

tandard deviation difference between treated and untreatedroups, a modest difference by most standards). Assessingepression symptom reductions separately for studies thatsed a nonpharmacologic intervention yielded a meanohen d value of 0.29 in depression reduction.

ISCUSSION

his meta-analysis provides evidence supporting a moderateo high prevalence of depression among patients with HF,n increased risk of mortality and clinical events among HFatients with depressive symptoms, and reductions in de-ressive symptoms resulting from a variety of treatmentnterventions. The relationship between depression and HFs a relatively new but rapidly expanding area of interest inardiology research. In fact, very few articles were publishedn this topic before 1990, but the key word searches useddentified more than 100 articles published on the topics ofepression and other psychosocial factors in HF patientsince that time. This increase in research attention coincidesith the growth in HF as a health care problem in the U.S.uring the past 10 years, HF was the fastest growing form

f cardiovascular disease, and accounted for more than 1illion annual hospital admissions and an estimated $60

illion in annual health care expenses (69). Because patientsith HF face high rates of debilitation and mortality, the

tudy of depression characteristics in this population is aritical research area in the pursuit of improved quantity anduality of their lives. Abundant evidence suggests thatepression is under-recognized and undertreated in cardiacopulations (3,10,70); this review begins to address some ofhe consequences of this treatment milieu.

epression prevalence in HF. The overall aggregatedoint estimate prevalence rate calculated in this study21.6%) indicates that HF patients experience clinicallyignificant depression at a rate similar to the 15% to 20%evels cited for CAD patients (71,72), and at 2 to 3 times

the rate of the general population (73). Considering theTab

Co (

Ko

Lad

Les

Lu

Rad

CB



tpseHtsnrh

wedcoFqolebpaiaowwditlmqt

mlcrtctkfiapdNicmiba

DmoswsdNddcdtedsitiHrpoaotatt

imoasfdttaotidlafskafEatm


ypically worse prognosis for individuals with HF as op-osed to CAD or myocardial infarction, this rate of depres-ion may seem modest. However, this combined depressionstimate is much lower than that suggested in a number ofF studies (48), a discrepancy that seems from this review

o be a result of defining depression on the basis of elevatedymptom severity from questionnaires rather than a diag-ostic interview. Individual study prevalence estimatesanged from 9% to 60%, suggesting a considerable degree ofeterogeneity.Explaining the heterogeneity of reported depression rates

as a secondary goal of this review. The moderatorsxplored in association with depression prevalence includedepression assessment method, conservative versus liberalutoff used to classify depression’s presence, inpatient versusutpatient samples, HF severity, ethnicity, age, and gender.rom these analyses, depression assessment methods (i.e.,uestionnaire versus diagnostic interview) and the liberalityf their cutoffs as well as HF severity seem to have theargest impact on reported depression rates. With onexception (17), differences in depression and HF severityased on patient age were not reported in a form that couldrovide an effective assessment of this relationship. Meanges between study samples were likewise too similar todentify age-related trends. Depression rates may also differccording to HF characteristics that affect symptom severityr degree of disability (e.g., systolic vs. diastolic HF, or HFith preserved vs. depressed ejection fraction). At present,e have almost no information concerning depressionifferences across subclasses of HF patients. Although

nteractions among the above moderator factors are poten-ially very interesting for understanding depression preva-ence in HF and could be approached using multilevel

odeling methods (74), we were not able to address theseuestions systematically because of inconsistent reporting ofhe moderator factors across studies.

Awareness of which demographic characteristics and/orethods of assessment are likely to result in a higher or

ower detection rate of depression may help researchers andlinicians alike. For example, researchers examining thiselationship could design their studies with greater precisiono answer their specific research questions. Likewise, clini-ians (e.g., primary care physicians, cardiologists, psychia-rists, and psychologists) would be equipped with thenowledge of the idiosyncrasies associated with the identi-cation of major depression versus depressive symptomsmong their patients; hence, clinicians could assess theiratients with different criteria depending on the severity ofepression they intend to identify and attempt to treat.umerous studies from the depression and CAD literature

ndicate that depression severity is linearly associated withlinical outcomes (3–5), suggesting that the presence of aajor depressive episode is not necessary as a standard for

ntervening. No single method or cutoff is necessarily theest; rather, each is different and should be used with
ppreciation of its implications and limitations. t

epression and clinical outcomes. Across a range ofortality, health care use, and associated clinical event

utcomes investigated in the prospective studies of depres-ion and HF, aggregate results indicated a substantiallyorse prognosis for HF patients with more severe depres-

ive symptoms. Nearly all of the clinic outcome studies ofepression adjusted for multiple covariates (e.g., age,YHA functional class, ejection fraction), suggesting that

epression effects are reasonably robust to demographic andisease severity characteristics. Although these findings areonsistent with depression effects on mortality in cardiacisease patients described in recent reviews (7,8), the mor-ality results were further reinforced by equally large differ-nces in hospital readmission and health care use byepressed versus nondepressed HF patients. This patternuggests that any interventional efforts targeting depressionn HF should consider a broad category of clinical outcomeso accurately assess potential benefits. Importantly, thencreased event risk associated with depression within the

F population has only been examined in studies withelatively small sample sizes, many of which were marginallyowered to detect clinically significant effects, particularlyver shorter time intervals. We can likewise only speculatet this stage regarding differences in the depression and HFutcomes relationship based on factors such as NYHA class,reatment adherence, or age. Thus, although this meta-nalysis was able to quantify a more precise magnitude ofhe increased risk of clinical events among depressed pa-ients, a clear need for additional research remains.

Unfortunately, outcome variables of interest such as HFncidence and health care costs did not lend themselves to

eta-analytic procedures because of an insufficient numberf studies. The data from existing studies are suggestive ofhigher risk of HF development among those with more

evere depressive symptoms, as well as higher costs resultingrom an increased use of health care services. We hope thatrawing attention to the current paucity of data regardinghese topics will provide direction to researchers working inhis area. Finally, in the context of prospective studies,lmost no information concerning changes in depressionver time is available. Because individuals’ depression symp-oms are likely to change over time, depression shoulddeally be assessed at multiple intervals throughout studies’urations to understand possible biobehavioral pathways

inking depression with HF events (75,76). Heart failurend depression share several biological mechanisms. Heartailure (75–77) and depression (78–81) are associated withympathetic activation and elevated proinflammatory cyto-ines, including interleukin (IL)-6, tumor necrosis factorlpha, and IL-1�. The additive effects of the inflammationound in depression likely adversely affect the heart (82–84).xercise programs reduce IL-6 and tumor necrosis factor

lpha levels in HF (85,86). Pharmacologic and exercisereatments for depression reduce depressive symptoms andight reduce accompanying inflammation (87–89), poten-

ially producing more favorable HF-related clinical out- by on August 9, 2008 cc.org


coHDnmspdswtphpftttsltsstmtsSpHapm

ntcmktntcbsStammapcgm

picawrhsHdTrsas

RV3T

R

1

1

1

1


omes in the patients. At the very least, reducing symptomsf depression could be expected to improve adherence toF treatment regimens (90,91).epression treatment studies among HF patients. Fi-

ally, the treatment/intervention section proved to be theost difficult to quantify because of the following factors:

mall number of identified studies documenting changes inretreatment versus posttreatment depression severity, briefurations, small sample sizes (many were pilot studies), andubstantial heterogeneity in the types of interventions. Weere not able to quantify the quality or appropriateness of

he interventions for the target patient groups. Some of theharmacologic interventions used, such as digoxin, do notave specific antidepressant effects, and may underestimateotential treatment benefits achievable via interventionsocused more specifically on depression symptoms. Unlikehe larger depression and CAD literature (10,92), no studyo date has investigated the effects of a depression interven-ion on objective clinical outcomes such as survival orecondary cardiac events in an HF population. Despite theseimitations, a general pattern of decreased depressive symp-oms and increased physical abilities (as measured by thetandard 6-min walk test) was observed across the treatmenttudies. Although these preliminary results are promising,he most important aspect of this section is the need forore studies with larger sample sizes, replicable interven-

ions (especially interventions with a behavioral or psycho-ocial component), and longer durations.tudy limitations. A large number of articles describesychosocial characteristics other than depression amongF patients. Anxiety, social support and social isolation,

nd quality of life are each examples of other topics ofsychosocial research in the HF literature (93–95) thaterit review but that are beyond the scope of this review.We did not include unpublished articles or articles from

on–peer-reviewed journals, which are more likely to con-ain negative associations (96). Additional articles thatontain findings relevant to the areas covered in this articleay have been missed because of a mismatch between our

ey word and abstract searches and the key words in theiritles or abstracts. Lastly, the protocol descriptions in aumber of studies suggested that additional informationhat could have been useful in the meta-analytic tests wasollected, but we were limited to the data as reportedecause of incomplete reporting or inability to contact thetudy investigators.ummary. We conducted a focused quantitative review of

he depression and HF literature concerning questionsbout depression prevalence, clinical outcomes, and treat-ent impact. This article is among the first to applyeta-analytic methods to the growing field of depression

nd HF. Across published studies concerning depressionrevalence, associations with clinical outcomes, andhanges resulting from treatment interventions, severaleneral conclusions can be drawn: 1) depression is com-
on among patients with HF, with approximately 1 in 5
1


atients meeting criteria for major depression based onnterview methods, and substantially higher rates of clini-ally significant depression are present among patientsssessed with questionnaires (vs. diagnostic interviews) orith more severe HF; 2) rates of mortality, clinical events,

ehospitalization, and general health care use are markedlyigher among HF patients reporting more severe depres-ion; and 3) studies describing depression treatments amongF patients are too small and heterogeneous to permit

efinitive conclusions regarding intervention effectiveness.hese results identify areas requiring further development,

aise novel questions, and provide information on depres-ion prevalence that can help researchers design studies withppropriate depression measures and adequately poweredample sizes.

eprint requests and correspondence: Dr. Thomas Rutledge,A San Diego Healthcare System, Psychology Service (116B),350 La Jolla Village Drive, San Diego, California 92161. E-mail:[email protected].

EFERENCES

1. Lett HS, Blumenthal JA, Babyak MA, et al. Depression as a risk factorfor coronary artery disease: evidence, mechanisms, and treatment.Psychosom Med 2004;66:305–15.

2. Davidson KW, Rieckmann N, Lesperance F. Psychological theories ofdepression: potential application for the prevention of acute coronarysyndrome recurrence. Psychosom Med 2004;66:165–73.

3. Musselman DL, Evans DL, Nemeroff CB. The relationship ofdepression to cardiovascular disease: epidemiology, biology, and treat-ment. Arch Gen Psychiatry 1998;55:580–92.

4. Shimbo D, Chaplin W, Crossman D, Haas D, Davidson KW. Role ofdepression and inflammation in incident coronary heart disease events.Am J Cardiol 2005;96:1016–21.

5. Rugulies R. Depression as a predictor for the development of coronaryheart disease: a systematic review and meta-analysis of the literature.Am J Prev Med 2003;23:51–61.

6. Wulsin LR, Singal BM. Do depressive symptoms increase the risk forthe onset of coronary disease? A systematic quantitative review.Psychosom Med 2003;65:201–10.

7. Barth J, Schumacher M, Herrman-Lingen C. Depression as a riskfactor for mortality in patients with coronary heart disease. PsychosomMed 2004;66:802–13.

8. Van Melle JP, De Jonge P, Spijkerman TA, et al. Prognosticassociation of depression following myocardial infarction with mortal-ity and cardiovascular events: a meta-analysis. Psychosom Med 2004;66:814–22.

9. Zheng D, Macera CA, Croft JB, Giles WH, Davis D, Scott WK.Major depression and all-cause mortality among white adults in theUnited States. Ann Epidemiol 1997;7:213–8.

0. The ENRICHD Investigators. Effects of treating depression and lowperceived social support on clinical events after myocardial infarction:the enhancing recovery in coronary heart disease (ENRICHD) ran-domized trial. JAMA 2003;289:3106–16.

1. Sullivan MD, Levy WC, Crane BA, Russo JE, Spertus JA. Usefulnessof depression to predict time to combined end point of transplant ordeath for outpatients with advanced heart failure. Am J Cardiol2004;94:1577–80.

2. Jiang W, Alexander J, Christopher E, et al. Relationship of depressionto increased risk of mortality and rehospitalization in patients withcongestive heart failure. Arch Intern Med 2001;161:1849–56.

3. Junger J, Schellberg D, Muller-Tasch T, et al. Depression increasinglypredicts mortality in the course of congestive heart failure. Eur J HeartFail 2005;7:261–7.

4. Rumsfeld JS, Havranek E, Masoudi FA, et al., for the CardiovascularOutcomes Research Consortium. Depressive symptoms are the stron-



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

4

4

4

5

5

5

5

5

5

5

5

5

5

6

6

6

6


gest predictors of short-term declines in health status in patients withheart failure. J Am Coll Cardiol 2003;42:1811–7.

5. Sullivan M, Simon G, Spertus J, Russo J. Depression-related costs inheart failure care. Arch Intern Med 2002;162:1860–6.

6. Cline CMJ, Israelsson BYA, Willenheimer RB, Broms K, ErhardtLR. Cost effective management programme for heart failure reduceshospitalization. Heart 1998;80:442–6.

7. Freedland KE, Rich MW, Skala JA, Carney RM, Davila-Roman VG,Jaffe AS. Prevalence of depression in hospitalized patients withcongestive heart failure. Psychosom Med 2003;65:119–28.

8. Gottlieb SS, Khatta M, Friedmann E, et al. The influence of age,gender, and race on the prevalence of depression in heart failurepatients. J Am Coll Cardiol 2004;43:1542–9.

9. Konstam V, Moser DK, De Jong MJ. Depression and anxiety in heartfailure. J Card Fail 2005;11:455–63.

0. Joynt KE, Whellan DJ, O’Connor CM. Why is depression bad for thefailing heart? A review of the mechanistic relationship betweendepression and heart failure. J Card Fail 2004;10:258–71.

1. Spitzer RL, Williams JBW, Gibbon M, First MB. The StructuredClinical Interview for DSM-III-R (SCID), I: history, rationale, anddescription. Arch Gen Psychiatry 1992;49:624–9.

2. Spitzer RL, Williams JBW, Gibbon M, First MB. Structured ClinicalInterview for DSM-III-R—Non-Patient Edition (SCID-NP, Version1.0). Washington, DC: American Psychiatric Press, 1990.

3. Robins L, Helzer J, Croughan J, Williams J, Spitzer R. NationalInstitute of Mental Health Diagnostic Interview Schedule: history,characteristics, validity. Arch Gen Psychiatry 1981;38:381–9.

4. Carney RM, Rich MW, teVelde A, Saini J, Clark K, Jaffe AS. Majordepressive disorder in coronary artery disease. Am J Cardiol 1987;60:1273–5.

5. Kessler RC, Abelson J, Demler O, et al. Clinical calibration ofDSM-IV diagnoses in the World Mental Health (WMH) version ofthe World Health Organization (WHO) Composite InternationalDiagnostic Interview (WMH-CIDI). Int J Methods Psychiatr Res2004;13:122–39.

6. Spitzer RL, Williams J, Kroenke K, et al. Utility of a new procedurefor diagnosing mental disorders in primary care: the PRIME-MD1000 study. JAMA 1994;272:1749–56.

7. Beck AT. Depression Inventory. Philadelphia, PA: Center for Cog-nitive Therapy, 1978.

8. Zung WWK. A self-rating depression scale. Arch Gen Psychiatry1965;12:63–70.

9. Yesavage JA, Brink TL, Rose TL, et al. Development and validationof a geriatric depression screening scale: a preliminary report. J Psy-chiatr Res 1982–1983;17:37–49.

0. Radloff L. The CES-D Scale: a self-report depression scale forresearch in the general population. Appl Psycho1 Meas 1977;1:385–90.

1. Zigmond AS, Snaith RP. The Hospital Anxiety and Depression Scale.Acta Psychiatr Scand 1983;67:361–70.

2. Zimmerman M, Coryell M, Corenthal C, Wilson S. A self-reportscale to diagnose major depressive disorder. Arch Gen Psychiatry1986;43:1076–86.

3. Hamilton M. Development of a rating scale for primary depressiveillness. Br J Soc Clin Psycho1 1967;6:278–46.

4. Lipman RS, Covi L, Shapiro AK. The Hopkins symptom checklist(HSCL)—factors derived from the HSCL-90. J Affect Disord 1979;1:9–24.

5. Nagel R, Lynch D, Tamburrino M. Validity of the medical outcomesstudy depression screener in family practice training centers andcommunity settings. Fam Med 1998;30:362–5.

6. Zuckerman M, Lubin B. The Multiple Affect Adjective Check List.San Diego, CA: Educational and Industrial Testing Service, 1965.

7. Zang J, Yu KF. What’s the relative risk? A method of correcting theodds ratio in cohort studies of common outcomes. JAMA 1998;280:1690–1.

8. Vaccarino V, Kasl SV, Abramson J, Krumholz HM. Depressivesymptoms and risk of functional decline and death in patients withheart failure. J Am Coll Cardiol 2001;38:199–205.

9. Pihl E, Jacobsson A, Fridlund B, Stromberg A, Martensson J.Depression and health-related quality of life in elderly patientssuffering from heart failure and their spouses: a comparative study. Eur
J Heart Fail 2005;7:583–9.

0. Turvey CL, Schultz K, Arndt S, Wallace RB, Herzog R. Prevalenceand correlates of depressive symptoms in a community sample ofpeople suffering from heart failure. J Am Geriatr Soc 2002;50:2003–8.

1. Skotzko CE, Krichten C, Zietowski G, et al. Depression is commonand precludes accurate assessment of functional status in elderlypatients with congestive heart failure. J Card Fail 2000;6:300–5.

2. Sullivan M, Levy WC, Russo JE, Spertus JA. Depression and healthstatus in patients with advanced heart failure: a prospective study intertiary care. J Card Fail 2004;10:390–6.

3. Westlake C, Dracup K, Fonarow G, Hamilton M. Depression inpatients with heart failure. J Card Fail 2005;11:30–5.

4. Williams SA, Kasl SV, Heiat A, Abramson JL, Krumholz HM,Vaccarino V. Depression and risk of heart failure among the elderly: aprospective community-based study. Psychosom Med 2002;64:6–12.

5. Yu DS, Lee DT, Woo J, Thompson DR. Correlates of psychologicaldistress in elderly patients with congestive heart failure. J PsychosomRes 2004;57:573–81.

6. Havranek EP, Ware MG, Lowes BD. Prevalence of depression incongestive heart failure. Am J Cardiol 1999;84:348–50.

7. Haworth JE, Moniz-Cook E, Clark AL, Wang M, Waddington R,Cleland JG. Prevalence and predictors of anxiety and depression in asample of chronic heart failure patients with left ventricular systolicdysfunction. Eur J Heart Fail 2005;7:803–8.

8. Koenig HG. Depression in hospitalized older patients with congestiveheart failure. Gen Hosp Psychiatry 1998;20:29–43.

9. de Denus S, Spinler SA, Jessup M, Kao A. History of depression as apredictor of adverse outcomes in patients hospitalized for decompen-sated heart failure. Pharmacotherapy 2004;24:1306–10.

0. Faris R, Purcell H, Henein MY, Coats AJ. Clinical depression iscommon and significantly associated with reduced survival in patientswith non-ischaemic heart failure. Eur J Heart Fail 2002;4:541–51.

1. Fraticelli A, Gesuita R, Vespa A, Paciaroni E. Congestive heart failurein the elderly requiring hospital admission. Special Congestive HeartFailure in the Elderly. Arch Gerontol Geriatr 1996;23:225–38.

2. Parissis JT, Adamopoulos S, Rigas A, et al. Comparison of circulatingproinflammatory cytokines and soluble apoptosis mediators in patientswith chronic heart failure with versus without symptoms of depression.Am J Cardiol 2004;94:1326–8.

3. Murberg TA, Bru E, Aarsland T, Svebak S. Functional status anddepression among men and women with congestive heart failure. IntJ Psychiatry Med 1998;28:273–91.

4. Pihl E, Jacobsson A, Fridlund B, Stromberg A, Martensson J.Depression and health-related quality of life in elderly patientssuffering from heart failure and their spouses: a comparative study. EurJ Heart Fail 2005;7:583–9.

5. Turvey CL, Klein DM, Pies CJ, Arndt S. Attitudes about impairmentand depression in elders suffering from chronic heart failure. IntJ Psychiatry Med 2003;33:117–32.

6. Lespérance F, Frasure-Smith N, Laliberté MA, et al. An open-labelstudy of Nefazodone treatment of major depression in patients withcongestive heart failure. Can J Psychiatry 2003;48:695–701.

7. Friedman MM, Griffin JA. Relationship of physical symptoms andphysical functioning to depression in patients with heart failure. HeartLung 2001;30:98–104.

8. Fulop G, Strain JJ, Stettin G. Congestive heart failure and depressionin older adults: clinical course and health services use 6 months afterhospitalization. Psychosomatics 2003;44:367–73.

9. Lane D, Carroll D, Ring C, Beevers DG, Lip GYH. Mortality andquality of life 12 months after myocardial infarction: effects ofdepression and anxiety. Psychosom Med 2001;63:221–30.

0. Freedland KE, Carney RM, Rich MW, et al. Depression in elderlypatients with congestive heart failure. J Geriatr Psychiatry 1991;24:59–71.

1. Abramson J, Berger A, Krumholz HM, Vaccarino V. Depression andrisk of heart failure among older persons with isolated systolichypertension. Arch Intern Med 2001;161:1725–30.

2. Himelhoch S, Weller WE, Wu AW, Anderson GF, Cooper LA.Chronic medical illness, depression, and use of acute medical servicesamong Medicare beneficiaries. Med Care 2004;42:512–21.

3. Murberg TA, Bru E, Svebak S, Tveteras R, Aarsland T. Depressedmood and subjective health symptoms as predictors of mortality in
patients with congestive heart failure: a two-years follow-up study. IntJ Psychiatry Med 1999;29:311–26.


6

6

6

6

6

6

7

7

7

7

7

7

7

7

7

7

8

8

8

8

8

8

8

8

8

8

9

9

9

9

9

9

9


4. Corvera-Tindel T, Doering LV, Gomez T, Dracup K. Predictors ofnoncompliance to exercise training in heart failure. J Cardiovasc Nurs2004;19:269–77.

5. Kostis JB, Rosen RC, Cosgrove NM, Shindler DM, Wilson AC.Nonpharmacologic therapy improves functional and emotional statusin congestive heart failure. Chest 1994;106:996–1001.

6. Lader E, Egan D, Hunsberger S, Garg R, Czajkowski S, McSherry F.The effect of digoxin on the quality of life in patients with heart failure.J Card Fail 2003;9:4–12.

7. Luskin F, Reitz M, Newell K, Quinn TG, Haskell W. A controlledpilot study of stress management training of elderly patients withcongestive heart failure. Prev Cardiol 2002;5:168–72.

8. Radzewitz A, Miche E, Herrmann G, et al. Exercise and musclestrength training and their effect on quality of life in patients withchronic heart failure. Eur J Heart Fail 2002;4:627–34.

9. Galbreath AD, Krasuski RA, Smith B, et al. Long-term health careand cost outcomes of disease management in a large, randomized,community-based population with heart failure. Circulation 2004;110:3518–26.

0. Gilberg K, Laouri M, Wade S, Isonaka S. Analysis of medication usepatterns: apparent overuse of antibiotics and underuse of prescriptiondrugs for asthma, depression, and CHF. J Manag Care Pharm2003;9:232–7.

1. Schleifer SJ, Macari-Hinson MM, Coyle DA, et al. The nature andcourse of depression following myocardial infarction. Arch Intern Med1989;149:1785–9.

2. Rozanski A, Blumenthal JA, Kaplan J. Impact of psychological factorson the pathogenesis of cardiovascular disease and implications fortherapy. Circulation 1999;99:2192–217.

3. Evans DL, Charney DS, Lewis L, et al. Mood disorders in themedically ill: scientific review and recommendations. Biol Psych2005;58:175–89.

4. Thompson SG. Why sources of heterogeneity in meta-analysis shouldbe investigated. BMJ 1994;309:135–5.

5. Parissis JT, Venetsanou KF, Mentzikof DG, Ziras NG, Kefalas CG,Karas SM. Tumor necrosis factor-alpha serum activity during treat-ment of acute decompensation of cachectic and non-cachectic patientswith advanced congestive heart failure. Scand Cardiovasc J 1999;33:344–50.

6. Hasper D, Hummel M, Kleber FX, Reindl I, Volk HD. Systemicinflammation in patients with heart failure. Eur Heart J 1998;19:761–5.

7. Tsutamoto T, Wada A, Maeda K, et al. Angiotensin II type 1 receptorantagonist decreases plasma levels of tumor necrosis factor alpha,interleukin-6 and soluble adhesion molecules in patients with chronicheart failure. J Am Coll Cardiol 2000;35:714–21.

8. Linden W, Stossel C, Maurice J. Psychosocial interventions forpatients with coronary artery disease: a meta-analysis. Arch Intern
Med 1996;156:745–52.

9. McEwen BS. The neurobiology of stress: from serendipity to clinicalrelevance. Brain Res 2000;886:172–89.

0. Leonard BE. Stress, norepinephrine and depression. J PsychiatryNeurosci 2001;26 Suppl:S11–6.

1. Irwin M. Psychoneuroimmunology of depression: clinical implica-tions. Brain, Behavior, and Immunity 2002;16:1–16.

2. Oral H, Kapadia S, Nakano M, et al. Tumor necrosis factor-alpha andthe failing human heart. Clin Cardiol 1995;18:IV20–7.

3. Paulus WJ. Cytokines and heart failure. Heart Fail Monit 2000;1:50–6.

4. Roberts AB, Roche NS, Winokur TS, Burmester JK, Sporn MB. Roleof transforming growth factor-beta in maintenance of function ofcultured neonatal cardiac myocytes. Autocrine action and reversal ofdamaging effects of interleukin-1. J Clin Invest 1992;90:2056–62.

5. LeMaitre JP, Harris S, Fox KA, Denvir M. Change in circulatingcytokines after 2 forms of exercise training in chronic stable heartfailure. Am Heart J 2004;147:100–5.

6. Adamopoulos S, Parissis J, Karatzas D, et al. Physical trainingmodulates proinflammatory cytokines and the soluble Fas/soluble Fasligand system in patients with chronic heart failure. J Am Coll Cardiol2002;39:653–63.

7. Stein MB. Sweating away the blues: can exercise treat depression?Am J Prev Med 2005;28:140–1.

8. Dunn AL, Trivedi MH, Kampert JB, Clark CG, Chambliss HO.Exercise treatment for depression: efficacy and dose response. Am JPrev Med 2005;28:1–8.

9. Babyak M, Blumenthal JA, Herman S, et al. Exercise treatment formajor depression: maintenance of therapeutic benefit at 10 months.Psychosom Med 2000;62:633–8.

0. Powell LH, Catellier D, Freedland KE, et al; ENRICHD Investiga-tors. Depression and heart failure in patients with a new myocardialinfarction. Am Heart J 2005;149:851–5.

1. Richardson LG. Psychosocial issues in patients with congestive heartfailure. Prog Cardiovasc Nurs 2003;18:19–27.

2. Frasure-Smith N, Lesperance F. Depression and other psychologicalrisks following myocardial infarction. Arch Gen Psychiatry 2003;60:627–36.

3. Yu DS, Lee DT, Woo J. Health-related quality of life in elderlyChinese patients with heart failure. Res Nurs Health 2004;27:332–44.

4. Jiang W, Kuchibhatla M, Cuffe MS, et al. Prognostic value of anxietyand depression in patients with chronic heart failure. Circulation2004;110:3452–6.

5. Murberg TA, Bru E. Social relationships and mortality in patientswith congestive heart failure. J Psychosom Res 2001;51:521–7.

6. Rosenthal R, Rosnow RL. Essentials of Behavioral Research: Methods
and Data Analysis. New York, NY: McGraw-Hill, 2000.


doi:10.1016/j.jacc.2006.06.055 2006;48;1527-1537; originally published online Sep 25, 2006; J. Am. Coll. Cardiol.

J. Mills Thomas Rutledge, Veronica A. Reis, Sarah E. Linke, Barry H. Greenberg, and Paul

Intervention Effects, and Associations With Clinical OutcomesDepression in Heart Failure: A Meta-Analytic Review of Prevalence,

This information is current as of August 9, 2008

& ServicesUpdated Information

http://content.onlinejacc.org/cgi/content/full/48/8/1527including high-resolution figures, can be found at:

References

http://content.onlinejacc.org/cgi/content/full/48/8/1527#BIBLfree at: This article cites 87 articles, 34 of which you can access for

Citations

rticleshttp://content.onlinejacc.org/cgi/content/full/48/8/1527#otheraThis article has been cited by 4 HighWire-hosted articles:

Rights & Permissions

http://content.onlinejacc.org/misc/permissions.dtltables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,

Reprints http://content.onlinejacc.org/misc/reprints.dtl

Information about ordering reprints can be found online:


http://content.onlinejacc.org/cgi/content/full/48/8/1527

http://content.onlinejacc.org/cgi/content/full/48/8/1527#BIBL

http://content.onlinejacc.org/cgi/content/full/48/8/1527#otherarticles

http://content.onlinejacc.org/misc/permissions.dtl

http://content.onlinejacc.org/misc/reprints.dtl


depression in heart failure: a meta-analytic review …zunis.org/jacc sota 7-06 to 8-08/depression...

Documents