Is depression associated with increased oxidative stress? A systematic review and meta-analysis

(1)

Psychoneuroendocrinology(2015)51,164—175

Availableonlineatwww.sciencedirect.com

ScienceDirect

j o ur na l ho me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / p s y n e u e n

Is

depression

associated

with

increased

oxidative

stress?

A

systematic

review

and

meta-analysis

Catherine

N. Black

a,d,∗

,

Mariska

Bot

a,d

,

Peter

G. Scheffer

b

,

Pim

Cuijpers

c,d

,

Brenda

W.J.H.

Penninx

a,d

a_Department_of_Psychiatry,_VU_University_Medical_Center,_Amsterdam,_The_Netherlands b_Department_of_Clinical_Chemistry,_Metabolic_Laboratory,_VU_University_Medical_Center,

Amsterdam,TheNetherlands

c_Department_of_Clinical_Psychology,_VU_University_Amsterdam,_The_Netherlands d_EMGO_Institute_for_Health_and_Care_Research,_Amsterdam,_The_Netherlands

Received25July2014;receivedinrevisedform23September2014;accepted24September2014

KEYWORDS Depression; Majordepressive disorder; Bipolardisorder; Oxidativestress; 8-Hydroxy-2 -deoxyguanosine (8-OHdG); F2-isoprostanes Summary

Background:Ithasbeensuggestedthatdepressedpersonshaveincreasedoxidativestressand decreasedanti-oxidantdefences.8-Hydroxy-2-deoxyguanosine(8-OHdG)andF2-isoprostanes, measuresofoxidativeDNAandlipiddamagerespectively,areamongthemostreliableoxidative stressmarkers,butstudiesontheirassociationwithdepressionshowconﬂictingresults.This meta-analysisquantiﬁestheassociationbetweendepressionandthesemarkersandexplores factorsthatmayexplaininconsistenciesintheresults.

Methods:AsystematicliteraturesearchwasconductedinPubMed,EMBASEandPsycINFO. Stud-iesassessingtheassociationof8-OHdGorF2-isoprostaneswithelevateddepressivesymptoms, majordepressivedisorder(MDD)orbipolardisorder(BD)werepooled intwo random-effect models.

Results:Thepooledeffectsize(Hedges’g)fortheassociationofdepressionwithoxidativestress was0.31(p=0.01,I2₌_75%)_for_8-OHdG₍₁₀_studies,₁₃₀₈_subjects)_and_0.48_(p₌_0.001,_I2₌_73%)

forF2-isoprostanes (8studies,2471subjects),indicating thatbothmarkersareincreasedin depression.Therewasnoindicationofpublicationbiasforeithermarker.TheF2-isoprostane resultsdidnotdifferbytypeofdepression,biologicalspecimen,laboratorymethodorquality, howeversubgroupanalysesinthe8-OHdGstudiesshowedsigniﬁcantlystrongerassociationsin plasma/serumvs.urinesamples(p<0.01),inmeasurementsperformedwithimmuno-assayvs. chromatography—massspectrometry(p<0.01)andweakerassociationsinhighqualitystudies vs.low(p=0.02).

∗_{Corresponding}_author_at:_Department_of_Psychiatry,_VU_University_Medical_Center,_Postbus_74077,₁₀₇₀_BB_Amsterdam,_The_Netherlands. Tel.:+31207884597.

E-mailaddress:[email protected](C.N.Black).

(2)

Oxidativestressindepression:Ameta-analysis 165 Conclusion: This meta-analysis ﬁnds thatoxidative stress, as measured by 8-OHdG and F2-isoprostanes,isincreasedindepression.Larger-scalestudiesareneededtoextendtheevidence onoxidativestressindepression,andexaminethepotentialimpactoftreatment.

1. Introduction

Depressionis aleadingcauseof morbidityworldwide(Vos

etal.,2012).Depressionishighlyprevalent(Kessleretal., 2003)andhasaprofound impactonfunctioningand qual-ity of life (Bijl and Ravelli, 2000) as well as on somatic health.Sufferersareathigherriskofdiseasesthatare usu-allyassociated withincreasingage such ascardiovascular disease (Nicholson et al., 2006), obesity (Luppino et al., 2010),diabetes(Mezuketal.,2008),cancer(Chidaetal., 2008),cognitiveimpairment(Barnesetal.,2006)andhave ahigherall-causemortalityrate(Cuijpersetal.,2014).It ishypothesizedthatincreasedmetabolicstressand accel-erated cellular ageing may be underlying pathways that contributetothispoorerphysicalhealthinindividualswith depression(Wolkowitzetal.,2011b).Afastgrowingbodyof evidencesuggeststheinvolvementofaspeciﬁccomponent ofmetabolicstress,oxidativestress,inthepathophysiology ofdepression(Maesetal.,2011).

Oxidative stress refers to the biologically damaging effects of free radicals(Valko et al., 2007). The produc-tionof free radicals,or reactive oxygen species(ROS), is a normalprocess inaerobic metabolismand ROS perform anumberofphysiologicalrolesincellularsignallingandin thedefence againstpathogens. However,when presentin excess,ROScausedamagetolipids,proteinsandDNA,and canultimatelyresultincelldeath.Oxidativestressisa well-recognizedmechanism in ageing and disease. Ithas been showntoplayaroleinthepathophysiologyof—among oth-ers—cardiovasculardisease,diabetesmellitus,cancerand Alzheimer’sdisease(Valkoetal.,2007).Additionally,there isevidencesuggestingthatoxidativestressmaybeincreased in anumber ofpsychiatricdisorders,includingdepression (Pandyaetal.,2013).

Arecentmeta-analysispoolingdatafromstudieswith dif-ferentoxidative stressmarkerssuggestsoxidativestressis increasedandantioxidantdefencesaredecreasedin depres-sion(Paltaetal.,2014).Inlinewiththeseﬁndings,increased nitric oxide (NO) and lipid peroxidation, as measured by thiobarbituricacidicreactivesubstance(TBARS)assay,have alsobeen foundinpatientswithbipolardisorder, however thesepatients didnotdifferfromcontrols inanti-oxidant enzymes levels (Andreazza et al., 2008). Overall, these studies suggestthat oxidative stressis increasedin major depressivedisorderandbipolardisorder.

Thereisawiderangeofoxidativestressbiomarkersand laboratorytechniquesavailable,eachofwhichhasitsown strengthsandlimitations(Dalle-Donneetal.,2006).Todate thereisnoconsensusonthe mostappropriatebiomarkers of oxidative stress in general andthe validity ofmany of thoseinuseistostillbeestablished.ROShaveashort half-life,makingmeasurementdifﬁcult.Levelsofantioxidants,

vitamins or anti-oxidant enzymes are informative, but reflect only one side of redox homeostasis, leaving the question unanswered whether decreased levels are actu-allyalsoindicative ofincreasedoxidativedamage.Studies showquiteconsistentlythatlipidperoxidationreflectedby malondialdehyde(MDA)measured withtheTBARSassayis increasedindepression (Paltaetal., 2014)andin bipolar disorder(Andreazzaetal.,2008).Howeverthiscommonly usedmethodalsohaswellrecognizedlimitations:MDAisnot aspecificproductoflipidperoxidation,andtheTBARSassay itselfcangenerateMDA,causing overestimationof levels. MDAtherefore cannotbeconsidered anoptimal represen-tationofoxidativestress invivo(MeagherandFitzGerald, 2000;Dalle-Donneetal.,2006).

The current studyfocusses ontwoimportant measures of oxidative damage thathave alreadybeen widely stud-iedinsomaticdiseaseandarethesubjectofanincreasing numberofrecentpublicationsondepression:8-hydroxy-2 -deoxyguanosine(8-OHdG)andF2-isoprostanes.Themajority of the currently available literature on these markers in depression was not included, or not yet available for inclusion, in the previous meta-analyses on this subject. 8-OHdG and F2-isoprostanes reﬂect oxidative damage to DNAandlipidsrespectively. 8-OHdGis anoxidized deriva-tive of deoxyguanosine and it is both the most abundant andmostinvestigatedDNAlesion.Ithasrecognized muta-genicpropertiesandhasbeenlinkedto—amongothers— thedevelopmentof cancer (Valavanidiset al.,2009). F2-isoprostanes,oxidizedderivativesofarachidonicacid,have come to be considered the preferred approach to assess oxidativestressinvivoandlipidperoxidationinparticular (Niki,2014).

Several studies have found elevated levels of F2-isoprostanes(Dimopoulos etal., 2008;Yageret al.,2010; Chungetal.,2013)and8-OHdG(Irieetal.,2005;Forlenza andMiller,2006)inpatientswithdepression,butthese find-ingshavenotbeenconsistent(Yietal.,2012;Rawdinetal., 2013).Earlierstudiesdidnotsystematicallyexploretowhat extent the (conflicting) findings are due to e.g. the lab-oratory methods, biological specimens used for oxidative stress,or the extent towhich studiestook potential con-founderssuchashealthandlifestylefactorsintoaccount. The present study extends the current evidence-base by systematicallymeta-analysingtheliteratureontworobust markers of oxidative stress, 8-OHdG and F2-isoprostanes, andtheirassociationwithdepression(majordepressive dis-order,bipolardisorderandelevateddepressivesymptoms). Inaddition,byconductingsubgroupanalysesbasedontype ofdepression,biologicalspecimen,laboratorymethodused tomeasureoxidativestress,correctionforconfoundersand thequalityofstudies,thisstudyaimstoidentityfactorsthat contributetotheinconsistentfindingsofindividualstudies.

(3)

166 C.N.Blacketal.

2. Methods

2.1. Literaturesearchandstudyselection

Systematicsearchesoftheliteraturewereconductedinthe databasesPubMed,EMBASEandPsycINFOuptoJanuary8th 2014withsearchtermscoveringmajordepressivedisorder, bipolardisorderanddepressivesymptomscombinedwith 8-OHdGandF2-isoprostanesrespectively,takingintoaccount awiderangeofsynonymsusedforthesemarkers.Afulllist ofsearch terms is reported in AppendixA.No limitations in the search strategy were set. The search results were reviewedby two independentreviewers (CNB andMB) by screeningtitleandabstract,followedbyafulltextreview. Disagreementsweresettledbydiscussion.

Studieswereeligibleforinclusionifthey:

(1) containedameasurementof8-OHdGorF2-isoprostanes inanybodyﬂuidortissueinlivehumanadultsubjects; (2) deﬁnedmajordepressivedisorder(MDD)orbipolar dis-order (BD) according to DSM-IV or ICD-10 criteria or assesseddepressivesymptomsusingavalidated instru-ment;

(3) reported(orwereabletoprovide)sufﬁcientinformation tocalculatean effectsizefor thedifferencebetween levelsoftheoxidativestressmarkersincontrolsubjects andsubjectswithdepression.

Anassessmentofthereferencesoftheincludedstudies aswellasasearchoftheircitationsinthePubMeddatabase wasperformedtoidentifyanyadditionalstudies.

2.2. Dataextractionandriskofbiasassessment

Two authors (CNB and MB) independently extracted the study characteristics (including among others biological specimen,age andsexdistribution)andresultsfromeach studyusingapredesignedcollectionform(seeAppendixB). Theresultsextractedincludedmeans, standarddeviations of oxidative stress markers in patient and control groups (oralternativeresultssufficienttocalculateaneffectsize) andan assessmentofcorrectionfor potentialconfounders (age,sex, ethnicity,socio-economicstatus,smoking, alco-holuse,body mass index(BMI),physical activity, somatic diseaseandantidepressantormoodstabilizeruse).Authors ofstudiesthatdidnotcontainsufficientinformation(means, standarddeviationsor standarderrorsandnumberof sub-jects) to calculate an effect size were contacted with a request for additional data. Authors of studies who only reportedresultsondepressivesymptomsascontinuous vari-ableswererequestedtoperformadditionalanalysesusinga dichotomizedclassificationofadepressedandcontrolgroup basedontheappropriatecut-offfortheinstrumentusedto assessdepressivesymptoms.Authorsofstudieswhose origi-nalarticlesdidnotcontainadjustedresultswerecontacted witharequesttoperformadditionalanalysesadjustingfor as many of the following confounders if available: age, sex, socio-economic status or education, ethnicity, smok-ing,alcoholuse,BMI;physicalactivity,presenceofsomatic diseaseandantidepressantor moodstabilizeruse.If mul-tiple adjusted analyses had been conducted, the results

corrected forthelargestnumberofpotentialconfounders wereincludedinthemeta-analysis.

Thequalityofthestudies wasdeterminedby assessing theriskofbiasbytwoindependentreviewers(CNBandMB) inthreedomains:selectionbias,informationbiasand con-founding.Disagreementsweresettledbydiscussion.Riskof biasassessmentinobservationalstudiesrequires consider-ationof therisksspeciﬁctothesubjectofstudy.Forthis purpose we adapted a tool by Hayden et al. (2006) (see AppendixB).Lowriskofbiaswasdeﬁnedasascoreof4.5 orhigher(ona5pointscale).

2.3. Meta-analysis

Analyseswereperformedwithcomprehensivemeta-analysis (CMA) software version 2.2.064. Effect sizes (Hedges’ g)

werecalculatedandpooledusingarandomeffectsmodel,as considerableheterogeneity wasexpected.Ap-value<0.05 wasconsideredsigniﬁcant.

Possible publication bias was tested by inspecting the funnel plot, by the statistical signiﬁcance of the Egger’s testoftheintercept(SterneandEgger,2001)andDuvaland Tweedie’s(2000)trimandﬁllprocedure.

Toexamineheterogeneity,theI2_-statistic_was_calculated.

Values of 25%, 50% and 75% indicate low, moderate and highheterogeneity,respectively(Higginsetal.,2003).The 95% conﬁdenceintervals aroundI2 ₍_Ioannidis_et_al., ₂₀₀₇₎

were calculated using the non-central chi-squared-based approachwithintheheterogimoduleforSTATAversion11.0 forMac.

When a minimum of three samples per subgroup was available,analyseswereperformedcategorizedby:typeof depression(MDD,BD,depressivesymptomsabovethe ques-tionnaire cut-off level), laboratory method for measuring oxidative stress(chromatography[coupled toeither mass-spectrometry or electrochemical detection] vs. immuno-assay),biologicalspecimenused(urine,blood[product]or other), lowor highriskof biasscore,andadjustedversus unadjustedresults.Studiesthataccountedforatleastage, sexandonelife-stylevariable(eitherbymatching, restric-tion,adjustmentortestingfor baselinegroupdifferences) weredeﬁnedascorrectedforconfounding.

3. Results

3.1. Searchresults,studycharacteristicsandrisk

ofbiasscoring

3.1.1. 8-OHdG

Afterremovalofduplicates,79recordswereassessedbased ontitleandabstract.The full-textof 39 oftheserecords wasretrievedforfurtherassessment.Ofthesestudies6did notreportsufﬁcientinformationintheoriginalarticlesto calculatean effectsizeandnonewasprovidedbyauthors onrequest(Irieetal.,2001,2002,2003;Maesetal.,2009; Iidaetal.,2011;Ceylan etal.,2013).In2studiesboth 8-OHdG and depression were measured but the association betweenthetwowasnotthefocusofstudyandtherefore not analyzed (Wu etal.,2009; Ceprnja et al.,2011), nor wasthisdataprovidedonrequest.Intotal10studiesmet

(4)

Oxidativestressindepression:Ameta-analysis 167

8-OHdG in depression

F2-isoprostanes in depression

PubMed 41 records EMBASE 73 records PsycInfo 15 records

129 records in total 79 records after duplicates removed

79 records screened on title and abstract

40 records excluded:

21 no humans 12 no depression 3 no 8-OHdG 3 review 1 no adults

39 full-text articles assessed for eligibility

10 studies included in the

meta-analysis

29 records excluded:

2 no 8-OHdG 8 no depression 2 post-mortem 3 duplicate data 5 review 8 no data on association depression and 8-OHdG

291 records in total 221 records after duplicates removed

221 records screened on title and abstract

191 records excluded: 115 no humans 65 no depression 6 no F2-isops 4 review 1 post-mortem

30 full-text articles assessed for eligibility

8 studies included in the meta-analysis 22 records excluded: 8 no F2-isoprostanes 3 no depression 1 duplicate data 8 review/letter 2 no data on association depression and F2-isops or insufficient data for effect size Identific a ti o n Screenin g E ligibi li ty Include d Identific a ti o n Screenin g E ligibi li ty Include d Pubmed 140 EMBASE 145 PsycInfo 6

Figure1 Flowchartoftheselectionofstudieson8-OHdGor

F2-isoprostanesindepression(majordepressivedisorder,

bipo-lardisorderanddepressivesymptoms).

allcriteriaforinclusion;4studiesondepressivesymptoms (Kupperetal.,2009;Weietal.,2009a,b;Yietal.,2012),2 studiesonMDD(Irieetal.,2005;ForlenzaandMiller,2006), 3studiesonBD(Ceylanetal.,2012;Soeiro-de-Souzaetal., 2013;Huzayyinetal.,2014)and1studywithbothMDDand BD patients (Jorgensen et al., 2013) (Fig. 1). Two of the included studies (Yi et al., 2012; Jorgensen etal., 2013) providedadditionaldataonrequest.

The10studies(Irieetal.,2005;ForlenzaandMiller,2006; Kupper etal., 2009; Wei et al.,2009a,b; Yiet al.,2012; Ceylanetal.,2012;Soeiro-de-Souzaetal.,2013;Jorgensen etal.,2013;Huzayyinetal.,2014)included atotalof579 subjects with depression (332 with depressive symptoms [scoringabovethecut-offoftheinstrumentused],141with MDD, 106 withBD) and 729 controls. All included studies werepublishedbetween2005and2014.Thestudiesinclude samplesfromthegeneralpopulation(N=2),psychiatric in-andoutpatientclinics(N=4),hospitaloncology(N=2)and heart failuredepartments(N=1)and1 unreportedsource

(Table 1). Riskof biaswasscored aslow in 3studies and highin7studies(AppendixC).

3.1.2. F2-isoprostanes

After removal of duplicates, 221 records were assessed based on title and abstract. The full-text of 30 of these recordswasretrievedforfurtherassessment.Ofthese stud-ies1 did not reportsufﬁcient informationin the original articletocalculateaneffectsizeandnonewasprovidedby authorsonrequest(Freund-Levietal.,2011).Inonestudy bothF2-isoprostanesanddepressionweremeasuredbutthe associationbetweenthetwowasnotthefocusofstudyand therefore not analyzed nor provided on request ( Janicki-Devertsetal.,2009).ThesamplesofRawdinetal.(2013) andWolkowitzetal.(2011a)overlap,leaving8original sam-pleseligibleforinclusion,3ondepressivesymptoms(Chung etal.,2009;Segaletal.,2012;Milaneschietal.,2013)and 5onMDD(Fig.1).Sixoftheincludedstudiesprovided addi-tionaldataupon request(Chungetal., 2009,2013;Yager etal., 2010; Wolkowitzet al.,2011a; Segal etal., 2012; Milaneschietal.,2013).

The 8 studies (Dimopoulos et al., 2008; Chung et al., 2009, 2013; Yager et al., 2010; Wolkowitz et al., 2011a; Segaletal., 2012;Pomaraetal.,2012; Milaneschietal., 2013)includeatotalof293subjectswithdepression(144 subjectswithdepressivesymptoms,149withMDD)and2178 controls.Allincludedstudieswerepublishedbetween2008 and2013. The studiesinclude samplesof adultsfromthe generalpopulation(N=2),elderlyadultsformthegeneral population(N=3),psychiatricoutpatients(N=1),systemic lupuserythematosus(SLE)patients(N=1)andﬁbromyalgia patients(N=1)(Table2).Riskofbiaswasscoredaslowin3 studiesandhighin5studies(AppendixC).

3.2. Meta-analyses

3.2.1. 8-OHdG

Theoveralleffectsize(Hedges’g)includingall10studieson 8-OHdGinthemeta-analysiswas0.31(95%CI0.06,0.56),

I2₌_75% _(95% _CI_58—86%)_indicating _8-OHdG _is _{signiﬁcantly}

increasedin depression(Fig.2aand Table3).Egger’stest forpublicationbiaswasnotsigniﬁcant(p=0.69).Theeffect sizeadjusted for publicationbias by Duval andTweedie’s trim and ﬁll procedure (two trimmed studies) increased marginallyto0.37 (95%CI0.13,0.61)indicating noor lit-tleeffects of publicationbias (seeAppendix D for funnel plot).

Subgroupanalysesinthe8-OHdG studiesshowed signif-icantlylargereffectsizes forstudiesthatwereconducted inplasmaorserumvs.thosein urine(p<0.001)and stud-iesthatwereperformedwithimmuno-assaysvs.thosewith chromatography(p=0.006).Studieswithalowriskofbias showedasigniﬁcantlylowereffectsizecomparedto stud-ieswithahigherriskofbias(p=0.02).Subgroupanalysesby typeofdepression(depressivesymptomsorBDvs.MDD)and resultscorrectedanduncorrectedforconfoundersrevealed nosigniﬁcantdifferences(Table3).

3.2.2. F2-isoprostanes

The overall effect size (Hedges’ g) including all 8 stud-iesonF2-isoprostanes in themeta-analysiswas0.48 (95%

(5)

168

C.N.

Black

et

al.

Table1 Studycharacteristicsoftheincludedstudiesontheassociationof8-OHdGwithmajordepressivedisorder,bipolardisorderanddepressivesymptoms.

8-OHdG Study design Sample description Depressive disorder Patients N Controls N Diagnostic method Biological specimen 8-OHdG analysis Confoundersd _Overall correction confounding Effectsize Hedges’g(95%CI) Socio demo-graphics Life-style Somatic disease AD\mod stabiliz-ers Forlenzaand Miller(2006) CC General population

MDD 62 85 DSM-IV(DISH) Serum ELISA + + + + + 0.47(0.14,0.80)

Irieetal.(2005) CC Psychiatric outpatients

MDD 30 60 DSM-IV LeukocytesHPLC/ED − − − − − 0.76(0.31,1.21)

Jorgensen etal. (2013)

CC Psychiatric department

MDDa ₂₆ ₂₇ _DSM-IV_(MINI) _Urine _UPLC/MS _± _± ₋ ₊ ₊ _−0.07₍_−0.81,_0.67)

CC Psychiatric department MDDb ₂₃ ₂₇ _ICD-10 psychiatrist Urine UPLC/MS ± ± + − + −0.10(−0.85,0.65) CC Psychiatric department BDc ₆ ₂₇ _ICD-10 psychiatrist Urine UPLC/MS ± ± + − + −0.15(−1.13,0.82) Ceylanetal. (2012)

CC Notreported BD 36 14 DSM-IV(SCID) Bloodnos GC/MS − − − − − −0.23(−0.83,0.38)

Soeiro-de-Souza etal.(2013)

C Psychiatric outpatients

BD 50 50 DSM-IV(SCID) Plasma ELISA ± − − + − 1.15(0.74,1.57)

Huzayyinetal. (2014)

CC Specialist psychiatric clinic

BD 14 16 DSM-IV(SADS-L) Lymph-oblasts ELISA ± − − − − 0.17(−0.53,0.87) Kupperetal. (2009) C Heartfailure outpatients

DS 38 72 BDI≥10 Serum ELISA − − − − − 0.14(−0.25,0.53)

Weietal.(2009a) CC Oncology department

DS 52 30 HAM-D≥20 Serum ELISA ± ± + − + 0.81(0.35,1.28)

Weietal.(2009b) CC Oncology department

DS 63 43 HAM-D≥20 Serum ELISA + − + − + 0.64(0.25,1.04)

Yietal.(2012)

Males C Municipal workers

DS 105 196 CES-D≥16 Urine HPLC/ED ± + − + + −0.01(−0.25—0.23) Females C Municipal

workers

DS 74 136 CES-D≥16 Urine HPLC/ED ± + + + + −0.09(−0.37,0.19) BD,bipolardisorder;BDI,BeckDepressionInventory;C,cohort;CC,case—control;CES-D,CentreforEpidemiologicalStudiesDepressionScale;DISH,DepressionInterviewandstructured HamiltonInterview;DS,depressivesymptoms;DSM-IV,DiagnosticandStatisticalManualofMentalDisordersIV;ELISA,enzyme-linkedimmunosorbentassayGC/MS,gaschromatography/mass spectrometryHAMD,HamiltonDepressionRatingScale;HPLC/ED,high-performanceliquidchromatography/electrochemicaldetector;ICD-10,InternationalStatisticalClassiﬁcationof DiseasesandRelatedHealthProblems;MDD,majordepressivedisorder;MINI,MiniInternationalNeuropsychiatricInterview;NOS,nototherwisespeciﬁed;SADS-L,ScheduleforAffective DisordersandSchizophrenia,lifetimeversion;SCID,structuredclinicalinterviewforDSM-IV;UPL/MS,ultraperformanceliquidchromatography/massspectrometry.

a _Subsample_of_patients_in_Jorgensen_et_al.₍₂₀₁₃₎_described_in_the_article_as_M-DEP_(moderately_depressed)_patients.

b _Subsample_of_patients_in_Jorgensen_et_al.₍₂₀₁₃₎_described_in_the_article_as_S-DEP_(severely_depressed_patients_unipolar_only). c _Subsample_of_patients_in_Jorgensen_et_al.₍₂₀₁₃₎_described_in_the_article_as_S-DEP_(severely_depressed,_bipolar_only).

d _{Socio-demographics}_include_age,_sex,_{socio-economic}_status_(income,_education_or_other),_ethnicity._Lifestyle:_smoking,_alcohol,_BMI,_physical_activity._Somatic_disease: _presence

of(chronic)disease(cardiovascular,infectious,auto-immuneormalignancy)thatmayinﬂuenceoxidativestresslevels.Antidepressant/moodstabilizers:currentuse.+,confoundersis accountedfor;−,confounderisnotaccountedfor;±,some,butnotalloftheconfoundersinthecategoryhavebeenaccountedfor.Studiesaredeﬁnedascorrected(+)ifthehave accountedforage,sexandatleastonlife-stylevariable.

(6)

Oxidative stress in depression: A meta-analysis 169

Table2 StudycharacteristicsoftheincludedstudiesontheassociationbetweenF2-isoprostanesandmajordepressivedisorderanddepressivesymptoms.

F2-isops Study design Sample description Depressive disorder

PatientsN ControlsN Diagnostic method Biological specimen F2-isoprostane analysis Confoundersa _Overall correction confounding Effectsize Hedges’g(95%CI) Socio demographics Life-style Somatic disease AD\mood stabilizers Chungetal. (2013) CC General population MDD 18 36 DSM-IV (SCID) Urine GC/MS ± ± + + + 1.12(0.53,1.72) Dimopoulos etal.(2008) CC General population>60 years MDD 33 33 DSM-IV psychiatrist Plasma ELISA ± ± + + + 1.11(0.60,1.63) Pomaraetal. (2012) CC General population>60 years MDD 28 19 DSM-IV (SCID) CFS ELISA ± ± − − + 0.88(0.28,1.48) Yageretal. (2010) CC General popluation MDD 57 74 DSM-IV (DISH) Serum ELISA + + + + + 0.45(0.10,0.80) Wolkowitz etal.(2011a) CC Psychiatric outpatients MDD 13 14 DSM-IV (SCID) Plasma GC/MS + + + + + −0.16(−0.89, 0.58) Chungetal. (2009) CC Fibromyalgia patients DS 28 20 CES-D≥16 Urine GC/MS − − + − − 0.37(−0.20,0.94) Milaneschietal.(2013) Males C General population 70—79years DS 31 996 GDS≥5 and/orAD use Urine RIA + + + − + 0.30(−0.06,0.66)

Females C DS 52 896 Urine RIA + + + − + −0.02(−0.30,

0.26)

Segaletal. (2012)

CC SLEpatients DS 33 90 CES-D≥16 Plasma GC/MS ± ± − − + 0.38(−0.02,0.78)

AD,antidepressant;C,cohort;CC,case—control;CES-D,CentreforEpidemiologicalStudiesDepressionScale;CFS,cerebrospinalﬂuid;DISH,DepressionInterviewandstructuredHamilton Interview;DS,depressive symptoms;DSM-IV,Diagnosticand StatisticalManualofMentalDisordersIV;ELISA,enzyme-linkedimmunosorbentassay;GC/MS,gaschromatography/mass spectrometry;GDS,geriatricdepressionscale;HAMD,HamiltonDepressionRatingScale;ICD-10,InternationalStatisticalClassiﬁcationofDiseasesandRelatedHealthProblems;MDD, majordepressivedisorder;RIA,radioimmunoassay;SCID,structuredclinicalinterviewforDSM-IV;SLE,systemiclupuserythematosus.

a _{Socio-demographics}_include_age,_sex,_{socio-economic}_status_(income,_education_or_other),_ethnicity._Lifestyle:_smoking,_alcohol,_BMI,_physical_activity._Somatic_disease: _presence

of(chronic)disease(cardiovascular,infectious,auto-immuneormalignancy)thatmayinﬂuenceoxidativestresslevels.Antidepressant/moodstabilizers:currentuse.+,confoundersis accountedfor;_−,confounderisnotaccountedfor;_±,some,butnotalloftheconfoundersinthecategoryhavebeenaccountedfor.Studiesaredeﬁnedascorrected(+)ifthehave accountedforage,sexandatleastonlifestylevariable.

(7)

170

C.N.

Black

et

al.

Table3 Resultsofmeta-analysesontheassociationof8-OHdGandF2-isoprostaneswithdepressivesymptoms,majordepressivedisorderandbipolardisorder.

8-OHdG Nstudies Hedges’g 95%CI p I2 _95%_CI

Allstudies(depressivesymptoms,MDD,BD) 10 0.31 0.06,0.56 0.01 75% 58—86%

Egger’stestforpublicationbiasa ₁₀ _0.69

Subgroupanalyses Nsamples Hedges’g 95%CI pf _I2 _95%_CI

Depressionb _Depressive_symptoms ₄ _0.27 _−0.06,_0.60 _0.67 _78% _48—91%

MDD 3 0.37 0.01,0.74 Reference 48% 0—83%

BD 4 0.28 _−0.48,1.05 0.83 83% 55—93%

Biologicalspecimenc _Urine ₂ _−0.05 _−0.22,_0.12 _0% _0—79%

Plasma/serum 6 0.52 0.18,0.87 <0.01 75% 43—89%

Laboratorymethod Chromatography 4 0.05 −0.20,0.31 50% 0—79%

Immunoassay 6 0.58 −0.28,0.88 <0.01 66% 19—86%

Correctionforconfounders Unadjusted 5 0.43 _−0.07,0.92 80% 53—92%

Adjusted 5 0.23 −0.04,0.51 0.50 69% 36—85%

Riskofbiasscored _{‘‘Low’’}_risk_of_bias ₃ _0.06 _−0.15,_0.27 _35% _0—74%

‘‘High’’riskofbias 7 0.53 0.19,0.87 0.02 72% 40—87%

F2-isoprostanese _N_studies _Hedges’_g _95%_CI _p _I2 _95%_CI

Allstudies(depressivesymptoms,MDD) 8 0.48 0.19,0.77 0.001 73% 47—86%

Egger’stestforpublicationbiasa ₈ _0.13

Subgroupanalyses Nsamples Hedges’g 95%CI pf _I2 _95%_CI

Depression Depressivesymptoms 3 0.24 −0.05,0.53 56% 0—86%

MDD 5 0.70 0.28,1.12 0.08 66% 13—87%

Biologicalspecimenc _Urine ₃ _0.41 _−0.06,_0.89 _80% _48—92%

Plasma/serum 4 0.49 0.08,0.89 0.82 67% 3—89%

Laboratorymethod Immunoassay 4 0.51 0.10,0.92 81% 57—92%

Chromatography 4 0.45 0.00,0.90 0.84 61% 0—87%

Correctionforconfounders NA

Riskofbiasscored _{‘‘Low’’}_risk_of_bias ₃ _0.51 _−0.10,_1.11 _74% _40—89%

‘‘High’’riskofbias 5 0.46 0.11,0.81 0.90 77% 24—93%

BD,bipolardisorder;DS,depressivesymptoms;MD,majordepressivedisorder;NA,notapplicable;noorinsufﬁcientstudiestoperformthesubgroupanalysis.

a _See_Appendix_D_for_funnel_plots.

b _Jorgensen_et_al.₍₂₀₁₃₎_is_counted_in_two_categories_because_it_describes_stratiﬁed_analyses_of_both_MDD_and_BD_patients.

c _studies_using_specimens_from_others_sources_were_excluded_from_this_analysis₍_Irie_et_al.,_2005;_Huzayyin_et_al.,_2014;_Pomara_et_al.,₂₀₁₂_). d _See_Appendix_C_for_risk_of_bias_scores.

e _Results_of_sensitivity_analyses,_excluding_Pomara_et_al.₍₂₀₁₂₎_with_specimens_from_CSF_{(cerebrospinal}_ﬂuid):_all_studies:_Hedges’_g_0.44,_p_0.004;_Egger’s_test:_p_0.22._MDD:_Hedges’_g

0.66,pbetweensubgroups0.17;immunoassay:Hedges’g0.44,pbetweensubgroups0.98,‘‘High’’riskofbias:Hedges’g0.39,pbetweensubgroups0.75.

(8)

Figure2 (a)Meta-analysis of8-OHdGindepression(major

depressivedisorder[MDD],bipolardisorder[BD]anddepressive

symptoms[DS])witheffectsizesHedges’gand95%conﬁdence

intervalsforthecomparisonof8-OHdGlevelswithcontrols.(b).

Meta-analysisofF2-isoprostanesindepression(major

depres-sivedisorder[MDD]anddepressivesymptoms[DS])witheffect

sizesHedges’gand95%conﬁdenceintervalsforthecomparison

ofF2-isoprostanelevelswithcontrols.

CI 0.19, 0.77), I2₌_73% _(95% _CI _47—86%), _indicating _that

F2-isoprostanes are significantly increased in depression (Fig. 2b and Table 3). Egger’s test was not significant (p=0.13)andtheeffectsizeadjusted forpublicationbias byDuvalandTweedie’strimandfillprocedure(notrimmed studies)wasunchanged,indicatingnosignificanteffectsof publicationbias(seeAppendixDforfunnelplot).

Asensitivityanalysiswasperformedexcludingonestudy (Pomara et al., 2012) as this study measured oxidative stresscentrally(incerebrospinalfluid)asopposedto periph-eral measurements (blood and urine) used in all others. Exclusion of this study did not affect the overall results (Hedges’ g0.44, p=0.004;Table 3). Subgroup analysesin F2-isoprostanes showeda trend level differencebetween studiesassessingdepressivesymptomsvs.MDDwiththeMDD studiesshowingalargereffectsize(p=0.08).Subgroup anal-yses by biological specimen used, laboratory method for oxidativestressmeasurementandriskofbiasscorerevealed nosignificant differences (Table 3). Nosubgroup analyses wereperformedbasedoncorrectionforconfoundersasonly onestudywasdefinedasuncorrected.

4. Discussion

This meta-analysisfound that both oxidative stress mark-ers,8-OHdGandF2-isoprostanes,areincreasedinsubjects withdepression(majordepressivedisorder,bipolardisorder anddepressivesymptoms)comparedtocontrols,witheffect sizesinthesmalltomoderaterange(Cohen,1988).In addi-tion,subgroupanalysesofthe8-OHdGstudiesrevealedthat someof the variation in theresults may be explainedby thetypeofthebiologicalspecimenand/or thelaboratory methodforoxidativestressmeasurement,andthatstudies withlowerriskofbiasreportedsigniﬁcantlysmallereffect sizes.Findings for F2-isoprostanes however, didnotdiffer whenanalyzedbytypeofdepression,biologicalspecimen, laboratorymethodorquality.

The results indicate that depression is associated with increased oxidative damage to DNA and lipids (reflected by 8-OHdG and F2-isoprostanes respectively). These find-ings are in line with previous meta-analytic studies that reporteddecreasedanti-oxidantsandanti-oxidantenzymes in unipolar depression, and increased oxidative stress in both uni- andbipolar depression (Andreazza etal., 2008; Paltaetal., 2014).Previously DNA damagein bipolar dis-orderhas been reported measured by increasedlevels of DNA strand breakage with the comet assay (Andreazza et al., 2007). Oxidative damage to proteins determined bythe protein carbonylassayhas notyet been aswidely studied, but increased levels have been reported in MDD andBD (Kapczinski etal., 2011; Magalhaes etal., 2012), howevernot consistently (Andreazza et al.,2009; Gubert et al., 2013). The oxidative stress markers used in this meta-analysisareamongthemostrobustmarkersof oxida-tive stress currently available and have recognized roles in the pathophysiology various somatic diseases such as cardiovascular disease, cancer and diabetes (Dalle-Donne etal.,2006;Valavanidisetal.,2009).8-OHdGisa biologi-callyimportantmutagenicDNAlesion,whileF2-isoprostanes areknown to be increased in atherosclerotic lesions and may also be biologically active in the pathogenesis of atherosclerosis (Ho et al., 2013). It should benoted that although both markers reflect oxidative damage they are not necessarily associated with each other. One study included in our meta-analysis (Yager et al., 2010) found nocorrelation between the two, suggesting that 8-OHdG and F2-isoprostanes reflect specific aspects of oxidative imbalance.

Although the main findings of this study are unlikely tobegreatly influencedby publicationbias, considerable heterogeneitybetweenstudieswasfoundforboth8-OHdG andF2-isoprostanes.This heterogeneitymay beexplained byseveral factors. In the studies that examined 8-OHdG, patients in three bipolar disorder studies (Ceylan et al., 2012;Soeiro-de-Souza etal., 2013;Huzayyinetal.,2014) werenotallcurrentlydepressed,butsomewereinamanic oreuthymicstate. Thismayaccountfor thefactthattwo ofthesestudies’resultshavethehighestandlowesteffect sizes included in the meta-analysis. In addition, the sub-groupanalyses in8-OHdGstudiesdemonstratedsignificant differences between studies based on risk of bias, with thestudieswithlowerriskof biasfindingsmallereffects. There was also a significant difference in effect size in the subgroup analyses by biological specimen (higher in

(9)

172 C.N.Blacketal. plasma/serumthaninurine)andbylaboratorymethodused

tomeasureoxidativestress(higherinimmuno-assaysthan chromatography).Withonlyoneexceptionhowever,allthe measurements using chromatography were done in urine samples,whereasallthe immuno-assaysfor the measure-mentof8-OHdGinweredoneinplasmaorserum.Therefore, itcannot be determinedwhether thisdifferenceis based onthebiologicalspecimenorlaboratorytechnique.8-OHdG levelsdeterminedbyimmuno-assaysarehigherthanthose measuredbychromatography,withthelatterconsideredthe goldstandard.Thecorrelationbetweenthetwomethodsis generallyhigh(Yoshidaetal.,2002)andsomaynot neces-sarilyaffectthestrengthoftheassociationwithdepression. Ithasbeen reportedthat urinary8-OHdGlevelsaremore stablethan thosein plasmaor serum, andtherefore pos-siblymorereliable(Matsumotoetal.,2008).Althoughthe subgroup analyses may help to explain the heterogeneity observed,theﬁndingsshouldbeinterpretedwith consider-ablecaution,asthenumberofstudiesinthesubgroupsis small.

Thismeta-analysisconfirmsoxidativestressmarkersare increasedinsubjectswithdepressionincross-sectional stud-ies,buttheunderlyingmechanismsexplainingthislinkneed tobeexaminedfurther.Manybehavioural factorsthatare relatedtoincreasedexposuretoROS(smoking,alcoholuse, overweight,physical activity)(Maritim etal., 2003; Dalle-Donneetal.,2006;Valkoetal.,2007)arealsoassociated withdepression (Glassmanetal., 1990; Abu-Omar etal., 2004;Sullivanetal.,2005;Luppinoetal.,2010).Itcannot beruledoutthattheassociationispartiallydrivenbythese orotherlifestyleconfounders.Themajorityoftheincluded studiestooksome,butfewtookallofthesepotential con-foundersintoaccount.Theresultsofstudiesthatdidcorrect forconfoundersdidnotdiffersignificantlyfromstudiesthat didnot,strengtheningthe observationthatan association between depression and oxidative stress is present inde-pendentoftheselife-stylefactors.Theobservedincreased levels of oxidative stress in depression might be under-stoodwithintheconceptsofallostasisandallostaticload. Theformerreferstothephysiologicaladaptationto physi-cal,psychological,socialandenvironmentalstressors.The latterrefers tothephysical ‘‘wearand tear’’induced by prolonged exposure to the stress response (McEwen and Wingfield,2003).Depression hasbeen foundtobe associ-atedwithincreased‘‘wearandtear’’oracceleratedcellular ageingreflectedbydecreasedtelomerelengthindepressed patients (Verhoeven et al., 2013). Dysregulations in the major stress systems (hypothalamic-pituitary-adrenal axis activity, autonomic nervous system function and inflam-matory functions) have been demonstrated in depression (Penninxet al.,2013) and thesecould becontributing to increasedoxidativestress.Oxidativestressiscloselyrelated totheinflammatorypathwayinparticular.Pro-inflammatory cytokinesareproducedin reactiontooxidative stressand oxidativestressinturnamplifiestheinflammatoryresponse. High cortisol levels have been associated with increased levelsofoxidative damage(Joergensenetal.,2011).The damage caused by this allostatic load experienced dur-ingmoodepisodesishypothesized torenderanindividual morevulnerabletodeveloping afollowing episode andat higherrisktodevelop somaticdisease(Maes etal.,2011; Grandeetal.,2012).Thebrainisparticularlyvulnerableto

oxidative damage due to its high oxygen consumption and low anti-oxidants defences. There is evidence from post-mortem studies suggesting that in depression oxida-tive stress is increased (Wang et al., 2009; Che et al., 2010;Micheletal.,2012)andanti-oxidantsaredecreased (Gawryluketal.,2011)inthebrain.

Thismeta-analysis’strengthliesin thefocusonrobust oxidative stress markersand inthe comprehensive search thatwasconductedtoidentifyallstudiesontheassociation of 8-OHdG and F2-isoprostanes with depression. Through additional data requests the number of studies eligible for inclusion was increased,and where possible adjusted datawasobtained,mitigatingtheeffectsofpossible pub-lication bias and increasing the reliability of the overall results.

Therearealsoanumberofimportantlimitations.As is apparentfromthenumberofincludedstudiesandsubjects theevidenceon8-OHdGandF2-isoprostanesindepression islimited.Thestudiesthatwereexcludedbecausetheyonly reportedondepressivesymptomsasacontinuousmeasure (Irieetal.,2001,2003;Iidaetal.,2011)allfoundapositive association.Thereforeitismostlikelythattheywouldnot havealteredtheoverallresult.Furthermore,thecovariates used in the adjusted analyses variedwidely. Therefore it wasnotpossibletoassesstheeffectsofadjustmentonthe results satisfactorily. Informationondiet andanti-oxidant supplementusewasnotavailableformoststudiesand there-fore it cannot be ruled out this factor may have had an effect.Inaddition,theuseofantidepressantsormood stabi-lizerscouldalsobesourcesofconfoundingthatwereoften not investigated. There is some evidence tosuggest that antidepressants,lithiumandothermoodstabilizersprotect againstoxidativestress(Behretal.,2012;Khairovaetal., 2012;deSousaetal.,2014)Studiesaddressingtheireffects onF2-isoprostaneshoweverfoundanincrease(Chungetal., 2013)ornoeffectonoxidativestresslevels(Rawdinetal., 2013).

In conclusion, the ﬁnding that both 8-OHdG and F2-isoprostanes areincreased in depressionstrongly suggests thatdepressionisaccompaniedbyincreasedoxidative dam-age. This ﬁnding supports the hypothesis that increased metabolicstressispresentindepression,whichcould poten-tiallycontributetothehigherriskofsomaticmorbidityand mortalityinsufferers.Thereisaneedforfuturelargerscale studiesonoxidativestressindepression,inwhichtherole oftreatmenteffectsshouldbeaddressed.

Role

of

the

funding

sources

CatherineN.BlackandBrendaW.J.H.Penninxaresupported throughan NWO-VICI grant(number91811602).The grant providerhadnoinvolvementinthestudydesign;inthedata collection,analysisorinterpretationofdata;inthewriting of thereport; or inthe decision tosubmitthearticle for publication.

Conﬂict

of

interest

Allauthorsconfirmtheyhavenofinancialorotherconflicts ofinteresttodisclose.

(10)

Acknowledgements

We would like to express our thanks and acknowledge thecontributionofallthefollowingauthorswhoprovided us with additional data for the meta-analysis. Cecilia P. Chung, MD, MPH, Assistant Professor of Medicine, Divi-sionofRheumatology,VanderbiltUniversityMedicalCenter. AndersJørgensen,MD,PhD,PsychiatricCentreCopenhagen, University Hospital of Copenhagen. Yuri Milaneschi PhD, DepartmentofPsychiatryandEMGOInstituteforHealthand CareResearch,VUUniversityMedicalCenter/GGZinGeest, Amsterdam, The Netherlands & Longitudinal Studies Sec-tion,ClinicalResearchBranch,NationalInstituteonAging, Baltimore, Maryland, United States of America. Gregory Miller,PhDProfessor,Departments ofPsychologyand Med-ical Social Sciences Faculty Fellow, Cells to Society: The Center on Social Disparities and Health at The Institute for Policy Research, Northwestern University. Barbara M. Segal,MD,University ofMinnesota,Minneapolis,MN, USA. C.MichaelStein,MBChB,ProfessorofMedicine,Professorof Pharmacology, VanderbiltUniversity MedicalCenter.Owen M. Wolkowitz, MD, Professor of Psychiatry, University of CaliforniaSan Francisco (UCSF) School of Medicine. Siyan Yi,MD,MHSc,PhD,ResearchFellow,AsiaHealthPolicy,The WalterH.ShorensteinAsia-PaciﬁcResearchCenter,Freeman SpogliInstituteforInternationalStudiesStanfordUniversity. TetsuyaMizoue,DepartmentofEpidemiologyandPrevention ClinicalResearchCenter,NationalCenterforGlobalHealth andMedicine.

Appendices

A,

B,

C,

D. Supplementary

data

Supplementary data associated with this article can be found,intheonlineversion,athttp://dx.doi.org/10.1016/ j.psyneuen.2014.09.025.

References

Abu-Omar,K.,Rutten,A.,Lehtinen, V.,2004. Mentalhealthand physicalactivityintheEuropeanUnion.Soz.Praventivmed.49, 301—309.

Andreazza, A.C., Kauer-Sant’anna, M., Frey, B.N., Bond, D.J., Kapczinski,F.,Young,L.T.,Yatham,L.N.,2008.Oxidativestress markersinbipolardisorder:ameta-analysis.J.Affect.Disord. 111,135—144.

Andreazza,A.C.,Frey,B.N.,Erdtmann,B.,Salvador,M.,Rombaldi, F.,Santin,A.,Goncalves,C.A.,Kapczinski,F.,2007.DNAdamage inbipolardisorder.PsychiatryRes.153,27—32.

Andreazza,A.C.,Kapczinski,F.,Kauer-Sant’anna,M.,Walz, J.C., Bond,D.J.,Goncalves, C.A.,Young,L.T.,Yatham,L.N.,2009. 3-Nitrotyrosineandglutathioneantioxidantsysteminpatients intheearlyand latestages ofbipolardisorder.J. Psychiatry Neurosci.34,263—271.

Barnes, D.E., Alexopoulos, G.S., Lopez, O.L., Williamson, J.D., Yaffe,K.,2006.Depressivesymptoms,vasculardisease,andmild cognitiveimpairment:ﬁndingsfromtheCardiovascularHealth Study.Arch.Gen.Psychiatry63,273—279.

Behr,G.A.,Moreira,J.C.F.,Frey,B.N.,2012.Preclinicalandclinical evidenceofantioxidanteffectsofantidepressantagents: impli-cationsfor thepathophysiology ofmajordepressive disorder. Oxid.Med.CellLongev.2012,609421.

Bijl, R.V., Ravelli, A., 2000. Psychiatric morbidity, service use, and need for care in thegeneral population: results of The

NetherlandsMentalHealthSurveyandIncidenceStudy.Am.J. PublicHealth90,602—607.

Ceprnja,M.,Derek,L.,Unic,A.,Blazev,M.,Fistonic,M., Kozaric-Kovacic, D., Franic, M., Romic, Z., 2011. Oxidative stress markers inpatients withpost-traumatic stress disorder.Coll. Antropol.35,1155—1160.

Ceylan,D.,Tuna,G.,Kirkali,G.,Dizdaronulllu,M.,Tunca,Z., Yesi-lyurt,S.,Ozerdem,A.,2012.DNAdamageinbipolardisorder:a quantitativeassessmentbyusinggaschromatography—isotope dilution mass spectrometry (GC—MS) null. Bipolar Disord. 14 (Suppl.1),66.

Ceylan,D.,Tuna,G.,Kirkali,G.,Tunca,Z.,Dizdaronulllu,M.,Can, G.,Arat,E.,Ozerdem,A.,2013.Increasedoxidativebase dam-ageto DNAinbipolardisorder: across-sectionalcasecontrol study.BipolarDisord.15(1),55.

Che,Y.,Wang,J.F.,Shao, L.,Young, T.,2010. Oxidativedamage toRNAbutnotDNAinthehippocampusofpatientswithmajor mentalillness.J.PsychiatryNeurosci.35,296—302.

Chida, Y., Hamer,M., Wardle,J., Steptoe,A., 2008. Do stress-relatedpsychosocialfactorscontributetocancerincidenceand survival?Nat.Clin.Pract.Oncol.5,466—475.

Chung,C.P.,Schmidt,D.,Stein,C.M.,Morrow,J.D.,Salomon,R.M., 2013. Increased oxidative stress in patients with depression and its relationship to treatment. PsychiatryRes. 206, 213— 216.

Chung,C.P.,Titova,D.,Oeser,A.,Randels,M.,Avalos,I.,Milne, G.L., Morrow, J.D., Stein, C.M., 2009. Oxidative stress in ﬁbromyalgiaanditsrelationshiptosymptoms.Clin.Rheumatol. 28,435—438.

Cohen,J.,1988.StatisticalPowerAnalysisfortheBehavioral Sci-ences.LawrenceErlbaum,Hillsdale,NJ.

Cuijpers,P.,Vogelzangs,N.,Twisk,J.,Kleiboer,A.,Li,J.,Penninx, B.W., 2014. Comprehensivemeta-analysisof excessmortality in depression in thegeneral community versus patients with speciﬁcillnesses.Am.J.Psychiatry171,453—462.

Dalle-Donne,I.,Rossi,R.,Colombo,R.,Giustarini,D.,Milzani,A., 2006.Biomarkersofoxidativedamageinhumandisease.Clin. Chem.52,601—623.

deSousa, R.T.,Zarate, C.A.J., Zanetti,M.V., Costa,A.C.,Talib, L.L., Gattaz, W.F., hado-Vieira,R., 2014. Oxidative stress in earlystagebipolardisorderandtheassociationwithresponse tolithium.J.Psychiatr.Res.50,36—41.

Dimopoulos,N.,Piperi, C.,Psarra,V.,Lea, R.W.,Kalofoutis,A., 2008.Increasedplasmalevelsof8-iso-PGF2alphaandIL-6inan elderlypopulationwithdepression.PsychiatryRes.161,59—66. Duval,S.,Tweedie,R.,2000.Trimandﬁll: asimple funnel-plot-basedmethodoftestingand adjustingfor publicationbiasin meta-analysis.Biometrics56,455—463.

Forlenza, M.J., Miller, G.E., 2006. Increased serum levels of 8-hydroxy-2-deoxyguanosine in clinical depression. Psychosom. Med.68,1—7.

Freund-Levi,Y.,Cederholm,T.,Basu,S.,2011.Oxidativestressand cox-mediatedinﬂammationinpatientswithAlzheimer’sdisease andeffectsofN-3fattyacidsupplementationtheomegADstudy. Clin.Nutr.6(Suppl.1),9.

Gawryluk, J.W., Wang, J.F., Andreazza, A.C., Shao, L., Young, L.T., 2011. Decreased levels of glutathione,the major brain antioxidant, in post-mortem prefrontal cortex from patients withpsychiatric disorders.Int.J.Neuropsychopharmacol.14, 123—130.

Glassman,A.H.,Helzer,J.E.,Covey,L.S.,Cottler,L.B.,Stetner,F., Tipp,J.E.,Johnson,J.,1990.Smoking,smokingcessation,and majordepression.J.Am.Med.Assoc.264,1546—1549. Grande,I.,Magalhaes,P.V.,Kunz,M.,Vieta,E.,Kapczinski,F.,2012.

Mediatorsofallostasisandsystemictoxicityinbipolardisorder. Physiol.Behav.106,46—50.

Gubert,C., Stertz, L., Pfaffenseller,B.,Panizzutti, B.S.,Rezin, G.T.,Massuda,R.,Streck,E.L.,Gama,C.S.,Kapczinski,F.,Kunz,

(11)

174 C.N.Blacketal. M.,2013.Mitochondrialactivityandoxidativestressmarkersin

peripheralbloodmononuclearcellsofpatientswithbipolar dis-order,schizophrenia,andhealthysubjects.J.Psychiatr.Res.47, 1396—1402.

Hayden,J.A.,Cote,P.,Bombardier,C.,2006.Evaluationofthe qual-ityofprognosisstudiesinsystematicreviews.Ann.Intern.Med. 144,427—437.

Higgins,J.P.T.,Thompson,S.G.,Deeks,J.J.,Altman, D.G.,2003. Measuring inconsistency in meta-analyses. Br. Med. J. 327, 557—560.

Ho,E.,KarimiGalougahi,K.,Liu,C.,Bhindi,R.,Figtree,G.,2013. Biologicalmarkersofoxidativestress:applicationsto cardiovas-cularresearchandpractice.RedoxBiol.1,483—491.

Huzayyin, A.A., Andreazza, A.C., Turecki, G., Cruceanu, C., Rouleau,G.A.,Alda, M., Young,L.T., 2014. Decreased global methylationin patientswithbipolardisorderwho respondto lithium.Int.J.Neuropsychopharmacol.17,561—569.

Iida,T.,Chikamura,C.,Inoue,K.,Ito,Y.,Ishikawa,H.,Teradaira, R.,Ono,Y.,2011. AssociationofSTAIand SDSscores with 8-hydroxydeoxyguanosineand serotonin levelsin young women withdepressive symptoms.J.NeuropsychiatryClin.Neurosci. 23,E10.

Ioannidis,J.P.A.,Patsopoulos, N.A.,Evangelou, E., 2007. Uncer-taintyinheterogeneityestimatesinmeta-analyses.Br.Med.J. 335,914—916.

Irie,M.,Asami,S.,Nagata,S.,Ikeda,M.,Miyata, M.,Kasai, H., 2001.PsychosocialfactorsasapotentialtriggerofoxidativeDNA damageinhumanleukocytes.Jpn.J.CancerRes.92,367—376. Irie,M., Asami, S.,Ikeda, M., Kasai,H.,2003. Depressive state relatestofemaleoxidativeDNAdamagevianeutrophil activa-tion.Biochem.Biophys.Res.Commun.311,1014—1018. Irie, M., Asami, S., Nagata, S., Miyata, M., Kasai, H., 2002.

Psychological mediationof a type ofoxidative DNA damage, 8-hydroxydeoxyguanosine, in peripheral blood leukocytes of non-smokingandnon-drinkingworkers.Psychother.Psychosom. 71,90—96.

Irie,M.,Miyata,M.,Kasai,H.,2005.Depressionandpossible can-cerrisk dueto oxidativeDNA damage. J.Psychiatr. Res.39, 553—560.

Janicki-Deverts,D.,Cohen,S.,Matthews,K.A.,Gross,M.D.,Jacobs, D.R.J.,2009.Socioeconomicstatus,antioxidantmicronutrients, andcorrelatesofoxidativedamage: theCoronaryArtery Risk DevelopmentinYoungAdults(CARDIA)study.Psychosom.Med. 71,541—548.

Joergensen,A.,Broedbaek,K.,Weimann, A.,Semba,R.D., Fer-rucci,L., Joergensen,M.B., Poulsen, H.E., 2011. Association between urinary excretion of cortisol and markers of oxida-tivelydamagedDNAandRNAinhumans.PLoSONE6,e20795, http://dx.doi.org/10.1371/journal.pone.0020795.

Jorgensen,A.,Krogh,J.,Miskowiak,K.,Bolwig,T.G.,Kessing,L.V., Fink-Jensen,A., Nordentoft,M.,Henriksen, T.,Weimann,A., Poulsen,H.E.,Jorgensen,M.B.,2013.Systemicoxidatively gen-eratedDNA/RNAdamageinclinicaldepression:associationsto symptomseverityandresponsetoelectroconvulsivetherapy.J. Affect.Disord.149,355—362.

Kapczinski, F., Dal-Pizzol, F., Teixeira, A.L., Magalhaes, P.V.S., Kauer-Sant’anna,M.,Klamt,F.,Moreira, J.C.,deBittencourt Pasquali,M.A.,Fries,G.R.,Quevedo,J.,Gama,C.S.,Post,R., 2011.Peripheralbiomarkersandillnessactivityinbipolar disor-der.J.Psychiatr.Res.45,156—161.

Kessler,R.C.,Berglund,P.,Demler,O.,Jin,R.,Koretz,D., Merikan-gas, K.R., Rush, A.J., Walters, E.E., Wang, P.S., 2003. The epidemiology of major depressive disorder: results from the NationalComorbiditySurveyReplication(NCS-R).J.Am.Med. Assoc.289,3095—3105.

Khairova,R.,Pawar,R.,Salvadore, G.,Juruena,M.F.,de Sousa, R.T.,Soeiro-de-Souza,M.G.,Salvador,M.,Zarate,C.A.,Gattaz,

W.F.,hado-Vieira,R.,2012.Effectsoflithiumonoxidativestress parametersinhealthysubjects.Mol.Med.Rep.5,680—682. Kupper,N.,Gidron,Y.,Winter,J., Denollet,J.,2009.Association

between type D personality, depression, and oxidative stress in patients with chronic heart failure. Psychosom. Med. 71, 973—980.

Luppino,F.S.,de Wit,L.M.,Bouvy, P.F.,Stijnen,T.,Cuijpers,P., Penninx,B.W.J.H.,Zitman,F.G.,2010.Overweight,obesity,and depression:asystematicreviewandmeta-analysisof longitudi-nalstudies.Arch.Gen.Psychiatry67,220—229.

Maes, M., Galecki, P., Chang, Y.S., Berk, M., 2011. A review on the oxidative and nitrosative stress (O&NS) pathways in major depression and their possible contribution to the (neuro)degenerativeprocessesinthatillness.Prog. Neuropsy-chopharmacol.Biol.Psychiatry35,676—692.

Maes, M., Mihaylova,I., Kubera,M., Uytterhoeven,M., Vrydags, N.,Bosmans,E.,2009.Increased8-hydroxy-deoxyguanosine,a markerofoxidativedamagetoDNA, inmajordepressionand myalgicencephalomyelitis/chronicfatiguesyndrome. Neuroen-docrinol.Lett.30,715—722.

Magalhaes,P.V.S.,Jansen,K.,Pinheiro,R.T.,Colpo,G.D.,daMotta, L.L., Klamt, F., da Silva, R.A., Kapczinski, F., 2012. Periph-eraloxidativedamageinearly-stagemooddisorders:anested population-basedcase—controlstudy.Int.J. Neuropsychophar-macol.15,1043—1050.

Maritim,A.C.,Sanders,R.A.,Watkins,J.B.,2003.Diabetes, oxida-tivestress,andantioxidants:areview.J.Biochem.Mol.Toxicol. 17,24—38.

Matsumoto,Y.,Ogawa,Y.,Yoshida,R.,Shimamori,A.,Kasai,H., Ohta, H.,2008. The stability ofthe oxidative stress marker, urinary8-hydroxy-2-deoxyguanosine(8-OHdG),whenstoredat roomtemperature.J.Occup.Health50,366—372.

McEwen,B.S.,Wingﬁeld,J.C.,2003.Theconceptofallostasisin biologyandbiomedicine.Horm.Behav.43,2—15.

Meagher,E.A.,FitzGerald, G.A., 2000. Indicesof lipid peroxida-tioninvivo:strengthsandlimitations.FreeRad.Biol.Med.28, 1745—1750.

Mezuk,B.,Eaton,W.W.,Albrecht,S.,Golden,S.H.,2008. Depres-sion and type 2 diabetes over thelifespan: a meta-analysis. DiabetesCare31,2383—2390.

Michel,T.M.,Pulschen,D.,Thome,J.,2012.Theroleofoxidative stressindepressivedisorders.Curr.Pharm.Des.18,5890—5899. Milaneschi,Y.,Cesari,M.,Simonsick,E.M.,Vogelzangs,N.,Kanaya, A.M.,Yaffe,K.,Patrignani,P.,Metti,A.,Kritchevsky,S.B.,Pahor, M.,Ferrucci,L.,Penninx,B.W.J.H.,2013.Lipidperoxidationand depressedmoodincommunity-dwellingoldermenandwomen. PLoSONE8,e65406.

Nicholson,A.,Kuper,H.,Hemingway,H.,2006.Depressionasan aetiologicand prognosticfactor in coronaryheartdisease: a meta-analysisof6362eventsamong146538participantsin54 observationalstudies.Eur.HeartJ.27,2763—2774.

Niki,E.,2014.Biomarkersoflipidperoxidationinclinicalmaterial. Biochim.Biophys.Acta1840,809—817.

Palta,P.,Samuel,L.J.,Miller,E.R.,Szanton,S.L.,2014.Depression and oxidativestress:resultsfrom ameta-analysisof observa-tionalstudies.Psychosom.Med.76,12—19.

Pandya, C.D., Howell, K.R., Pillai, A., 2013. Antioxidants as potentialtherapeuticsforneuropsychiatricdisorders.Prog. Neu-ropsychopharmacol.Biol.Psychiatry46,214—223.

Penninx, B.W.J.H., Milaneschi, Y., Lamers, F., Vogelzangs, N., 2013.Understandingthesomaticconsequencesofdepression: biologicalmechanismsandtheroleofdepressionsymptom pro-ﬁle.BMCMed.11,129.

Pomara,N.,Bruno,D.,Sarreal,A.S.,Hernando,R.T.,Nierenberg, J., Petkova, E., Sidtis, J.J., Wisniewski, T.M., Mehta, P.D., Pratico,D.,Zetterberg,H.,Blennow,K.,2012.LowerCSF amy-loid beta peptides and higher F2-isoprostanes in cognitively

(12)

Oxidativestressindepression:Ameta-analysis 175 intactelderly individualswithmajordepressive disorder.Am.

J.Psychiatry169,523—530.

Rawdin, B.J.,Mellon, S.H.,Dhabhar, F.S.,Epel, E.S., Puterman, E.,Su,Y.,Burke,H.M., Reus,V.I.,Rosser,R.,Hamilton,S.P., Nelson,J.C.,Wolkowitz,O.M.,2013.Dysregulatedrelationship ofinﬂammationandoxidativestressinmajordepression.Brain Behav.Immun.31,143—152.

Segal,B.M.,Thomas,W.,Zhu,X.,Diebes,A.,McElvain,G., Baech-ler,E.,Gross,M.,2012.Oxidativestressandfatigueinsystemic lupuserythematosus.Lupus21,984—992.

Soeiro-de-Souza, M.G., Andreazza, A.C., Carvalho, A.F., hado-Vieira,R.,Young,L.T., Moreno,R.A.,2013. Numberofmanic episodesisassociatedwithelevatedDNAoxidationinbipolarI disorder.Int.J.Neuropsychopharmacol.16,1505—1512. Sterne, J.A.,Egger, M.,2001. Funnel plotsfor detecting bias in

meta-analysis:guidelinesonchoiceofaxis.J.Clin.Epidemiol. 54,1046—1055.

Sullivan,L.E.,Fiellin,D.A.,O’Connor,P.G.,2005.Theprevalence andimpactofalcoholproblemsinmajordepression:a system-aticreview.Am.J.Med.118,330—341.

Valavanidis, A., Vlachogianni, T., Fiotakis, C., 2009. 8-Hydroxy-2-deoxyguanosine(8-OHdG):a criticalbiomarkerofoxidative stress and carcinogenesis.J. Environ. Sci. HealthC:Environ. Carcinog.Ecotoxicol.Rev.27,120—139.

Valko,M.,Leibfritz,D.,Moncol,J.,Cronin,M.T.D.,Mazur,M.,Telser, J., 2007.Freeradicalsand antioxidantsinnormal physiologi-calfunctionsandhumandisease.Int.J.Biochem.CellBiol.39, 44—84.

Verhoeven,J.,Révész,D.,Epel,E.,Lin,J.,Wolkowitz,O.,Penninx, B.,2013. Major depressive disorder and acceleratedcellular aging:resultsfroma largepsychiatriccohortstudy.Mol. Psy-chiatry,http://dx.doi.org/10.1038/mp.2013.151.

Vos, T., Flaxman, A.D., Naghavi, M., Lozano, R., Michaud, C., Ezzati,M.,Shibuya,K.,Salomon,J.A.,Abdalla,S.,Aboyans,V., Abraham,J.,Ackerman,I.,Aggarwal,R.,Ahn,S.Y.,Ali,M.K., Alvarado,M.,Anderson,H.R.,Anderson,L.M.,Andrews,K.G., Atkinson, C.,Baddour, L.M.,Bahalim, A.N., Barker-Collo,S., Barrero, L.H.,Bartels, D.H.,Basanez, M.G.,Baxter, A.,Bell, M.L.,Benjamin,E.J.,Bennett,D.,Bernabe,E.,Bhalla,K., Bhan-dari,B.,Bikbov,B.,BinAbdulhak,A.,Birbeck,G.,Black,J.A., Blencowe,H.,Blore,J.D.,Blyth,F.,Bolliger,I.,Bonaventure,A., Boufous,S.,Bourne,R.,Boussinesq,M.,Braithwaite,T.,Brayne, C., Bridgett,L., Brooker,S., Brooks,P., Brugha, T.S., Bryan-Hancock, C.,Bucello,C., Buchbinder,R.,Buckle,G., Budke, C.M.,Burch,M.,Burney,P.,Burstein,R.,Calabria,B.,Campbell, B.,Canter,C.E.,Carabin,H.,Carapetis,J.,Carmona,L.,Cella, C.,Charlson,F.,Chen,H.,Cheng,A.T.-A.,Chou,D.,Chugh,S.S., Coffeng,L.E.,Colan,S.D.,Colquhoun,S.,Colson,K.E.,Condon, J.,Connor,M.D.,Cooper,L.T.,Corriere,M.,Cortinovis,M.,de Vaccaro,K.C.,Couser,W.,Cowie,B.C.,Criqui,M.H.,Cross,M., Dabhadkar,K.C.,Dahiya,M., Dahodwala,N.,msere-Derry,J., Danaei,G.,Davis,A.,DeLeo, D.,Degenhardt,L.,Dellavalle, R.,Delossantos,A.,Denenberg,J.,Derrett,S.,DesJarlais,D.C., Dharmaratne,S.D.,Dherani,M.,az-Torne,C.,Dolk,H.,Dorsey, E.R.,Driscoll,T.,Duber, H.,Ebel,B.,Edmond,K.,Elbaz,A., Ali,S.E.,Erskine,H.,Erwin,P.J.,Espindola,P.,Ewoigbokhan, S.E.,Farzadfar, F.,Feigin,V.,Felson, D.T.,Ferrari, A.,Ferri, C.P.,Fevre,E.M.,Finucane,M.M.,Flaxman,S.,Flood,L., Fore-man,K.,Forouzanfar,M.H.,Fowkes,F.G.,Franklin,R.,Fransen, M., Freeman, M.K., Gabbe,B.J., Gabriel,S.E., Gakidou,E., Ganatra,H.A.,Garcia,B.,Gaspari,F.,Gillum,R.F.,Gmel,G., Gosselin,R.,Grainger,R.,Groeger,J.,Guillemin,F.,Gunnell,

D.,Gupta,R.,Haagsma,J.,Hagan,H.,Halasa,Y.A.,Hall,W., Haring,D.,Haro,J.M.,Harrison,J.E.,Havmoeller,R.,Hay,R.J., Higashi,H.,Hill,C.,Hoen,B.,Hoffman,H.,Hotez,P.J.,Hoy, D., Huang,J.J., Ibeanusi, S.E., Jacobsen,K.H., James,S.L., Jarvis,D.,Jasrasaria,R.,Jayaraman,S.,Johns,N.,Jonas,J.B., Karthikeyan,G.,Kassebaum,N.,Kawakami,N.,Keren,A.,Khoo, J.P.,King,C.H.,Knowlton,L.M.,Kobusingye,O.,Koranteng,A., Krishnamurthi,R.,Lalloo,R.,Laslett,L.L.,Lathlean,T.,Leasher, J.L.,Lee,Y.Y.,Leigh,J.,Lim,S.S.,Limb,E.,Lin,J.K.,Lipnick, M.,Lipshultz,S.E.,Liu,W.,Loane,M.,Ohno,S.L.,Lyons,R.,Ma, J.,Mabweijano,J.,MacIntyre,M.F.,Malekzadeh,R.,Mallinger, L., Manivannan, S., Marcenes, W.,March, L., Margolis, D.J., Marks,G.B.,Marks,R.,Matsumori,A.,Matzopoulos,R.,Mayosi, B.M.,McAnulty,J.H.,McDermott,M.M.,McGill,N.,McGrath,J., Medina-Mora,M.E.,Meltzer,M.,Mensah,G.A.,Merriman,T.R., Meyer,A.C.,Miglioli,V.,Miller,M.,Miller,T.R.,Mitchell,P.B., Mocumbi,A.O.,Mofﬁtt,T.E.,Mokdad,A.A.,Monasta,L., Mon-tico,M.,Moradi-Lakeh,M.,Moran,A.,Morawska,L.,Mori,R., Murdoch,M.E.,Mwaniki,M.K.,Naidoo,K.,Nair,M.N.,Naldi,L., Narayan,K.M.V.,Nelson,P.K.,Nelson,R.G.,Nevitt,M.C., New-ton, C.R., Nolte,S.,Norman,P., Norman,R.,O’Donnell, M., O’Hanlon,S.,Olives,C.,Omer,S.B.,Ortblad,K.,Osborne,R., Ozgediz,D.,Page,A.,Pahari,B.,Pandian,J.D.,Rivero,A.P., Pat-ten,S.B.,Pearce,N.,Padilla,R.P.,Perez-Ruiz,F.,2012.Years livedwithdisability(YLDs)for1160sequelaeof289diseasesand injuries1990—2010:asystematicanalysisfortheGlobalBurden ofDiseaseStudy2010.Lancet380,2163—2196.

Wang,J.F.,Shao,L.,Sun,X.,Young,L.T.,2009.Increasedoxidative stressintheanteriorcingulatecortexofsubjectswithbipolar disorderandschizophrenia.BipolarDisord.11,523—529. Wei,Y.C., Zhou, F.L., He,D.L., Bai,J.R., Ding,H.,Wang,X.Y.,

Nan,K.J.,2009a.Oxidativestress indepressivepatientswith gastric adenocarcinoma. Int. J. Neuropsychopharmacol. 12, 1089—1096.

Wei,Y.C.,Zhou,F.L.,He,D.L.,Bai,J.R.,Hui,L.y.,Wang,X.Y.,Nan, K.J., 2009b.The levelofoxidative stress and theexpression ofgenesinvolvedinDNA-damagesignalingpathwaysin depres-sivepatientswithcolorectalcarcinoma.J.Psychosom.Res.66, 259—266.

Wolkowitz,O.M.,Mellon,S.H.,Epel, E.S.,Lin,J., Dhabhar,F.S., Su,Y.,Reus,V.I.,Rosser,R.,Burke,H.M.,Kupferman,E., Com-pagnone, M.,Nelson,J.C.,Blackburn,E.H.,2011a.Leukocyte telomerelengthinmajordepression:correlationswith chronic-ity, inﬂammationand oxidative stress —preliminary ﬁndings. PLoSONE6,e17837.

Wolkowitz,O.M.,Reus,V.I.,Mellon,S.H.,2011b. Ofsound mind andbody:depression,disease,andacceleratedaging.Dialogues Clin.Neurosci.13,25—39.

Wu,I.C.,Shiesh,S.C., Kuo,P.H.,Lin, X.Z.,2009. Highoxidative stressiscorrelatedwithfrailtyinelderlyChinese.J.Am.Geriatr. Soc.57,1666—1671.

Yager, S., Forlenza, M.J., Miller, G.E., 2010. Depression and oxidative damage to lipids. Psychoneuroendocrinology 35, 1356—1362.

Yi, S., Nanri, A., Matsushita, Y., Kasai, H., Kawai, K., Mizoue, T., 2012. Depressive symptomsand oxidative DNA damagein Japanesemunicipalemployees.PsychiatryRes.200,318—322. Yoshida,R.,Ogawa,Y.,Kasai,H.,2002.Urinary

8-oxo-7,8-dihydro-2-deoxyguanosinevaluesmeasuredbyanELISAcorrelatedwell withmeasurementsbyhigh-performanceliquidchromatography withelectrochemicaldetection.CancerEpidemiol.Biomarkers Prev.11,1076—1081.