Bidirectional associations between treatment-resistant depression and general medical conditions

(1)

University of Groningen

Bidirectional associations between treatment-resistant depression and general medical

conditions

Madsen, Kathrine Bang; Momen, Natalie C; Petersen, Liselotte Vogdrup; Plana-Ripoll,

Oleguer; Haarman, Bartholomeus C M; Drexhage, Hemmo; Mortensen, Preben Bo; McGrath,

John J; Munk-Olsen, Trine

Published in:

European Neuropsychopharmacology

DOI:

10.1016/j.euroneuro.2021.04.021

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from

it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date:

2021

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Madsen, K. B., Momen, N. C., Petersen, L. V., Plana-Ripoll, O., Haarman, B. C. M., Drexhage, H.,

Mortensen, P. B., McGrath, J. J., & Munk-Olsen, T. (2021). Bidirectional associations between

treatment-resistant depression and general medical conditions. European Neuropsychopharmacology, 51, 7-19.

https://doi.org/10.1016/j.euroneuro.2021.04.021

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

www.elsevier.com/locate/euroneuro

Bidirectional

associations

between

treatment-resistant

depression

and

general

medical

conditions

Kathrine

Bang

Madsen

a,b,∗

,

Natalie

C. Momen

a

,

Liselotte

Vogdrup

Petersen

a,b

,

Oleguer

Plana-Ripoll

a

,

Bartholomeus

C.M.

Haarman

c

,

Hemmo

Drexhage

d

,

Preben

Bo

Mortensen

a,b,e

_,

_John

_J.

_McGrath

a,f,g

_,

Trine

Munk-Olsen

a,b

a_National_Centre_for_{Register-based}_Research,_Business_and_Social_Sciences,_Aarhus_University,_Denmark b_iPSYCH,_the_Lundbeck_Foundation_Initiative_for_Integrative_Psychiatric_Research,_Denmark

c_Department_of_Psychiatry,_University_Medical_Center_Groningen,_University_of_Groningen,_Groningen,

theNetherlands

d_Department_of_Immunology,_Erasmus_MC,_University_Medical_Center_Rotterdam,_Rotterdam,_the

Netherlands

e_CIRRAU_-_Centre_for_Integrated_{Register-based}_Research,_Aarhus_University,_Aarhus,_Denmark f_Queensland_Brain_Institute,_University_of_Queensland,_St_Lucia,_Queensland,_Australia

g_Queensland_Centre_for_Mental_Health_Research,_The_Park_Centre_for_Mental_Health,_Queensland,

Australia

Received 3February2021;receivedinrevisedform27April2021;accepted29April2021

KEYWORDS Treatment-resistant depression; Generalmedical conditions; Antidepressants; Abstract

Depression is associated with general medical conditions (GMCs), but it is not known if treatment-resistantdepression(TRD)affects GMCriskandviceversa.Weestimated bidirec-tionalassociationsbetweenTRDandGMCs(priorandsubsequent).Allindividualsaged18–69 years,bornandlivinginDenmark,withafirst-timeprescriptionforanantidepressantbetween 2005and2012wereidentifiedintheDanishPrescriptionRegistry(N=154,513).TRDwas de-finedasatleasttwo shiftsintreatmentregimes. Forprior GMCs,we estimatedoddsratios

∗_{Corresponding}_author_at:_National_Centre_for_{Register-based}_Research,_Business_and_Social_Sciences,_Aarhus_University,_Denmark.

E-mail address: kathrine.bang@econ.au.dk(K.B.Madsen).

https://doi.org/10.1016/j.euroneuro.2021.04.021

(3)

Population-based study;

Register-basedstudy

(ORs)usingconditionallogisticregressioncomparingTRDpatientswithmatchednon-TRD con-trolsadjustedforotherGMCsandnumberofotherGMCs.ForsubsequentGMCs,weusedCox regressiontocalculatehazardratios(HRs)inTRDvs.non-TRDpatientsadjustedforageatﬁrst prescription,calendar time,otherGMCsandnumber ofotherGMCs.PatientswithTRDhad higherprevalenceofpriorGMCsrelatedtotheimmuneorneurologicalsystems; musculoskele-tal disorders(women aOR:1.35, 95% CI:1.26–1.46,men aOR: 1.30, 95%CI:1.19–1.42)and migraine(womenaOR:1.22,95%CI:1.09–1.36,menaOR:1.25,95%CI:1.00–1.56).Subsequent GMCswererelatedtoabroaderspectrum;cardiovascular(womenaHR:1.43,95%CI:1.32–1.54, menaHR:1.31,95%CI:1.19–1.43),endocrine(womenaHR:1.52,95%CI:1.37–1.67,menaHR: 1.24,95%CI:1.07–1.44),andneurologicaldisorders(womenaHR:1.24,95%CI:1.13–1.35,men aHR:1.19,95%CI:1.07–1.34).

OurstudypresentsabroadoverviewofcomorbidmedicalconditionsinpatientswithTRDand furtherstudiesareneededtoexploretheassociationsindetail.

1. Introduction

Extensive evidence has shown that general medical con-ditions (GMCs) are associated with an increased risk for developing depression (Egede 2007 ; Patten 2001 ) and, vice versa,depressionis associatedwithan increasedrisk for developing GMCs (Momen et al., 2020 ; Scott et al., 2016 ; Tegethoff et al., 2016 ). The co-occurrence of de-pression and GMCs suggests bidirectional associations be-tween the two, further hinting at a shared aetiology. This has been proposed for depression andcardiovascular disorder, chronic obstructive pulmonary disease, rheuma-toid arthritis and diabetes type 1 and 2, explained by inﬂammatory and oxidative / nitrosative stress pathways (Maes et al., 2011 ;Miller et al., 2009 ).Despitenot know-ing what develops ﬁrst (temporality of events), the co-occurrence of depression with other medical conditions is associated with greater depression symptom severity and GMC severity, decreased treatment adherence and lower remission rates compared to individuals suffer-ing from depression without GMCs (IsHak et al., 2018 ;

Kronish et al., 2006 ; Moussavi et al., 2007 ; Rush et al., 2008 ).

Both depression severityas well as treatment-resistant depression (TRD) influence the complexity of the bidi-rectional association between depression and GMCs (Amital et al., 2013 ; Niles et al., 2015 ). TRD is broadly defined as the occurrence of an insufficient clinical re-sponsetoatleasttwoadequateregimesofantidepressants (Gaynes et al., 2019 ), and has a poor prognosis in terms of increased mortalityanddisability (BangMadsen et al., 2020a ,Madsen et al., 2020b ).Asofyet,itisunknownwhy somepeoplerespondtoantidepressanttreatmentandwhy somepeopledonot,andnospecificsetofneurobiological markersorgeneticprofilehaveprovenusefulinpredicting response or nonresponse (Berlim and Turecki 2007 ). How-ever,medicalconditionsrelatedtotheimmunesystemand metabolism aswell assome types of medicationssuch as immunosuppressants,steroidsandsedativeshavebeen con-sideredapotentialcontributingfactortotheoccurrenceof TRD (Berlim and Turecki 2007 ; Fagiolini and Kupfer 2003 ;

Keitner et al., 1991 ).Nevertheless,studieshaveyielded in-consistent results.Co-morbiddepression and

hypercholes-terolemia has been found tobe associated with poor re-sponse to antidepressant treatment (Papakostas et al., 2003 ;Sonawalla et al., 2002 ),andarthritisand cardiovas-cular problems tobeassociatedwith aworse outcome of depression(Oslin et al., 2002 ).Incontrast,astudy assess-ingtheeffectofGMCcomorbidityonresponsetonext-step antidepressanttreatmentsamongTRDpatientsshowedthat medicalconditionswerenotassociatedwiththelikelihood ofremission,butthesampleincludedonly97subjectsand medical conditionswereexamined withacombinedscore (Perlis et al., 2004 ).Also, anotherstudy found no signiﬁ-cantdifferencebetweenTRDandnon-TRDpatientswith re-gardstotheprevalenceofanyICD-10categoryofGMCs,this studyalsoincludedonlyarelativelylimitednumberof pa-tients(N=702),limitingstatisticalpowertodetectGMCs (Amital et al., 2013 ).

Sofar,studiesconductedonTRDandmedicalconditions havemainlybeenclinicalstudiesbasedondatasourceswith insufﬁcientinformationonthesequenceoftheevents stud-ied,aswellaslimitedsamplesizesandfollow-uptime.Asa directconsequence,nostudieshavebeenpositionedto en-lightenthebidirectionalassociationbetweenTRDandGMC and,moreimportantly,havenotansweredtheoverarching questions:i) Dopreviousmedicalconditionsinﬂuencethe treatment outcome of a depressive episode and ii) does TRD increase the risk of subsequent medical conditions?

To specifically investigate the outlined bidirectional asso-ciations,theaimofthisstudywastwofold:first,to exam-ineanypotentialdifferenceinpriormedicalconditions be-tween TRDandnon-TRDpatientsand second,toestimate thedifferenceinriskofsubsequentmedicalconditions be-tweenthosewithTRDandnon-TRDfollowingtheirfirst de-pressiveepisode.

2. Methods

2.1. Studypopulation

For the present study, we linked information from the Danish Civil Registration System (Pedersen 2011 ), the Danish National Prescription Register (Kildemoes et al., 2011 ), the Danish Psychiatric Central Research Register

(4)

(Mors et al., 2011 ),andtheDanishNationalPatientRegister (Lynge et al., 2011 ). We identified the study population in the Danish National Prescription Registry, including all individualsbornandlivinginDenmarkwhofilledtheirfirst prescription for an antidepressant drug (ATC-code N06A) with the indication “for depression” or “for prevention of depression” (indication codes 168and 270) aged18–69 years between January 1, 2005 and December 31, 2012 (N=175,639).

Inordertocapture onlyfirst-timeantidepressant users, a set of four exclusioncriteria were applied:1) Individu-alswereexcludediftheyhadfilledaprescriptionfora po-tentialadd-onmedicationfordepression,includinglithium (N05AN01), risperidone (N05AX08),olanzapine (N05AH03), aripiprazole (N05AX12) and quetiapine (N05AH04), before theirfirstprescriptionforanantidepressantdrug(n=430; 0.2%). 2)Linkingeach individualtotheDanish Psychiatric Central ResearchRegistry andthe DanishNational Patient Registry, individuals were excluded if they had a hospi-tal in- or outpatient contact before filling their first an-tidepressantprescription foroneofthe followingdisorder categories: Organic,including symptomatic,mental disor-ders(ICD-10:F00–09,ICD-8:290.09,290.10,290.11,290.18, 290.19, 292.x9, 293.x9, 294.x9, 309.x9), Schizophrenia, schizotypal anddelusional disorders (ICD-10:F20–29, ICD-8: 295.x9,296.89,297.x9,298.29–298.99, 299.04,299.05, 299.09,301.83),andBipolardisorders(ICD-10:F30–31, ICD-8:296.19,296.39,298.19)(n=2991;1.7%)(Pedersen et al., 2014 ).3)Individualswereexcludediftheyhadahospital in-oroutpatientcontactwithadiagnosisofsingleorrecurrent depressiveepisode(ICD-10:F32,F33,ICD-8:296.09,296.29, 298.09,300.49)morethan30dayspriortotheirfirst antide-pressant(n=3374;1.9%);otherwisetheywereincludedin the sample with follow-upbeginning on thedate of their firstprescription.4)Individualswereexcludediftheirfirst treatmentregimelastedlessthanfourweeks(n=14,331; 8.2%).Thefinalstudypopulationincluded154,513 individu-alswithdepression.

Toanswer thetwoaims,twodifferentapproacheswere used resultingin twodifferent subsetsof the population, whichareshowninFigure 1 anddescribedin detailinthe following.

2.2. Ethics

According toDanish law,informedconsentis notrequired for register-based studies. The Danish Data Protection AgencyandtheDanishHealthDataAuthorityapprovedthe currentstudy.Alldatawerede-identiﬁedandnot recogniz-ableatanindividuallevel.

2.3. Deﬁnitionoftreatment-resistantdepression

We deﬁned TRD based on the established deﬁnition of at least twoshifts in treatment regimes, which has been usedinrecent studiesofTRD(BangMadsen et al., 2020a ;

Conway et al., 2017 ; Gaynes et al., 2019 ; Madsen et al., 2020b ).Shifts wereidentiﬁedanddeﬁnedusingthree cri-teria: 1) Shift in medication out of class, e.g. from se-lective serotonin reuptake inhibitors (SSRI) to

serotonin-norepinephrine reuptake inhibitors (SNRI). 2) Augmenta-tion with other psychotropic drugs; lithium, risperidone, olanzapine, aripiprazole and quetiapine or a combination of two different antidepressant drug classes at the same time. 3) In- or outpatient contact with single or recur-rent depressiveepisode. Shifts were only counted ifthey occurred within the same episode of continuous medi-cal treatment within two years from first antidepressant prescription (Bang Madsen et al., 2020a ; Kubitz et al., 2013 , Madsen et al., 2020b ). Hence, patientswere classi-fiedashavingTRD iftheye.g., hadtwoshifts in medica-tionclassor oneshift in medicationclassAND augmenta-tion/combination OR a hospital diagnosis of single or re-currentdepressiveepisode.TRDwasclassifiedonthedate that the criteria were met. For further information, the definitionofdepressiveepisodeandmedicationshiftsusing dispensedprescriptionsfromTheDanishNational Prescrip-tionRegistryhavebeendescribedindetailelsewhere(Bang

Madsen et al., 2020a ,Madsen et al., 2020b ).

2.4. Generalmedicalconditions

Information aboutGMCswasobtainedusingcriteriabased on previous Danish research on coexisting conditions go-ingbackto1995(theyear prescriptiondatabecame com-plete) (Momen et al., 2020 ; Prior et al., 2016 ). As de-scribed in Momen et al., the criteria focus on 31 medi-calconditions,withinninebroadcategories:cardiovascular, endocrine, pulmonary, gastrointestinal, urogenital, mus-culoskeletal, hematologic, neurologic and oncologic con-ditions (Momen et al., 2020 ). We identified individuals with GMCs by combining data from three sources: diag-nosesmadeduringinpatientadmissionsandoutpatientand emergency visits from the Danish National Patient Reg-istry(Lynge et al., 2011 ),prescriptionsfordisease-specific medications in the Danish National Prescription Registry (Kildemoes et al., 2011 ),anddiagnosesrecordedascauses ofdeathintheDanishRegisterofCausesofDeath(onlyfor incidentGMCs)(Helweg-Larsen 2011 ).Forexample,chronic pulmonary disease was identified by the ICD-10 diagnosis J40-J47andbyprescriptionsforobstructiveairwaydisease drugs (ATC code R03). The diagnoses (ICD-10 codes) and drugs(Anatomical TherapeuticChemicalclassification sys-tem codes) that were included in the definition of each medical conditionareprovidedin SupplementaryTable 1. Thediagnosisdateofthemedicalconditionofinterestwas thedateoffirsthospitaldiagnosis,thedateoftherelevant repeatprescription,orthedateofdeathfromthemedical condition,whicheveroccurredfirst.

2.5. Covariates

Covariates included GMCs other than the GMC of interest and number of other GMCs (0,2,3,4+) deﬁned and cate-gorized as above.Age at ﬁrstantidepressant prescription wasobtainedfromtheDanishNationalPrescriptionRegistry (Kildemoes et al., 2011 )anddateofbirth, sexand migra-tionstatuswereobtainedfromtheDanishCivilRegistration System(Pedersen 2011 ).

(5)

Fig.1 Foraim1,wedesignedamatchedcase-controlstudyandestimatedassociationsbetweenpriormedicalconditions.Foraim 2,wedesignedacohortstudyandestimatedincidenceratesofmedicalconditionsaccordingtothepresenceorabsenceofTRD.

2.6. Statisticalanalyses

CharacteristicsofTRDandnon-TRDpatientswere summa-rizedbytheuseofdescriptivestatistics.

Prior medical conditions (aim 1): We estimated associ-ations betweenprior medicalconditionsand TRDby com-paringmedicalconditionsgoingbackto1995betweenTRD casesandamatchedreferencegroupofdepressionpatients (non-TRD) (Fig. 1 ).For each TRDcase, fivecontrols were randomly selectedfromthestudypopulation matching on date of birth (+/− 60 days), sexand age at first antide-pressant prescriptionfill (+/− 6 months).Example:a pa-tient wasdefinedtreatment-resistantonAugust 20,2008, he/she was matched with five patients of same age and sex,whoredeemedtheirfirstantidepressantsonthesame dateasthecase,andwhoonAugust20,2008hadnot be-come TRD. From this date and backward we look to see iftheybetween 1995andthe indexdatehave hadanyof theGMCs.Associationswereexaminedusingconditional lo-gistic regressionmodels. Modelswere unadjustedand ad-justedforotherpriorGMCsandnumberofotherpriorGMCs and estimates arepresented asoddsratios (OR) for each sex.

Subsequent medicalconditions(aim2):Toevaluatethe association between TRD and subsequent medical condi-tions,individualswithapriormedicalconditionofinterest wereexcluded (Fig. 1 ).Wecomparedincidenceratesofa diagnosisofamedicalconditionaccordingtothepresence or absenceof TRDusing Cox proportionalhazards models with age asthe underlying time scale. Patientswere fol-lowedfromthedateoffirstantidepressantprescription un-tilemigration,death,theGMCofinterestorDecember31, 2015,whichevercamefirst.Patientswerecensoredifthey, aftertheirfirstprescriptionandwithintwoyears,hada hos-pitalcontactwiththeICD-10diagnosesorganicmental dis-orders (F00-09), schizophrenia, schizotypal anddelusional

disorders (F20-29), manic episode (F30), and bipolar dis-order(F31)onthedate of admission.TRDwastreatedas atime-varyingvariable,meaningthatanindividualmoved fromtheunexposed(non-TRD)totheexposedgroup(TRD) when theyfulfilled therequirements for thedefinition of TRD.Models wereadjusted for birthyear, age atfirst an-tidepressantprescription,otherpreviousGMCs,andnumber ofpreviousGMCs.Hazardratios(HR)arepresentedforeach sex.

StatisticalanalyseswereperformedinStata15.1 (Stata-Corp,CollegeStation,TX,USA).

3. Results

Intheperiodfrom2005to2012,154,513patientswith first-time depression redeemed an antidepressant prescription at ages 18–69 years. Of the total study population, 8294 (5.4%)metthedefined criteriafor TRDduringthe follow-up. There were more women in both groups; TRD (60%), non-TRD (58%),and a larger proportionof the TRD group werebetween18and29yearsatthetimeoftheirfirst an-tidepressantprescriptionfillcomparedtonon-TRD(37%vs. 26%)andfewerwerebetween50and69years(21%vs.31%),

Table 1 .

ThetemporalassociationsofTRDandtheninebroad cat-egories of GMCs areshown in Fig. 2 . Detailedresults for each of the 31 medical conditions are providedin Tables 2 (priorGMCs)and3 (subsequentGMCs).

3.1. Priormedicalconditionsoccurringbefore TRD-Women

ForwomenwithTRD,theprevalenceofmusculoskeletal dis-ordersbeforetheybecametreatment-resistantwaslarger

(6)

Fig.2 Temporalassociationsoftreatment-resistantdepressionandninebroadcategoriesofgeneralmedicalconditionsinwomen andmen.TotheleftisshowntheoddsforpatientswithTRDofhavinghadamedicalconditionbeforetheybecame treatment-resistantcomparedwithmatchednon-TRDcontrols,presentedasoddsratiosadjustedforotherGMCsandnumberofotherGMCs. TotherightisshowntheriskforpatientswithTRDofsubsequentGMCscomparedtoallnon-TRDpatients,presentedashazard ratiosadjustedfor birthyear,ageatﬁrstprescription,previousGMCsandnumberofpreviousGMCs(patientswithpriormedical conditionswereexcluded).

comparedtomatchednon-TRDcontrols(aOR:1.35,95%CI: 1.26–1.46)(Table 2 ).This wasdrivenby painfulconditions (aOR:1.38,95%CI:1.28–1.48)andconnectivetissue disor-der(aOR:1.23,95%CI:1.00–1.51).Theprevalenceofprior pulmonarydisorderswasslightlylarger(aOR:1.08,95%CI: 1.01–1.15)amongwomenwithTRD,buttheassociationsof anyof thedisordersinthecategorywerenotsigniﬁcantly increasedindividually.Largerprevalenceofprior neurolog-ical disorders (aOR: 1.19, 95% CI: 1.09–1.29) was mainly drivenbymigraine(aOR:1.22,95%CI:1.09–1.36),whilefor multiple sclerosis the prevalence wassmaller (aOR:0.46, 95% CI: 0.24–0.88). For endocrine disorders, associations were in opposite directions;diabetes wasmore prevalent (aOR:1.23,95%CI:1.03–1.46),andthyroiddisorderwasless prevalent(aOR:0.87,95%CI:0.75–1.00).Prioroncological disorderswere moreprevalentin non-TRDversus TRD pa-tients(aOR:0.76,95%CI:0.62–0.93).

3.2. Subsequentmedicalconditionsoccurring afterﬁrstmedicallytreateddepressiveepisode -Women

ForwomenwithTRD,theriskwasincreasedforsubsequent cardiovascular disorders (aHR: 1.43, 95% CI: 1.32–1.54), mainlyhypertension(aHR:1.47,95%CI:1.40–1.52)and dys-lipidemia(aHR:1.42,95%CI:1.27–1.58);andforendocrine disorders(aHR:1.52,95%CI:1.37–1.67),wheretherisksof diabetes(aHR:1.63,95%CI:1.42–1.87)andthyroiddisorder (aHR: 1.38, 95% CI: 1.21–1.58) were increased (Table 3 ). The risk of subsequent neurological disorders(aHR: 1.24, 95% CI: 1.13–1.35) was mainly driven by migraine (aHR: 1.22,95%CI:1.08–1.39),epilepsy(aHR:1.70,95%CI:1.25– 2.30) and neuropathies (aHR: 1.19, 95%CI: 1.03–1.37). In addition, for women with TRD the risk of subsequent pulmonarydisorderswasslightlyincreased(aHR.1.12,95%

(7)

Table1 DemographiccharacteristicsofTRDandnon-TRD pa-tients(N₌154,513). TRD n₌8294 Non-TRD n₌146,219 n(%) n(%) Sex Female 4938(60) 85,522(58) Birthyear 1936–1944 345(4) 10,469(7) 1945–1954 916(11) 24,743(17) 1955–1964 1414(17) 28,318(19) 1965–1974 1826(22) 31,934(22) 1975–1984 1945(24) 29,506(20) 1985–1994 1848(22) 21,246(15) YearofﬁrstADprescription 2005–2006 1934(24) 36,580(25) 2007–2008 2077(25) 37,594(26) 2009–2010 2349(28) 40,620(28) 2011–2012 1934(23) 31,425(21)

AgeatﬁrstADprescription(years)

18–29 3044(37) 38,446(26)

30–49 3495(42) 62,352(43)

50–69 1755(21) 45,421(31)

CI: 1.02–1.24). Finally, the risk of ulcer/chronic gastritis (aHR:1.55,95%CI:1.22–1.98)andpainfulcondition(aHR: 1.23,95%CI:1.15–1.32)wasincreased.

3.3. Priormedicalconditionsoccurringbefore TRD– men

FormenwithTRD,theprevalence ofmusculoskeletal dis-orders at the time they became treatment-resistant was larger comparedto matched controls (aOR: 1.30, 95% CI: 1.19–1.42),drivenbypainfulconditions(aOR:1.32,95%CI: 1.20–1.43)(Table 2 ).Whilenoassociationwasfoundfor car-diovasculardisordersasacategory(aOR:1.00,95%CI:0.90– 1.12),anegativeassociationwasfoundforperipheralartery occlusion (aOR: 0.64, 95%CI: 0.45–0.92) and stroke (aOR: 0.68, 95% CI:0.55–0.86). Forneurological disorders(aOR: 1.03,95%CI:0.92–1.16),theprevalenceofmigraine(aOR: 1.25,95%CI:1.00–1.56)andneuropathies(aOR:1.27,95% CI:1.07–1.52)werelargerinmenwithTRDversusnon-TRD.

3.4. Subsequentmedicalconditionsoccurring afterﬁrstmedicallytreateddepressiveepisode -Men

FormenwithTRD,theriskwasincreasedforcardiovascular disorders(aHR:1.31,95%CI:1.19–1.43),mainly hyperten-sion(aHR:1.36,95%CI:1.22–1.51)anddyslipidemia(aHR: 1.24,95%CI:1.10–1.40);andforendocrinedisorders(aHR: 1.24,95%CI:1.07–1.44),drivenbydiabetes(aHR:1.36,95% CI:1.15–1.60)(Table 3 ).Theincreasedriskof musculoskele-taldisorders(aHR:1.48,95%CI:1.08–2.04)wasdueto con-nective tissuedisorder (aHR:1.95,95% CI:1.43–2.65)and painful condition (aHR:1.12, 95% CI:1.02–1.22), and the

increased risk of hematological disorders(aHR: 1.38, 95% CI:1.10–1.74)wasduetoanemias(aHR:1.39,95%CI:1.10– 1.77).Forneurologicaldisorders(aHR:1.19,95%CI:1.07– 1.34), Parkinsons disease (aHR: 2.15, 95% CI: 1.34–3.43), multiplesclerosis(aHR:2.07,95%CI:1.04–4.13)andhearing problems(aHR:1.28,95% CI:1.03–1.59)weretheprimary causes.

4. Discussion

Toourknowledge,thisis theﬁrststudytoinvestigatethe bidirectional associations between TRD and a broad and comprehensive range of GMCs. We found that for both womenandmenwithTRDthe prevalenceof prior muscu-loskeletal disorders (connective tissue disorder in women andpainfulconditionsinboth)andmigrainebeforeTRD on-set was largerthan in matched controls with depression. We further found differences between sexes; for women with TRD, prior diabetes was more prevalent and onco-logical disorders were less prevalent, while for menwith TRD, prior neuropathies were more prevalent.For subse-quentmedical conditionswherewe exploredriskof GMCs after ﬁrst depressive episode, cardiovascular (hyperten-sionand dyslipidemia), endocrine(diabetes), neurological (mainlyepilepsyinwomenandParkinson’sdiseaseinmen) andmusculoskeletaldisorders(connectivetissuedisorderin menandpainfulconditioninboth)wereincreasedforboth sexes withTRD. In addition,for women withTRD, subse-quent chronic pulmonary disease and gastric ulcers were increased,while for men withTRD, increasedsubsequent hematologicaldisorders(anemias)wasfound.

Thecurrent studyis exploratoryandoffersan overview of the bidirectional associations between TRD and GMCs, andcanbeusedtogeneratehypothesesforfuturestudies thatconsidertheassociationsindetail.Wedonotpropose acausalrelationshipbetweenTRDandmedicalconditions. ThepresenceofbothTRDandamedicalconditionmaybe confounded byprevious exposures(e.g.,childhoodabuse, socioeconomicfactors,andsharedenvironmentalrisk fac-tors)orsharedgeneticfactors.However,wespeculatethat speciﬁcexplanationsforourobservationsmayinclude sev-eralfactorsrelatedtothemetabolism,theimmunesystem, andpotentialsharedgeneticvulnerability.

Ourobservedassociationsshouldbeconsideredinlightof previousstudies suggestingthatapooreroutcomeand re-sponse totreatment characterize depressedpatientswith concurrentillness irrespectiveofthe coexistingdiagnosis. This suggests that anyadditional burden, ratherthan the burden caused by speciﬁc illness, is the determining fac-tor in predicting a poorer prognosis (Black et al., 1987 ;

Coryell et al., 1985 ;Keitner et al., 1991 ).However,other studiessuggestthatonlycertainsomaticdisorders,suchas neurologicaldisordersandconnectivetissuedisorders,play aroleinthetreatmentresponseofdepression(Berlim and Turecki 2007 ;Oslin et al., 2002 ),whichisinaccordancewith ourresults.

Cardiovascularproblemssuchasheartdisease, hyperten-sionanddyslipidemiahavebeen reportedtobe accompa-niedbyco-morbid depressionandsomestudies haveeven foundthatco-occurringcardiovasculardisordersare associ-atedwithpoorresponsetoantidepressanttreatmentin

(8)

pa-Table2 PriormedicaldisordersinwomenandmenwithTRD.

WOMEN MEN

nofTRDcases OR aOR(95%CI) nofTRDcases OR aOR(95%CI)

Cardiovascular 929 1.09 1.01(0.92;1.11) 817 1.04 1.00(0.90;1.12)

Hypertension 781 1.01 1.02(0.93;1.13) 615 1.09 1.06(0.95;1.19)

Dyslipidemia 253 1.02 0.96(0.82;1.12) 348 0.95 0.92(0.79;1.06)

Ischemicheartdisease 81 1.05 0.99(0.77;1.27) 137 0.86 0.83(0.68;1.00)

Atrialﬁbrilation 17 0.89 0.88(0.52;1.49) 48 0.78 0.77(0.56;1.05)

Heartfailure 11 0.72 0.69(0.36;1.31) 32 0.89 0.88(0.59;1.30)

Peripheralarteryocclusivedisease 35 1.16 1.09(0.75;1.58) 35 0.67 0.64(0.45;0.92)

Stroke 79 1.10 1.05(0.82;1.35) 96 0.70 0.68(0.55;0.86) Endocrine 386 1.01 0.99(0.88;1.11) 215 1.02 0.97(0.83;1.14) Diabetes 172 1.26 1.23(1.03;1.46) 158 1.00 0.96(0.79;1.16) Thyroiddisorder 231 0.88 0.87(0.75;1.00) 36 0.90 0.89(0.62;1.27) Gout NA NA NA 25 1.09 0.99(0.64;1.54) Pulmonary 1795 1.12 1.08(1.01;1.15) 883 1.04 1.02(0.94;1.11)

Chronicpulmonarydisease 1053 1.12 1.08(0.99;1.16) 465 1.04 1.02(0.91;1.14)

Allergy 1170 1.08 1.05(0.98;1.13) 595 1.00 0.99(0.89;1.09)

Gastrointestinal 158 1.09 1.03(0.87;1.23) 178 1.11 1.09(0.92;1.29)

Ulcer/chronicgastritis 59 1.31 1.22(0.91;1.63) 56 0.87 0.84(0.63;1.13)

Chronicliverdisease 23 1.14 1.08(0.68;1.70) 40 0.94 0.93(0.66;1.31)

Inﬂammatoryboweldisease 51 0.82 0.79(0.59;1.07) 53 1.33 1.31(0.96;1.78)

Diverticulardiseaseofintestine 32 1.41 1.33(0.89;1.97) 46 1.58 1.55(1.10;2.17)

Urogenital 13 0.77 0.73(0.40;1.31) 108 1.01 1.01(0.81;1.27)

Chronickidneydisease 13 0.77 0.73(0.40;1.32) 20 0.95 0.96(0.59;1.56)

Prostatedisorders NA NA NA 91 1.01 1.01(0.79;1.29)

Musculoskeletal 1533 1.39 1.35(1.26;1.46) 999 1.29 1.30(1.19;1.42)

Connectivetissuedisorder 117 1.26 1.23(1.00;1.51) 42 1.12 1.11(0.79;1.55)

Osteoporosis 64 1.11 1.10(0.83;1.46) 19 0.88 0.87(0.53;1.43) Painfulcondition 1479 1.41 1.38(1.28;1.48) 983 1.31 1.32(1.20;1.43) Hematological 58 0.88 0.86(0.65;1.15) 42 0.84 0.84(0.59;1.19) HIV/AIDS NA NA NA 8 0.91 0.93(0.44;1.98) Anemias NA 0.85 0.83(0.62;1.12) 34 0.82 0.81(0.56;1.19) Oncological 121 0.79 0.76(0.62;0.93) 111 0.85 0.81(0.65;1.01) Neurological 827 1.26 1.19(1.09;1.29) 416 1.07 1.03(0.92;1.16) Visionproblem 47 1.16 1.13(0.81;1.56) 32 0.87 0.85(0.58;1.25) Hearingproblem 109 1.19 1.17(0.95;1.45) 93 0.83 0.81(0.65;1.02) Migraine 446 1.30 1.22(1.09;1.36) 105 1.30 1.25(1.00;1.56) Epilepsy 60 1.12 1.09(0.82;1.45) 42 0.83 0.82(0.59;1.14) Parkinson’sdisease NA NA NA NA NA NA Multiplesclerosis 10 0.47 0.46(0.24;0.88) NA NA NA Neuropathies 222 1.19 1.12(0.96;1.31) 172 1.32 1.27(1.07;1.52)

NAwheretherearelessthan5observationsorifitispossibletocalculatethenumberinrowlinebasedonthenumbersintheotherrows.

tientswithdepression(Oslin et al., 2002 ;Papakostas et al., 2003 ;Sonawalla et al., 2002 ).Incontrast,wefoundno as-sociationsbetweenTRDandprevalentcardiovacular disor-ders,whichisinlinewiththestudybyAmital et al. (2013) , but patients with TRD had an increased risk of both sub-sequenthypertensionand dyslipidemia.This findingmight suggest that although cardiovascular conditions may not necessarily have an impact on the likelihood to respond to treatment, untreated/insufficiently treated depression mightplayaroleintheoccurrenceofcertaincardiovascular conditions. However, caution should be exercised regard-ingdirectionalityof thisassociation,because cardiovascu-lar diseases can remain indolentfor a prolongedduration beforesymptomsmanifest(Li et al., 2020 ).Also,giventhat we restrictedonfirst-timedepressiveepisode,thisfinding

mightbeduetoaneffectofage,sincethepopulationis rel-ativelyyoungattimeofﬁrsttreatment andcardiovascular conditionshavealateronset.

Migraine shares common genetic variant risks with de-pression (de Boer, van den Maagdenberg, and Terwindt, 2019 ),andseveralstudieshavedemonstratedthatmigraine is often associated with depression (Minen et al., 2016 ;

Oedegaard et al., 2006 ),and that this may be a bidirec-tional association (Breslau et al., 1991 ). We found that for both women and men with TRD, migraine was more prevalent. This is in line with the study by Amital, who found that patients withTRDwere more likelyto experi-enceprevalentmigraine,butthestudyincludedaverysmall sampleandtheassociationwasnotstatisticallysigniﬁcant (Amital et al., 2013 ).Evidencesuggestthatmigraineurswho

(9)

K.B. Madsen, N.C. Momen, L.V . Pe te rs e n et al. WOMEN MEN Nofobs Risktime (years) nofTRD failures aHR∗ aHR∗∗95% CI Nof obs Risktime (years) nofTRD

failures aHR∗ aHR∗∗95%CI

Cardiovascular 70,386 439,688 720 1.51 1.43 (1.32;1.54) 47,252 281,713 491 1.35 1.31 (1.19;1.43) Hypertension 73,726 472,566 532 1.47 1.47 (1.40;1.52) 51,868 319,49 367 1.39 1.36 (1.22;1.51) Dyslipidemia 85,531 545,613 356 1.47 1.42 (1.27;1.58) 55,456 339,416 281 1.26 1.24 (1.10;1.40)

Ischemicheartdisease 88,481 603,656 80 1.11 1.05

(0.84;1.32) 60,495 389,954 108 1.09 1.07 (0.88;1.29) Atrialﬁbrilation 89,778 616,331 30 0.87 0.86 (0.59;1.23) 62,632 407,671 57 0.98 0.97 (0.74;1.26) Heartfailure 90,071 619,584 26 1.45 1.39 (0.94;2.07) 63,200 412,509 38 0.98 0.95 (0.69;1.32) Peripheralartery occlusivedisease 89,601 614,868 40 1.19 1.14 (0.83;1.56) 62,865 409,266 48 0.97 0.95 (0.71;1.26) Stroke 88,612 606,917 65 1.26 1.23 (0.95;1.57) 61,027 396,841 70 0.98 0.97 (0.76;1.23) Endocrine 82,670 548,338 419 1.57 1.52 (1.37;1.67) 59,563 381,487 186 1.33 1.24 (1.07;1.44) Diabetes 87,431 593,064 218 1.75 1.63 (1.42;1.87) 60,659 391,143 151 1.45 1.36 (1.15;1.60) Thyroiddisorder 85,345 574,311 235 1.41 1.38 (1.21;1.58) 63,178 411,453 41 1.19 1.15 (0.84;1.57) Gout NA NA NA NA NA 63,509 415,036 14 0.81 0.73 (0.43;1.24) Pulmonary 59,809 381,377 438 1.17 1.12 (1.02;1.24) 48,328 299,501 253 1.07 1.04 (0.92;1.19) Chronicpulmonary disease 72,853 484,837 259 1.19 1.15 (1.01;1.30) 55,821 357,450 160 1.14 1.10 (0.94;1.29) Allergy 70,332 458,639 368 1.15 1.09 (0.98;1.22) 53,4 335,965 165 0.99 0.96 (0.82;1.12) Gastrointestinal 87,369 592,700 152 1.18 1.13 (0.96;1.33) 60,883 392,025 147 1.23 1.17 (0.99;1.38) Ulcer/chronicgastritis 89,329 612,094 71 1.67 1.55 (1.22;1.98) 62,707 408,560 56 1.40 1.31 (0.99;1.71)

Chronicliverdisease 89,949 618,597 24 1.08 1.05

(0.69;1.59) 63,187 412,385 45 1.10 1.06 (0.79;1.43) Inﬂammatorybowel disease 89,334 613,324 38 1.19 1.14 (0.82;1.58) 63,305 413,35 23 1.36 1.31 (0.86;2.00)

( continuedonnextpage)

(10)

European Neuropsychopharmacology 51 (2021) 7–19 WOMEN MEN Nofobs Risktime (years) nofTRD failures aHR∗ aHR∗∗95% CI Nof obs Risktime (years) nofTRD failures aHR∗ aHR∗∗95%CI Diverticulardiseaseof intestine 89,798 615,395 46 1.01 0.98 (0.73;1.32) 63,419 412,771 48 1.13 1.09 (0.81;1.45) Urogenital 90,096 619,993 19 1.12 1.02 (0.65;1.63) 61,703 397,144 104 1.03 1.02 (0.84;1.24)

Chronickidneydisease 90,096 619,993 19 1.12 1.02

(0.65;1.63) 63,59 415,673 29 1.18 1.11 (0.77;1.62) Prostatedisorders NA NA NA NA NA 62,021 399,849 83 0.99 0.99 (0.79;1.23) Musculoskeletal 65,190 396,952 157 1.16 1.15 (0.93;1.43) 47,781 284,558 74 1.48 1.48 (1.08;2.04) Connectivetissue disorder 88,544 605,407 78 1.28 1.22 (0.97;1.54) 63,309 413,315 44 2.02 1.95 (1.43;2.65) Osteoporosis 88,790 604,081 87 1.04 1.03 (0.83;1.28) 63,602 414,580 40 1.23 1.18 (0.86;1.62) Painfulcondition 66,573 408,861 840 1.27 1.23 (1.15;1.32) 48,187 287,864 475 1.13 1.12 (1.02;1.22) Hematological 89,224 611,281 96 1.29 1.19 (0.97;1.47) 63,118 411,894 77 1.42 1.38 (1.10;1.74) HIV/AIDS NA NA NA NA NA NA NA NA NA NA Anemias 89,248 611,458 96 1.29 1.19 (0.97;1.46) 63,241 412,925 NA 1.43 1.39 (1.10;1.77) Oncological 86,659 596,091 122 1.07 1.06 (0.88;1.27) 61,204 403,031 104 1.01 0.99 (0.82;1.21) Neurological 76,858 495,416 537 1.29 1.24 (1.13;1.35) 56,669 352,833 306 1.24 1.19 (1.07;1.34) Visionproblem 89,375 609,747 76 1.15 1.15 (0.91;1.44) 63,198 410,625 52 1.00 0.99 (0.75;1.31) Hearingproblem 88,403 604,020 75 1.22 1.19 (0.95;1.51) 61,777 399,45 87 1.29 1.28 (1.03;1.59) Migraine 83,462 560,191 257 1.33 1.22 (1.08;1.39) 62,522 406,252 55 1.39 1.27 (0.96;1.67) Epilepsy 89,526 614,966 45 1.83 1.70 (1.25;2.30) 63,270 412,361 39 1.24 1.18 (0.85;1.62) Parkinson’sdisease 90,341 622,017 7 1.85 1.86 (0.87;4.00) 63,872 417,661 19 2.22 2.15 (1.34;3.43) Multiplesclerosis 89,982 619,029 16 1.18 1.13 (0.68;1.87) 63,770 417,459 9 2.07 2.08 (1.04;4.13) Neuropathies 86,982 586,562 205 1.28 1.19 (1.03;1.37) 61,590 396,219 121 1.20 1.12 (0.94;1.35)

NA where there are less than 5 observations or if it is possible to calculate the number in row line based on the numbers in the other rows.

∗_{adjusted for birth}_{year and age}_at_ﬁrst_{prescription.}

∗∗_adjusted_for_birth_{year, age}_{at ﬁrst}_{prescription,}_{other GMCs and number of}_other_GMCs.

(11)

suffer fromdepressionaremorelikelytoberefractoryto migrainetreatments(Minen et al., 2016 ).Ourstudysuggest thatthismayalsobethecaseforantidepressanttreatment. Autoinflammatorydiseasessuchasarthritisanddiabetes have beenassociatedwithaworseoutcome ofdepression (Oslin et al., 2002 ).Asforconnectivetissuedisorder, includ-ing arthritis, compared tothose with non-TRD, we found thatpriordiabeteswasmoreprevalentinwomenwithTRD, and for both women and men with TRD the risk of sub-sequent diabetes was increased. Although certain health behaviorsandriskfactorspartiallyexplaintheassociation of depression and diabetes, mechanisms acting on a cel-lular level (presence of a proinflammatory state/immune dysregulation)maycontributetothisassociationproposing a shared aetiology (Gibney and Drexhage 2013 ; Malhi and Mann 2018 ). TRD hasbeen linked toa state of increased low-grade-inflammation (Arteaga-Henríquez et al., 2019 ;

Benedetti et al., 2017 ), while the glucose intolerance of type 2 diabetes (the most prevalentform of diabetes) is consideredtobeduetothestateoflow-grade-inflammation inducedbyobesity,morespecificallybytheincreased pro-duction of pro-inflammatory cytokinesby the adipose tis-sue (Baldeón et al., 2014 ). Also,connective tissue disor-dersarecharacterizedbyanincreasedstate of low-grade-inflammation, more specically by a state of low-grade-inflammation induced by a high production of type II in-terferons (IFNs)(Maria et al., 2016 ).TheIFNsnotonly in-duce the expression of a multitude of pro-inflammatory compounds, but alsothe expression of the enzyme IDO-1 affecting the tryptophan catabolism and serotonin levels (Maria et al., 2016 ).

Several studies have found a higher incidence of a number of medical conditions in patients withdepression (Momen et al., 2020 ;Scott et al., 2016 ;Tegethoff et al., 2016 ).Ourstudyadds totheexistingknowledgeby show-ingthatcomparedtodepressionpatients,patientswithTRD have an increased risk of subsequent diseases related to the cardiovascular,endocrine,pulmonary,musculoskeletal andneurological systemssuggestingthatpoor responseto antidepressant treatment might have an influence onthe broadspectrum oflater GMCs.However,another possibil-ityisthatsomemedicalconditionshadbeenpresentbefore depressiononset,butfirstdiagnosedafterdepressiononset. Forinstance,theriskofprevalentconnectivetissue disor-derwasincreasedforwomenwithTRD,whileformenwith TRDonlyincidentconnectivetissuedisorderwasincreased. Forboth menandwomenwithTRDthe useof analgesics, whichisusedformusculoskeletalconditions,wasincreased both beforeandafterdepressiononset.Thismightreflect sex differencesin help-seeking behavior or differences in onset of depression between men and women. However, it may alsobe explainedby differences in sexhormones, whereestrogensareknownfortheircollagen-increasing ef-fectinconnectivetissue(Chidi-Ogbolu and Baar 2018 )and depression-preventive properties (Keyes et al., 2013 ). Es-trogensalsoregulatebasalandstimulatedHPAaxisactivity (Weiser and Handa 2009 ).The effectsof aprolongedhigh stressstateonbodilysystemshasrepeatedlybeenreported asaconfoundingfactor,increasingboththeriskofmedical conditionsaswellasdepression(Menke 2019 ).The mech-anisms underlyingtheseeffectsareslowlyunraveled.Just recently,itwasfoundthatchronicstressandHPAaxis

reg-ulation waslinked withepigenetic signatures at immune-related genes, thereby providing a possible explanation howaberrantHPAaxis functionmay contributeto height-ened inﬂammation and disease risk (Palma-Gudiel et al., 2020 ).

4.1. Strengthsandlimitations

Ourstudyhasseveralstrengths,includingthelargesample sizeanduseofnationwideregisterdatawhichiscollected onauniformbasisandwithoutanylosstofollow-up.In addi-tion,theregisterinformationallowsustostudythe tempo-ralityofthemedicalconditionsaswehavetheexactdates formedicationuseanddiagnosesgiven,whichalsoensures thattherearenoproblemscausedbyrecallorself-reporting bias.Also,weareabletopickupfatalGMCs,whichisnot possibleinsurveystudiesofGMCs.

Theproportionofpatients withTRD(5.4%)in ourstudy wassimilar towhat has been reportedin US Claimsdata (6%) (Kubitz et al., 2013 ) but was smaller than what has recentlybeenreportedinregister-basedScandinavian stud-ies (13–14%) (Gronemann et al., 2018 ; Hägg et al., 2020 ;

Reutfors et al., 2018 ) and international studies (9–20%) (Fife et al., 2017 ;Li et al., 2019 ).Themainreasonforthis arethe criteriaby whichthe populationsare drawnfrom theregisters.Theabove-mentionedstudiesinvestigate pop-ulations of patients who were all diagnosed with depres-sioninspecializedcare.Thisrestrictionmightbenecessary whentheindicationfortheprescriptionofantidepressants isnotavailableandmayincreasethevalidityofthe diagno-sis(adiagnosisinprimarycarevs.adiagnosisinspecialized psychiatric care). However,therestriction reduces gener-alizability tothe majority of depressivepatients whoare firsttreatedfordepressioninprimarycare(Musliner et al., 2019 ). Ourstudy included allfirst-time pharmacologically treateddepressivepatients,includingthosewhoneverhad contacttospecializedcare,improvingthegeneralizability to the broader population of individuals with depression. AnotherreasonmaybethatdifferentTRDdefinitionswere appliedinthestudies.ArecentstudybyGronemannetal. examiningtreatmentpatternsinpatientswithahospital di-agnosisofMDDshowedthatthosewhoswitchedtreatment most often switched withinATC class (to another SSRIor SNRI)(Gronemann et al., 2021 ).Incontrasttothestudies byGronemannetal.andReutforsetal.wechosenotto in-cludemedicationswitcheswithinthesameATC class(e.g. fromoneSSRItoanotherSSRI) toavoid misclassifying pa-tientsastreatment-resistantwheninfacttheswitchcould havebeen madein responsetoside effects(Boyce et al., 2020 ).

Anumberoflimitationsshouldalsobeacknowledged.In registers,dataisprimarilycollectedforadministrative pur-posesandtherefore the datadoes notinformonwhether thepatients adheretotreatment, whetherthetreatment reduces depressive symptoms, or the reasons for which thepatientsshift medication(Taipale and Tiihonen 2021 ). Therefore,we use shifts in treatment trialsasa proxy of failedregimes.Similarly,theexactdoseofmedicationwas not available in the registers. Instead, we used each pa-tient’spurchasepatterntoestimatetheindividualduration oftrialsforrecordingofshiftsinmedicationclassasapplied

(12)

inourpreviousstudy(Madsen et al., 2020a )anddescribed inthestudybyTanskanen et al. (2014) .Further,thereare nostandardmethods fordeﬁningTRD,butinourprevious studies,wehaveevaluatedourTRDdeﬁnitionbyrestricting theinclusionofTRDpatientsandthepatternofassociations withoutcomesremainedrobust(BangMadsen et al., 2020a ,

Madsen et al., 2020b ).Wealsoreducedthepossibilitythat patients withmoreseverepsychopathology wereincluded inourstudybyexcludingindividualswithadiagnosisof or-ganicmentaldisorder,severemooddisorderor schizophre-nia beforethe inception of ourstudy, andcensoring indi-vidualswhoreceivedthediagnosesafterstartoffollow-up. Also,GMCcategoriescomprisearangeofdiagnoses,and as-certainmentofGMCsreliesonpatientsseekingtreatment, allofwhichjointlycanlimitgeneralizabilitytospecific dis-ordersinspecificpopulations.Inaddition,therecouldbea potentialriskforsurveillancebiasaspatientswithTRDare followedupmorefrequently,whichsubsequentlycanlead tothedetecting of otherillnesses. However,in our previ-ousstudy,wefoundahigherall-causemortalityinpatients withTRDwhichisinlinewithourfindingsthatpatientswith TRD have an increasedrisk ofa broadspectrum ofGMC’s (Madsen et al., 2020a ).Finally,whenusingregister informa-tion,wehadnoinformationonpotentialconfounderssuch as smoking, alcohol consumption, diet, BMI, and physical activity.

Overall, the present study contributes with a broad overview of comorbid medical conditions in patients with TRDanddemonstratedthatpatients withTRDshowa dis-tinct patternofco-occurrencewithbothprevalentand in-cidentmedicalconditionscomparedwithnon-TRDpatients. WomenandmenwithTRDhadhigherprevalenceof con-ditionsrelatedonlytotheimmuneorneurologicalsystems, whiletheriskofsubsequentmedicalconditionswasrelated toamuch broaderspectrumofdisorders.Futureresearch is needed todetermine theexact mechanismsbehindthe bidirectionalassociations betweenTRD andgeneral medi-calconditions.

Contributors

KBM,TMOandLVPdesignedthestudy.KBM,LVP,NMandOPR conductedtheanalysisandKBMwrotetheﬁrstdraftofthe manuscript.OPR,NM,JJMG,BCMH,HD,LVP,PBMandTMO madesigniﬁcantcontributionstointerpretationofthe anal-ysisandwritingofthestudy.Allauthorsshareresponsibility forthecontentofthemanuscript.Themanuscripthasbeen approvedbyallauthors.

Declaration

of

Competing

Interest

Theauthorsdeclarenoconﬂictsofinterest.

Acknowledgement

None

Supplementary

materials

Supplementarymaterialassociatedwiththisarticlecanbe found, in the online version, at doi:10.1016/j.euroneuro. 2021.04.021 .

References

Amital,D.,Fostick,L.,Silberman,A.,Calati,R.,Spindelegger,C., Serretti, A., Juven-Wetzler, A., Souery, D., Mendlewicz, J., Montgomery,S.,Kasper,S.,Zohar,J.,2013.Physical co-morbid-ityamongtreatmentresistantvs.treatmentresponsivepatients withmajordepressivedisorder.Eur.Neuropsychopharmacol.:J. Eur.Coll.Neuropsychopharmacol.23(8),895–901.

Arteaga-Henríquez,G.,Simon,M.S.,Burger,B.,Weidinger,E., Wi-jkhuijs, A., Arolt,V., Birkenhager,T.K.,Musil, R., Müller, N., Drexhage,H.A.,2019.Low-gradeinﬂammationasapredictorof antidepressantandanti-inﬂammatorytherapyresponseinMDD Patients:asystematicreviewoftheliteratureincombination withananalysisofexperimentaldatacollectedinthe EU-MOOD-INFLAMEconsortium.FrontPsychiatry10,458.

Baldeón,R.,Lucy,K.W.,Wit,H.de,Ozcan,B.,Oudenaren,A.van, Sempértegui, F., Sijbrands, E., Grosse, L., Freire, W., Drex-hage, H.A., Leenen, P.J.M., 2014. Decreased serum level of MiR-146aassignofchronicinﬂammationinType2diabetic pa-tients.PLoSONE9(12),e115209.

Benedetti,F., Poletti,S., Hoogenboezem,T.A., Locatelli,C., de Wit,H.,Wijkhuijs,A.J.M.,Colombo,C.,Drexhage,H.A., 2017. Higherbaseline proinﬂammatorycytokines mark poor antide-pressantresponseinbipolardisorder.J.Clin.Psychiatry78(8), e986–e993.

Berlim,M.T.,Turecki,G.,2007.Deﬁnition,assessment,andstaging oftreatment-resistantrefractorymajordepression:areviewof currentconceptsandmethods.Can.J.Psychiatry52(1),46–54. Black,D.W.,Winokur,G.,Nasrallah,A.,1987.Treatmentand out-comeinsecondarydepression:anaturalisticstudyof1087 pa-tients.J.Clin.Psychiatry48(11),438–441.

deBoer,I.,Maagdenberg,A.M.J.M.vanden,Terwindt,G.M.,2019. Advance ingenetics of migraine. Curr.Opin. Neurol. 32 (3), 413–421.

Boyce, P., Hopwood, M., Morris, G., Hamilton, A., Bassett, D., Baune, B.T., Mulder, R., Porter, R., Parker, G., Singh, A.B., Outhred,T.,Das,P., Malhi,G.S., 2020.Switching antidepres-santsinthetreatmentofmajordepression:when,howandwhat toswitchto?J.Affect.Disord.261,160–163.

Breslau,N.,Davis,G.C.,Andreski,P.,1991.Migraine,psychiatric disorders, and suicide attempts: an epidemiologic study of youngadults.PsychiatryRes.37(1),11–23.

Chidi-Ogbolu, N., Baar, K., 2018. Effect of estrogen on muscu-loskeletalperformanceandinjuryrisk.Front.Physiol.9,1834. Conway,C.R.,George,M.S.,Sackeim,H.A.,2017.Towardan

evi-dence-based,operationaldeﬁnitionoftreatment-resistant de-pression:whenenoughisenough.JAMAPsychiatry74(1),9–10. Coryell,W.,Zimmerman,M.,Pfohl,B.,1985.Short-termprognosis inprimaryandsecondarymajordepression.J.Affect.Disord.9 (3),265–270.

Egede, L.E., 2007. Major depression in individuals with chronic medicaldisorders:prevalence,correlatesandassociationwith healthresourceutilization,lostproductivityandfunctional dis-ability.Gen.Hosp.Psychiatry29(5),409–416.

Fagiolini,A.andD.J.Kupfer.2003.“Istreatment-resistant depres-sionauniquesubtypeofdepression?” 3223(03).

Fife, D., Feng, Y., Wang, M.Y.-H., Chang, C.-J., Liu, C.-Y., Juang,H.-T.,Furnback,W.,Singh,J.,Wang,B.,2017. Epidemi-ologyofpharmaceuticallytreateddepressionandtreatment re-sistantdepressioninTaiwan.PsychiatryRes.252,277–283.

(13)

Gaynes, B.N., Lux, L., Gartlehner, G., Asher, G., Forman-Hoff-man, V., Green, J., Boland, E., Weber, R.P., Randolph, C., Bann, C.,Coker-Schwimmer, E., Viswanathan,M., Lohr, K.N., 2019.Deﬁningtreatment-resistantdepression.DepressAnxiety. Gibney, S.M.,Drexhage, H.A., 2013.Evidencefora dysregulated immunesystemintheetiologyofpsychiatricdisorders.J. Neu-roimmunePharmacol.:Off.J.eSoc.NeuroImmunePharmacol.8 (4),900–920.

Gronemann,F.H.,Jorgensen,M.B.,Nordentoft,M.,Andersen,P.K., Osler,M.,2018.Incidenceof,riskfactorsfor,andchangesover time in treatment-resistant depression in Denmark: a regis-ter-basedcohortstudy.J.Clin.Psychiatry79(4).

Gronemann,F.H.,Petersen,J.,Alulis,S.,Jensen,K.J., Riise,J., Ankarfeldt, M.Z., Solem, E.J., Bødker, N., Osler, M., 2021. “Treatment patternsin patientswithtreatment-resistant de-pressionindanish patientswithmajordepressive disorder.J. Affect.Disord.287,204–213.

Hägg, D., Brenner, P., Reutfors, J., Li, G., DiBernardo, A., Bodén,R.,Brandt,L.,2020.Aregister-basedapproachto iden-tifyingtreatment-resistantdepression-comparisonwithclinical deﬁnitions.PLoSOne15(7),e0236434.

Helweg-Larsen,K., 2011.TheDanishregisterofcausesofdeath. Scand.J.PublicHealth39(7Suppl),26–29.

IsHak, W.W., Steiner, A.J., Klimowicz, A., Kauzor, K., Dang, J., Vanle, B., Elzahaby, C., Reid, M., Sumner, L., Danovitch, I., 2018.Majordepressioncomorbidwithmedicalconditions: anal-ysis of quality of life, functioning, and depressive symptom severity.Psychopharmacol.Bull.48(1),8–25.

Keitner,G.I.,Ryan,C.E.,Miller,I.W.,Kohn,R.,Epstein,N.B.,1991. 12-monthoutcome ofpatientswithmajordepressionand co-morbid psychiatricor medical illness(compound depression). Am.J.Psychiatry148(3),345–350.

Keyes, K.M., Cheslack-Postava, K., Westhoff, C., Heim, C.M., Haloossim, M., Walsh, K., Koenen, K., 2013. Association of hormonalcontraceptiveusewithreduced levelsofdepressive symptoms: a nationalstudy ofsexually active women in the UnitedStates.Am.J.Epidemiol.178(9),1378–1388.

Kildemoes,H.W.,Sorensen,H.T.,Hallas,J.,2011.TheDanish na-tionalprescriptionregistry.Scand.J.PublicHealth39(7Suppl), 38–41.

Kronish, I.M., Rieckmann, N., Halm, E.A., Shimbo, D., Vorch-heimer, D., Haas, D.C., Davidson, K.W., 2006. Persistent de-pressionaffectsadherenceto secondarypreventionbehaviors afteracutecoronarysyndromes.J.Gen.Intern.Med.21(11), 1178–1183.

Kubitz,N.,Mehra,M.,Potluri,R.C.,Garg,N.,Cossrow,N.,2013. Characterizationoftreatmentresistantdepressionepisodesina cohortofpatientsfromaUScommercialclaimsdatabase.PLoS One8(10),e76882.

Li,C.H.,Woo,T.,Ganesananthan,S.,2020.Whatistheassociation betweendepressionandcardiovasculardisease?JAMA Psychia-try.

Li,G.,Fife,D., Wang,G., Sheehan,J.J.,Boden, R., Brandt,L., Brenner,P.,Reutfors,J.,DiBernardo,A.,2019.All-cause mor-talityinpatientswithtreatment-resistantdepression:acohort studyintheUSpopulation.Ann.Gen.Psychiatry18,23. Lynge,E.,Sandegaard,J.L.,Rebolj,M.,2011.Thedanishnational

patientregister.Scand.J.PublicHealth39(7Suppl),30–33. Madsen, B., Kathrine, Petersen, L.V., Plana-Ripoll, O.,

Musliner, K.L., Debost, J.-C.P., Gronemann, F.H., Mortensen, P.Bo, Munk-Olsen, T., 2020a. Early labor force exits in patients with treatment-resistant depression: an assessment of work years lost in a danish nationwide reg-ister-based cohort study. Ther. Adv. Psychopharmacol. 10, 2045125320973791.

Madsen,K.B.,Plana-Ripoll,O.,Musliner,K.L.,Debost,J.-C.P., Pe-tersen,L.V.,Munk-Olsen,T.,2020b.Cause-speciﬁclifeyearslost inindividualswithtreatment-resistantdepression:adanish

na-tionwideregister-basedcohortstudy.J.Affect.Disord.280(Pt A),250–257.

Maes,M.,Kubera,M.,Obuchowiczwa,E.,Goehler,L.,Brzeszcz,J., 2011. Depression’s multiple comorbidities explained by (neuro)inﬂammatory and oxidative & nitrosative stress path-ways.NeuroEndocrinol.Lett.32(1),7–24.

Malhi,G.S.,Mann,J.J.,2018.“Depression.”.Lancet(Lond.,Engl.) 392(10161),2299–2312.

Maria,N.I.,Helden-Meeuwsen,C.G.V.,Brkic,Z.,Paulissen,S.M.J., Steenwijk,E.C.,Dalm,V.A.,Daele,P.L.V.,MartinvanHagen,P., Kroese, F.G.M., Roon, J.A.G.V., Harkin, A., Dik, W.A., Drex-hage,H.A., Lubberts,E., 2016. Associationofincreased treg cell levels with elevated indoleamine 2,3-dioxygenase activ-ityand animbalancedkynurenine pathwayin interferon-pos-itiveprimarysjögren’ssyndrome.ArthritisRheumatol.68(7), 1688–1699.

Menke,A.,2019.IstheHPAaxisastargetfordepressionoutdated, oristhereanewhope?Front.Psychiatry10,101.

Miller,G., Chen,E.,Cole, S.W., 2009.Health psychology: devel-opingbiologicallyplausiblemodelslinkingthesocialworldand physicalhealth.Annu.Rev.Psychol.60,501–524.

Minen, M.T., Dhaem, O.B.D., Diest, A.K.V., Powers, S., Schwedt, T.J., Lipton, R., Silbersweig, D., 2016. Migraine anditspsychiatriccomorbidities.J.Neurol.Neurosurg. Psychi-atr.87(7),741–749.

Momen, N.C., Plana-Ripoll, O., Agerbo, E., Benros, M.E., Børglum, A.D., Christensen, M.K., Dalsgaard, S., Degen-hardt, L., Jonge, P.de, Debost, J.-C.P.G., Fenger-Grøn, M., Gunn, J.M., Iburg, K.M., Kessing, L.V, Kessler, R.C., Laursen, T.M., Lim, C.C.W., Mors, O., Mortensen, P.B., Musliner,K.L.,Nordentoft,M.,Pedersen,C.B.,Petersen,L.V, Ribe, A.R., Roest, A.M., Saha, S., Schork, A.J., Scott, K.M., Sievert,C.,Sørensen,H.J.,Stedman,T.J.,Vestergaard,M., Vil-hjalmsson,B.,Werge,T.,Weye,N.,Whiteford,H.A.,Prior,A., McGrath, J.J., 2020. Association between mental disorders andsubsequentmedicalconditions.N.Engl.J.Med.382(18), 1721–1731.

Mors,O.,Perto,G.P.,Mortensen,P.Bo,2011.TheDanishpsychiatric centralresearchregister.Scand.J.PublicHealth39(7Suppl), 54–57.

Moussavi,S., Chatterji,S.,Verdes,E.,Tandon, A.,Patel,V., Us-tun, B., 2007. Depression, chronic diseases, and decrements inhealth: resultsfrom the worldhealth surveys.Lancet 370 (9590),851–858.

Musliner,K.L.,Liu,X.,Gasse,C.,Christensen,K.S.,Wimberley,T., Munk-Olsen,T.,2019.Incidenceofmedicallytreateddepression inDenmarkamongindividuals15-44yearsold:acomprehensive overviewbasedonpopulationregisters.ActaPsychiatr.Scand. 139(6),548–557.

Niles,A.N.,Dour,H.J.,Stanton,A.L.,Roy-Byrne,P.P.,Stein,M.B., Sullivan,G.,Sherbourne,C.D.,Rose,R.D.,Craske,M.G.,2015. Anxiety and depressive symptoms and medical illness among adults with anxiety disorders. J. Psychosom. Res. 78 (2), 109–115.

Oedegaard, K.J., Neckelmann, D., Mykletun, A., Dahl, A.A., Zwart,J.A.,Hagen,K.,Fasmer,O.B.,2006.Migrainewithand withoutAura:associationwithdepressionandanxietydisorder inapopulation-basedstudy.TheHUNTstudy.Cephalalgia:Int. J.Headache26(1),1–6.

Oslin,D.W.,Datto,C.J.,Kallan,M.J.,Katz,I.R.,Edell,W.S., Ten-Have,T.,2002. Associationbetweenmedicalcomorbidityand treatmentoutcomesinlate-lifedepression.J.Am.Geriatr.Soc. 50(5),823–828.

Palma-Gudiel,H., Prather, A.A., Lin, J., Oxendine,J.D., Guinti-vano, J., Xia, K., Rubinow, D.R., Wolkowitz, O., Epel, E.S., Zannas, A.S., 2020. HPA axis regulation and epigenetic pro-grammingofimmune-relatedgenesinchronicallystressedand non-stressedmid-lifewomen.BrainBehav.Immun..

(14)

Papakostas, G.I., Petersen, T., Sonawalla, S.B., Merens, W., Iosifescu,D.V, Alpert,J.E.,Fava,M.,Nierenberg,A.A.,2003. Serumcholesterolintreatment-resistantdepression. Neuropsy-chobiology47(3),146–151.

Patten, S.B., 2001. Long-term medicalconditions and major de-pressionina Canadianpopulationstudyatwaves 1and 2.J. Affect.Disord.63(1–3),35–41.

Pedersen,C.B.,2011.TheDanishcivilregistrationsystem.Scand. J.PublicHealth39(7Suppl),22–25.

Pedersen,C.B.,Mors,O.,Bertelsen,A.,Waltoft,B.L.,Agerbo,E., McGrath,J.J.,Mortensen,P.B.,Eaton,W.W.,2014.A compre-hensivenationwidestudyoftheincidencerateandlifetimerisk fortreatedmentaldisorders.JAMAPsychiatry71(5),573–581. Perlis, R.H., Iosifescu, D.V, Alpert,J., Nierenberg, A.A.,

Rosen-baum,J.F.,Fava,M., 2004.Effect ofmedicalcomorbidityon responsetoﬂuoxetineaugmentationordoseincreasein outpa-tientswithtreatment-resistantdepression.Psychosomatics45 (3),224–229.

Prior, A., Fenger-Grøn, M., Larsen, K.K., Larsen, F.B., Robin-son, K.M., Nielsen, M.G., Christensen, K.S., Mercer, S.W., Vestergaard, M., 2016. The association between perceived stress and mortality among people with multimorbidity: a prospective population-basedcohortstudy.Am. J.Epidemiol. 184(3),199–210.

Reutfors, J., Andersson, T.M.L., Brenner, P., Brandt, L., DiBernardo, A., Li, G., Hagg, D., Wingard, L., Boden, R., 2018. Mortality in treatment-resistantunipolar depression: a register-basedcohortstudyinSweden.J.Affect.Disord.238, 674–679.

Rush,A.J.,Wisniewski,S.R.,Warden,D.,Luther,J.F.,Davis,L.L., Fava, M., Nierenberg, A.A., Trivedi, M.H., 2008. Selecting

amongsecond-stepantidepressantmedicationmonotherapies: predictive valueof clinical, demographic, orﬁrst-step treat-mentfeatures.Arch.Gen.Psychiatry65(8),870–880.

Scott,K.M., Lim,C.,Al-Hamzawi,A., Alonso,J., Bruffaerts,R., Caldas-de-Almeida,J.M.,Florescu,S.,Girolamo,G.d.,Hu,C., Jonge,P.d.,Kawakami,N.,Medina-Mora,M.E.,Moskalewicz,J., Navarro-Mateu,F.,O’Neill,S.,Piazza,M.,Posada-Villa,J., Tor-res,Y.,Kessler,R.C.,2016.Associationofmentaldisorderswith subsequent chronic physical conditions: world mental health surveysfrom17countries.JAMAPsychiatry73(2),150–158. Sonawalla, S.B., Papakostas, G.I., Petersen, T.J., Yeung, A.S.,

Smith, M.M., Sickinger, A.H., Gordon, J., Israel, J.A., Ted-low,J.R.,Lamon-Fava,S.,Fava,Ma.,2002.Elevatedcholesterol levelsassociatedwithnonresponseto ﬂuoxetinetreatmentin majordepressivedisorder.Psychosomatics43(4),310–316. Taipale,H.,Tiihonen,J.,2021.Registry-basedstudies:whatthey

cantellus,andwhattheycannot.Eur.Neuropsychopharmacol.: J.Eur.Coll.Neuropsychopharmacol.45,35–37.

Tanskanen, A., Taipale,H.,Koponen, M.,Tolppanen,A.-M., Har-tikainen,S.,Ahonen,R.,Tiihonen,J., 2014.Fromprescriptions todruguseperiods-thingstonotice.BMCRes.Notes7,796. Tegethoff,M.,Stalujanis,E.,Belardi,A.,Meinlschmidt,G.,2016.

Chronologyofonsetofmentaldisordersandphysicaldiseasesin mental-physicalcomorbidity-anationalrepresentativesurvey ofadolescents.PLoSOne11(10),e0165196.

Weiser, M.J., Handa,R.J., 2009. Estrogenimpairs glucocorticoid dependentnegativefeedbackonthe hypothalamic-pituitary-a-drenalaxisviaestrogen receptoralphawithin the hypothala-mus.Neuroscience159(2),883–895.