INTRODUCTION
Acco ding o he psychobiological model o empe amen
and cha ac e 1 he human empe amen can be di ided in o
ou di e en independen dimensions and cha ac e in o
h ee di e en dimensions. The empe amen al ea u es a e: 1)
he beha io in ela ion o new o pleasu e-p oducing s imuli
(no el y seeking, NS), 2) beha io al inhibi ion in ela ion o
18 Copy igh
©
2014 Ko ean Neu opsychia ic Associa ion
issues which may lead o nega i e consequences (ha m a oid-
ance, HA), 3) con inuing o beha io ha has ea lie been suc-
cess ul in he hope o ewa d ( ewa d dependence, RD), and
he endency o main ain ce ain beha io despi e us a ion
(pe sis ence, P).1 These empe amen dimensions a e sugges -
ed o be connec ed wi h cen al neu o ansmi e ci cui s in
he cen al ne ous sys em: dopamine (no el y seeking), se-
o onin (ha m a oidance) and no epineph ine ( ewa d de-
pendence).2 The h ee dimensions o cha ac e ma u e in adul -
hood and in luence pe sonal and social e ec i eness by insigh
lea ning abou sel -concep s. Sel -concep s a y acco ding o
he ex en o which a pe son iden i ies he sel as 1) an au ono-
mous indi idual (sel -di ec edness, SD), 2) an in eg al pa o
humani y (coope a i eness, C), and 3) an in eg al pa o he
uni e se as a whole (sel - anscendence, ST).1
The Tempe amen and Cha ac e In en o y (TCI) has been
A Clus e Model o Tempe amen as an Indica o o An idep essan
Response and Symp om Se e i y in Majo Dep ession
Vesa Paa onen1,2 , Olli Kampman1,4, A i Illi1,6, Me ja Viikki1,5, Eija Se älä-Soikkeli3, and Esa Leinonen1,2
1Uni e si y o Tampe e School o Medicine, Tampe e, Finland
2Tampe e Uni e si y Hospi al, Depa men o Psychia y, Tampe e, Finland
3Kan a-Häme Cen al Hospi al, Depa men o Psychia y, Hämeenlinna, Finland
4Seinäjoki Hospi al Dis ic , Depa men o Psychia y, Seinäjoki, Finland
5Tampe e Men al Heal h Cen e, Tampe e, Finland
6Sa akun a Hospi al Dis ic , Depa men o Psychia y, Ha ja al a, Finland
Objec i eaaNo enough is known abou which pa ien s su e ing om majo dep essi e diso de bene i om an idep essan d ug ea -
men . Indi idual empe amen is ela i ely s able o e a pe son’s li espan and is hough o be la gely biologically p ede ined. We assessed
how empe amen p o iles a e ela ed o dep ession and p edic he e icacy o an idep essan ea men .
Me hodsaaWe ec ui ed one hund ed Finnish ou pa ien s (aged 19 o 72) su e ing om majo dep essi e diso de , o whom 86 comple -
ed he 6-week s udy. We assessed hei empe amen ea u es wi h he Tempe amen and Cha ac e In en o y and used clus e analysis o
de e mine he pa ien ’s empe amen p o ile. We also ca ego ized he pa ien s acco ding o he ege a i e symp oms o majo dep essi e
diso de .
Resul saaThe e was an associa ion be ween skewed empe amen p o ile and se e i y o majo dep essi e diso de , bu he empe amen
p o iles alone did no p edic an idep essan ea men esponse. Those wi h highe baseline ege a i e symp oms sco e had modes ea -
men esponse. Ou model wi h baseline Mon gome y Åsbe g Dep ession Ra ing Scale (MADRS) ege a i e symp oms, age and empe a-
men clus e s as explana o y a iables explained 20% o he a iance in he endpoin MADRS sco es.
ConclusionaaThe empe amen clus e s we e associa ed bo h wi h se e i y o dep ession and an idep essi e ea men esponse o de-
p ession. The e ec o he empe amen p o ile alone was modes bu , combined wi h ege a i e symp oms o dep ession, hei explana-
o y powe was mo e ma ked sugges ing ha he e could be an associa ion o hese wo in he biological basis o MDD.
Psychia y In es ig 2014;11:18-23
Key Wo dsaa
Dep essi e diso de , Tempe amen , TCI, An idep essi e agen s, T ea men esponse.
Recei ed: Ma ch 5, 2013 Re ised: June 14, 2013
Accep ed: July 3, 2013 A ailable online: Janua y 21, 2014

Co espondence: Vesa Paa onen, MD
Uni e si y o Tampe e School o Medicine, A o Building, FI-33014 Tampe e,
Finland
Tel: +358-407389051, Fax: +358-335516164, E-mail: esa.paa onen@u a. i
cc This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons
A ibu ion Non-Comme cial License (h p://c ea i ecommons.o g/licenses/by-
nc/3.0) which pe mi s un es ic ed non-comme cial use, dis ibu ion, and ep oduc-
ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
P in ISSN 1738-3684 / On-line ISSN 1976-3026
OPEN ACCESS
h p://dx.doi.o g/10.4306/pi.2014.11.1.18
ORIGINAL ARTICLE
online © ML Comm
V Paa onen e al.
www.psychia yin es iga ion.o g 19
used in gene al popula ion s udies and in s udies including
pa ien s wi h majo dep essi e diso de (MDD) o assess how
di e en empe amen dimensions a e associa ed wi h his
diso de .1 HA has epo edly been highe in MDD pa ien s
han in gene al popula ion3-9 and has been s a e dependen
in MDD.5,6 HA has also been associa ed wi h dep essi e symp-
oms in gene al popula ion.10-12 HA, RD and NS ha e been
ound o ha e ai -like cha ac e is ics ha a e ela ed o he
amilial occu ence o dep ession.4 High HA has been epo -
ed o p edic poo ea men ou come.8,13-15
HA also seems o ha e a ai -like cha ac e is ic, e lec ing
gene ic suscep ibili y in dep ession p one subjec s.4,16 I seems
ha HA is bo h a s a e- and ai -dependen a iable in MDD.17
Howe e , one s udy using he Hospi al Anxie y and Dep es-
sion Scale (HAD), which excludes soma ic symp oms, ound
no associa ion be ween HA and MDD.18 This inding may
sugges ha HA is connec ed speci ically o he soma ic com-
ponen o MDD. The con ibu ion o o he empe amen di-
mensions in MDD is somewha con o e sial o limi ed. Low
RD may be associa ed wi h MDD and dep essi e symp oms in
gene al popula ion,4 bu he esul s a e unequi ocal.10,19 NS
seems o be s a e dependen in MDD and al oge he lowe in
MDD pa ien s.4,11 Howe e , high NS has been associa ed wi h
his o y o suicide a emp s in gene al popula ion.20 In one
s udy P was a s a e ma ke in dep ession.17
In a s udy by G ucza e al.20 di e en combina ions o em-
pe amen dimensions we e associa ed wi h di e en dep es-
si e symp oms. I has been p oposed ha he symp oms o
MDD consis o clus e s, which a e linked o dis inc gene ic
mechanisms which when combined in one indi idual, can
lead o a diagnosable psychopa hology.21 Suzuki e al.22 p o-
posed a h ee- ac o model o he MADRS o di e en ia e he
ege a i e symp oms (soma ic symp oms) obse ed in a pa-
ien g oup. This h ee- ac o model has been used in some
s udies wi h MDD pa ien samples.23-25 I has been p oposed
ha he ege a i e symp oms a e connec ed o enhanced ex-
p ession o 5HT2A ecep o s.25 We ound no s udies add ess-
ing he associa ion o empe amen and ege a i e symp oms
o MDD. The p esen s udy analyzes i empe amen p o iles
in associa ion wi h ege a i e symp oms explain he an ide-
p essan ea men esponse in MDD pa ien s and i he se-
e i y o dep ession is associa ed wi h cu en empe amen
clus e s.
METHODS
A hund ed Finnish ou pa ien s we e ec ui ed om sec-
onda y ou pa ien se ices, p ima y heal h ca e and by news-
pape ad e isemen s du ing he yea s 2002–2006 in he a ea
o Tampe e in sou he n Finland. The s udy was app o ed by
he local Human Subjec s Re iew Commi ee and subjec s
pa icipa ed ha ing gi en in o med, olun a y, w i en con-
sen . The ec ui men esul ed in 41 emale and 59 male ou -
pa ien s, aged 19–72 y s (mean 40.7 yea s, SD±14.0). Pa ien s
me he c i e ia o majo dep essi e episode acco ding o
he Diagnos ic and S a is ical Manual o Men al Diso de s,
ou h edi ion (DSM-IV). All pa ien s we e diagnosed by a
psychia is and he se e i y o hei dep ession was e alua -
ed wi h he Mon gome y-Åsbe g Dep ession Ra ing Scale
(MADRS). Those pa ien s who sco ed 20 o highe a base-
line MADRS we e included in he s udy. Pa ien s wi h se e e
soma ic illness o medica ion a ec ing hei mood, o he sig-
ni ican psychia ic diso de s (bipola illness, psychosis o
se e e pe sonali y diso de s) o pa ien s wi h alcohol o sub-
s ance abuse we e excluded om he s udy. Eigh y-six pa ien s
comple ed he en i e s udy acco ding o he p o ocol and we e
included in he inal analysis.26 S udy da a was collec ed on
h ee occasions. A he i s isi basic sociodemog aphic da a
was collec ed: gende , age, ma i al s a us, educa ion, wo kplace
be o e he sick lea e, soma ic illnesses and hei medica ions,
o he psychia ic diso de s, and possible use o psycho ophic
medica ions. The baseline MADRS o m was sco ed and he
pa ien s comple ed he empe amen sec ion o he TCI ques-
ionnai e o assess he empe amen p o ile.1 All pa ien s we e
p esc ibed ei he ci alop am, luoxe ine o pa oxe ine. Anx-
ioly ics and seda i e hypno ics as adju an ea men and o h-
e medica ion o concomi an gene al medical condi ions
we e allowed. A he second isi , h ee weeks a e ini ia ion
o ea men , pa ien s’ adhe ence o ea men and he dosage
o he medica ions we e checked. Compliance was e alua ed
by a medica ion dia y kep by he pa ien . T ea men compli-
ance was deemed su icien i he pa ien had aken he medi-
ca ion on a leas 80% o he days in he s udy pe iod. A he
hi d isi , six weeks a e he ini ia ion, pa ien s’ adhe ence
was moni o ed again and he MADRS and TCI o ms we e
comple ed again (endpoin da a). In he case o possible d op-
ou s he necessa y pa ien in o ma ion on he easons o
d opou was also collec ed. The empe amen p o iles we e
de e mined om he baseline TCI da a. The ege a i e symp-
oms we e assessed as he sum o ques ions h ee o i e in he
MADRS.22 These a e impai ed sleep, impai ed appe i e, and
inne ension.
S a is ical me hods
A wo-s ep clus e analysis was used o he de ini ion o he
pa ien ’s empe amen p o ile. In ou clus e model, we decid-
ed o use h ee empe amen dimensions, NS, HA and RD
wi h hei baseline sco es. In he s a is ical analysis he pa-
ien s we e di ided in o h ee clus e s.
The di e ences in con inuous a iables (MADRS o al sco e,
20 Psychia y In es ig 2014;11:18-23
A Clus e Model o Tempe amen in Majo Dep ession
MADRS ac o sco es and age) be ween he clus e s we e cal-
cula ed wi h ANOVA. The di e ence in MADRS change be-
ween low and high ege a i e symp om g oups was analyzed
wi h - es . Di e ences be ween g ouping a iables we e cal-
cula ed wi h χ2-s a is ics. Non-pa ame ic es s we e used in
compa isons in o dinal a iables be ween di e en clus e s
(used medica ions, pa ien compliance). Pea son’s co ela ion
coe icien s we e calcula ed be ween MADRS o al sco e and
ege a i e symp om sco e, bo h a baseline and endpoin .
In he mul i a ia e analysis all a iables included in he
s udy and likely o ha e an impac on ei he dep ession se e i-
y o ea men esponse we e used in he models. The e ec o
backg ound a iables on clus e s was analyzed wi h a mul i-
nominal logis ic eg ession model. Gende , age, se e i y o
dep ession, an idep essan aken and dose, subjec i e adhe -
ence o ea men , and ea lie dep ession episode we e used as
explana o y a iables. A linea eg ession model (ANCOVA)
was used o es ing he e ec s o empe amen clus e s and
o he a iables on MADRS endpoin sco es. The i s model
included empe amen clus e s and age, and he second model
empe amen clus e s, age and he MADRS ege a i e symp-
oms (ques ions 3–5) a baseline as explana o y a iables. All
analyses we e pe o med wi h SPSS o Windows so wa e
( e sion 17.0).
RESULTS
The clus e analysis esul ed in he ollowing clus e s: LNS/
HHA/LRD, INS/HHA/HRD and HNS/LHA/(HRD) wi h H
indica ing high le el, L low le el and I in e media e le el on
he empe amen dimensions. In he hi d clus e RD did no
each s a is ical signi icance in he clus e ing model. In he
i s clus e we disco e ed he mos obus slope in he NS
and HA a baseline. In he second clus e he e we e ele a ed
poin s in HA. The esul s o he clus e analysis and MADRS
sco es in each clus e a e p esen ed in Table 1.
The e we e no di e ences in he dis ibu ions o gende
be ween he clus e s (p=0.23, chi-squa e es ). The pa ien s in
clus e 1 we e olde han in o he clus e s (age mean±SD,
clus e 1=45.9±11.4, clus e 2=37.4±14.4, clus e 3=38.2±15.3;
p=0.03, ANOVA). The e was no di e ence be ween he clus-
e s in he dosages o he medica ions aken in weeks one o
h ee (p=0.48, K uskal Wallis es ), no in compliance o ea -
men (p=0.69). Gende , se e i y o dep ession, an idep es-
san aken and dose, adhe ence o ea men , and ea lie de-
p ession episode had no e ec on he clus e s in he mul i-
nominal eg ession model. Age o he pa ien s had a ma ginal
e ec on clus e s (p=0.051) in he mul inominal eg ession
model.
The co ela ions be ween MADRS ege a i e symp om
sco e wi h MADRS o al sco e we e a baseline 0.73, (p<0.001)
and a endpoin 0.76, (p<0.001). The MADRS ege a i e
symp om sco e a baseline had a mode a e co ela ion wi h
MADRS endpoin sco es ( =0.38, p<0.001), and a non-signi i-
can co ela ion wi h MADRS sco e change ( =0.13, p=0.26).
We also analyzed he MADRS sco e change be ween pa ien s
wi h low (1–7, n=50) and high (8 o mo e, n=48) ege a i e
symp oms. The di e ence was close o signi ican [MADRS
Table 1. Resul s o he clus e analysis. All sco es excep esponse pe cen ages a e indica ed as mean±SD
Clus e LNS/HHA/LRD, N=33 INS/HHA/HRD, N=35 HNS/LHA/(HRD), N=30
TCI baseline sco e
NS 13.9±5.2 19.6±7.1 25.7±4.9
HA 26.8±6.1 27.2±3.6 16.0±4.2
RD 11.8±2.2 18.2±2.5 16.6±3.6
Baseline MADRS*28.3±6.1 27.3±5.7 25.0±4.4
Fac o 1 (dyspho ia) 8.5±2.6 7.8±2.2 7.4±1.4
Fac o 2 ( e a da ion) 11.7±2.2 12.1±2.7 11.0±2.9
Fac o 3 ( ege a i e symp oms) 7.8±3.1 7.4±2.8 6.4±2.7
Endpoin MADRS** 14.1±9.1 13.5±8.3 8.3±5.5
Fac o 1 (dyspho ia) 3.7±3.2 3.9±2.7 2.5±1.9
Fac o 2 ( e a da ion)***** 6.3±4.5 6.0±3.7 3.4±3.0
Fac o 3 ( ege a i e symp oms) 3.5±2.3 3.7±2.9 2.4±1.9
MADRS sco e change*** 14.2±7.4 14.3±8.0 16.7±6.0
Response (pe cen age decline in MADRS)**** 51.6% 51.9% 66.7%
*p=0.05 be ween g oups (ANOVA), **p=0.01 be ween g oups (ANOVA), ***p=0.36 be ween g oups (ANOVA), ****p=0.04 be ween g oups
(ANOVA), *****p=0.01 be ween g oups (ANOVA). TCI: Tempe amen and Cha ac e In en o y, MADRS: Mon gome y-Åsbe g Dep ession
Ra ing Scale, NS: no el y seeking, HA: ha m a oidance, RD: ewa d dependence, wi h H indica ing high le el, L low le el and I in e media e
le el on he empe amen dimensions
V Paa onen e al.
www.psychia yin es iga ion.o g 21
change, mean (±SD), low symp oms=13.5 (±5.5), high symp-
oms=16.6 (±8.5), p=0.052, - es ]. The e we e no di e ences
in he MADRS ege a i e symp om sco e a baseline o a end-
poin be ween he di e en clus e s (p=0.17, baseline; p=0.14,
endpoin , ANOVA). The e was a non-signi ican co ela ion
be ween baseline MADRS dyspho ia symp oms and MADRS
sco e change ( =0.17, p=0.11). MADRS endpoin sco es we e
used as he ou come a iable in wo linea eg ession models.
In he i s model age and empe amen clus e s we e used as
explaining a iables. This model explained 10% o he a i-
ance in he MADRS endpoin sco es (p=0.04; powe 0.69).
The clus e s explained 9% (p=0.02), and age explained 1%
(p=0.36). In he second model, baseline MADRS ege a i e
symp oms, age and empe amen clus e s we e used as ex-
plana o y a iables. This model explained 20% o he a iance
in he MADRS endpoin sco es (p=0.001; powe 0.96 o he
comple e model). In his model he baseline ege a i e symp-
oms explained 12% (p=0.001), age 0.2% (p=0.70) and em-
pe amen clus e s 5% (p=0.12). Using he del a sco es o
MADRS as an ou come a iable, and age and empe amen
clus e s as explana o y a iables ( i s model), and ege a i e
symp oms, age and empe amen clus e s as explana o y a i-
ables (second model) esul ed in non-signi ican models ( i s
model: ηp2=0.085, p=0.061, powe =0.61; second model: ηp2=
0.096, p=0.087, powe =0.60).
DISCUSSION
Ou main hypo hesis was ha empe amen clus e s in pa-
ien s wi h MDD explain he ea men esponse. In p ac ice
his means ha di e en empe amen p o iles could unc-
ion as a classi ying ac o and ha MDD pa ien s could be di-
ided in o di e en g oups wi h di e en ou comes o an i-
dep essan ea men . In ou s udy we used p ima ily he MA-
DRS endpoin sco es as an ou come a iable in he mul i-
a ia e analyses. Using he del a sco es o MADRS as an ou -
come a iable esul ed in non-signi ican models al hough
he e was a end owa ds a be e esponse in pa ien s wi h
high ege a i e symp oms. The p esen esul s sugges ha
he combined e ec o ege a i e symp oms and empe amen
clus e s is impo an in ela ion o he dep ession ea men
ou come when measu ed as pos - ea men symp oms. How-
e e , hese ac o s showed a non-signi ican e ec when p e-
dic ing he change in dep ession sco es du ing ea men . This
inding may be due o bo h he empe amen clus e s and p e-
ea men ege a i e symp oms ep esen ing dep essi e ai s
less connec ed wi h he magni ude o symp om alle ia ion
du ing ea men .
The app oach o using TCI empe amen clus e s o p e-
dic ing he esponse o an idep essan ea men in MDD pa-
ien s is no el. In se e al s udies indi idual dimensions o em-
pe amen ha e been used as p ecu so s. Two ea lie s udies
wi h gene al popula ion samples ha e used combina ions o
high o low empe amen ai s o p edic ing di e en clini-
cal ea u es, bu in hese s udies no clus e analysis me hod
was used in classi ying he empe amen ai s.18,20 This s udy
did no include he cha ac e dimensions o he TCI (SD, C,
ST) in he explana o y model. Adding he cha ac e ai s o
he p edic o s in he s a is ical model migh ha e inc eased
i s p edic i e alue ega ding an idep essan esponse, since
many s udies ha e demons a ed ha SD exhibi s a s a e/ ai
ma ke in dep ession.9,27,28
The e we e some limi a ions conce ning ou pa ien sam-
ple and s udy se ing. In con as o some ea lie s udies, ou
pa ien sample comp ised solely ou pa ien s. This may ha e
esul ed in lowe in ensi y o symp oms as e lec ed by he
MADRS sco es. Tempe amen p o iles could ha e had mo e
explana o y powe i he pa ien sample had included inpa-
ien s wi h mo e se e e dep ession. The pa ien s we e deemed
complian wi h medica ion i hey ook he p esc ibed medica-
ion a leas 80% o he ime, which can be ega ded as a mod-
e a e le el o ea men compliance, and he da a we e col-
lec ed om pa ien epo s, which in some cases may p oduce
un eliable esul s. No did he pa ien s ecei e any speci ic
psychological ea men du ing he s udy, bu we e ea ed in
a s anda d seconda y ou pa ien se ing.
As he ela ionship be ween empe amen and ege a i e
symp oms o dep ession has no p e iously been s udied, a
pos -hoc analysis wi h ege a i e symp oms was pe o med
in his s udy. This was done by sepa a ing he ege a i e symp-
oms om he o he symp oms o dep ession (dyspho ia and
e a da ion) which was based on he s udy by Suzuki e al.22
I has been p oposed ha he ege a i e symp oms a e con-
nec ed o enhanced exp ession o 5-HT2A ecep o s.25
To assess pa ien s’ empe amen p o iles we used Cloninge ’s
TCI, which has been widely used, alida ed and shown o be
eliable in s udies on gene al popula ion and MDD pa ien s.1
Tempe amen p o iles we e de e mined by clus e ing he dis-
ibu ions in he h ee empe amen dimensions. Due o he
limi ed sample size, he numbe o clus e s was de e mined
as h ee in he analysis o yield g oups o easonable size. The
clus e ing me hod was able o di e en ia e be ween he h ee
combina ions o empe amen ai s, al hough in he hi d
clus e he di e ence on he dimension RD did no each s a-
is ical signi icance. I has been sugges ed ha high RD co e-
la es nega i ely wi h dep essi e symp oms, bu he e idence
is con adic o y.3,4,9,10 The clus e s di e ed on he dimension
NS as i was low in clus e one, in e media e in clus e wo and
high in clus e h ee. The hi d clus e (HNS/LHA/HRD) p ob-
ably e lec s mo e impulsi e dep ession, and di e ges subs an-
22 Psychia y In es ig 2014;11:18-23
A Clus e Model o Tempe amen in Majo Dep ession
ially om he ypical empe amen p o ile o an MDD pa ien ,
and in his s udy om he o he wo empe amen clus e s.
This may explain why his subg oup o pa ien s eco e ed be -
e om he e a da ion symp oms han did he o he pa ien s.
In addi ion, his g oup showed a highe pe cen age o MADRS
changes. This di e ence in esponse could be due o he p e-
dic i e e ec o HA on MDD emission shown in ou ea lie
epo .29 In he i s clus e we disco e ed he mos dis inc
sloping in he dis ibu ion o he NS and HA dimensions a
baseline. Acco ding o ea lie epo s he subg oup in his
clus e has an inc eased isk o MDD and hei dep ession is
mo e disease-like.4-6,9 NS has been nega i ely associa ed wi h
dep essi e mood s a e and ends o be a a low le el in MDD
pa ien s.4 The clus e s migh hus e lec an unde lying ac o
explaining clinically di e en symp oms p o iles and cou se
o dep ession.
In ou s udy he empe amen clus e s we e associa ed wi h
bo h baseline and endpoin dep essi e symp oms and wi h
he ea men esponse. The indings sugges ed ha dep ession
was mos se e e and di icul o ea in clus e one pa ien s.
In clus e s one and wo, in which HA sco es we e high, he e-
sponse in pe cen age decline o MADRS sco es was lowe han
in clus e h ee. These indings concu wi h hose o ea lie
s udies on he associa ion be ween HA and dep ession.4,5,7-9,15,30
Al hough he dep ession ege a i e symp om sco e is only
a subscale o MADRS, i may be conside ed a sepa a e dimen-
sion in dep ession symp oma ology.22 MADRS o al sco es
and ege a i e symp oms showed a s ong co ela ion a bo h
baseline and endpoin . Howe e , he co ela ion be ween
MADRS endpoin o al sco e and baseline ege a i e symp-
om sco e was much lowe , sugges ing ha he ege a i e
symp oms a e a sepa a e en i y wi hin dep essi e symp om-
a ology.22 The e o e we conside ed i jus i ied o s udy he im-
pac o baseline ege a i e symp oms on o al symp oms a
endpoin .
In he linea eg ession models ou aim was o p edic he
ea men esponse in MDD pa ien s. The i s model was de-
signed o e eal he impac o empe amen clus e s on ea -
men ou come. When pa ien ’s age was also aken in o ac-
coun as an explana o y a iable, he empe amen clus e s
had only modes explana o y powe . Age as such did no unc-
ion as an explaining a iable in his model a a signi ican
le el. In he second model we wan ed o asce ain i he e
was an in e ac ion wi h empe amen clus e s and ege a i e
symp oms o MDD. The e o e we added he ege a i e symp-
om sco es o he model as an explana o y a iable. In his
model, he ege a i e symp oms explained abou wice as
much as he clus e s o he a iance o endpoin MADRS
sco es. Howe e , he whole model explained as much as one
i h o he a iance in esponse o SSRI ea men . The ole
o he in e ac ion be ween empe amen clus e s and ege a-
i e symp oms on ea men esul has o be in e p e ed cau-
iously, as he impac o clus e s on ea men esponse in he
inal model was ma ginal. I seems ha he ege a i e symp-
oms o dep ession, in addi ion o a ce ain empe amen p o-
ile, is a ma ked p edic o o an idep essi e ea men ou -
come. I is, howe e , possible ha he ege a i e symp oms
alone ha e a mo e ma ked impac in bo h se e i y and e-
sponse o dep ession compa ed o empe amen . E en hough
he di e ences be ween he clus e s in dep ession se e i y
we e ma ginal, ou indings sugges an associa ion be ween
skewed empe amen p o ile and se e i y o MDD. I is pos-
sible ha he empe amen p o ile can unc ion as a p edis-
posing ac o o dep ession o ha e an impac on he clinical
p o ile and cou se o dep ession.
In conclusion ou s udy showed ha MDD pa ien s could
be di ided in o di e en empe amen clus e s wi h di e en
se e i y and ou comes o an idep essan ea men . The ege-
a i e symp oms o dep ession combined wi h empe amen
p o iles and age p edic ed an idep essan ea men esponse.
The e ec o he empe amen p o ile alone was modes bu ,
combined wi h ege a i e symp oms o dep ession hei ex-
plana o y powe was mo e ma ked, sugges ing ha he e
could be an associa ion be ween hese wo in he biological
basis o MDD.
Acknowledgmen s
This s udy was suppo ed by esea ch g an s by he Lilly Founda ion
Finland, and by medical unds o Kan a-Häme Cen al Hospi al and Tam-
pe e Uni e si y Hospi al Dis ic s.
REFERENCES
1. Cloninge CR, S akic DM, P zybeck TR. A psychobiological model
o empe amen and cha ac e . A ch Gen Psychia y 1993;50:975-990.
2. Cloninge CR. A uni ied biosocial heo y o pe sonali y and i s ole in
he de elopmen o anxie y s a es. Psychia De 1986;4:167-226.
3. de Win e RFP, Wol e beek R, Spinho en P, Zi man FG, Goekoop JG.
Cha ac e and empe amen in majo dep essi e diso de and a highly
anxious- e a ded sub ype de i ed om melancholia. Comp Psychia-
y 2007;48:426-435.
4. Fa me A, Mahmood A, Redman K, Ha is T, Sadle S, McGu in P. A
sib-pai s udy o he Tempe amen and Cha ac e In en o y scales in
majo dep ession. A ch Gen Psychia y 2003;60:490-496.
5. Hansenne M, Regge s J, Pin o E, Kji i K, Ajamie A, Ansseau M. Tem-
pe amen and Cha ac e In en o y (TCI) and dep ession. J Psychia
Res 1999;33:31-36.
6. Hi ano S, Sa o T, Na i a T, Kusunoki K, Ozaki N, Kimu a S, e al. E al-
ua ing he s a e dependency o he Tempe amen and Cha ac e In-
en o y dimensions in pa ien s wi h majo dep ession: a me hodologi-
cal con ibu ion. J A ec Diso d 2002;69:31-38.
7. Kimu a S, Sa o T, Takahashi T, Na i a T, Hi ano S, Go o M. Typus mel-
ancholicus and he Tempe amen and Cha ac e In en o y pe sonali y
dimensions in pa ien s wi h majo dep ession. Psychia y Clin Neu o-
sci 2000;54:181-189.
8. Ma ijnissen G, Tuinie S, Sijben AE, Ve hoe en WM. The empe a-
men and cha ac e in en o y in majo dep ession. J A ec Diso d

V Paa onen e al.
www.psychia yin es iga ion.o g 23
2002;70:219-223.
9. Rich e J, Polak T, Eisenmann M. Dep essi e mood and pe sonali y in
e ms o empe amen and cha ac e among he no mal popula ion
and dep essi e inpa ien s. Pe s Indi Di e 2003;35:917-927.
10. Nai o M, Kijima N, Ki amu a T. Tempe amen and Cha ac e In en-
o y (TCI) as p edic o s o dep ession among Japanese college s uden s.
J Clin Psychol 2000;56:1579-1585.
11. Ju ado D, Gu pegui M, Mo eno O, Fe nandez MC, Luna JD, Gal ez R.
Associa ion o pe sonali y and wo k condi ions wi h dep essi e symp-
oms. Eu Psychia y 2005;20:213-222.
12. Jylha P, Isome sa E. Tempe amen , cha ac e and symp oms o anxie y
and dep ession in he gene al popula ion. Eu Psychia y 2006;21:389-
395.
13. Ab ams KY, Yune SK, Kim SJ, Jeon HJ, Han SJ, Hwang J, e al. T ai and
s a e aspec s o ha m a oidance and i s implica ion o ea men in ma-
jo dep essi e diso de , dys hymic diso de , and dep essi e pe sonali y
diso de . Psychia y Clin Neu osci 2004;58:240-248.
14. Sa o T, Hi ano S, Na i a T, Kusunoki K, Ka o J, Go o M, e al. Tempe a-
men and cha ac e in en o y dimensions as a p edic o o esponse o
an idep essan ea men in majo dep ession. J A ec Diso d 1999;56:
153-161.
15. Kampman O, Pou anen O. Can onse and eco e y in dep ession be
p edic ed by empe amen ? A sys ema ic e iew and me a-analysis. J
A ec Diso d 2011;135:20-27.
16. Sa o T, Na i a T, Hi ano S, Kusunoki K, Go o M, Sakado K, e al. Fac o
alidi y o he empe amen and cha ac e in en o y in pa ien s wi h
majo dep ession. Comp Psychia y 2001;42:337-341.
17. Hansenne M, Bianchi J. Emo ional in elligence and pe sonali y in ma-
jo dep ession: ai e sus s a e e ec s. Psychia y Res 2009;166:63-68.
18. Ma sudai a T, Ki amu a T. Pe sonali y ai s as isk ac o s o dep es-
sion and anxie y among Japanese s uden s. J Clin Psychol 2006;62:97-
109.
19. Pei son AR, Heuche JW. The ela ionship be ween pe sonali y and
mood: Compa ison o he BDI and he TCI. Pe s Indi Di e 2001;30:
391-399.
20. G ucza RA, P zybeck TR, Spi znagel EL, Cloninge CR. Pe sonali y
and dep essi e symp oms: a mul i-dimensional analysis. J A ec Diso d
2003;74:123-130.
21. Ginsbu g BE, We ick TM, Escoba JI, Kugelmass S, T eano JJ, Wend -
land L. Molecula gene ics o psychopa hologies: a sea ch o simple
answe s o complex p oblems. Beha Gene 1996;26:325-333.
22. Suzuki A, Aoshima T, Fukasawa T, Yoshida K, Higuchi H, Shimizu T,
e al. A h ee- ac o model o he MADRS in majo dep essi e diso -
de . Dep ess Anxie y 2005;21:95-97.
23. Higuchi H, Sa o K, Yoshida K, Takahashi H, Kama a M, O ani K, e al.
P edic o s o an idep essan esponse o lu oxamine ob ained using
he h ee- ac o s uc u es o he Mon gome y and Asbe g Dep ession
Ra ing Scale o majo dep essi e diso de s in Japanese pa ien s. Psy-
chia y Clin Neu osci 2008;62:301-306.
24. Higuchi H, Sa o K, Yoshida K, Takahashi H, Kama a M, O ani K, e al.
No p edic o s o an idep essan pa ien esponse o milnacip an we e
ob ained using he h ee- ac o s uc u es o he Mon gome y and As-
be g Dep ession Ra ing Scale in Japanese pa ien s wi h majo dep es-
si e diso de s. Psychia y Clin Neu osci 2008;62:197-202.
25. Kama a M, Suzuki A, Yoshida K, Takahashi H, Higuchi H, O ani K.
Gene ic polymo phisms in he se o one gic sys em and symp om clus-
e s o majo dep essi e diso de . J A ec Diso d 2011;135:374-376.
26. Illi A, Se ala-Soikkeli E, Viikki M, Pou anen O, Huh ala H, Mononen N,
e al. 5-HTR1A, 5-HTR2A, 5-HTR6, TPH1 and TPH2 polymo phisms
and majo dep ession. Neu o epo 2009;20:1125-1128.
27. Fa me RF, Seeley JR. Tempe amen and cha ac e p edic o s o de-
p essed mood o e a 4-yea in e al. Dep ess Anxie y 2009;26:371-381.
28. Spi lehouse JK, Pea son JF, Lu y SE, Mulde RT, Ca e JD, McKenzie
JM, e al. Measu es o empe amen and cha ac e a e di e en ially im-
pac ed on by dep ession se e i y. J A ec Diso d 2010;126:140-146.
29. Kampman O, Pou anen O, Illi A, Se ala-Soikkeli E, Viikki M, Nuoli i -
a T, e al. Tempe amen p o iles, majo dep ession, and esponse o
ea men wi h SSRIs in psychia ic ou pa ien s. Eu Psychia y 2012;
27:245-249.
30. Kusunoki K, Sa o T, Taga C, Yoshida T, Komo i K, Na i a T, e al. Low
no el y-seeking di e en ia es obsessi e-compulsi e diso de om ma-
jo dep ession. Ac a Psychia Scand 2000;101:403-405.