Eur. J. Psychiat. Vol. 24, N.

° 2, (98-113)
2010

Keywords: Schizophrenia; Affective disorder; Cour-

se; Prodomal symptoms; Symptom dimensions.

The early Kraepelin’s dichotomy of schizophrenia

and affective disorder – Evidence
of separate diseases?a
H. Häfner
Professor emeritus of Psychiatry,
University of Heidelberg
Head, Schizophrenia Research Unit,
Central Institute of Mental Health
Mannheim
GERMANY

ABSTRACT – Background and Objectives: Testing Kraepelin’s dichotomy model, we stud-

ied the separability of schizophrenia and affective disorders by their symptoms and course.
Methods: To this end symptoms and illness course were assessed retrospectively in in-
dividually matched untreated probands with schizophrenia and depression (n=130 each)
from first admission back to illness onset in comparison with 130 “healthy” controls. In a
second study these same variables were studied prospectively in 107 patients with schizo-
phrenia over a homogenised follow-up of 134 months (11.2 years). The actual mean
length of the follow-up period was 12.3 years.
Results: The symptom most frequently marking the onset of both schizophrenia and
depression was depressive mood. Both disorders exhibited the same prodromal core syn-
drome. It was not until the emergence of positive symptoms that the disorders became sep-
arable by the international classification systems. Depression remained the most frequent
syndrome over the course of schizophrenia.
Conclusions: Obviously, depression does not represent comorbidity, but an integral
part of psychosis. A dimensional disease model based on (successively emerging) hierar-
chical symptom patterns of the human brain with increasing brain dysfunction in the
course of schizophrenia and several neuro-degenerative disorders, not unknown to the
later Kraepelin, is offered as an explanation.

Received: 6 July 2009

Accepted: 18 December 2009

a Based on a paper presented at the International Symposium about Current Issues and Controversies
“Beyond the Kraepelinean Nosology” held in Barcelona, April 3-4, 2008.
 THE EARLY KRAEPELIN’S DICHOTOMY OF SCHIZOPHRENIA AND AFFECTIVE... 99

Introduction molecular level systematic genome-linkage

studies, focussed linkage studies and associ-
ation studies have yielded a limited number
With biological discriminative criteria of gene loci and haploid genes associated
lacking the early Kraepelin1 defined the two with an increased risk for a disorder of ei-
major psychoses, dementia praecox and ther psychotic or affective type, frequently
manic-depressive insanity, by their symp- also for both types6, 11-13.
toms and course. Although discarded by
Kraepelin2 in his later days, this dichotomy The most important gene products associ-
of two discrete disease entities has survived ated with susceptibility for both schizophre-
in our diagnostic classification systems and nia and affective disorder are NRG1 (neu-
clinical practice. For this reason the issue regulin 1), on chromosome 814,15, DAOA
deserves our attention. (diaminooxydase activator) on the G72/G30
gene on chromosome 1316,17 and in rare cases
Studying the course of schizophrenia and
DISC1 (disrupted in schizophrenia 1)18. Cra-
its main symptom dimensions –psychotic,
ddock & Owen19 stress that these genes
negative, depressive and manic– over a peri-
straddle the Kraepelinian divide carrying
od extending from illness onset to about 12
risk for both types of disorder. A plausible
years after first admission we will draw
explanation could be offered by a polyge-
some epidemiological comparisons be-
netic model, as in the case of the risk for
tween the two disease constructs.
coronary heart disease, associated both with
Kendel & Brockington3 studied the em- high blood pressure and increased blood fat
pirical separability of the constructs of levels. According to such a model, underly-
schizophrenia and affective disorder in the ing the symptom patterns in question are
population at large. They found no “point of several neurophysiological dysfunctions, in
rarity”, i.e. no decrease in symptom fre- the causation of which several genes may
quencies to chance value between the char- play a role20.
acteristic syndromes. The question whether
A psychological risk factor demonstrated
modern neurobiological findings might pro-
by van Os et al.21 is the neuroticism person-
vide indicators for the validity of the di-
ality trait. It is associated with a risk for
chotomy needs to be addressed at different
both syndromes. Environmental factors are
levels. The apparently disease-specific effi-
more difficult to prove, as they may con-
cacy of pharmacotherapy is not a proof of
tribute to the manifestation of illness by in-
the validity of these disease constructs. Ever
teracting with the genetic predisposition.
since Freyhan4 numerous authors have de-
Growing up in an urban environment22, mi-
monstrated that antidepressant and antipsy-
gration23 and being member of an ethnic mi-
chotic substances act on target symptoms or
nority group24 have been confirmed as risk
syndromes independently of the disease
factors. But here it is difficult to distinguish
constructs involved.
between social selection and social causa-
On the genetic level, twin and family tion, because the genetically determined
studies and studies of the offspring of wo- neural dysfunctions increase not only the
men with schizophrenia have demonstrated risk for psychosis, but also the probability
different degrees of familial cosegregation of more or less deviant personality traits,
of the risk for bipolar, unipolar, schizoaffec- which, on their part, affect social bonding
tive and schizophrenic psychoses5-10. On the and expansive behaviour –“novelty seek-
100 H. HÄFNER

ing”–, thus influencing selection forces in the illness courses of the two syndromes
migration processes. The pioneer of psychi- does not make matters easier either.
atric epidemiology, Ørnulf Ødegaard25, ear-
If we proceed from operational defini-
ly demonstrated the role such factors may
tions of the diagnoses adjusted for severity
play in the migration of individuals at risk.
of illness29 the long-term courses of the dis-
To conclude thus far: not a single power- orders are even harder to distinguish. Test-
ful and reliable criterion has been found yet ing the stability of six diagnostic clusters,
that clearly discriminates between the two the long-term follow-up study of Marneros
Kraepelinean disease entities. et al.28, covering a mean follow-up period of
23 years (range: 10 to 50 yrs.) following ad-
The course-related criterion the early mission to hospital and based on diagnoses
Kraepelin used for distinguishing between not adjusted for severity of illness, found that
the two disease constructs held that manic- “uniform”, chronic syndromes such as schiz-
depressive illness usually runs a remitting ophrenia and major depression –which the
type of course without affecting the patients’ authors called melancholy– were compara-
intelligence, whereas dementia praecox is tively stable. In contrast, “multiple“ syn-
characterised by a therapy-resistant mental dromes, such as bipolar and schizoaffective
defect growing more and more severe with disorders, showed a high degree of instabili-
each “florid” episode of psychosis and fi- ty. The problem we are faced with here is
nally resulting in dementia. It was this tra- that the probability of change reflects the
jectory of a descending staircase that the au- number of syndromes and their degree of
thors of the NIMH follow-up study26, too, chronicity in the long-term course. To con-
proposed, although their empirical data did clude, attempts to validate Kraepelin’s di-
not justify it. chotomy by illness course-related criteria
have also failed.
Differences in the social course of the
two disorders have been studied repeatedly
and used as a criterion for their discreteness.
Tsuang et al.27 in their Iowa-500 Study sho-
wed that schizophrenia has the most un- Own studies
favourable, bipolar affective disorder the
most favourable course –apart from surgical For this reason, we set out to compare
controls– with schizoaffective psychoses oc- schizophrenia and depression after illness
cupying an intermediate position. A more re- onset and to analyse the long-term course of
cent study conducted by Marneros et al.28, the affective and non-affective symptom di-
too, came to a similar conclusion on the mensions of schizophrenia.
basis of proportions of patients in indepen-
dent versus dependent living situations. But
this frequently confirmed finding, too, fails Material and methods
to provide a reliable criterion for distinguish-
ing between the two disorders, because their As described in our first publications from
social courses tend to reflect differences in the ABC (Age, Beginning, Course) Schizo-
the severity of illness in these diagnostic phrenia Study30,31, we collected data on all
groups and are subject to interaction with the first admissions for schizophrenia spectrum
environment. The high degree of variation in disorder (ICD -International Classification
 THE EARLY KRAEPELIN’S DICHOTOMY OF SCHIZOPHRENIA AND AFFECTIVE... 101

of Diseases -9: 295, 297, 298.3, 298.4) in At five-year follow-up 103 of the 130
age range 12 to 59 years in a two-year peri- first-admission patients and at 12.3-year fol-
od from a semi-rural, semi-urban popula- low-up 107 patients from the initial first-
tion of 1.5 million (Heidelberg, Mannheim, episode sample of 232 patients could be
Ludwigshafen, Rhine-Neckar District, east- reached and interviewed. The patients’ mean
ern Palatinate). Of the 276 patients inter- age at the final follow-up was 42 years (ran-
viewed 232 were in their first psychotic ge: 29 to 67). 24 patients (10.3%) had died,
episodes. To distinguish prodromal signs 15 (= 6.5%) of them of suicide. The follow-
from symptoms generally occurring in the up sample was representative of the initial
sample, which we tested on demographic
population, we also assessed 130 “healthy“
and some illness-related data, but found no
controls from the population of origin,
significant differences40.
matched by age and sex with a representa-
tive subsample of 130 first admissions for
schizophrenia and 130 equally matched first
admissions for moderately severe (F32.10,
Results
32.11) and severe depression (F32.2, 32.30,
32.31). Immediately on hospital admission
the patients went through interviews by the Early course
PSE – Present State Examination –32, SANS
Of the 115 patients diagnosed as suffer-
– Scale for the Assessment of Negative
ing from schizophrenia 80% and of the 115
Symptoms –33, PIRS – Psychiatric Impair-
patients diagnosed at first admission as suf-
ments Rating Schedule –34 and DAS – Psychi- fering from severe or moderately severe de-
atric Disability Assessment Schedule –35,36. pression 79% had not previously received
Two to 4 weeks later, after acute symptoms any antidepressant or antipsychotic medica-
had remitted, they were administered an tion41. Hence, we were able to study symp-
IRAOS – Interview for the Retrospective tomatology more or less unaffected by spe-
Assessment of Schizophrenia – interview37-39. cially targeted drugs.
In the controlled study of illness course we
assessed 115 first illness episodes (from Table 1 gives the initial symptoms most
among the 130 first admissions) retrospec- frequent in the two illness groups, a total of
tively back to illness onset and followed up 13. Eight of these 13 symptoms had almost
equal frequencies and similar rankings in
this subsample prospectively at six cross-
both the schizophrenia and the depression
sections over five years and at a 7th after a
group. As expected, the period prevalences
mean of 12.3 years after first admission. The
of the 10 most frequent symptoms, assessed
healthy controls were assessed using the over the entire early illness course, –these
same instruments as with the probands. At symptoms reflect the main prodromal sym-
the final follow-up matched controls from ptomatology of schizophrenia and depres-
the population of the study area were inter- sion– were significantly higher in the two
viewed on the phone using shortened ver- illness groups, schizophrenia and depres-
sions of the instruments in order to obtain sion, than for healthy controls. (Figure 1). The
standard values for symptom variables and most frequent prodromal symptoms –rest-
social parameters, for example on changes in lessness, anxiety, difficulties of concentra-
the unemployment rate of the population. tion, disturbed appetite– showed no signifi-
102 H. HÄFNER

Table 1
The ten most frequent initial symptoms of schizophrenia (Sz) and depression (Dep) and the prodromal core
syndrome of these disorders (grey background) –symptoms with rank 1 to 10 in either group–
Symptom Schizophrenia Depression Sz vs. Dep
% Rank % Rank
Worrying 19.2 4 14.1 5 n.s.
Headaches, other aches and pains 10.3 – 13.2 8 n.s.
Difficulties of thinking, concentration 17.1 5 16.5 3 n.s.
Loss of self-confidence 11.9 8 14.0 6 n.s.
Social withdrawal, suspiciousness 11.6 9 13.3 7 n.s.
Disturbed appetite, sleep 15.0 6 21.9 2 n.s.
Loss of energy/ slowness 13.5 7 8.5 5 n.s.
Loss of libido 4.1 – 8.5 5 n.s.
Nervousness, restlessness 21.9 2 6.2 – ***
Anxiety 23.2 1 15.4 4 o

Depressive mood 20.6 3 34.9 1 *

Oversensitivity 3.3 – 9.3 9 o

Blunted affect 11.1 10 0.8 – ***

McNemar test: n.s. = not significant; o p < 0.10; * p < 0.05; *** p < 0.001.
Source: see reference 41, modified.

Figure 1. Frequency of symptoms (period prevalences %) in patients with schizophrenia (Sz), depression (Dep) and
healthy controls (HC) Symptoms with ranks 1 to 10 and prevalences > 5% in any of the three groups.
McNemar test: n.s. = not significant; * p < 0.05; ** p < 0.01, *** p < 0.001.
 THE EARLY KRAEPELIN’S DICHOTOMY OF SCHIZOPHRENIA AND AFFECTIVE... 103

cant differences in frequency between the back at the inseparable prepsychotic prodro-
two groups. The remaining symptoms, ex- mal stage of the two diagnoses. After that
cept the psychotic symptoms depicted on stage the psychotic symptoms at first admis-
the right-hand side in the Figure, while sig- sion were also included in the comparisons,
nificantly differing in frequency, have pre- and, in accordance with the diagnostic defi-
valences fairly similar in size and ranking, nitions of the two groups, a clear-cut dis-
e.g. depressed mood in the depression group tinction between them emerged.
100% (rank: 1), in the schizophrenia group
84.9% (rank: 5), worrying: 94.6% (rank: 4)
versus 74.6% (rank: 9), loss of energy/slow- Long-term course
ness 93.8% (rank: 5) versus 82.5% (rank: 6),
Outcome analysis
social withdrawal /suspiciousness 90.8%
(rank: 6) versus 79.8% (rank: 8), lack of self- A comparison of patients at first admis-
confidence 89.2% (rank: 7) versus 68.3% sion and at 12.3-year follow-up showed an
(rank: 10), reduced free-time activities 89.1% impressive decrease in the proportion of
(rank: 8) versus 63.5% (-). The symptoms in- those presenting symptoms and social dis-
cluded in the parenthesis are the ones that ability (Table 2). But this result, too, was an
were already counted among the initial symp- artefact. At first admission practically all
toms. That illustrates not only the high degree patients were in a psychotic episode exhibit-
of similarity between the prodromal symp- ing maximum symptom scores. At follow-
toms of the two disorders, but also the stability up we had a random selection of patients in
of these symptoms in the early illness course. episodes and intervals with a mean of 22%
in psychotic episodes.
The non-psychotic symptom profiles of
schizophrenia and depression before the A comparison of marriage and partnership
onset of psychotic symptoms seem to be al- showed that the proportion of patients living
most identical. It is the two delusional alone had decreased considerably and the pro-
symptoms, which show high prevalence portion of married, especially among women
rates in schizophrenia, but in depression with schizophrenia, had increased markedly,
very low rates hardly different from those but so had also the proportion of divorced pa-
for healthy controls, that represent a criteri- tients (Table 3). Compared with healthy con-
on for discriminating between the two dis- trols, however, patients suffered from pro-
orders. Since their onset does not occur until nounced social disability. A male-female
late at the prodromal stage, usually in the comparison showed that, despite their illness,
last year preceding the climax of the first clearly more women than men were able to
psychotic episode, the non-psychotic pro- find a partner and tie the knot (Table 4). But
dromal stages in the two disorders remain the high rate of divorce also indicated that a
inseparable until then. lot of these marriages did not last for long.
The high rate of single male patients is re-
But the apparently reliable discrimination
flected in the above-average proportion of
between the two diagnostic groups on the
men among psychiatric home residents.
basis of positive symptoms is an artefact,
because in view of the hypotheses we in- These sex differences are accounted for by
tended to test we had selected the probands normal sex differences in social behaviour
on the basis of their diagnosis. In the course rather than by the biological illness, which
of our retrospective analyses we landed affects men and women in the same way.
104 H. HÄFNER

Table 2
Clinical characteristics at first admission and long-term follow-up (12.3 yrs.) (%)
ABC sample
First admission % Follow-up %
Sociale impairment (DAS total score > = 2) 57.9 19.6
Delusions, hallucinations (PSE-DAH > = 2) 95.3 15.9
Speech, behaviour (PSE-BSO > = 2) 95.3 51.4
Affective flattening (SANS global > 2) 29.5 5.6
Alogia / paralogia (SANS global > 2) 17.1 2.8
Abulia / apathy (SANS global > 2) 39.4 23.6
Anhedonia (SANS global > 2) 37.9 16.3
Attention (SANS global > 2) 29.8 9.9
Source: see reference 43.

Table 3
Marital status at first admission and long-term follow-up (12.3 yrs.) compared with “healthy” controls (%)
ABC sample Controls
(N = 107) (N = 107)
% %
First admission Follow-up Follow-up
Single 71.0 46.7 21.5
Married 26.1 34.5 69.2
Divorced 1.9 13.1 6.5
Widowed 0.9 2.8 2.8
Unknown 0 2.8 0
Source: see reference 43.

Analysis of illness course

To prevent the high degree of variation in During the 134-month period following
the durations of illness and its course from first admission we counted a total of 333 psy-
distorting the results we based our analyses chotic relapses –defined by deterioration in
instead of the mean follow-up period of psychotic symptoms of at least 14 days’ du-
12.3 years on the shortest follow-up period ration, preceded and followed by four weeks
of 11.2 years or 134 months. The uniform in which symptoms were absent or reduced–
duration of illness and monthly steps of a mean of three and a range of 0 to 29 per pa-
analysis allowed us to closely monitor the tient. The psychotic episodes usually con-
presence of symptoms and changes in social tained a certain proportion of depressive
and other parameters. symptoms. 73 (range 0 to 11/patient) or 18%
 THE EARLY KRAEPELIN’S DICHOTOMY OF SCHIZOPHRENIA AND AFFECTIVE... 105

Table 4
Marital status and living situation at long-term follow-up (12.3 yrs.) (% men versus women)
Marital status Men Women
Single 68.8 28.8
Married 18.8 47.4
Divorced 10.4 15.3
Widowed – 5.1
Unknown 2.1 3.4
Living situation Men % Women %
Alone 22.9 15.3
Sheltered (supervised home) 18.8 3.4
Total 41.7 18.7
Source: see reference 43.

of the relapses were mainly depressive in total of 19 partly heterogeneous single sym-
type with no psychotic symptoms occurring. ptoms, we refrained from defining a nega-
tive core syndrome.
We studied the course of clinical symptom
categories on the basis of five traditional syn- Again, a depressive symptom, depressed
dromes. The months patients spent with de- mood, turned out to be the predominant
pressive symptoms clearly predominated. symptom of schizophrenia (Table 6). A com-
Months spent with negative, positive and parison of episodes and intervals showed that
manic symptoms were clearly rarer (Table 5). the depressive core syndrome was present in
44% (median: 36%, range 0-100) of psychot-
To avoid overlap between the symptom
ic episodes and in 34.5% (median: 6%; range
categories we introduced discrete core syn-
0-100)) of intervals. The mean-median dif-
dromes in our analyses. In each group we
ference indicates a skewed distribution.
selected the most frequent non-overlapping
symptoms. The depressive core syndrome The montly prevalences of the three core
consisted of four symptoms different from syndromes over the entire follow-up period
negative symptoms, the manic syndrome of 134 months are illustrated in Figure 2. The
consisted of five symptoms separate from retrospectively assessed prevalences for de-
psychotic symptoms. To define a psychotic pressive symptoms were validated on pros-
core syndrome –a difficult task given the pective data gathered at seven cross-sections.
great number of such symptoms– we select- The black triangles stand for PSE-CAT-
ed the four most frequent ones from among EGO-based frequencies of severe and mod-
Kurt Schneider’s symptoms of first rank (cf. erately severe depression, the black quad-
Table 6) excluding those with prevalences rangles for those of severe depression. The
below 1%b. Since in the IRAOS interview prevalences of all the five syndromes de-
the complex negative syndrome comprises a clined more or less steeply after the first

b Delusional perception; delusion of influence; experience of externally made actions, impulses and
feelings; delusional mood.
106 H. HÄFNER

Table 5
Number of months (raw data) spent with symptoms from the main clinical categories in the 134-month
(11.2-year) follow-up period
Symptom category Mean SD
Depressive 76.9 56.2
Manic 9.0 24.8
Negative 45.1 54.5
Positive 26.7 42.6
Disorganization 6.3 19.2
Source: see reference 43.

Table 6
Mean duration (in months) of presence of depressive, manic and psychotic core symptoms in the long-term
(11.2-year, 134-month) course of schizophrenia
Depressive Depressive Loss of Feelings Suicidal
symptoms mood self-confidence of guilt thoughts/attempt
Mean number of months 30.4 27.9 8.2 4.2
with symptom

Manic symptoms: Elated Reduced need Pressure Hyper-activity Flight of ideas

mood for sleep of speech
Mean number of 5.8 3.8 3.6 0.8 0.7
months with symptom

Psychotic Verbal Thought Thought withdrawal Thought echo

symptoms: hallucination insertion
Mean number of 4.8 2.4 1.7 1.2
months with symptom
Source: see reference 43.

episode. In the subsequent course they all ed our conclusion that unlike age at onset or
showed a plateau. Depression remained the the medium-term social course45, the disor-
predominant syndrome throughout the fol- der as such is the same in men and women.
low-up period. There was visible neither a
Negative symptoms, assessed on the basis
trend of increase or decrease in symptoms
of raw, not necessarily non-overlapping da-
nor a staircase-like deteriorating course, as
ta, showed a different course (Figure 4): a
postulated by Kraepelin.
slow, pronounced initial decrease followed
A comparison between males and fe- by a stable plateau after about five years, as
males showed no difference in the frequen- was the case with the other symptom di-
cy and course of psychotic and depressive mensions, too. The male prevalences were
symptoms (Figure 3). This finding support- slower to decline than the female ones. Ac-
 THE EARLY KRAEPELIN’S DICHOTOMY OF SCHIZOPHRENIA AND AFFECTIVE... 107

Figure 2. Long-term course of the depressive, manic and positive symptom dimensions in schizophrenia (n=107).
Source: see reference 43.

Figure 3. Monthly prevalence of positive and depressive symptoms over 134 months
after first admission – for men and women. Source: see reference 43.
108 H. HÄFNER

Figure 4. Monthly prevalence of negative symptoms over 134 months after first admission – for men and women.
Source: see reference 43.

counting for this sex difference, too, might toms persisting over the entire follow-up pe-
be not the disorder as such, but typically riod: 7% of the depressive symptoms, 6% of
male and female behaviours. the negative and only 1% of the positive
symptoms were of that type. The majority
Behind these mean values, which convey
of exacerbations were of medium or short
the impression of a homogeneous illness
term, deterioration in the depressive core
course, there lies a high degree of interindi-
syndrome unfolding over 20.0 months (me-
vidual variability. 19% of the cases experi-
dian 5 months), in the positive core syn-
enced only one single episode and no relapse
drome over 6.3 months (median: 2) and in
episodes, symptoms or signs of disability.
negative symptoms over 23.2 months (me-
But there were also extremely poor courses
dian: 5). This result, which reflects the high
involving cognitive impairment and consid-
degree of interindividual variability, indi-
erable functional disability.
cates that a clearly chronic course of any of
We also tried to assess the individual time- the three syndromes, of the psychotic syn-
spans in which deterioration occurred in the drome in particular, is extremely rare. The
three symptom dimensions – the manic di- longest span of deterioration was shown by
mension was excluded because of too small the depressive core syndrome, when the ne-
values. As Figures 5 and 6 show, only a small gative symptoms, not fully comparable, we-
proportion of the patients experienced symp- re left out of consideration.
THE EARLY KRAEPELIN’S DICHOTOMY OF SCHIZOPHRENIA AND AFFECTIVE... 109

Figure 5. Duration of spells with depressive and positive symptoms.

Source: see reference 43.

Figure 6. Duration of spells with negative symptom.

Source: see reference 43.
110 H. HÄFNER

These results strongly suggest that schiz- state and dementia. However, social impair-
ophrenia does not represent primarily a stat- ment is pronounced. The maximum of so-
ic encephalopathy46,47, but a dynamic pro- cial decline or social stagnation occurs at
cess originating in a neurodevelopmental the early stage of schizophrenia.
disorder. That process unfolds with bouts of
To conclude, the majority of the psycho-
increasing dysfunction followed by neuro-
pathology currently diagnosed as schizo-
plastic compensation permitting functional
phrenia involves both psychotic and depres-
recovery. The process is characterised by
sive symptoms, with the latter being clearly
asynchronous waves of exacerbation, trig-
more frequent. After illness onset different
gered by intrinsic and extrinsic factors, and
types of symptoms occur successively with
remission of its main symptom dimensions,
depression leading the way in the early ac-
a process we do not yet fully understand.
tive stage characterised by an accumulation
Social course was assessed on the basis of of symptoms and social consequences. The
the proportion of patients in full-time em- first episode is followed by asynchronous
ployment, a well-comparable measure (Figu- waves of exacerbation and intervals. The
re 7). At initial assessment the figure was combination of positive, manic and negative
about 30%, several years later it had in- symptoms and different proportions of con-
creased to over 40% and at the final follow- comitant depressive symptoms –the hierar-
up fallen back to 30%, compared with 70% chy of symptom frequencies remaining un-
for the healthy controls. On average, the dis- changed– makes up what is traditionally
order does not obviously lead to a defect understood as schizophrenia. Since our cur-

Figure 7. Employment status and income over 134 months (homogenised) (means).
Source: see reference 43.
 THE EARLY KRAEPELIN’S DICHOTOMY OF SCHIZOPHRENIA AND AFFECTIVE... 111

rent therapeutic instruments are not diagno- tary based on issues debated at the World Congress of Psy-
chiatric Genetics, Boston, October 12-18, 2005. Am J Med
sis-specific, i.e. not targeted at schizophre-
Genet B Neuropsychiatr 2006; 141B: 319-322.
nia, but its syndromes, it should seriously be
considered whether a specific therapy direct- 8. Li D, He L. Association study between the dystro-
brevin binding protein I gene (DTNBP1) and schizophre-
ed at current symptoms would not be more
nia: a meta-analysis. Schizophr Res 2007; 96: 112-118.
efficacious than the traditional diagnosis-
based treatment. What we do already know 9. Li D, He L. G72/G30 genes and schizophrenia: a sys-
tematic meta-analysis of association studies. Genetics
is that focussing therapy and preventive 2007; 175: 917-922.
measures exclusively on psychotic symp-
10. Sanders AR, Duan J, Levinson DF, Shi J, He D, Hou
toms while neglecting the leading symptom
C, et al. No significant association of 14 candidate genes
dimension, depression, does not provide the with schizophrenia in a large European ancestry sample:
best possible treatment. This regimen both implications for psychiatric genetics. Am J Psychiatry
ignores an increased risk for suicide and fails 2008; 165: 497-506.
to relieve the distress caused by depression. 11. Berrettini WH. Are schizophrenic and bipolar disor-
ders related? A review of family and molecular studies.
Biol Psychiatry 2000; 48: 531-538.

Acknowledgment 12. Craddock N, O’Donovan MC, Owen MJ. Genetics

of schizophrenia and bipolar disorder: dissecting psy-
chosis. J Med Genet 2005; 42: 193-204.
The ABC study was funded by the German
13. Van den Bogaert A, Del-Favero J, van Broeckhoven
Research Association (Deutsche Forschungs- C. Major affective disorders and schizophrenia: a common
gemeinschaft) as part of the Special Research molecular signature? Hum Mutat 2006; 27: 833-853.
Branch (Sonderforschungsbereich) 258 at 14. Hattori E, Liu C, Badner JA, Bonner TI, Christian
the Central Institute of Mental Health until SL, Maheshwari M, et al. Polymorphisms at the G72/G30
December 1998 and from January 1999 to gene locus, on 13q33, are associated with bipolar disorder
Sept. 2002 as an independent project. in two independent pedigree series. Am J Hum Genet
2003; 72: 1131-1140.

15. Detera-Wadleigh SD, McMahon FJ. G72/G30 in

schizophrenia and bipolar disorder: review and meta-
References analysis. Biol Psychiatry 2006; 60: 106-114.

16. Chumakov I, Blumenfeld M, Guerassimenko O,

1. Kraepelin E. Psychiatrie, 5th edn. Leipzig: Barth; 1896. Cavarec L, Palicio M, Abderrahim H, et al. Genetic and
2. Kraepelin E. Die Erscheinungsformen des Irreseins. physiological data implicatining the new human gene G72
Zschr Ges Neurol Psychiat 1920; 62: 1-29. and the gene for D-amino acid oxidase in schizophrenia.
Proc Natl Acad Sci USA 2002; 99: 13675-13680.
3. Kendell RE, Brockington IF. The identification of dis-
ease entities and the relationship between schizophrenic and 17. Williams NM, Green EK, Macgregor S, Dwyer S,
affective psychoses. Br J Psychiatry 1980; 137: 324-331. Norton N, Williams H, et al. Variation at the DAOA/G30
locus influences susceptibility to major mood episodes but
4. Freyhan FA. Psychomotilität, extrapyramidale Syn- not psychosis in schizophrenia and bipolar disorder. Arch
drome und Wirkungsweisen neuroleptischer Therapien Gen Psychiatry 2006; 63: 366-373.
Nervenarzt 1957; 28: 504-509.
18. Hennah W, Thomson P, Peltonen L, Porteous D.
5. Owen MJ, Craddock N, O’Donovan MC. Schizo-
Genes and schizophrenia: beyond schizophrenia: the role
phrenia: genes at last? Trends Genet 2005; 21: 518-525.
of DISC1 in major mental illness. Schizophr Bull 2006;
6. Owen MJ, Craddock N, Jablensky A. The genetic de- 32: 409-416.
construction of psychosis. Schizophr Bull 2007; 33: 905-911.
19. Craddock N, Owen MJ. Rethinking psychosis: the
7. De Lisi LE, Faraone SV. When is a “positive” associ- disadvantages of a dichotomous classification now out-
ation truly a “positive” in psychiatric genetics? A commen- weigh the advantages. World Psychiatry 2007; 6: 20-27.
112 H. HÄFNER

20. Kane JM, Lencz T. Cognitive deficits in schizophre- 34. Biehl H, Maurer K, Jablensky A, Cooper JE, Tomov
nia: short-term and long-term. World Psychiatry 2008; 7: T. The WHO Psychological Impairments Rating Schedule
29-30. (WHO/PIRS). I. Introducing a new instrument for rating
observed behaviour and the rationale of the psychological
21. Van Os J, Janssen I, Hanssen M, Bak M, Myin-Ger-
impairment concept. Br J Psychiatry 1989; 155 (Suppl. 7):
meys I, Marcelis M, et al. Cognitive epidemiology: psy-
68-70.
chological and social risk mechanisms for psychosis. In:
Häfner H, editor. Risk and protective factors in schizophre- 35. WHO - World Health Organization. Psychiatric Dis-
nia. Darmstadt: Steinkopff; 2002; p. 207-228. ability Assessment Schedule. Geneva: WHO; 1988.

22. Van Os J, Krabbendam L, Myin-Germeys I, De- 36. Jung E, Krumm B, Biehl H, Maurer K, Bauer-
lespaul P. The schizophrenia envirome. Curr Opin Psychia- Schubart C. Mannheimer Skala zur Einschätzung von
try 2005; 18:141-145. sozialer Behinderung (DAS-M). Weinheim: Beltz; 1989.

23. Selten JP, Cantor-Graae E, Kahn RS. Migration and 37. Häfner H, Riecher-Rössler A, Hambrecht M, Maur-
schizophrenia. Curr Opin Psychiatry 2007; 20: 111-115. er K, Meissner S, Schmidtke A, et al. IRAOS: an instru-
ment for the assessment of onset and early course of schiz-
24. W, Selten JP, Susser E, Laan W, Mackenbach JP,
ophrenia. Schizophr Res 1992; 6: 209-223.
Hoek HW. Discrimination and the incidence of psychotic
disorders among ethnic minorities in The Netherlands. Int 38. Häfner H, Löffler W, Maurer K, Riecher-RösslerA,
J Epidemiol 2007; 36:761-768. Stein A. IRAOS. Interview für die retrospektive Erfassung
des Erkrankungsbeginns und –verlaufs bei Schizophrenie
25. Ødegaard Ø. Emigration and insanity. A study of
und anderen Psychosen. Bern: Hans Huber Verlag; 1999.
mental disease among the Norwegian-born population of
Minnesota. Acta Psychiat (Kbh) 1932; Suppl 4. 39. Häfner H, Löffler W, Maurer K, Riecher-Rössler A,
Stein A. IRAOS – Interview for the retrospective assess-
26. Breier A, Schreiber JL, Dyer J, Pickar D. National
ment of the onset and course of schizophrenia and other
Institute of Mental Health longitudinal study of chronic
psychoses. Göttingen: Hogrefe & Huber; 2003.
schizophrenia: prognosis and predictors of outcome. Arch
Gen Psychiatry 1991; 48: 239-246. 40. An der Heiden W, Könnecke R, Maurer K, Ropeter D,
Häfner H. Depression in the long-term course of schizophre-
27. Tsuang MT, Woolson RF, Fleming JA. Long-term
nia. Eur Arch Psychiatry Clin Neurosci 2005; 255: 174-184.
outcome of major psychoses. I. Schizophrenia and affective
disorders compared with psychiatrically symptom-free sur- 41. Häfner H, Maurer K, Trendler G, an der Heiden W,
gical conditions. Arch Gen Psychiatry 1979; 36: 1295-1301. Schmidt M, Könnecke R. Schizophrenia and depression:
challenging the paradigm of two separate diseases - a con-
28. Marneros A, Deister A, Rohde A. Affektive,
trolled study of schizophrenia, depression and healthy con-
schizoaffektive und schizophrene Psychosen. Berlin: Hei-
trols. Schizophr Res 2005; 77: 11-24.
delberg: Springer-Verlag; 1991.
42. Häfner H, an der Heiden W, Maurer K. Evidence for
29. Häfner H, Maurer K, Ruhrmann S, Bechdolf S,
separate diseases? Stages of one disease or different com-
Klosterkötter J, Wagner M, et al. Early detection and sec-
binations of symptom dimensions? Eur Arch Psychiatry
ondary prevention of psychosis: facts and visions. Eur
Clin Neurosci 2008; 258 (Suppl.2): 85-96.
Arch Psychiatry Clin Neurosci 2004; 254: 117-128.
43. Häfner H. Psychose, depression und manische
30. Häfner H, Maurer K, Löffler W, Riecher-Rössler A.
Symptomatik – Leitsyndrome eigener Krankheiten oder
The influence of age and sex on the onset and early course
Kontinuum? In: Möller H-J, Müller N, editors. Schizo-
of schizophrenia. Br J Psychiatry 1993; 162: 80-86.
phrenie – Zukunftsperspektiven in Klinik und Forschung.
31. Häfner H, Nowotny B, Löffler W, an der Heiden W, Wien: Springer; 2010 (in print).
Maurer K. When and how does schizophrenia lead to social
44. Häfner H, an der Heiden W, Behrens S, Gattaz WF,
deficits? Eur Arch Psychiatr Clin Neurosci 1995; 246: 17-28.
Hambrecht M, Löffler W, et al. Causes and consequences
32. Wing JK, Cooper JE, Sartorius N. Measurement and of the gender difference in age at onset of schizophrenia.
classification of psychiatric symptoms: An instruction Schizophr Bull 1998; 24: 99-113.
manual for the PSE and CATEGO Program. London:
45. Häfner H, Maurer K, Löffler W, an der Heiden W,
Cambridge University Press; 1974.
Munk-J∆rgensen P, Hambrecht M, et al. The ABC schizo-
33. Andreasen NC. The scale for the assessment of nega- phrenia study: a preliminary overview of the results. Soc
tive symptoms (SANS). Iowa City: University of Iowa; 1983. Psychiatry Psychiatr Epidemiol 1998; 33: 380-386.
 THE EARLY KRAEPELIN’S DICHOTOMY OF SCHIZOPHRENIA AND AFFECTIVE... 113

46. Gold S, Arndt S, Nopoulos P, O’Leary DS, An- Address for correspondence:
dreasen NC. Longitudinal study of cognitive function in Prof. Dr. Dr. h.c. mult. Heinz Häfner
first-episode and recent-onset schizophrenia. Am J Psychi- Central Institute of Mental Health
atry 1999; 156: 1342-1348. J5 D-68159 Mannheim
Germany
47. Goldberg TE, David A, Gold JM. Neurocognitive
Tel.: +49 621 1703 2951
deficits in schizophrenia. In: Hirsch SR, Weinberger D, ed-
Fax: +49 621 1703 2955
itos. Schizophrenia, 2nd edition. Oxford: Blackwell Pub-
E-mail: [email protected]
lishing; 2003; p. 168-184.