вход по аккаунту


Could cognitive vulnerability identify high-risk subjects for schizophrenia.

код для вставкиСкачать
American Journal of Medical Genetics (Neuropsychiatric Genetics) 114:893– 897 (2002)
Invited Paper
Could Cognitive Vulnerability Identify High-Risk
Subjects for Schizophrenia?
Yves Sarfati* and Marie-Christine Hardy-Baylé
University Department of Psychiatry, Versailles Hospital, Versailles, France
This review puts into questions the possible
role of cognitive vulnerability markers in
prediction and prevention of schizophrenia.
Until recently, none of the identified cognitive anomalies has been proved to be definitive. However, as new promising candidates
are emerging (DS-CPT, CPT-IP, P suppression, Saccadic Eye Movements), the predictive value of these trait–type anomalies may
be criticized regarding four issues, which
are discussed : technical, metrological, theoritical, and clinical. As things stand, the
existence of a cognitive vulnerability marker, which testify to a permanant pathological trait, does not constitute a suffisant
factor to identify and treat subjects who are
at risk for schizophenia. ß 2002 Wiley-Liss, Inc.
KEY WORDS: cognitive vulnerability; neuropsychological
markers; schizophrenia; prediction
Of all the psychiatric pathologies, schizophrenia is
undoubtedly the one that has been most frequently
examined in studies bearing on the question of cognitive
vulnerability. These studies are situated at the meeting
point of two vast research fields; namely the study of
vulnerability markers, on the one hand, and the study
of cognitive anomalies on the other. We shall start by
Proceedings of the November, 2000 Lyon Meeting ‘‘Schizophrenia: From Prediction to Prevention.’’
*Correspondence to: Yves Sarfati, M.D., Ph.D., Consultation de
Psychiatrie et Psychologie, Centre Hospitalier de Versailles,
Hôpital André Mignot, 177, rue de Versailles, 78157 Le Chesnay,
France. E-mail:
Received 16 July 2001; Accepted 16 November 2001
DOI 10.1002/ajmg.10251
ß 2002 Wiley-Liss, Inc.
specifying the scope of these two domains in order to
define the precise meaning of the expression ‘‘cognitive
vulnerability’’, which has been widely used in recent
years and has therefore been subject to a number of
distortions of meaning. We shall then consider whether
the identification of cognitive vulnerability makes it
possible to identify subjects who are at risk of schizophrenia and permit their preventive treatment.
The term ‘‘vulnerability marker’’ can be understood
in two different ways. In the broader sense, a vulnerability marker can be considered to be any sign that
indicates an increased empirical probability that a
given individual may suffer from a pathological
episode. Consequently, these are epidemiological factors, such as ‘‘at risk of schizophrenia’’, i.e., factors
associated with a relatively high level of risk of being
affected by the illness when compared with the general
population. Given this interpretation, vulnerability
markers for schizophrenia include clinical items as
varied as drug addiction, social isolation, academic
performance, etc. [Amminger et al., 1999; Cannon et al.,
1999; Hodges et al., 1999]. Viewed within this broad
perspective, the vulnerability markers are non-specific
factors with multiple determinants and for which
there generally exists only a statistical relation with
Taken in a narrower sense, much more precise
conditions must be satisfied before we can discuss
vulnerability markers. These conditions, set out in the
very first models of vulnerability to schizophrenia
[Zubin and Spring, 1977; Neuchterlein and Dawson,
1984] and restated on many subsequent occasions
[Azorin, 1997], are as follows:
1. A vulnerability marker must be specific in nature,
i.e., it must exhibit a different distribution in
schizophrenic patients than in the population at
large, including psychiatric populations;
2. A vulnerability marker must possess a high level of
heritability, i.e., it should constitute a biological trait
that is correlated with a genetic susceptibility to
schizophrenia, even if, in occasional cases, the
Sarfati and Hardy-Baylé
vulnerability can be acquired. In other words, it
should be observed more frequently in first-degree
relatives of schizophrenic patients than in other
populations, but is not confounded with gene markers (defined as genetic loci with a known chromosomal assignment);
3. The marker should, in principle, be stable over time,
thus indicating that it is independent of the clinical
profile and, in particular, it should be present before,
during, and after an acute episode;
4. The marker should be replicable by different
research teams.
When fulfilled, these conditions theoretically define
the ideal marker, referred to as an A-marker [Zubin
et al., 1985], stable marker [Neuchterlein, 1987], or,
more generally, trait marker. It should be noted that
some models also emphasize the existence of other
biological variables that do not strictly fulfill the four
conditions described above, but that may nevertheless
be of great interest in the field of prediction and
prevention. For example, Neuchterlein [1987] describes
mediating vulnerability factors, which could be preexistant but covary with symptomatology, and may be
useful predictors of psychotic episodes. They could
constitute mediators between vulnerability markers
and symptoms, but their theoretical status remains
controversial [Azorin, 1997]. In our argument here, we
shall therefore limit ourselves to the strict definition of
vulnerability markers that testify to a permanent
pathological trait.
Cognitive anomalies rank among the best candidates
as vulnerability markers for schizophrenia. Studies
undertaken since the 1980s, first in cognitive neuropsychology and then in the cognitive neurophysiology of
schizophrenia, have made it possible to reveal a
number of anomalies, which, it is hoped, may meet
the conditions that have to be fulfilled by vulnerability
markers. Tables I and II summarize our current state of
knowledge concerning the main cognitive candidates
with regard to the four conditions that have to be
fulfilled by a vulnerability marker: specificity, heritability, stability, and replication.
A chronological examination of the large number of
studies dedicated to the cognitive vulnerability markers indicates that several different anomalies have
been proposed as markers of schizophrenia. However,
none of these anomalies have proved to be definitive
because they do not satisfy the conditions, because the
results have been obtained too recently, or because the identified anomalies characterize only one
schizophrenic subgroup that is not distinguished by
any clinical characteristic [Hardy-Baylé, 1997]. Now, as
in the past, promising new candidates are emerging.
These relate either to increasingly refined and accurate
neuropsychological variables, such as the degraded
version of the Continuous Performance Test (DS-CPT)
when the stimulus is isolated, or the identical pair
version (CPT-IP) when the stimuli are pairwise
[Laurent et al., 2000], or to neurophysiological variables, such as P50 suppression [Clementz et al., 1998]
or saccadic eye movements [Crawford et al., 1998].
However, the initial results that have been obtained
still need to be replicated, as is also the case for smooth
pursuit eye movements, which represent a potentially
fruitful avenue of investigation.
As we have seen, cognitive vulnerability is defined as
the fact of exhibiting one or more anomalies in the
above-mentioned tests, which are thought to be the best
adapted for revealing the trait-type anomalies of
schizophrenia and, in particular, for indicating its
onset. The question ‘‘Could cognitive vulnerability
identify high-risk subjects for schizophrenia?’’ can be
reformulated as follows: the cognitive neuropsychological and/or neurophysiological anomalies mentioned
above are thought to be present in ill subjects before
the appearance of the symptoms and to be present in
the patients’ relatives with a greater frequency than in
the general population. Leaving all else aside, their
presence should therefore make it possible to predict an
TABLE I. Current State of Knowledge Concerning the Neuropsychological Vulnerability Markers
for Schizophrenia in the Light of the Following Characteristics: Specificity Compared With Other
Groups, Heritability in First-Degree Relatives, Stability Over Time, and Replications by Different
Research Teams*
DS-CPT only
CPT-IP only
Digit span test
Yes (forward
Yes (backward
Span of
apprehension task
Backward masking
Trail-making B
Yes (difficult
*CPT, continuous performance test; WCST, wisconsin card-sorting test.
In progress for
In progress for
Cognitive Vulnerability
TABLE II. Current State of Knowledge Concerning the Neurophysiological Vulnerability Markers
for Schizophrenia in the Light of the Following Characteristics: Specificity Compared With Other
Groups, Heritability in First-Degree Relatives, Stability Over Time, and Replications by Different
Research Teams*
Smooth pursuit
eye movements
Saccadic eye
Phasic SCOR
In progress
In progress
In progress
*SCOR, skin conductance response.
increased risk of schizophrenia relative to the population at large.
Now, in the face of this evidence, we should like to
show in what ways, given our current state of knowledge, these tests may fail to be of relevance for
predictive purposes. Our objections will be structured
on the basis of four issues:
1. A technical objection. As we have just pointed out,
very few current tests are able to reveal an anomaly
that precisely fulfills the four conditions required of a
vulnerability marker: specificity, heritability, stability, and replication (see Tables I and II). This is the
least important of our objections, because it can be
argued that the most promising candidate markers,
such as smooth pursuit eye movement, are currently
being studied and may be confirmed in the very near
2. A metrological objection. Two of the four conditions
that a vulnerability marker needs to fulfill actually
turn it into a sign that is totally independent of
clinical observation. On the one hand, there is its
stability: the sign must be present before the illness;
must be present before, during, and after a symptomatic episode; and must not vary whether or not
treatment is administered. On the other, there is its
presence in healthy subjects: parents, cousins, or
children of schizophrenic patients. These two characteristics intrinsically deprive it of part of its value
as a predictor of a morbid process because they make
it impossible to associate the existence of the anomaly
with the presence of any given clinical sign. The
specificity of the association between the marker of
cognitive vulnerability and the illness is considerably
reduced by the existence of all the asymptomatic
carriers of the marker.
3. A theoretical objection. It is necessary to reposition
the interest shown in the cognitive vulnerability
markers within the more general framework of
models of vulnerability. The models of vulnerability
are etiological models and therefore understand
cognitive markers as an endophenotype, that is to
say as a variable that is better defined than the
clinical signs and which can assist genetic analysis.
In other words, cognitive markers are often confounded with genuine genetic markers that must be
independent of the environment and which reveal an
innate, congenital (as Walker puts it), or genetic (as
Neuchterlein puts it) vulnerability [Neuchterlein
et al., 1999; Walker, 1999]. Genetic research actually
needs cognitive markers that are innate and insensitive to interaction with the environment. This
undoubtedly explains why cognitive neurophysiology
(and in particular the study of genetically determined
eye movements) is currently gaining ground on
cognitive neuropsychology, which is more influenced by developmental and environmental factors.
However, between an endophenotypic marker, which
would be closely linked to the genome, and the
observed pathology, there still remains the gap that
Bleuler [1911] said would be so difficult to cross.
4. A clinical objection. The validity of the candidate
markers is often tested among patients exhibiting a
pathology relating to the schizophrenia spectrum,
i.e., among patients with a high level of clinical risk.
This research involving subjects exhibiting a high
level of clinical risk (schizophreniform disorder,
schizotypal personality disorder) is often presented
as being equivalent to research conducted among
subjects with a high level of biological risk (firstdegree relationship to schizophrenic subjects). However, the research conducted among subjects with a
high level of clinical risk is not comparable to that
involving subjects with a high level of genetic risk.
These cannot be considered as two interchangeable
modes of validating a cognitive marker. Nevertheless, the comparison that is often made between
these two types of research maintains the illusion
that the same type of genetic marker might be
expressed in the form of a continuum of varied
phenotypical manifestations, a possibility about
which we have no information. And, in particular, it
fosters the illusion of a relationship between the
genetic aspect (the marker) and the clinical aspect
(the schizophrenia spectrum, which suddenly is
considered to constitute a single entity). As things
stand, it is the marker, once it has been discovered,
that will enable us to determine the still vague border
between the pathologies of the schizophrenia spectrum and the schizophrenias themselves: in this case,
it is possible to speak of ‘‘laboratory tests’’. The
inverse operation, i.e., using the schizophrenia
spectrum in order to identify the vulnerability
markers, is equivalent either to making clinical
Sarfati and Hardy-Baylé
speculations (the pathologies that are related to
schizophrenia have a genetic etiology close to that
of schizophrenia) or genetic speculations, as Holzman and colleagues demonstrated with the single
major locus hypothesis [Matthysse et al., 1986].
Neuchterlein, in line with the previously adopted
logic, has recently pointed out that ‘‘Neurocognitive
and psychophysiological abnormalities that can be
detected among biological relatives of schizophrenia
patients who do not manifest even schizotypical
personality disorder or other symptomatically defined schizophrenia spectrum disorders may particularly extend the relevant phenotype for genetic
transmission studies’’ [Neuchterlein et al., 1999].
The models of vulnerability to schizophrenia have
made it possible to precisely define the notion of
cognitive vulnerability in the form of trait markers of
the illness. Thus, the existence of a cognitive vulnerability has no direct link to clinical observations and is
unable to account for them. Other cognitive models of
schizophrenia that do not possess the explicit objective
of identifying trait markers attempt to establish links
between cognitive anomalies and the symptoms of the
illness. These are the so-called comprehensive models:
these consider the cognitive anomaly to possess a state
marker that explains the emergence of the symptom
but that says nothing about its etiology.
Recent electrophysiological data obtained through
the study of cognitive evoked potentials for each of the
two approaches makes it possible to illustrate the logic
of the two models. Depending on the vulnerability
model adopted, genetic linkage analyses have revealed
an association between a deficit in P50 wave elimination and an alpha-7 subunit in the nicotine receptor
[Freedman et al., 1997]. Within this approach, a trait
marker representing a constant anomaly in a patient
makes it possible to identify a genetic vulnerability
without providing an enhanced clinical characterization of the patient.
In contrast, and in accordance with the psychopathological models, symptom-directed analyses in groups of
disorganized patients have revealed a link between
schizophrenic disorganization and an absence of
N400 modulation. The anomaly appears in step with
the clinical pattern of disorganization, disappears
with it, and does not appear in non-disorganized
schizophrenic patients [Sarfati and Passerieux, 2000].
Within this approach, a state marker, representing an
anomaly that varies in patients, permits a better
understanding of the emergence of clinical anomalies
without providing a better explanation of their etiopathogenesis.
The existence of a cognitive vulnerability to schizophrenia in a subject should not be thought of as a
necessary and sufficient factor for the emergence of a
schizophrenic disorder. It cannot, therefore, be sufficient for the prevention of the disorder and the
commencement of treatment before the appearance of
the first clinical signs.
The presence of a cognitive anomaly can only assist in
individual diagnoses and, consequently, the provision
of treatment as of the moment that the symptamatology
has emerged. Considered from the strictly cognitive
viewpoint, the presence of one or more markers in a
subject is not in itself sufficient to justify preventive
treatment given that such subjects may remain asymptomatic all their lives. This comment would remain true
even if the subject exhibited personality traits or
clinical dimensions of the schizophrenic spectrum
because the models of vulnerability do not make it
possible to amalgamate clinical presentation with endophenotypical markers. As things stand, the co-existence of a clinical risk (subject within the schizophrenic
spectrum) and genetic risk (presence of a cognitive
vulnerability) constitute nothing more than an accumulation of illness risk factors, without it currently
being possible to say anything about the causal link
between the two risks or about their possible accumulated effects. The preventive treatment of subjects
exhibiting an accumulation of risk factors, including
cognitive factors, relating to schizophrenia is therefore
based on a statistical risk-oriented approach to treatment but does not adhere to the logic of models of
vulnerability to schizophrenia.
Amminger GP, Pape S, Rock D, Roberts SA, Ott SL, Squires-Wheeler E,
Kestenbaum C, Erlenmeyer-Kimling L. 1999. Relationship between
childhood behavioral disturbance and later schizophrenia in the New
York high-risk project. Am J Psychiatry 156:525–530.
Azorin JM. 1997. Les modèles de vulnerabilité dans la schizophrenie. Paris:
Bleuler E. 1911. Dementia praecox oder die Gruppe der Schizophrenien.
Leipzig: Deuticke.
Cannon M, Jones P, Huttunen MO, Tanskanen A, Huttunen T, RabeHesketh S, Murray RM. 1999. School performance in Finnish children
and later development of schizophrenia. Arch Gen Psychiatry 56:457–
Clementz BA, Geyer MA, Braff DL. 1998. Poor P50 suppression among
schizophrenia patients and their first-degree biological relatives. Am J
Psychiatry 155:1691–1694.
Crawford TJ, Sharma T, Puri BK, Murray RM, Berridge DM, Lewis SW.
1998. Saccadic eye movements in families multiply affected with
schizophrenia: the Maudsley family study. Am J Psychiatry 155:1703–
Freedman R, Coon H, Myles-Worsley M, Orr-Urtreger A, Olincy A, Davis A,
Polymeropoulos M, Holik J, Hopkins J, Hoff M, Rosenthal J, Waldo MC,
Reinherr F, Wender P, Yaw J, Young DA, Breese CR, Adams C,
Patterson D, Adler LE, Kruglyak L, Leonard S, Byerley W. 1997.
Linkage of a neurophysiological deficit in schizophrenia to a chromosome 15 locus. Proc Natl Acad Sci USA 94:587–592.
Hardy-Baylé MC. 1997. Neuropsychologie cognitive de la schizophrenie. In:
Encyclopedie Medico-Chirurgicale. Paris: Elsevier. 37-285-A10.
Hodges A, Byrnes M, Grant E, Johnstone E. 1999. People at risk of
schizophrenia: sample characteristics of the first 100 cases in the
Edimburgh high-risk study. Br J Psychiatry 174:547–553.
Laurent A, Rochet T, D’Amato T, Anchisi AM, Daumal M, Favre P,
Bougerol T, Dalery J. 2000. Vulnerabilité à la schizophrénie. III:
interêts et limites du test de performances continues à paires
identiques (CPT-IP). Encephale 26:48–55.
Cognitive Vulnerability
Matthysse S, Holzman PS, Lange K. 1986. The genetic transmission of
schizophrenia: application of Mendelian latent structure analysis to eye
tracking dysfunctions in schizophrenia and affective disorders. J Psy
Res 20:57–65.
Sarfati Y, Passerieux C. 2000. Une approche cognitive de la désorganisation schizophrénique: vers une échelle pathogénique de la désorganisation. In: Trognon A, Musiol M, editors. Eléments de psychopathologie
cognitive. Paris: Armand Colin. p 195–220.
Neuchterlein KH. 1987. Vulnerability models for schizophrenia: state of
the art. In: Häfner, Gattaz WF, Janzarik W, editors. Search for the
causes of schizophrenia. Heidelberg: Springer. p 297–316.
Neuchterlein KH, Dawson ME. 1984. A heuristic vulnerability/stress
model of schizophrenic episodes. Schizophr Bull 10:300–312.
Neuchterlein KH, Subotnik KL, Asarnow RF. 1999. Dimension of
schizotypy in symptomatic, neurocognitive and psychophysiological
indicators. In: Maj M, Sartorius N, editors. Schizophrenia. Chichester:
Wiley. p 384–386.
Walker EF. 1999. Exploring schizophrenia across the continuum. In:
Maj M, Sartorius N, editors. Schizophrenia. Chichester: Wiley. p 379–
Zubin J, Spring B. 1977. Vulnerability: a new view of schizophrenia. J
Abnorm Psychol 86:103–126.
Zubin J, Steinhauer SR, Day R, van Kammen D. 1985. Schizophrenia at the
cross-road: a blueprint for the 80’s. Comprehensive Psychiatr 26:217–
Без категории
Размер файла
56 Кб
vulnerability, identify, high, cognitive, risk, schizophrenia, coulda, subjects
Пожаловаться на содержимое документа