Poster Session II

S253

with the tests, i.e. CVLT and Forward Digit Span, were specifically F87. SERUM PROLACTIN LEVELS AND
related to JTC measured by Fish Task 60:40. These variables turned out COGNITIVE OUTCOME IN FIRST EPISODE
to be insignificant predictors of JTC 80:20 and JTC as the DACOBS
subscale. JTC measured using Fish Task (60:40) was correlated only
PSYCHOSIS: A PROSPECTIVE 1-YEAR
with disorganization. FOLLOW-UP STUDY
Discussion: The results from the present study suggest that the relationship
Meritxell Tost Bonet*,1, Itziar Montalvo1, Francesc Estrada1,
between decision making during the reasoning task and neuropsychologi-
Maria Isabel Ahuir Perez1, Ángel Cabezas2, Montse Paños1,
cal functioning is modulated by task demands. In particular, verbal work-
ing memory deficits are implicated in more hasty decision making when the Montserrat Sole2, José Antonio Monreal1,
task is demanding. vanessa Sanchez-gistau2, Elisabet Vilella2, Diego Palao1,
Javier Labad1
1
Parc Taulí Hospital Universitari, I3PT, Universitat Autònoma
F86. SOCIAL COGNITION IN HOMICIDE de Barcelona; 2Early Intervention Service, Hospital Universitari
Institut Pere Mata, IISPV, Universitat Rovira i Virgili
OFFENDERS WITH SCHIZOPHRENIA
Katharina Engelstad*,1, Bjørn Rishovd Rund2, Background: Recent studies in patients with psychotic disorders or prolac-
Anne-Kari Torgalsboen2, Anja Vaskinn3 tinomas suggest that increased prolactin levels may have negative effects on
1
Vestre Viken Hospital Trust; 2University of Oslo; 3Oslo University cognition. Most previous studies including patients with psychotic disor-
Hospital ders are cross-sectional, and longitudinal studies are lacking. We aimed to
conduct an observational, prospective study to explore whether prolactin
levels during the first year of treatment are associated with changes in cog-
Background: Schizophrenia is characterized by impairments in social
nitive tasks. Our hypothesis would be that those patients with increased
and non-social cognition that act as strong predictors of the poor out-
prolactin levels would show a poorer cognitive outcome than those patients
come of the disorder, including violence. In fact, it has been suggested
with lower prolactin levels.
that the inclusion of (social) cognition in risk assessment tools may
Methods: We studied 60 patients (24.5  ±  6.8  years; 36% women) with a
increase their accuracy. A premise for this line of thinking is that indi-
first episode psychosis (FEP) attending the Early Psychosis Programme
viduals with schizophrenia and a history of violent offences should
from two institutions (Parc Taulí Hospital Universitari, Sabadell, Spain;
present with a different social cognitive profile than individuals with
Hospital Universitari Institut Pere Mata). Ethical approval was obtained
schizophrenia without such a history. In this study, we compare social
from the local Ethics Committees of both institutions. Clinical diagnoses
and non-social cognition in homicide offenders with schizophrenia
for a FEP were generated with the OPCRIT checklist v.4.0 after a semis-
(HOS) and in individuals with schizophrenia without a history of inter-
tructured interview by a psychiatrist. The MATRICS Consensus Cognitive
personal violence (non-HOS).
Battery (MCCB) was administered to explore neuropsychological function-
Methods: Twenty-six HOS and 28 non-HOS were included. They
ing at two-time points (baseline, 1 year). The MCCB contains 10 tests to
underwent a comprehensive clinical and neuropsychological assess-
measure cognitive performance in 7 cognitive domains. Three fasting blood
ment protocol where the MATRICS Consensus Cognitive Battery
samples (baseline, 6  months, 12  months) were obtained in the morning
(MCCB) and four tests of social cognition were applied. Facial emo-
between 8:30 h and 9:30 h in resting conditions, to determine unstimulated
tion perception was assessed with Pictures of Facial Affect (PFA) where
plasma prolactin. Serum prolactin levels were determined with immunoas-
portrait photographs of people expressing one of six emotions are pre-
says standardized against the 3rd International Reference Standard 84/500.
sented. Emotion perception from bodies was measured with Emotion
Statistical analyses were performed with SPSS version 21.0. As prolactin
in Biological Motion (EmoBio), a point-light task consisting of short
levels might be increased by stress, particularly at the onset of the FEP,
movie clips where lighted dots, indicative of one of four emotions,
we calculated mean prolactin levels over the follow-up period taking into
move across the computer screen. Theory of mind (ToM) was indexed
account prolactin values at 6 and 12  months. A  general linear model for
by two tests. The Hinting Task consists of ten short stories where a hint
repeated measures was performed in order to test whether longitudinal
is dropped. The ecologically valid Movie for the Assessment of Social
changes in cognitive tasks differed by mean prolactin levels. All analyses
Cognition (MASC) shows four characters who meet for dinner. The
were adjusted for gender. A p value < 0.05 (two-tailed) was considered to
movie is stopped several times, and the test taker is instructed to answer
be significant. For descriptive purposes, patients in the fourth quartile for
questions about the thoughts, intentions and emotions of a given char-
serum prolactin (>47.8 ng/ml for men; >54.3 ng/ml for women) where com-
acter. This yields scores for cognitive and affective ToM. In addition,
pared with those with lower prolactin levels.
the multiple-choice response format provides information on overmen-
Results: Patients improved in all 10 MCCB cognitive tasks one year later
talizing and undermentalizing errors.
(p<0.05 for all tasks). When exploring the interaction between time x pro-
Results: Preliminary analyses, using independent samples t-tests, showed
lactin in the GLM analyses, significant interactions were found for three
significant differences for the overall scores on the EmoBio and MASC
cognitive tasks related with processing speed (BACS-SC [F= 5.9, p= 0.018];
tests, with non-HOS outperforming HOS participants. Follow-up analy-
Fluency [F= 5.6, p= 0.022]) and reasoning and problem solving (NAB
ses for the EmoBio test revealed no statistically significant differences for
mazes [F= 4.8, p= 0.033]). Percent change over the follow-up period in
any specific emotion, although the group difference for the recognition of
these cognitive tasks was greater for patients with lower prolactin levels,
happy, fearful or neutral body movements approached a medium effect size.
when compared to those in the highest quartile: BACS-SC (11.8% vs 0.6%),
On the MASC test, HOS presented with reduced cognitive and affective
Fluency (8.5% vs -7.4%) and NAB mazes (10.7% vs -1.1%).
ToM, compared to non-HOS, and also committed more undermentalizing
Discussion: Our study is in accordance with previous cross-sectional stud-
errors. These effect sizes were large.
ies reporting a negative effect of prolactin levels on cognition and adds
Discussion: Our preliminary analyses show that homicide offenders with
new information with a prospective design. A  limitation of our study is
schizophrenia have reduced emotion perception and theory of mind com-
that patients were treated with different antipsychotics based on the clini-
pared to individuals with schizophrenia without a history of interpersonal
cal routine practice. Antipsychotic-induced hyperprolactinaemia, which is
violence. This lends support to the idea that social cognitive deficits can
caused by tuberoinfundibular blockade of D2 receptors, may be reflect-
be a risk factor for interpersonal violence in persons with a diagnosis of
ing the blockade of D2 receptors in other dopaminergic pathways (stri-
schizophrenia.
atum and mesocortical pathways) that may also affect cognitive abilities.

Abstracts for the Sixth Biennial SIRS Conference

Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/44/suppl_1/S253/4957683
by guest
on 09 August 2018
 S254 Poster Session II
Further clinical trials are needed to reduce the heterogeneity of the treat-
ment effects and to confirm the potential negative effects of prolactin on
cognitive abilities.
F88. MANIPULATION OF THE GUT MICROBIOTA
WITH A PREBIOTIC IN SCHIZOPHRENIA:
A DOUBLE-BLINDED RANDOMIZED PLACEBO-
CONTROLLED CROSS-OVER STUDY
Amy Kao*,1, Phil Burnet1, Belinda Lennox1
1
University of Oxford
Background: Neurocognitive impairment is increasingly recognized as
a fundamental symptomatology in schizophrenia with more than 80%
of patients exhibiting significant deficits, even at first episode of illness.
Current pharmacotherapies do not alleviate cognitive symptoms, and often
cause severe metabolic dysfunction and weight gain that readily become
major health concerns. Identifying adjunctive interventions that improve
cognition without disrupting the therapeutic actions of anti-psychotic med-
ications, and can mitigate the metabolic side-effects of these drugs, would
be highly beneficial to patients and improve prognosis. We have recently
demonstrated that the manipulation of the rat gut microbiota with a prebi-
otic (dietary fibre that grows beneficial enteric bacteria) improves cognitive
flexibility and prevents olanzapine-mediated weight gain. We therefore aim
to explore whether these actions of the prebiotic translate to medicated sta-
ble patients with schizophrenia.
Methods: A total of 40 patients with psychosis aged 18–65 will be enrolled
in a 24-week maltodextrin-controlled cross-over experimental medicine
study. Participants will receive either a 12-week treatment with a prebiotic
(active compound) followed by a 12-week maltodextrin supplement (pla-
cebo), or in the reverse order. The order of supplements that participants
receive is randomized. The primary outcome is to examine the influence of
prebiotic supplementation on neurocognitive functioning, which is meas-
ured using a tablet-based neuropsychometric test battery. We will also
examine the impact on clinical metabolic measures such as total weight and
visceral adiposity. The concentration of immune-related serum proteins as
well as neuroendocrine hormones will be evaluated. All measurements will
take place at baseline, at 12-week cross-over, and at the end of the 24-week
study. A within-subjects repeated measures analysis will be performed, and
co-variates (gender, weight, medication) identified. This trial is registered
with ClinicalTrials.gov, identifier number NCT03153046.
Results: We have currently screened 36 patients, of whom 30 were eligi-
ble for the study (67% male). At baseline, the average age of all recruited
was 36.41  ±  11.42. The overall cognitive score was -2.03  ±  0.53 where
the subtests included verbal memory (29.09  ±  8.75), digit sequencing
(15.41 ± 2.77), token motor (55.74 ± 24.14), semantic fluency (8.83 ± 3.46),
letter fluency (11.39  ±  3.65), symbol coding (35.61  ±  9.22) and tower of
London (15.1 ± 4.06). There was no difference in overall cognitive between
male (14.29 ± 2.46) and female (17.22 ± 2.30) patients. However, long-term
associative learning as measured by the digit sequencing subtest appeared
to show a significant difference between male (14.29  ±  2.46) and female
(17.22  ±  2.33; p=0.008) participants. No significant differences between
clinical metabolic measures were observed in baseline BMI (32.25 ± 6.79)
and abdominal obesity as measured by hip-to-waist ratio (0.94  ±  0.11).
The serum concentrations of immune and endocrine markers will also be
presented.
Discussion: This investigation, to our knowledge, is the first clinical study
to provide medicated schizophrenia patients with a prebiotic, as a poten-
tial means of improving cognition and managing secondary metabolic dys-
function. Although a potential link between commensal enteric bacteria
and schizophrenia have been suggested, earlier work with probiotics (live
cultures) did not support this association. However, since the current study
uses a prebiotic that proliferates multiple species of gut bacteria, it will pro-
vide more robust data that will support, or refute, the validity of manipulat-
ing the gut microbiome in the treatment of schizophrenia.
Abstracts for the Sixth Biennial SIRS Conference
Downloaded from https://academic.oup.com/schizophreniabulletin/article-abstract/44/suppl_1/S253/4957683
by guest
on 09 August 2018
F89. COGNITIVE IMPAIRMENT WORSENING
ACROSS AFFECTIVE TO PSYCHOSIS SPECTRUM:
A COHORT STUDY OF UNIPOLAR/BIPOLAR
DEPRESSION AND BIPOLAR SCHIZOAFFECTIVE
DISORDERS
Martin Roy*,1, Marie-Pierre Strippoli2, Caroline Vandeleur2,
Elsa Gilbert3, Enrique Castelao2, Jean-Michel Aubry4,
Pierre Marquet3, Martin Preisig2
1
CERVO Brain Research Centre; 2University Hospital of Lausanne;
3
Université Laval, CERVO Brain Research Centre; 4University
Hospital of Geneva
Background: Bipolar (BPD), schizoaffective bipolar (SAM) and major
depressive disorders (MDD) reveal large heterogeneity in terms of symp-
tom expression, course and treatment response. This heterogeneity could
be the source of a large variance of cognitive performance observed in
these subjects. The aim of the present analyses was to compare the cogni-
tive performance of patients with BPD, SAM, MDD and medical controls
with adjustment for a comprehensive array of potential confounders. To
go a step further we will simultaneously test the effects of multiple clin-
ical characteristics including lifetime history and duration of psychotic
symptoms, manic/hypomanic and depressive episodes, age of onset of dis-
order, current GAF score, time since remission of the last episode and pres-
ence of a depressive episode at the time of the assessment on the cognitive
performance.
Methods: Data stemmed from the Lausanne-Geneva Family and High-
Risk study. Patients with BPD (n=62), SAM (n=22) and MDD (n=51)
were interviewed every three years over a mean duration of follow-up of
12  years. All patients were assessed clinically with the semi-structured
Diagnostic Interview for Genetic Studies (DIGS). The cognitive assessment
was made with the MATRICS and the Victoria Stroop Test.
Results: The global cognitive index (excluding Stroop result) shows that
SAM subgroup had the lowest global score with 40.6 (SD=8.5), BPD 47.4
(SD=7.8) and MDD 49.7 (SD=8.7). A multiple linear regression account-
ing for several confounders such as comorbid psychiatric disorders and
medication confirms that only SAM and BPD are statistically different
from controls (p<0.001 and p<0.01 respectively). MDD did not differ from
controls (p>0.05). Overall, patients with BPD or SAM but not with MDD
showed poorer cognitive performance than controls in terms of the global
score and speed of processing, verbal learning, working memory, visual
learning, attention/vigilance and inhibition.
Discussion: Our data confirm cognitive impairment in patients with BPD
or SAM compared to controls after adjustment for a comprehensive array
of potential confounder variables. We were able to evaluate the specific-
ity of cognitive performance of psychotic, maniac and depressive dimen-
sions of the major mood disorders within the same sample. Furthermore,
these data stress that the presence of the “schizo dimension” concomitant
to mania and depression contributes to worsening the cognitive perfor-
mance in an additive manner.
F90. SOCIAL INFERENCE AND BELIEFS
DIFFER IN INDIVIDUALS WITH SUBCLINICAL
PERSECUTORY DELUSIONAL TENDENCIES
Katharina Wellstein*,1, Andreea Diaconescu2, Martin Bischof3,
Annia Rüesch Ranganadan2, Johannes Ullrich4,
Klaas Enno Stephan1
1
ETH Zurich; 2University of Zurich, ETH Zurich; 3Psychiatric
University Hospital Zurich; 4University of Zurich
Background: It has long been suspected that abnormalities in social infer-
ence (e.g., about the intentions of others) play a key role for persecutory