Youth at Risk of Psychosis: Neurocognitive profiles and
non-pharmacological interventions
Mémoire
Rossana Peredo Nunez de Arco
Maîtrise en épidémiologie
Maître ès sciences (M.Sc.)
Québec, Canada
iii
Résumé
Introduction: L'intérêt pour les premiers stades de la psychose a augmenté au cours des
dernières années, vu que cette maladie apparaît pendant l'adolescence. Ensuite, de nombreuses études ont révélé que l'identification et le traitement précoces peuvent retarder la transition vers un trouble mental, et aussi prévenir des effets néfastes sur le fonctionnement global. Afin d'identifier ces individus, certains critères cliniques ont déjà été développés, on sait ainsi que les enfants avec risque génétique de psychose s'engagent tôt dans une trajectoire cognitive déficiente. Même si les modèles de prédiction sont très prometteurs, le nombre de faux positifs est élevé, ce qui nuit au développement de traitements préventifs. L'objectif du premier article était d'identifier deux profils neurocognitifs parmi les descendants des parents avec psychose. Le deuxième article avait comme objectif d'évaluer l'effet sur la transition d'interventions non pharmacologiques, chez les individus à risque de psychose et leur effet sur les comorbidités non psychotiques.
Méthodologie: Une analyse de cluster hiérarchique a été effectuée afin d'identifier deux
profils neurocognitifs. Ensuite, une analyse systématique et méta-analyse d'essais contrôlés randomisés a été effectué pour analyser les interventions non pharmacologiques publiées jusqu'à cette date.
Résultats: L'analyse de cluster a montré l‟existence de deux sous-groupes de
descendants à risque élevé, l'un d'entre eux ayant montré une performance cognitive presque identique aux sujets témoins, tandis que l'autre ayant eu des résultats pires que les scores du groupe control. La méta-analyse a rapporté que les thérapies non pharmacologiques étaient associées à un risque réduit de transition vers la psychose.
Conclusion: Les interventions non pharmacologiques peuvent avoir du potentiel de
traitement chez les individus à risque de psychose. Toutefois on a besoin de plus d‟études concentrés à réduire les taux de retrait. Notre étude suggère que les interventions visant à renforcer l‟aspect neurocognitif devraient être abordées plus tôt. D'autres recherches de types longitudinales sont nécessaires.
iv
Abstract
Introduction: Interest in the early stages of psychosis has been increasing in the last
years, mainly because it appears mostly in adolescence. Also, numerous studies have reported that early identification and treatment may not only delay the transition to a frank mental disorder, but also prevent detrimental effects on global functioning. In order to identify these individuals, some clinical criteria have already been developed; it is known for example that children at genetic risk of psychosis engage early in a deficient cognitive trajectory. Even though models of prediction are very promising, the number of false positives is still high, which impairs the development of preventive treatments. The objective of the first article was to identify two neurocognitive profiles among offspring at genetic risk of psychosis. The objective of the second article was to assess, the effect of pharmacological interventions on transition to psychosis, compared to any no non-pharmacological treatment, in individuals at risk of psychosis and the effect of these interventions on non-psychotic comorbidities.
Methodology: First a hierarchical clustering analysis was performed in order to identify
the two neurocognitive profiles. Then a systematic review and meta-analysis of randomized controlled trials was conducted to analyse all non-pharmacological interventions published until now.
Results: The cluster analysis yielded two subgroups of high risk offspring, one of them
showing a cognitive performance almost identical to control subjects, whereas the other having performed worse than the control scores. The meta-analysis reported that non-pharmacological therapies were associated with a reduced risk of transition to psychosis.
Conclusion: Non-pharmacological interventions may have potential in the treatment of
individuals at risk of psychosis however; further research is needed accompanied by efforts to diminish withdrawal rates. Our study suggests that interventions with a neurocognitive target should be addressed earlier. Still further research is needed in longitudinal studies.
v
Table of Contents
Résumé iii
Abstract iv
Table of Contents v
List of Tables vii
List of Figures viii
List of Abbreviations and acronyms ix
Acknowledgments x
Preface xi
Introduction 1
CHAPTER I: Theoretical framework 2
1.1 Risk of psychosis 2
1.2 Prodromal state of psychosis 2
1.3 Prediction of psychosis and predictive markers 5
1.3.1 Physical activity in CHR 6
1.3.2 Neurocognition in CHR 6
1.4 Prognosis and the need of intervention 7
CHAPTER II: Article 1 9
Résumé 9 Abstract 10 Introduction 11 Methods 12 Results 14 Discussion 15 Conclusion 18 Acknowledgments 18 References 19 Annexes 21
CHAPTER III: Article 2 28
Résumé 28
Abstract 29
1. Introduction 30
2. Methods 31
2.1Eligibility criteria 31
2.2 Information sources and literature search 33
2.3 Study selection and data collection process 33
2.4 Risk of bias 34
2.5 Summary of measures and synthesis of results 34
2.6 Risk of bias across studies 35
2.7 Additional analyses 35
3. Results 35
3.1 Study Selection 35
3.2 Studies characteristics 35
3.3 Risk of bias within studies 36
vi
3.4.1 Primary Outcome 36
3.4.2 Secondary outcomes 37
3.4.3 Subgroup analysis and sensitivity analysis 38
4. Discussion 39 5. Conclusion 41 Acknowledgments 41 References 42 Annexes 45 Search strategy 57
Discussion and Conclusion 68
vii
List of tables
Article 1:
Table 1. Socio-demographic characteristics of the sample 21 Table 2. Comparisons of neurocognitive domains between Deficient HR,
Healthy HR and Control group 23
Table 3 a. Neurocognitive function means of HR1, HR2,
Control group and comparisons among groups of age 24
Table 3 b. Neurocognitive function means of HR1, HR2,
Control group and comparisons among overall age groups 24 Article 2:
Table 1. Population characteristics 46
Table 2. Intervention characteristics 48
viii
List of Figures
Article 1:
Figure 1. Cluster solution. Dendrograms by age 22
Figure 2 a. Processing speed values by age and by group 25
Figure 2 b. Verbal Memory values by age and by group 25
Figure 2 c. Visual Memory values by age and by group 26
Figure 2 d. Working memory values by age and by group 26
Figure 2 e. Executive functioning values by age and by group 27
Article 2:
Figure 1. Flow chart: Identification of included trials 45
Figure 2. Risk of Bias graph 50
Figure 3. Transition to psychosis at 6 months of follow-up. Non pharmacological intervention vs any no non-pharmacological intervention 51 Figure 4. Transition to psychosis at 12 months of follow-up. Non
pharmacological intervention vs any no non-pharmacological intervention 52 Figure 5. Transition to psychosis at 18 months or more of follow-up. Non
pharmacological intervention vs any no non-pharmacological intervention 53 Figure 6. Transition to psychosis stratified by duration of the intervention at 12
months of follow up 54
Figure 7. Transition to psychosis stratified by duration of the intervention at
ix
List of abbreviations and acronyms
APS: Attenuated Positive Psychotic symptoms BLIPS: Brief limited intermittent psychotic symptoms BP: Bipolar Disorder
BS: Basic Symptoms
BSIP: Basel Screening Instrument for Psychosis
CAARMS: Comprehensive Assessment of At-Risk Mental States CBT: Cognitive Behavioral Therapy
CCC: Cubic Clustering criterion CHR: Clinical High Risk
CI: Confidence Interval
COPS: Criteria of Prodromal Syndromes
DSM: Diagnostic and Statistical Manual of Mental Disorders
EC: Enhanced care
EPA: European Psychiatric Association
ERIraos: Early Recognition Inventory for the Retrospective Assessment of the Onset of schizophrenia
FFT: Family-focused therapy
GAF: Global Assessment of Functioning GRD: Genetic risk and deterioration syndrome HR: Genetic high risk
NDRL: Non-directive reflective listening OR: Odds Ratio
PACE: Personal Assessment and Crisis Evaluation
PRIME: Prevention through Risk Identification Management and Education RCT: Randomized clinical trials
SD: Standard Deviation
SIPS: Structured Interview for Prodromal Symptoms SMD: Standardised mean differences
SOPS: Scale of Prodromal Symptoms SSD: Schizophrenia spectrum diagnoses SZ: Schizophrenia
UHR: Ultra High Risk
VEM: Verbal episodic memory VISEM: Visual episodic memory
x
Acknowledgments
I would like to express my gratitude to my supervisor Chantal Mérette PhD. for her guidance, and for providing me with an excellent work environment conducive to serious research. Her enthusiasm, her immense knowledge and her accessibility made my journey very enriching and stimulating.
I would like to thank the rest of experts who were involved in this project: Dr. Michel Maziade, Elsa Gilbert PhD, Valérie Jomphe M.Sc. for their insightful comments and for answering all my questions.
I would also like to thank my parents and my husband for all their support and for encouraging me to follow my dreams.
xi
Preface
This project was submitted to Laval University for the degree of Master‟s in Epidemiology. This research was conducted under the supervision of Professor Chantal Mérette Ph.D., Professor at Laval University, Department of Psychiatry and Neuroscience of the Faculty of Medicine, and a researcher at the Centre de recherche de l‟Insititut universitaire en santé mentale de Québec (IUSMQ) where she is the Director of the Neuroscience Biostatistics Platform.
All of the work presented henceforth was conducted at the Centre de recherche de l‟Insititut Universitaire en Santé Mentale de Québec. The two following articles are presented according to the thesis with insertion of article option:
1.- Rossana Peredo, Michel Maziade, Elsa Gilbert, Valérie Jomphe, Thomas Paccalet, Chantal Mérette. Cluster analysis identifies two neurocognitive profiles among
offspring at genetic risk of psychosis. This article is being reviewed for a future
publication.
2.- Rossana Peredo, Geneviève Picher, Michel Maziade, Elsa Gilbert, Kaoutar Ennour-Idrissi, Chantal Mérette. Non-pharmacological interventions in individuals at risk of psychosis: a systematic review and meta-analysis of randomized controlled trials.
This article is being reviewed for a future publication.
I was the lead investigator for the two articles. The data for the first article was collected previously as part of a research project under the responsibility of my supervisor Chantal Mèrette, and it was approved by the ethics committee of the „Comité d‟éthique de la recherche Institut universitaire en santé mentale de Québec‟. I was responsible for part of the concept formation as well as the analysis and the manuscript composition. My supervisor was involved throughout the project concept formation, supervision and guidance of statistical analysis and manuscript edits. Michel Maziade, Elsa Gilbert, Thomas Paccalet, participated also in the supervision and conceptualization and Valérie Jomphe, supervised the statistical analysis.
xii For the second article I was responsible for all major areas of concept formation, data collection and analysis, as well as the manuscript composition. Chantal Mérette was also the supervisory author, Geneviève Picher and Kaoutar Ennour-Idrissi participated in the secondary data collection necessary for the meta-analysis, Michel Maziade, Elsa Gilbert, will supervise the manuscript for future publication.
1
Introduction
Schizophrenia (SZ) not only causes a terrible effect on individuals and their families but it also represents a substantial financial burden for the health care system, due to its early onset in life predominately during adolescence1. Although the causes of the disease are not completely known, its developmental course and prodromal phase are now better understood2-7. Screening tools for prodromal symptoms have been developed and are currently in use8-16; researchers have even suggested that approximately 30% of individuals at risk may transit to psychosis in the following two years17. It is possible, however, that the existence of a high number of false positives casts some doubt upon the real transition rate18-20. A recent challenge for researchers has been to identify markers that could increase the specificity of screening instruments21-23. Previous studies for example have studied the lifestyle choices in this population as possible risk factors24-26. Also researchers have observed cognitive impairments not only in patients with schizophrenia but also in children at genetic high risk (HR)27-29.
Based on this information, this project was divided into two parts. First we hypothesized that within a sample of offspring of individuals with SZ or Bipolar Disorder (BP), distinct neurological function profiles would exist, one being similar to that of a control group. Then in order to contribute to the knowledge about effective treatments that target not only psychological symptoms but also, cognitive deficiencies the objective of the second article was to assess the effect on transition to psychosis of non-pharmacological interventions compared to any no non-pharmacological treatment, in individuals at risk of psychosis and the effect of the interventions on effect on non-psychotic comorbidities.
2
CHAPTER I: Theoretical framework
1.1 Risk of Psychosis
Within all psychotic disorders, schizophrenia spectrum diagnoses (SSD) account for approximately two‐thirds. These types of disorders are more prevalent between ages 15-17 and rarely occur before age 141 which is catastrophic. However early identification is possible since there is clear evidence for heritability and familial transmission, and even though it has been very difficult to determine the specific causative genes2, it is known that the risk for schizophrenia (SZ) among first-degree relatives range from 6% to 13% and for second-degree relatives from 2 to 4%3,4.
Furthermore, some statistical models that includes familial risk and demographic factors have already been proposed to estimate the risk for SZwith a level of predictive accuracy comparable to that in other areas of preventive medicine4,5. Thus, offspring of parents with a severe mental illness are studied as a population at increased risk for psychiatric disorders. In fact a meta-analysis that included 33 studies showed that one third of offspring may develop severe mental illness by early adulthood6.
1.2 Prodromal state of psychosis
The first episode of psychosis may be preceded by a prodromal state recognized by numerous studies. This prodromal phase is characterized in most individuals by the presence of attenuated symptoms and impaired functioning that tend to accumulate exponentially until there is a transition to frank psychosis7. Yung and McGorry8 were the pioneers in proposing the diagnostic criteria for ‟‟Ultra High Risk‟‟ (UHR) status, and they established the first clinical service for potentially prodromal individuals9 in the mid-1990s: The Personal Assessment and Crisis Evaluation (PACE) Clinic in Australia. The UHR criteria7 are based on a combination of epidemiological evidence that includes being between 14 to 29 years old, having been referred to a specialized service for help and meet the criteria for one or more of the following groups:
Group 1.- Attenuated Positive Psychotic symptoms (APS): which includes the
presence of at least one of the following symptoms: ideas of reference, odd beliefs or magical thinking, perceptual disturbance, paranoid ideation, odd thinking and speech, odd behaviour and appearance. Also, that the frequency of symptoms is at least several times
3 a week, and must be present within the last year, with a duration of at least 1 week and no longer than 5 years.
Group 2: Brief limited intermittent psychotic symptoms (BLIPS): Include the
presence of at least one of the following transient psychotic symptoms: ideas of reference, magical thinking, perceptual disturbance, paranoid ideation, odd thinking or speech. With a duration of less than one week, at least several times per week. The symptoms must resolve spontaneously, and must have occurred within the last year.
Group 3.- Genetic risk and deterioration syndrome (GRD): Schizotypal personality
disorder in the identified individual, or a first-degree relative with a psychotic disorder. Significant decline in mental state or functioning, maintained for at least one month and no longer than 5 years and the decline in functioning must have occurred within the past year.
The first UHR psychometric instrument, created by the same authors Yung et al., is called Comprehensive Assessment of At-Risk Mental States (CAARMS)10 which is a semi structured interview designed for use by mental health professionals who are already able to assess and evaluate patients‟ information and designed for repeated use over time. This instrument uses subscales with scores that range from 0 to 6, that target the following:
Disorders of thought content (e.g. delusional mood, overvalued ideas and delusions)
Perceptual abnormalities (e.g. distortions, illusions and hallucinations)
Conceptual disorganization (e.g. subjectively experienced difficulties with forming thoughts and objective assessment of formal thought disorder)
Motor changes (e.g. subjectively experienced difficulties with movement and objective signs of catatonia)
Concentration and attention (assessing both the subjective experience and objective rating),
Emotion and affect (assessing subjective sense of change in emotions and objective rating of blunting of affect)
Subjectively impaired energy (a basic symptom)
4 Another instrument was proposed by Miller and the research team at Yale University: Prevention through Risk Identification Management and Education (PRIME). This instrument also aimed to identify the same three prodromal syndromes mentioned before called the Criteria of Prodromal Syndromes (COPS)11. This instrument which is the most used in North America and Europe9 is called the Structured Interview for Prodromal Symptoms (SIPS) and it includes the companion Scale of Prodromal Symptoms (SOPS). It is a semi structured diagnostic interview designed to be used by experienced clinicians with specific training. The particularities of the SIPS is that it includes the 19 –item SOPS, designed to measure the severity of prodromal symptoms and changes over time, a version of the Global Assessment of Functioning (GAF) with well-defined anchor points, a schizotypal personality disorder checklist in the Diagnostic and Statistical Manual of Mental Disorders IV (DSM-IV), a family history of mental illness and a checklist for the Criteria of Prodromal Syndromes12.
Another set of clinical criteria was also developed in Germany by the psychiatrist Gerd Huber13 called Basic Symptoms which are subjective subclinical disturbances in drive, affect, thinking, speech perception, motor action central vegetative functions and stress tolerance. They are distinct from frank psychotic symptoms because meanwhile psychotic patients experience symptoms as real and normal, at risk individuals recognize immediately and in a very spontaneous way, that those symptoms are not „normal‟, and they will gradually increase in number and severity till they transit to psychosis. Other instruments are the Early Recognition Inventory for the Retrospective Assessment of the Onset of schizophrenia (ERIraos) Developed by Häfner et al, used in German and Italian studies and the Basel Screening Instrument for Psychosis (BSIP) developed in the Early Detection of psychosis Clinic in Basel by Riecher-Rössler et al.10,14.
Till this date, there is no diagnostic category in the DSM8, although the criteria for attenuated psychosis syndrome (APS) has already been published in the DSM Fifth Edition (DSM-5) as a „syndrome that is characterized by psychotic-like symptoms that are below a threshold for full psychosis‟15. This definition has already been tested in clinical
samples and compared with CAARMS definitions concluding that both definitions of APS are diagnostically different but prognostically comparable, the only advantage of the DSM-5 definition is the relatively quicker administration compared to CAARMS16. No matter which instrument is used, the prodromal phase for psychosis is defined as the period
5 between a relatively asymptomatic phase and the apparition of a frankly psychotic episode that is characterized by escalating severity of symptoms and functional decline10.
1.3 Prediction of psychosis and predictive markers
According to a meta-analysis published in 2013, approximately 30 % of a Clinical High Risk (CHR) population would transition to psychosis in the two following years17. However some studies have suggested that there is a decline in the transition rates, one explanation could be the presence of a possible “dilution” effect because of the occurrence of more false positives that were never at risk, another explanation could be that treatments are more effective at a very early stage18. It has also been proposed that the increasing awareness of UHR symptoms not only among health professionals but also in the general population could result in faster referrals of young people, changing the referral pathways to UHR services19,20, but still there is no consensus about the cause of this phenomenon.
In order to decrease the risk of offering unnecessary treatment to false positives it is necessary to improve the predictive methods. This could be achieved by combining clinical markers with family history of psychosis, with neurocognitive or electrophysiological measures or with environmental factors9,21. Patterns of whole-brain neuroanatomical abnormalities for example, have been shown to serve as valuable biomarkers to guide early detection of psychosis, with a cross-validated classification accuracy of 88% for individuals with transition and 86% for those without transition9,22.
Indeed, one study systematically reviewed predictive models for psychosis onset including clinical, biological, neurocognitive, environmental and combined predictive models, suggesting that sequential testing in CHR individuals may improve psychosis prediction, and finding the best model with a probability of transition of 98% for a 3-stage sequential testing based on one combinatory model (EEG+ clinical) and two biological models (structural MRI and blood markers) 23.
Several attempts have been made to improve early-detection of prodromal phases of psychosis, and to confidently predict the transition to the disease, but still further studies are needed to clarify the knowledge in order to address the correct affliction and provide an appropriate and early treatment.
6
1.3.1 Physical Activity in CHR
One cohort analysis in adolescents at risk for psychosis suggested that adolescents who later develop psychosis are three to four times more likely to be physically inactive compared to those who do not develop psychosis, having also a worse cardiorespiratory fitness than their age mates. Even though there were no statistically significant differences between with or without a familial risk for psychosis, those who reported several prodromal symptoms were physically more inactive than subjects with few or no prodromal symptoms. The more symptoms the subjects had the less physically active they were24.
Another more recent exploratory study also corroborated these findings reporting that healthy controls reported higher levels of participation in indoor/outdoor activities and in strength and /or flexibility training, and contrary to a previous study they did not find any relationship between level of activity and symptoms. However they reported that the CHR group endorsed more barriers to exercise, related especially to esteem and self-perception. They also had fewer reasons for exercising, reporting fewer reasons related to self-confidence and self-perception than healthy controls25.
One study26 even examined the relationship between physical activity level, brain structure and symptoms in UHR individuals, suggesting that the UHR group showed less total physical activity, but also this group exhibited smaller medial temporal volumes when compared with healthy controls. Total level of physical activity was moderately correlated with parahippocampal gyri bilaterally and with occupational functioning but not positive symptomatology, suggesting that inactivity is associated with medial temporal lobe health.
1.3.2.- Neurocognition in CHR
A meta-analysis that included 21 studies showed that CHR individuals have mild to moderate globally distributed neuropsychological performance deficits that lie between first episode of psychosis and healthy subjects. Neuropsychological performance deficits are greater in CHR that transition to psychosis than those who don‟t transition27. Differences in
verbal episodic memory (VEM) and visual episodic memory (VISEM) could play an important role in genetic mechanisms and therefore also in prevention28 In addition, researchers have observed that cognitive performance in domains such as processing speed, attention, working memory and verbal memory, decreases gradually from first
7 degree relatives of patients with psychosis (siblings, parents or offspring) to ultra-high risk individuals (UHR) and then to first episode of psychosis patients29.
Although there are numerous studies that aim to demonstrate neurological performance as a potential marker for prediction models of risk, still there is a need for more research in this area. This topic will be discussed in the following chapter.
1.4 Prognosis and the need for intervention
Only 46% of subjects experienced a full remission of prodromal symptoms from baseline according to the meta-analyses mentioned earlier17.On the other hand it is been proposed that transition does not occur from healthy individuals to disorder directly but mostly from a common mental disorder with a certain degree of psychosis to one with a greater degree of psychosis30.
In this vein, a recent study that followed UHR individuals for 6 years with a transition rate of 28,4%, demonstrated that persistence or recurrence of non-psychotic comorbid mental disorders in the UHR state is associated with poor functional outcomes (defined by the GAF), rather than with transition to psychosis. Furthermore, among those who did not transition, 28,3% reported attenuated psychotic symptoms whereas 45.3% remained functionally impaired at follow-up (GAF < 60), and 58 % of patients were affected by at least one comorbid disorder at follow-up. Among those who already had a comorbid disorder at baseline, 61,5% had persistent or recurrent course of the comorbidity and incident comorbid disorders emerged in 45,4% of baseline UHR patients. The most common comorbid disorders were affective and anxiety disorders, particularly major depressive disorders and panic disorders31. A recent study showed that only 21% of their sample of 744 CHR did not have a comorbid diagnosis, and that the most common diagnoses were also anxiety and depressive disorders and comorbidity whereas cannabis use was essentially independent of clinical outcome32.
Some promising interventions have already been proposed and appear to be not only feasible but also effective, indicating clinically meaningful advantages9. However despite the need to establish optimal approaches, there is still no gold standard treatment to date.
The European Psychiatric Association (EPA) published a guidance paper about the recommendations on early intervention in CHR states of psychosis. First it is considered
8 that psychological, in particular Cognitive Behavioral Therapy (CBT), as well as pharmacological interventions are able to prevent or at least postpone a first episode in adult CHR individuals, even though any long-term antipsychotic treatment with a primarily preventive purpose is still not recommended (grade of recommendation:D).
Also, the EPA considers that an early intervention should not only aim to prevent the first episode of an affective or non-affective psychotic disorder but also the development or persistence of functional or vocational deficits. In this regard, they recommend that comorbidities like depression and anxiety should be treated according to their respective treatment guidelines. Besides they consider that the current evidence of psychological and pharmacological interventions in children and young adolescents is not sufficient to justify primarily preventive interventions (grade of recommendation:D), but specific psychological intervention to improve functioning should be provided as part of an overall treatment plan and complemented with other interventions for other psychosocial problems and co-morbid mental disorders33.
There are some ethical aspects that need to be taken into consideration. First there is a potential for harm like over-medicalization especially with antipsychotic medication that has very important side effects. There is also potential harm from psychosocial therapies, though considered to be somewhat benign, because of the stigma attached to psychiatric diagnoses, an important consequence to be avoided for young people at risk for psychosis7. This issue will be further developed in coming chapters.
9
CHAPTER 2: Article 1
Cluster analysis identifies two neurocognitive profiles among offspring at genetic risk of psychosis.
Peredo Rossana, Jomphe Valérie, Maziade Michel, Gilbert Elsa, Paccalet Thomas, Mérette Chantal
Résumé.-
Contexte: Les descendants de patients atteints de schizophrénie (SZ) ou de trouble
bipolaire (BP) présentent des déficiences en divers domaines cognitifs. La performance des enfants des parents avec psychose diminue graduellement vers la phase prodromique et enfin vers le premier épisode de psychose. L'objectif de cette étude est de séparer les descendants d'individus avec SZ ou BP dans deux sous-groupes selon leur profil neurologique afin de trouver un groupe ayant des performances cognitives saines ou proches au celui des sujets sains.
Méthodes: l'échantillon était constitué de 140 descendants (âge moyen de 16,17 ans), qui
ont été appariés selon le sexe et l'âge à 140 enfants non exposés. Une analyse hiérarchique moyenne de cluster a été effectuée pour séparer le groupe exposé par groupes d'âge selon le fonctionnement cognitif. Ensuite, les deux sous-groupes ont été comparés au groupe témoin non exposé.
Résultats: Deux sous-groupes HR1 et HR2 ont été obtenus. Les moyennes des domaines
cognitifs ont montré une différence statistiquement significative entre HR1 et HR2, tandis que le groupe HR1 a été très similaire au groupe témoin.
Conclusion: Nous avons observé un sous-groupe avec une performance cognitive très
semblable aux individus non exposés au risque, tandis que l'autre sous-groupe a été encore pire que les cas présentés dans la littérature. Néanmoins, d'autres études futures de type longitudinal sont nécessaires.
10
Abstract
Background: Offspring of patients with Schizophrenia (SZ) or Bipolar Disorder (BP) show
impairment in various cognitive domains. The performance gradually decreases from relatives of patients with psychosis to individuals at prodromal phase and finally to subjects at first episode of psychosis. The aim of this study is to separate offspring of individuals with SZ or BP into two subgroups according to their neurological profile in order to find a group with healthy or close to healthy cognitive performance.
Methods: The sample consisted of 140 offspring (mean age of 16.17 years), who were
matched to 140 non-exposed children by gender and age. An average hierarchical cluster analysis was performed to separate the exposed group by age group in terms of their cognitive functioning. Then, both subgroups were compared to the non-exposed control group.
Results: Two subgroups HR1 and HR2 were obtained in the offspring group. The means
of the cognitive domains showed a statistically significant difference between HR1 and HR2, whereas the HR1 group performed very similar to the control group.
Conclusion: We observed one subgroup of offspring at risk for a mental illness with
cognitive performance very similar to non at risk individuals, while the other subgroup performed even worse than what was presented in the literature. Nevertheless, further longitudinal studies are needed.
11
Introduction
Our previous studies concluded that subjects with schizophrenia have impairment in cognitive functions1, as reported by some other authors2,3; likewise cognitive impairment was also observed in subjects with bipolar disorder (BP)4,5. What is more, our studies6,7,8 as other publications9,10,11 have shown that offspring of parents with SZ or BP perform worse than healthy control subjects in various cognitive domains, suggesting that children at genetic high risk (HR) for a major mental health disorder engage early in childhood in a deficient cognitive trajectory.
It also has been well established that a proportion of individuals at Ultra High Risk (UHR) or Clinical High Risk (CHR) transition to psychosis within a short period of time12,13 and that cognitive performance decreases gradually from first degree relatives of patients with psychosis to UHR individuals, to patients with a first episode of psychosis14. These studies contribute to tracking down the development of major mental disorder and offer the opportunity to develop interventions aimed at preventing these diseases. Also, it is important to identify at risk individuals as early as possible in the disease process given that a longer duration of untreated „at risk‟ symptoms was found to be correlated with worse functioning15.
As emphasized earlier, in a recent meta-analysis10 various cognitive domains were found to be impaired in CHR subjects, with pooled effect sizes ranging from -0.30 to -0.85. Several domains were also affected in HR offspring with effect sizes of similar magnitudes8,9. However, theses deficits were obtained from data of the entire sampleof subjects at risk even though it is known that only a third of them will eventually transit toward a severe mental illness16 and fortunately, two thirds will never convert to SZ or BP. Hence, the effect size reported may represent a mixture of larger and smaller deficits, referring to those who will eventually convert versus those who won‟t, respectively. This present study addresses this issue by separating offspring of individuals with schizophrenia (SZ) or bipolar disorder (BP) into two subgroups according to their neurological profile. Our hypothesis was that distinct neurological function profiles would exist; one comparable to that of a control group and another with impaired cognitive functions.
12
Methods
Sample
This is a cross-sectional study that uses a cohort design recruitment strategy. The „exposed „cohort‟ was defined as subjects having a parent affected by either SZ or BP. We recruited 140 exposed subjects from 6 to 24 years old; then, each exposed participant was matched for gender and age more or less one year, with an individual in the non-exposed cohort.
The data was drawn from previous independent studies that targeted all multigenerational families densely affected by SZ or BP in the Eastern Québec (Canada) catchment area6,7,8. Written informed consent was obtained for all participants. Inclusion criteria were: having a parent with a definite diagnosis of SZ or BP disorder according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). Exclusion criteria were the presence of a diagnosis of DSM-IV psychotic disorder, BP or major depression, and brain and metabolic disorders known to cause neuropsychological impairments. We also excluded individuals of 25 years old or more due to the age of onset of SZ and BD being around that age. The individuals in the control group were selected by advertisements in local newspapers and in the population. The exclusion criteria for this group were the same as those in HR with the addition of any axis I DSM diagnosis or a positive family history of SZ or BP spectrum disorders. More details are published elsewhere6.
Measurement Variables for cluster analysis
Cognitive battery
The following cognitive domains were assessed: 1) processing speed: digit symbol substitution Task from the WAIS –III and Category Fluency: animal naming; 2) verbal episodic memory: California Verbal Learning Test (CVLT II) total recall trials 1-5 and delayed recall: 3) visual episodic memory : Rey Complex Figure (RFC) immediate and delayed recall; 4) Working memory : Digit span from the WAIS-III and Spatial Span, and 5) executive functioning Wisconsin Card Sorting Test: total errors and Tower of London (TOLDX): number of problems solved in minimum moves. To compose the cognitive domains, for each subtest z scores were calculated using data from published standardized norms and then, for each cognitive domain, the average of the two subtests z
13 scores were calculated and presented in percentiles. Further information published elsewhere1.
Statistical Analysis
Data were analysed using SAS 9.4. First HR individuals were compared in age, sex and socioeconomic status using a t test. Then a hierarchical agglomerative clustering analysis by age group was performed, using the Ward‟s method, based on Euclidean distance which is the sum of the squared differences over all of the five neurocognitive measures. Ward‟s method was preferred in order to minimize the within-group dispersion and to minimize the overlap18. However, the Average method was also conducted in order to verify that similar clustering would be obtained. The scores z of the cognitive measures were presented in percentiles; therefore, it was not necessary to standardize the data before the analysis.
In order to estimate the number of clusters, Pseudo F statistics and Pseudo t-Squared were analyzed. Peaks or large values of Pseudo F indicate close-knit and separated clusters. Pseudo T square index quantifies the difference between two clusters that are merged at a given step. Thus if the pseudo T-square has a distinct jump at step k of the hierarchical clustering then the clustering in step k+1 is selected as the optimal cluster19.It is advisable to look for consensus among Cubic Clustering criterion (CCC), pseudo F and Pseudo Tsquare, but because of the few number of clusters solutions that were eligible for a CCC in our sample it couldn‟t be evaluated (number of clusters not greater than one-fifth of the numbers of observations in all 5 variables20). Because these criteria are appropriate only for compact or slightly elongated clusters, preferably clusters that are roughly multivariate normal20 and given that we had a priori the hypothesis that there would be two groups, more than three groups would not be required at this stage.
The obtained subgroups were later compared using two-way ANOVA models to determine the mean differences between clusters and control group stratified by age slices. Assumptions of normality were tested for all variables according to the Shapiro-Wilk test where small values led to the rejection of normality, while being close to 1 indicated normality of the data, so the values <0.90 may be considered small and lead to the rejection of normality21. Also, Kurtosis and Skewness statistics should be near 0 and ideally between -2 and 2 to be considered a normal distribution.22,23 Homogeneity of
14 variance was tested by evaluating the residual plots of each variable. For ANOVAs that yielded significant results at p =0.05 we verified whether the group age interaction was significant, if so, groups were compared separately for each age group, otherwise group comparisons were made overall age groups. Finally Cohen‟s d effect size (ES) for differences in cognitive performance was determined between clusters and control group.
Results
140 exposed children were recruited and matched for gender and age (more or less one year) with 140 non-exposed children. Due to missing values, nine participants were excluded from analysis including their matches and remained unclustered. The final sample was composed of 131 individuals in the HR group and 131 in the control group. Group comparisons of demographic characteristics between HR and control showed no significant difference for age or sex. The control group had a statistically significant higher socioeconomic status than the HR group, see Table 1.
Cluster Analysis
A cluster analysis was performed to separate the exposed group in terms of their cognitive functioning. The cluster analysis yielded a two-cluster solution for every age group, see Figure 1. Pseudo F statistics and Pseudo T square statistics generated by the cluster analysis revealed a demarcation point between 1 and 2 cluster solutions, suggesting that a 2 cluster solution best distinguished the cases. An inspection of the dendrogram confirmed that there was no overlap. In order to evaluate the sensitivity of the cluster solution with respect to the cluster algorithm used, we repeated the analysis using a different clustering procedure, the Average Hierarchical, which yielded two clusters similar to the solutions given by the Ward‟s method. Then we performed an agreement statistic obtaining kappa = 0.69, which could be interpreted as substantial degree of agreement24. These findings support the internal consistency of the analysis. According to the clinical characteristics of the variables, clusters were termed as HR1 if their scores in cognitive functioning were better than the other group, and the group who performed worse was renamed HR2.
15
Neurocognitive characteristics of the clusters
Assumptions of normality and homoscedasticity were met according to the criteria previously mentioned. The overall F test was significant (p <.0001) for all five cognitive domains and the interaction age group was significant for Processing Speed (p = 0.03), VEM (p = 0.007) and VISEM (p = 0.0005) see Table 2.
The means of the three cognitive domains that were significant for interaction are presented in Table 3a for each age group, while the means of the other two domains are presented in Table 3b. The differences between the HR1 and HR2 were statistically significant on Processing speed and VEM from 6 to 20 years old, the difference for VISEM and Working memory was significant in groups from 11 to 22 years old, and Executive functioning presented a significant difference between HR groups between 16 to 22 years old. On the other hand, the HR1 group performed similar to the control group in all functions within almost all age groups. Figure 2a shows the differences and effect sizes in Processing speed between HR2 and Control where all p values were significant for all age groups. Figure 2b presents the effect sizes in VEM, where also all p values were significant. VISEM comparison between HR2 and Control was described in Figure 2c: significant values were observed in ages between 11 and 24. Working memory and Executive functioning did not show a significant interaction between groups and ages, but significant differences between HR2 and Control were observed; in Working memory an overall effect size of -1.3 was found between HR2 and Control (p<0.05) (Figure 2 d), Executive functioning had an overall effect size of -0.6 (p<0.05) between HR2 and Control, (Figure 2 e).
Discussion
One of the most striking results from our study was to detect a subgroup of HR offspring that showed a cognitive performance almost identical to control subjects of the same age. Indeed, our study was done in two phases. First we performed a cluster analysis among HR offspring in this phase the control group was ignored. Then once the two subgroups HR1 and HR2 were identified a second phase was performed in order to compare each group to the control group, and it was then that we observed that HR1 performed very similar or almost exactly to control subjects. These two steps guaranteed that the resemblance between HR1 and the control group was not a statistical artefact.
16 Our statistical approach differed from previous studies that also attempted to split CHR subjects into a more or less vulnerable group. One of these studies25 used a likelihood function to classify each CHR subject according to having either a greater resemblance with a group of healthy control subjects or rather with SZ patients. Their results also reported two groups referred to as CHR-HC and CHR-SZ which showed average cognitive performances that were slightly worse and slightly better than those of healthy and SZ subjects respectively. The advantage of this study was the opportunity to observe a gradual decline in cognition from healthy to SZ patients passing through CHR-HC and CHR-SZ. However, the advantage of our approach based on cluster analysis in HR offspring was that the two groups of HR subjects could be identified without the need of a group of SZ patients and even without a healthy subjects group that became useful only as a comparison group after the splitting up of HR offspring group. Another attempt to divide HR individuals was approached by separating them into two groups20: one group of CHR according to the Structured Interview for Prodromal Syndromes (SIPS) criteria and another group of first degree relatives that did not fulfill the UHR criteria. Even though some authors also found some gradual worsening from controls to first degree relatives to CHR individuals, their results for both subgroups differed significantly from the healthy control group on most cognitive domains, with effect sizes that varied from 0.26 to 0.8.
In our study, when the cognitive performance of the HR2 subgroup was compared to that of controls, the corresponding ES were larger than the studies mentioned above, ranging from 0.2 to 2.3 according to the age group and cognitive domain. But it is important to note that both studies used neurocognitive tests that differed from our study. The larger effects sizes that were statistically significant between HR2 and controls in our study were found in individuals between 11 to 20 years old for processing speed, verbal and visual memory, whereas the differences in 6 to 10 year old subjects were significant only for processing speed and verbal memory.
Although the good news is that some of our HR offspring were very proximate to healthy subjects in cognitive domains, the counterpart is that the other subgroup performed worse than the CHR individuals reported in the literature9. A meta-analysis10 of 32 studies; concluded that CHR performed significantly worse in most cognitive domains and showed that CHR converters performed worse than converters; but the group of non-converters would be contaminated, because they were still at true risk of converting in the future. This could explain why their ES (Hedges g) were also smaller than ours ranging
17 from -0.2 to -0.86. Another strength of this study, apart from the statistical approach, was the power given our sample size of 131 individuals. To our knowledge, this is the first study to present neurocognitive assessments and comparisons for different age groups in the HR population.
Our study, however, does have some limitations. First, this was a cross-sectional study, this means that each individual belonging to the „susceptible‟ cluster was also part of a specific group of age, therefore we cannot determine if individuals will remain susceptible or not over time since this type of design cannot distinguish between incidence and natural history26,27, this is why there is a need of follow up studies. Another limitation is that the only inclusion criterion apart from age, was having a parent with a definite diagnostics of SZ or BP, without taking into consideration the number of sporadic or familial form of illness. This is important to mention because offspring with a very high familial/genetic loading tends to affect the cognitive impairment in this population6. We compared the proportion of HR having more than one first degree affected parent, between HR1 (48%) and HR2 (52%) and found no differences (X2 = 0.81, p=0.37) suggesting that the loading of the disorder does not explain our results of finding two distinctly neurological profiles.
Additionally, both schizophrenia and bipolar offspring showed evidence of neurocognitive differences in relation to the control group, although our results must be taken with caution since neurodevelopmental insults may have a greater impact in offspring of SZ individuals than in offspring of BP parents28. We compared the proportion of HR having a parent having SZ, BPD or both and found no differences (X2 = 4.58, p=0.10) with rates of 15.79%, 80.26%, 3.95% respectively in the HR1 group, and HR2 with rates of 27.27%, 67.27%, 5.45% respectively.
Family socioeconomic circumstance is a potential confounder because it‟s been described that the link between brain structure and cognitive processes vary by family socioeconomic circumstance29 and a higher socioeconomic position is associated with less psychiatric diagnoses as well as higher returns to work and lower rates of recurrence30. Since the risk of psychiatric disorders is strongly associated with socioeconomic status31 we verified if socioeconomic status could explain our findings of two clusters by comparing the mean scores among clusters and control groups. We found that the means did not differ between the two clusters (t = 0.89 p=0.37) and both HR1 and HR2 differed from the control group (t test =-2.47 p value=0.01 and t=-3.19 p =0.0016 respectively). Hence our results revealed a
18 subgroup HR1 of at risk children that showed very similar performance to control, despite having 6 points less than the latter ( the mean score of socio economic status in HR1 = 42.4 and in control group= 48.4). Therefore, the socioeconomic status is not a confounding variable in our study. On the other hand, because all individuals were Caucasians originally from Quebec, external validity or generalizability should be taken with caution.
Finally, some neurocognitive dysfunctions have been reported as possible predictors of the progression to psychosis12,32, and global functioning in early psychosis34. We could infer that there already exists a gap within the group of at risk patients that could represent a distinction between those who have less chance of developing a psychiatric disease and those who could be at higher risk.
Attention/Processing Speed and Executive Function at baseline may predict global functional outcome of early psychosis. These neurocognitive tests are easy to incorporate in clinical settings and, if replicated in independent samples, may be included in routine clinical assessments for prediction of functional outcome in early psychosis.
Conclusion
To conclude, within the group of HR offspring we detected two subgroups different from each other in terms of neuropsychological performance; one of them was very similar to control scores whereas the other group showed an important gap compared to control individuals. This last group performed even worse than the ones reported previously in the literature, which could be explained at least partially by our approach that aimed to disentangle the two cognitive susceptibility profiles. Still, further research is needed in longitudinal studies to investigate whether these findings are associated with the transition to a psychiatric disorder in the following years. Nevertheless our study suggests that interventions with a neurocognitive target should be addressed earlier, due to the apparition of a breach in performance even at early stages in life.
Aknowledgements:
This work was supported by the Canadian Institute of Health Reseach (CIHR). Authors have nothing else to declare.
19
References
1.- Gilbert E. Mérette C, Jomphe V, Émond C, Rouleau N, Bouchard R, Roy M, Paccalet T, Maziade M. Cluster analysis of cognitive deficits may mark heterogeneity in schizophrenia in terms of outcome and response to treatment. Eur Arch Psychiatry Clin Nurosci [Internet]. 2014; 264 (4):333 -343
2.- Mortiz S, Klein J, Desler T, Lill H, Gallinat J. Neurocognitive deficits in schizophrenia. Are we making mountains out of molehills?. Psychological Medicine.2017:1-11
3.- Fatouros-Bergman H, Cervenka S, Flyckt L, Edman G, Farde L. Meta-analysis of cognitive performance in drug-naïve patients with schizophrenia. Schizophrenia Research.2014;158:156-162
4.- Bora E, Yücel M, Pantelis C, Berk M. Meta-analytic review of neurocognition in bipolar II disorder. Acta Psychiatrica Scandinavica. 2011;123(3):165–74.
5.- Bora E, Pantelis C. Meta-analysis of cognitive impairment in first-episode bipolar disorder: comparision with first-episode schizophrenia and healthy controls. Schizophrenia Bulletin.2015;41(5):1095-1104
6.- Maziade M, Rouleau N, Gingras N, Boutin J, Paradis M, Jomphe V, et al. Shared Neurocognitive Dysfunctions in Young Offspring at Extreme Risk for Schizophrenia or Bipolar Disorder in Eastern Quebec Multigenerational Families. Schizophrenia Bulletin. 2008; 35 (5) :919-930
7.- Maziade M, Rouleau N, Mérette C, Cellard C, Battaglia M, et al. Verbal and Visual Memory Impairments Among Young Offspring and Healthy Adult Relatives of Patients whith schizophrenia and bipolar disorder : Selective generational patterns indicate different developmental trajectoires. Schizophrenia Bulletin. 2010;37(6) :1218-28
8.- Maziade M, Rouleau N, Cellard C, Battaglia M, Paccalet T, et al. Young offspring at genetic risk of adult psychosis : The form of the trajectory of IQ or memory may orient to the right dysfunction at the right time. 2011;6(4):e19153
9.- Üçok A, Direk N, Koyuncu A, Keskin-Ergen Y, Yüksel Ç, et al. Cognitive deficits and familial high risk groups for psychosis are common as in first episode schizophrenia. Schizophrenia Research. 2013;151:265-269
10.- Hauser M, Zhang J, Sheridan E, Burdick K, Mogil R et al. neurospcyhological test performance to enhance identification of subjects at clinical high risk for psychosis and to be most promising for predictive algorithms for conversion to psychosis: A Meta-analysis.2017. JClinPsychiatry.2017:78(1):28-40
11.- de la Serna E, Sugranyes G, Sanchez-Gistau V, Rodriguez-Toscano E, Baeza I, et al. Neuropsychological characteristics of child and adolescent offspring of patients with schizophrenia or bipolar disorder. Schizophrenia Research. 2017;183:110-115
12.- Fusar-Poli P, Bonoldi I, Yung A, Borgwardt S, Kempton M, et al. Predicting Psychosis. Meta-analysis of transition outcomes in individuals at high clinical risk. Arch Gen Psychiatry.2012;69 (3):220-229
13.- Kempton M,Bonoldi I, Valmaggia L, McGuirre P, Fusar-Poli P.Speed of psychosis progression in people at ultra-high clinical risk: a complementary meta-analysis. JAMA Psychiatry. 2015;72 (6):622-23
14.- Hou C, Xiang Y, Wang Z, Everall I, Tang Y et al. Cognitive functioning in individuals at ultra-high risk for psychosis, first-degree relatives of patients with psychosis and patients with first-episode schizophrenia. Schizophrenia Research.2016;174:71-76
15.- Fusar-Poli P, Meneghelli A, Valmaggia L, Allen P, Galvan F et al. Duration of untreated prodromal symptoms and 12-month functional outcome of individuals at risk of psychosis.BJPsych. 2009; 194:181-182
20 16.- Rasic D, Hajek T, Aolda M, Uher R. Risk of mental illness in offspring of parents with schizophrenia, bipolar disorder, and major depressive disorder: A meta-analysis of Family High-Risk Studies. Schizophrenia bulletin. 2014; 40(1): 28-38
17.- 1.- Tan P, Steinbach M, Kumar V, Introduction to Data mining. Chap 8 : Cluster Analysis: Basic Concepts and Algorithms. Addison-Wesley Companion Book site. [Internet]. 2006. [cited 2017 Jul 12]; Available in:
http://www-users.cs.umn.edu/~kumar/dmbook/index.php
18.- Murtagh F, Legendre P. Ward‟s Hierarchical Agglomerative Clustering Method: Which Algoritms Implement Ward‟s Criterion?. Journal of Classification. 2014. 31:274-295
19.- Wilkinson L, Engelman L, Corter J. Coward M, Cluster Analysis Chapter 4 Systat Manuals. [Internet]. Available in:
https://www.kellogg.northwestern.edu/rs/software/systat/systat_manuals.aspx
20.- SAS Institute Inc. 2010. SAS/STAT® 9.22 User‟s Guide. Cary, NC: SAS Institute Inc 21.- Peng G, Lilly E, Indianapolis Company. Testing Normality of Data using SAS®. Paper PO04. Available in :
http://lexjansen.com/pharmasug/2004/posters/po04.pdf
22.- SAS Institute Inc. 2014. Base SAS® 9.4 Procedures Guide: Statistical Procedures, Third Edition. Cary, NC: SAS Institute Inc.
23.- George, D. & Mallery, M. (2010). Using SPSS for Windows step by step: a simple guide and reference. Boston, MA: Allyn & Bacon.
24.- Landis, J.R. and Koch G.G. The measurement of observer agreement for categorical data. Biometrics. 1977; 33: 159-174.
25.- Choi S, Kyeong S, Cho K, Yun J, Lee T, et al. Brain network characteristics separating individuals at clinical high risk for psychosis into normality or psychosis. Schizophrenia Research. 2017. [Internet]. Available in:
http://dx.doi.org/10.1016/j.schres.2017.03.028
26.- Rothman K, Greenland S, Lash T. Modern Epidemiology 3rd Edition. Timothy L. Lash Lippincott Williams & Wilkins. 2008
27.- Sedgwick P. Cross sectional studies: advantages and disadvantages. BMJ.2014;348:g2276
28.- Sugranyes G, de la Serna E, Borras R, Sanchez-Gistau V, Pariente J. Clinical, Cognitive, and Neuroimaging Evidence of a Neurodevelopmental Continuum in Offspring of Probands With Schizophrenia and Bipolar Disorder. Schizophrenia Bulletin. 2017; doi:10.1093/schbul/sbx002
29.- Brito, N, Piccolo L, Noble K. Brain and Cognition. 2017, http://dx.doi.org/10.1016/j.bandc.2017.03.007
30.-37.- Virtanen M, Kawachi I, Oksanen T, Salo P, Tuisku K et al. Socio-economic differences in long-term psychiatric work disability : prospective cohort study of onset, recovery and recurrence. Occup Environ Med 2011;68:791e798
31.- 36.- Agerbo E, Sullivan P, Vilhjálmsson B, Pedersen C, Mors O, et al. Polygenic risk score, parental socioeconomic status, family history of psychiatric disorders, and the risk for schizophrenia. A Danish population-based study and Meta-analysis. JAMA Psychiatry. 2015;72(7):635-41.
32.- Cannon T, Yu C, Addington J, Bearden C, Cadenhead K et al. An individualized risk calculator for research in prodromal psychosis. Am J Psychiatry. 2016;173 (10): 980-88 34.- Sawada K, Kanehara A, Sakakibara E, Eguchi S, Tada M, et al. Identifying neurocognitive markers for outcome prediction of global functioning in ultra-high-risk for psychosis and first episode psychosis. Psychiatry and Clinical Neurosciences. 2017; 71: 318–327
21
Annexes
Table 1. Socio-demographic characteristics of the sample
Characteristics HR group (N=131) Control group (N=131) Mean (SD) or n (%) Range Mean (SD) or n(%) Range t(p value) Age 16.17 (4.96) 6.13 - 24.49 16.27 (4.85) 6.50 - 24.38 0.18 (0.85) Gender (female) 65 (49.62) 65 (49.62) Socioeconomic statusa 41.32 (15.57) 22.08 - 71.62 48.93 (16.95) 22.08 - 78.34 3.69 (<0.05)
a: According to the Family Blishen index (13) according to the highest socioeconomic status of the two parents. This index is based on education and income and on a Canadian census of 514 occupational categories according to the Canadian Classification and Dictionary of Occupations.
22
Figure 1.- Cluster solution. Dendrograms by age.
Note: The cluster analysis was performed according to the Ward‟s method; it was based on the Euclidean
distance over all of the five neurocognitive measures: Processing speed, VEM, VISEM, Working memory and Executive functioning. This analysis permitted us to find two clusters after which we decided to name HR1 the cluster that performed better and HR2 the cluster that performed worse of the five variables.
23
Table 2. Comparisons of neurocognitive domains between Deficient HR, Healthy HR and Control group.
Neurocognitive domains Control vs HR 1 vs HR 2
Interaction a DF F P value F P value Processing speed 11 8.59 <.0001b 2.40 0.0287c VEM 11 13.50 <.0001b 3.02 0.0072c VISEM 11 7.91 <.0001b 4.14 0.0005c Working memory 11 7.99 <.0001b 1.74 0.1116 Executive functioning 11 5.36 <.0001b 0.70 0.6510 a: Interaction cluster * age group
b: This is the p value for the overall F test of the two way ANOVA.
24
Table 3 a. Neurocognitive function means of HR1, HR2, Control group and comparisons among groups of age
6 to 10 years old 11 to 15 years old 16 to 20 years old 21 to 24 years old HR1 N=11 HR2 N=6 Control N=17 p value HR1 N=30 HR2 N=9 Control N=39 p value HR1 N=23 HR2 N=15 Control N=37 p value HR1 N=12 HR2 N=25 Control N=37 p value Processing speed Mean (SD) 64.58 (23.94) 17.93 (13.43) 60.58 (22.29) 57.39 (18.18) 17.67 (12.84) 49.26 (25.63) 48.99 (17.03) 18.51 (12.92) 50.52 (24.11) 55.39 (16.74) 45.32 (19.67) 60.16 (20.98) HR1 vs HR2 <.05 <.05 <.05 0.17 HR1 vs control 0.62 0.11 0.78 0.49 HR2 vs control <.05 <.05 <.05 <.05 VEM Mean (SD) 72.55 (14.64) 35.58 (17.19) 79.26 (13.75) 74.70 (18.55) 27.91 (20.54) 74.85 (15.08) 64.30 (19.25) 25.37 (19.53) 71.38 (22.68) 64.00 (26.37) 53.38 (27.89) 75.60 (20.27) HR1 vs HR2 <.05 <.05 <.05 0.14 HR1 vs control 0.39 0.97 0.18 0.09 HR2 vs control <.05 <.05 <.05 <.05 VISEM Mean (SD) 21.86 (19.56) 30.33 (21.84) 36.06 (24.59) 35.13 (24.37) 15.50 (20.39) 41.91 (28.45) 41.87 (30.44) 13.27 (14.01) 53.46 (31.68) 76.38 (12.22) 15.74 (15.86) 48.16 (33.24) HR1 vs HR2 0.53 0.05 <.05 <.05 HR1 vs control 0.17 0.29 0.10 <.05 HR2 vs control 0.65 <.05 <.05 <.05
Table 3b. Neurocognitive function means of HR1, HR2, Control group and comparisons among overall age groups
HR1 HR2 Control P value Working memory Mean (IC) 55.6 (50.9 to 60.2) 31.3 (25.79 to 3676) 57.52 (53.97 to 60.05) HR1 vs HR2 <.05 HR1 vs control 0.50 HR2 vs control <.05 Executive functioning Mean (IC) 58.28 (53.33 to 63.23) 46.39 (40.54 to 52.25) 60.34 (56.56 to 64.11) HR1 vs HR2 0.02 HR1 vs control 0.51 HR2 vs control <.05
25
Figure 2 a. Processing speed values by age and by group.
26
Figure 2 c. Visual Memory values by age and by group.
27
28
CHAPTER 3: Article 2
Non-pharmacological interventions in individuals at risk of psychosis: a systematic review and meta-analysis of randomized controlled trials
Rossana Peredo, Michel Maziade, Elsa Gilbert, Geneviève Picher, Kaoutar Ennour-Idrissi, Chantal Mérette
Résumé.-
Introduction : L'identification précoce et le traitement des personnes à risque de psychose peuvent retarder la transition et prévenir les effets néfastes sur le fonctionnement global. Des études antérieures ont analysé la thérapie comportementale et cognitive (TCC) comme un possible traitement efficace. Cependant, récemment, certaines études ont signalé la possibilité d'une augmentation des faux positifs qui non seulement dilue les taux de transition, mais peut également représenter une menace s'ils sont exposés à des traitements psychosociaux qui peuvent causer la stigmatisation et augmenter le risque de transition. Cette étude examine l'effet des interventions non pharmacologiques sur la transition et évalue leur effet sur les comorbidités non psychotiques. Méthodes: La recherche a été faite en février 2017, dans le Registre central des essais contrôlés de Cochrane Central, MEDLINE et EMBASE. Les critères d'inclusion étaient : interventions non pharmacologiques comparées à n‟importe quel comparateur. Deux examinateurs indépendants ont extrait les données et évalué la qualité des essais; les désaccords ont été résolus par un troisième examinateur. Peto odds ratios ont été calculés pour des données dichotomiques et des modèles à effet aléatoire ont été utilisés pour l'analyse des résultats continus. Des analyses de sous-groupes et de sensibilité ont été effectuées. Résultats: 796 études ont été inclus, impliquant 969 participants, avec un âge moyen de 20,81 (SD = 5,38). À 6 mois de suivi, les interventions non pharmacologiques ont entraîné une incidence de transition plus faible, mais avec une hétérogénéité substantielle. À 12 mois de suivi, la BCT a également montré une incidence significativement plus faible, avec une hétérogénéité élevée. Une analyse de sensibilité a été faite avec une seule étude qui avais des résultats aberrants, les résultats n‟ont pas changé, mais l'hétérogénéité a chuté
à 0% dans les deux cas. Il n'y avait pas de résultats statistiquement significatifs dans les comorbidités non psychotiques. Conclusions: Les thérapies non pharmacologiques, en particulier la CBT, ont été associées à un risque réduit de transition vers la psychose. Cette intervention peut également être utilisée pour le traitement de la psychose. D'autres recherches sont nécessaires pour minimiser les taux de retrait.
