Structural brain abnormalities and aggressive behaviour in schizophrenia: Mega-analysis of data from 2095 patients and 2861 healthy controls via the ENIGMA consortium.

Background: Schizophrenia is associated with an increased risk of aggressive behaviour, which may partly be explained by illness-related changes in brain structure. However, previous studies have been limited by group-level analyses, small and selective samples of inpatients and long time lags between exposure and outcome. Methods: This cross-sectional study pooled data from 20 sites participating in the international ENIGMA-Schizophrenia Working Group. Sites acquired T1-weighted and diffusion-weighted magnetic resonance imaging scans in a total of 2095 patients with schizophrenia and 2861 healthy controls. Measures of grey matter volume and white matter microstructural integrity were extracted from the scans using harmonised protocols. For each measure, normative modelling was used to calculate how much patients deviated (in z-scores) from healthy controls at the individual level. Ordinal regression models were used to estimate the associations of these deviations with concurrent aggressive behaviour (as odds ratios [ORs] with 99% confidence intervals [CIs]). Mediation analyses were performed for positive symptoms (i.e., delusions, hallucinations and disorganised thinking), impulse control and illness insight. Aggression and potential mediators were assessed with the Positive and Negative Syndrome Scale, Scale for the Assessment of Positive Symptoms or Brief Psychiatric Rating Scale. Results: Aggressive behaviour was significantly associated with reductions in total cortical volume (OR [99% CI] = 0.88 [0.78, 0.98], p = .003) and global white matter integrity (OR [99% CI] = 0.72 [0.59, 0.88], p = 3.50 × 10-5) and additional reductions in dorsolateral prefrontal cortex volume (OR [99% CI] = 0.85 [0.74, 0.97], p =.002), inferior parietal lobule volume (OR [99% CI] = 0.76 [0.66, 0.87], p = 2.20 × 10-7) and internal capsule integrity (OR [99% CI] = 0.76 [0.63, 0.92], p = 2.90 × 10-4). Except for inferior parietal lobule volume, these associations were largely mediated by increased severity of positive symptoms and reduced impulse control. Conclusions: This study provides evidence that the co-occurrence of positive symptoms, poor impulse control and aggressive behaviour in schizophrenia has a neurobiological basis, which may inform the development of therapeutic interventions.

Validation and recalibration of OxMIV in predicting violent behaviour in patients with schizophrenia spectrum disorders.

Oxford Mental Illness and Violence (OxMIV) addresses the need in mental health services for a scalable, transparent and valid tool to predict violent behaviour in patients with severe mental illness. However, external validations are lacking. Therefore, we have used a Dutch sample of general psychiatric patients with schizophrenia spectrum disorders (N = 637) to evaluate the performance of OxMIV in predicting interpersonal violence over 3 years. The predictors and outcome were measured with standardized instruments and multiple sources of information. Patients were mostly male (n = 493, 77%) and, on average, 27 (SD = 7) years old. The outcome rate was 9% (n = 59). Discrimination, as measured by the area under the curve, was moderate at 0.67 (95% confidence interval 0.61-0.73). Calibration-in-the-large was adequate, with a ratio between predicted and observed events of 1.2 and a Brier score of 0.09. At the individual level, risks were systematically underestimated in the original model, which was remedied by recalibrating the intercept and slope of the model. Probability scores generated by the recalibrated model can be used as an adjunct to clinical decision-making in Dutch mental health services.

Use of illicit substances and violent behaviour in psychotic disorders: two nationwide case-control studies and meta-analyses.

BACKGROUND: Substance use disorder explains much of the excess risk of violent behaviour in psychotic disorders. However, it is unclear to what extent the pharmacological properties and subthreshold use of illicit substances are associated with violence. METHODS: Individuals with psychotic disorders were recruited for two nationwide projects: GROUP (N = 871) in the Netherlands and NEDEN (N = 921) in the United Kingdom. Substance use and violent behaviour were assessed with standardized instruments and multiple sources of information. First, we used logistic regression models to estimate the associations of daily and nondaily use with violence for cannabis, stimulants, depressants and hallucinogens in the GROUP and NEDEN samples separately. Adjustments were made for age, sex and educational level. We then combined the results in random-effects meta-analyses. RESULTS: Daily use, compared with nondaily or no use, and nondaily use, compared with no use, increased the pooled odds of violence in people with psychotic disorders for all substance categories. The increases were significant for daily use of cannabis [pooled odds ratio (pOR) 1.6, 95% confidence interval (CI) 1.2-2.0), stimulants (pOR 2.8, 95% CI 1.7-4.5) and depressants (pOR 2.2, 95% CI 1.1-4.5), and nondaily use of stimulants (pOR 1.6, 95% CI 1.2-2.0) and hallucinogens (pOR 1.5, 95% CI 1.1-2.1). Daily use of hallucinogens, which could only be analysed in the NEDEN sample, significantly increased the risk of violence (adjusted odds ratio 3.3, 95% CI 1.2-9.3). CONCLUSIONS: Strategies to prevent violent behaviour in psychotic disorders should target any substance use.

Cognition and violent behavior in psychotic disorders: A nationwide case-control study.

BACKGROUND: The excess risk of violence in psychotic disorders may partly be explained by impairments in executive functions (EFs) and theory of mind (ToM). However, previous studies have been limited by composite measures of EFs and small samples of inpatients. METHODS: Data were collected for the research project Genetic Risk and Outcome of Psychosis (GROUP). Patients with psychotic disorders (N = 891) were recruited from various care settings in the Netherlands. The following neuropsychological tests were administered (targeted cognitive function in parentheses): (i) Continuous Performance Test-HQ (inhibition); (ii) Response Shifting Task (cognitive flexibility); (iii) Wechsler Adult Intelligence Scale, Third Edition (WAIS-III) Block Design subtest (fluid intelligence); (iv) Neuropsychological Assessment Battery (NAB) Mazes Test (planning); (v) Degraded Facial Affect Recognition Task (affective ToM); and (vi) Hinting Task (cognitive ToM). Lifetime violence was ascertained from medical records and patient interviews. We used analysis of covariance to compare the mean scores of violent and nonviolent patients on each test, adjusting for age and sex. RESULTS: Violent patients performed significantly worse than nonviolent patients on the WAIS-III Block Design subtest (F [1, 847] = 5.12, p = .024), NAB Mazes Test (F [1, 499] = 5.32, p = .022) and Hinting Task (F [1, 839] = 9.38, p = .002). For the other tests, the between-group differences were nonsignificant. Violent behavior explained no more than 1% of the variance in performance on each test. CONCLUSION: Impairments in EFs and ToM are unlikely to provide useful targets for risk assessment and interventions.

Structural brain correlates of interpersonal violence: Systematic review and voxel-based meta-analysis of neuroimaging studies.

Owing to inconsistent nomenclature and results, we have undertaken a label-based review and anatomical likelihood estimation (ALE) meta-analysis of studies measuring the quantitative association between regional grey matter (GM) volume and interpersonal violence. Following PRISMA guidelines, we identified studies by searching 3 online databases (Embase, Medline, PsycInfo) and reference lists. Thirty-five studies were included in the label-based review, providing information for 1288 participants and 86 brain regions. Per region, 0-57% of the results indicated significant reductions in GM volume, while 0-23% indicated significant increases. The only region for which more than half of all results indicated significant reductions was the parietal lobe. However, these results were dispersed across subregions. The ALE meta-analysis, which included 6 whole-brain voxel-based morphometry studies totaling 278 participants and reporting 144 foci, showed no significant clusters of reduced GM volume. No material differences were observed when excluding experiments using reactive violence as outcome or subjects diagnosed with psychopathy. Possible explanations for these findings are phenomenological and etiological heterogeneity, and insufficient power in the label-based review and ALE meta-analysis to detect small effects. We recommend that future studies distinguish between subtypes of interpersonal violence, and investigate mediation by underlying emotional and cognitive processes.