Development of multivariable models to predict change in Body Mass Index within a clinical trial population of psychotic individuals

Weight change following diagnosis with psychosis: a retrospective cohort study in Greater Manchester, UK

INTRODUCTION

Weight gain in the months/years after diagnosis/treatment of severe enduring mental illness (SMI) is a major predictor of future diabetes, dysmetabolic profile and increased risk of cardiometabolic diseases. There is limited data on the longer-term profile of weight change in people with a history of SMI and how this may differ between individuals. We here report a retrospective study on weight change over the 5 years following an SMI diagnosis in Greater Manchester UK, an ethnically and culturally diverse community, with particular focus on comparing non-affective psychosis (NAP) vs affective psychosis (AP) diagnoses.

METHODS

We undertook an anonymised search in the Greater Manchester Care Record (GMCR). We reviewed the health records of anyone who had been diagnosed for the first time with first episode psychosis, schizophrenia, schizoaffective disorder, delusional disorder (non-affective psychosis = NAP) or affective psychosis (AP). We analysed body mass index (BMI) change in the 5-year period following the first prescription of antipsychotic medication. All individuals had taken an antipsychotic agent for at least 3 months. The 5-year follow-up point was anywhere between 2003 and 2023.

RESULTS

We identified 9125 people with the diagnoses above. NAP (n = 5618; 37.3% female) mean age 49.9 years; AP (n = 4131; 60.5% female) mean age 48.7 years. 27.0% of NAP were of non-White ethnicity vs 17.8% of AP individuals. A higher proportion of people diagnosed with NAP were in the highest quintile of social disadvantage 52.4% vs 39.5% for AP. There were no significant differences in baseline BMI profile. In a subsample with HbA1c data (n = 2103), mean HbA1c was higher in NAP at baseline (40.4 mmol/mol in NAP vs 36.7 mmol/mol for AP). At 5-year follow-up, there was similarity in both the overall % of individuals in the obese ≥ 30 kg/m2 category (39.8% NAP vs 39.7% AP), and % progressing from a normal healthy BMI transitioned to obese/overweight BMI (53.6% of NAP vs 55.6% with AP). 43.7% of those NAP with normal BMI remained at a healthy BMI vs 42.7% with AP. At 5-year follow-up for NAP, 83.1% of those with BMI ≥ 30 kg/m2 stayed in this category vs 81.5% of AP.

CONCLUSION

The results of this real-world longitudinal cohort study suggest that the changes in BMI with treatment of non-affective psychosis vs bipolar disorder are not significantly different, while 43% maintain a healthy weight in the first 5 years following antipsychotic prescription.

First episode psychosis and weight gain a longitudinal perspective in Cheshire UK: a comparison between individuals with nonaffective versus affective psychosis

Early weight gain following initiation of antipsychotic treatment predicts longer-term weight gain, with attendant long-term consequences including premature cardiovascular events/death. An important question is whether there is a difference in weight change over time between people with affective versus nonaffective psychosis. Here we describe the results of a real-world analysis of the BMI change in the months postdiagnosis with affective versus nonaffective psychosis.

Methods

We undertook an anonymised search across one Primary Care Network in Cheshire, UK with a total population of 32 301 individuals. We reviewed the health records of anyone who had been diagnosed over a 10-year period between June 2012 and June 2022 for the first time with first episode nonaffective psychosis versus psychosis associated with depression or bipolar affective disorder (affective psychosis).

Results

The overall % change in BMI was +8% in nonaffective psychosis individuals and +4% in those with a diagnosis of affective psychosis - however, the distribution was markedly skewed for nonaffective psychosis patients. Using caseness as >30% increase in BMI; affective = 4% cases and nonaffective = 13% cases, there was a three-fold difference in terms of increase in BMI. In regression analysis, the r² linking the initial BMI to % change in BMI was 0.13 for nonaffective psychosis and 0.14 for affective psychosis.

Conclusion

The differences observed here in the distribution of weight change over time between individuals with affective versus nonaffective psychosis may relate to underlying constitutional differences. The phenotypic and genetic factors underlying this difference remain to be defined.

The genetics of bipolar disorder with obesity and type 2 diabetes

BACKGROUND

Bipolar disorder (BD) presents with high obesity and type 2 diabetes (T2D) and pathophysiological and phenomenological abnormalities shared with cardiometabolic disorders. Genomic studies may help define if they share genetic liability. This selective review of BD with obesity and T2D will focus on genomic studies, stress their current limitations and guide future steps in developing the field.

METHODS

We searched electronic databases (PubMed, Scopus) until December 2021 to identify genome-wide association studies, polygenic risk score analyses, and functional genomics of BD accounting for body mass index (BMI), obesity, or T2D.

RESULTS

The first genome-wide association studies (GWAS) of BD accounting for obesity found a promising genome-wide association in an intronic gene variant of TCF7L2 that was further replicated. Polygenic risk scores of obesity and T2D have also been associated with BD, yet, no genetic correlations have been demonstrated. Finally, human-induced stem cell studies of the intronic variant in TCF7L2 show a potential biological impact of the products of this genetic variant in BD risk.

LIMITATIONS

The narrative nature of this review.

CONCLUSIONS

Findings from BD GWAS accounting for obesity and their functional testing, have prompted potential biological insights. Yet, BD, obesity, and T2D display high phenotypic, genetic, and population-related heterogeneity, limiting our ability to detect genetic associations. Further studies should refine cardiometabolic phenotypes, test gene-environmental interactions and add population diversity.

The COPILOT Raw Illumina Genotyping QC Protocol

The Illumina genotyping microarrays generate data in image format, which is processed by the platform-specific software GenomeStudio, followed by an array of complex bioinformatics analyses that rely on various software, different programming languages, and numerous dependencies to be installed and configured correctly. The entire process can be time-consuming, can lead to reproducibility errors, and can be a daunting task for bioinformaticians. To address this, we introduce the COPILOT protocol, which has been successfully used to transform raw Illumina genotype intensity data into high-quality analysis-ready data on tens of thousands of human patient samples that have been genotyped on a variety of Illumina genotyping arrays. This includes processing both mainstream and custom content genotyping chips with over 4 million markers per sample. The COPILOT QC protocol consists of two distinct tandem procedures to process raw Illumina genotyping data. The first protocol is an up-to-date process to systematically QC raw Illumina microarray genotyping data using the Illumina-specific GenomeStudio software. The second protocol takes the output from the first protocol and further processes the data through the COPILOT (Containerised wOrkflow for Processing ILlumina genOtyping daTa) containerized QC pipeline, to automate an array of complex bioinformatics analyses to improve data quality through a secondary clustering algorithm and to automatically identify typical Genome-Wide Association Study (GWAS) data issues, including gender discrepancies, heterozygosity outliers, related individuals, and population outliers, through ancestry estimation. The data is returned to the user in analysis-ready PLINK binary format and is accompanied by a comprehensive and interactive HTML summary report file which quickly helps the user understand the data and guides the user for further data analyses. The COPILOT protocol and containerized pipeline are also available at https://khp-informatics.github.io/COPILOT/index.html. © 2022 The Authors. Current Protocols published by Wiley Periodicals LLC. Basic Protocol 1: Processing raw Illumina genotyping data using GenomeStudio Basic Protocol 2: COPILOT: A containerised workflow for processing Illumina genotyping data.

Previous exposure to antipsychotic drug treatment is an effective predictor of metabolic disturbances experienced with current antipsychotic drug treatments

BACKGROUND

Antipsychotic drugs are associated with adverse events, but serious side effects are not frequent. This study aimed to ascertain whether previous exposure to antipsychotic treatment was associated with metabolic disturbances induced by current antipsychotic medication.

METHODS

A total of 115 antipsychotic-naïve patients, 65 patients with previous exposure to low-metabolic-risk antipsychotics, and 88 patients with previous exposure to high-metabolic-risk antipsychotics were enrolled in our case-control study. All patients were administered olanzapine. Body weight, body mass index (BMI), biochemical indicators of blood glucose and lipids, the proportion of patients who gained more than 7% of their body weight at baseline, and the percentage of dyslipidemia were evaluated. All assessments were conducted at baseline and at 4 and 6 weeks after treatment.

RESULTS

Olanzapine treatment resulted in a significant increase in body weight and BMI in antipsychotic-naïve patients compared with the other two groups (both p < 0.05). However, increases in lipid levels in the high-metabolic-risk antipsychotics group were significantly higher than that in the other two groups (both p < 0.05). A history of antipsychotics use was not associated with weight gain (all p > 0.05). Higher low-density lipoprotein cholesterol ≥3.37 mmol/L^-1 was observed in antipsychotics exposure group compared with no history of antipsychotics exposure (aOR, 1.75; 95% CI, 1.07-3.52). Particularly, a history of high-metabolic-risk antipsychotics use was associated with a higher risk of LDL-C ≥3.37 mmol/L-1(aOR, 2.18; 95% CI, 1.03-3.32) compare with other two groups.

CONCLUSIONS

A history of exposure to antipsychotics, particularly high-metabolic-risk antipsychotics, is associated with current antipsychotic-induced metabolic disturbances.

Key use cases for artificial intelligence to reduce the frequency of adverse drug events: a scoping review

Adverse drug events (ADEs) represent one of the most prevalent types of health-care-related harm, and there is substantial room for improvement in the way that they are currently predicted and detected. We conducted a scoping review to identify key use cases in which artificial intelligence (AI) could be leveraged to reduce the frequency of ADEs. We focused on modern machine learning techniques and natural language processing. 78 articles were included in the scoping review. Studies were heterogeneous and applied various AI techniques covering a wide range of medications and ADEs. We identified several key use cases in which AI could contribute to reducing the frequency and consequences of ADEs, through prediction to prevent ADEs and early detection to mitigate the effects. Most studies (73 [94%] of 78) assessed technical algorithm performance, and few studies evaluated the use of AI in clinical settings. Most articles (58 [74%] of 78) were published within the past 5 years, highlighting an emerging area of study. Availability of new types of data, such as genetic information, and access to unstructured clinical notes might further advance the field.

Accurate neuroimaging biomarkers to predict body mass index in adolescents: a longitudinal study

Obesity is often associated with cardiovascular complications. Adolescent obesity is a risk factor for cardiovascular disease in adulthood; thus, intensive management is warranted in adolescence. The brain state contributes to the development of obesity in addition to metabolic conditions, and hence neuroimaging is an important tool for accurately assessing an individual's risk of developing obesity. Here, we aimed to predict body mass index (BMI) progression in adolescents with neuroimaging features using machine learning approaches. From an open database, we adopted 76 resting-state functional magnetic resonance imaging (rs-fMRI) datasets from adolescents with longitudinal BMI scores. Functional connectivity analyses were performed on cortical surfaces and subcortical volumes. We identified baseline functional connectivity features in the prefrontal-, posterior cingulate-, sensorimotor-, and inferior parietal-cortices as significant determinants of BMI changes. A BMI prediction model based on the identified fMRI biomarkers exhibited a high accuracy (intra-class correlation = 0.98) in predicting BMI at the second visit (1~2 years later). The identified brain regions were significantly correlated with the eating disorder-, anxiety-, and depression-related scores. Based on these results, we concluded that these functional connectivity features in brain regions related to eating disorders and emotional processing could be important neuroimaging biomarkers for predicting BMI progression.

Evidence that genes involved in hedgehog signaling are associated with both bipolar disorder and high BMI

Patients with bipolar disorder (BD) show higher frequency of obesity and type 2 diabetes (T2D), but the underlying genetic determinants and molecular pathways are not well studied. Using large publicly available datasets, we (1) conducted a gene-based analysis using MAGMA to identify genes associated with BD and body mass index (BMI) or T2D and investigated their functional enrichment; and (2) performed two meta-analyses between BD and BMI, as well as BD and T2D using Metasoft. Target druggability was assessed using the Drug Gene Interaction Database (DGIdb). We identified 518 and 390 genes significantly associated with BD and BMI or BD and T2D, respectively. A total of 52 and 12 genes, respectively, were significant after multiple testing correction. Pathway analyses conducted on nominally significant targets showed that genes associated with BD and BMI were enriched for the Neuronal cell body Gene Ontology (GO) term (p = 1.0E-04; false discovery rate (FDR) = 0.025) and different pathways, including the Signaling by Hedgehog pathway (p = 4.8E-05, FDR = 0.02), while genes associated with BD and T2D showed no specific enrichment. The meta-analysis between BD and BMI identified 64 relevant single nucleotide polymorphisms (SNPs). While the majority of these were located in intergenic regions or in a locus on chromosome 16 near and in the NPIPL1 and SH2B1 genes (best SNP: rs4788101, p = 2.1E-24), five were located in the ETV5 gene (best SNP: rs1516725, p = 1E-24), which was previously associated with both BD and obesity, and one in the RPGRIP1L gene (rs1477199, p = 5.7E-09), which was also included in the Signaling by Hedgehog pathway. The meta-analysis between BD and T2D identified six significant SNPs, three of which were located in ALAS1 (best SNP: rs352165, p = 3.4E-08). Thirteen SNPs associated with BD and BMI, and one with BD and T2D, were located in genes which are part of the druggable genome. Our results support the hypothesis of shared genetic determinants between BD and BMI and point to genes involved in Hedgehog signaling as promising targets.

Antipsychotic-induced weight gain and birth weight in psychosis: A fetal programming model

Antipsychotic induced weight gain is a frequent reason for treatment discontinuation in psychosis, subsequently increasing the risk of relapse and negatively affecting patient well-being. The metabolic effect of weight gain and the subsequent risk of obesity constitute a major medical problem on the long term. Despite its consequences, to date few risk factors have been identified (age, gender, body mass index at baseline), with some authors suggesting the implication of early life stressful events, such as perinatal conditions. We aim to describe if a surrogate marker of intrauterine environment (birth weight) might predict weight gain in a cohort of 23 antipsychotic naïve patients at the onset of the psychotic disease evaluated during 16 weeks with olanzapine treatment and in another cohort of 24 psychosis-resistant patients initiating clozapine assessed for 18 weeks. Two independent linear mixed model analyses were performed in each cohort of patients, with prospective weight gain as the dependent variable, age, gender, body mass index, duration of treatment and time as independent variables. Only in naïve patients, weight gain due to antipsychotics was significantly associated with birth weight, while male gender and body mass index at baseline were associated in both cohorts of patients. Treatment-resistant psychotic patients under clozapine were older, had previous antipsychotic treatment and more years of disease, confounders that might have influence a non significant association. Our results suggest that early environmental events might be playing a role in weight evolution in naïve patients treated with antipsychotics.

Selecting Multiple Biomarker Subsets with Similarly Effective Binary Classification Performances

Biomarker detection is one of the more important biomedical questions for high-throughput 'omics' researchers, and almost all existing biomarker detection algorithms generate one biomarker subset with the optimized performance measurement for a given dataset. However, a recent study demonstrated the existence of multiple biomarker subsets with similarly effective or even identical classification performances. This protocol presents a simple and straightforward methodology for detecting biomarker subsets with binary classification performances, better than a user-defined cutoff. The protocol consists of data preparation and loading, baseline information summarization, parameter tuning, biomarker screening, result visualization and interpretation, biomarker gene annotations, and result and visualization exportation at publication quality. The proposed biomarker screening strategy is intuitive and demonstrates a general rule for developing biomarker detection algorithms. A user-friendly graphical user interface (GUI) was developed using the programming language Python, allowing biomedical researchers to have direct access to their results. The source code and manual of kSolutionVis can be downloaded from http://www.healthinformaticslab.org/supp/resources.php.

Is It Possible to Predict the Future in First-Episode Psychosis?

The outcome of first-episode psychosis (FEP) is highly variable, ranging from early sustained recovery to antipsychotic treatment resistance from the onset of illness. For clinicians, a possibility to predict patient outcomes would be highly valuable for the selection of antipsychotic treatment and in tailoring psychosocial treatments and psychoeducation. This selective review summarizes current knowledge of prognostic markers in FEP. We sought potential outcome predictors from clinical and sociodemographic factors, cognition, brain imaging, genetics, and blood-based biomarkers, and we considered different outcomes, like remission, recovery, physical comorbidities, and suicide risk. Based on the review, it is currently possible to predict the future for FEP patients to some extent. Some clinical features-like the longer duration of untreated psychosis (DUP), poor premorbid adjustment, the insidious mode of onset, the greater severity of negative symptoms, comorbid substance use disorders (SUDs), a history of suicide attempts and suicidal ideation and having non-affective psychosis-are associated with a worse outcome. Of the social and demographic factors, male gender, social disadvantage, neighborhood deprivation, dysfunctional family environment, and ethnicity may be relevant. Treatment non-adherence is a substantial risk factor for relapse, but a small minority of patients with acute onset of FEP and early remission may benefit from antipsychotic discontinuation. Cognitive functioning is associated with functional outcomes. Brain imaging currently has limited utility as an outcome predictor, but this may change with methodological advancements. Polygenic risk scores (PRSs) might be useful as one component of a predictive tool, and pharmacogenetic testing is already available and valuable for patients who have problems in treatment response or with side effects. Most blood-based biomarkers need further validation. None of the currently available predictive markers has adequate sensitivity or specificity used alone. However, personalized treatment of FEP will need predictive tools. We discuss some methodologies, such as machine learning (ML), and tools that could lead to the improved prediction and clinical utility of different prognostic markers in FEP. Combination of different markers in ML models with a user friendly interface, or novel findings from e.g., molecular genetics or neuroimaging, may result in computer-assisted clinical applications in the near future.