PRSet: Pathway-based polygenic risk score analyses and software.

Polygenic risk scores (PRSs) have been among the leading advances in biomedicine in recent years. As a proxy of genetic liability, PRSs are utilised across multiple fields and applications. While numerous statistical and machine learning methods have been developed to optimise their predictive accuracy, these typically distil genetic liability to a single number based on aggregation of an individual's genome-wide risk alleles. This results in a key loss of information about an individual's genetic profile, which could be critical given the functional sub-structure of the genome and the heterogeneity of complex disease. In this manuscript, we introduce a 'pathway polygenic' paradigm of disease risk, in which multiple genetic liabilities underlie complex diseases, rather than a single genome-wide liability. We describe a method and accompanying software, PRSet, for computing and analysing pathway-based PRSs, in which polygenic scores are calculated across genomic pathways for each individual. We evaluate the potential of pathway PRSs in two distinct ways, creating two major sections: (1) In the first section, we benchmark PRSet as a pathway enrichment tool, evaluating its capacity to capture GWAS signal in pathways. We find that for target sample sizes of >10,000 individuals, pathway PRSs have similar power for evaluating pathway enrichment as leading methods MAGMA and LD score regression, with the distinct advantage of providing individual-level estimates of genetic liability for each pathway -opening up a range of pathway-based PRS applications, (2) In the second section, we evaluate the performance of pathway PRSs for disease stratification. We show that using a supervised disease stratification approach, pathway PRSs (computed by PRSet) outperform two standard genome-wide PRSs (computed by C+T and lassosum) for classifying disease subtypes in 20 of 21 scenarios tested. As the definition and functional annotation of pathways becomes increasingly refined, we expect pathway PRSs to offer key insights into the heterogeneity of complex disease and treatment response, to generate biologically tractable therapeutic targets from polygenic signal, and, ultimately, to provide a powerful path to precision medicine.

Portability of 245 polygenic scores when derived from the UK Biobank and applied to 9 ancestry groups from the same cohort.

The low portability of polygenic scores (PGSs) across global populations is a major concern that must be addressed before PGSs can be used for everyone in the clinic. Indeed, prediction accuracy has been shown to decay as a function of the genetic distance between the training and test cohorts. However, such cohorts differ not only in their genetic distance but also in their geographical distance and their data collection and assaying, conflating multiple factors. In this study, we examine the extent to which PGSs are transferable between ancestries by deriving polygenic scores for 245 curated traits from the UK Biobank data and applying them in nine ancestry groups from the same cohort. By restricting both training and testing to the UK Biobank data, we reduce the risk of environmental and genotyping confounding from using different cohorts. We define the nine ancestry groups at a sub-continental level, based on a simple, robust, and effective method that we introduce here. We then apply two different predictive methods to derive polygenic scores for all 245 phenotypes and show a systematic and dramatic reduction in portability of PGSs trained using Northwestern European individuals and applied to nine ancestry groups. These analyses demonstrate that prediction already drops off within European ancestries and reduces globally in proportion to genetic distance. Altogether, our study provides unique and robust insights into the PGS portability problem.

Polygenic risk of social isolation behavior and its influence on psychopathology and personality.

Social isolation has been linked to a range of psychiatric issues, but the behavioral component that drives it is not well understood. Here, a genome-wide associations study (GWAS) was carried out to identify genetic variants that contribute specifically to social isolation behavior (SIB) in up to 449,609 participants from the UK Biobank. 17 loci were identified at genome-wide significance, contributing to a 4% SNP-based heritability estimate. Using the SIB GWAS, polygenic risk scores (PRS) were derived in ALSPAC, an independent, developmental cohort, and used to test for association with self-reported friendship scores, comprising items related to friendship quality and quantity, at age 12 and 18 to determine whether genetic predisposition manifests during childhood development. At age 18, friendship scores were associated with the SIB PRS, demonstrating that the genetic factors can predict related social traits in late adolescence. Linkage disequilibrium (LD) score correlation using the SIB GWAS demonstrated genetic correlations with autism spectrum disorder (ASD), schizophrenia, major depressive disorder (MDD), educational attainment, extraversion, and loneliness. However, no evidence of causality was found using a conservative Mendelian randomization approach between SIB and any of the traits in either direction. Genomic Structural Equation Modeling (SEM) revealed a common factor contributing to SIB, neuroticism, loneliness, MDD, and ASD, weakly correlated with a second common factor that contributes to psychiatric and psychotic traits. Our results show that SIB contributes a small heritable component, which is associated genetically with other social traits such as friendship as well as psychiatric disorders.

Using genetic data to strengthen causal inference in observational research.

Causal inference is essential across the biomedical, behavioural and social sciences.By progressing from confounded statistical associations to evidence of causal relationships, causal inference can reveal complex pathways underlying traits and diseases and help to prioritize targets for intervention. Recent progress in genetic epidemiology - including statistical innovation, massive genotyped data sets and novel computational tools for deep data mining - has fostered the intense development of methods exploiting genetic data and relatedness to strengthen causal inference in observational research. In this Review, we describe how such genetically informed methods differ in their rationale, applicability and inherent limitations and outline how they should be integrated in the future to offer a rich causal inference toolbox.

Genetic sensitivity analysis: Adjusting for genetic confounding in epidemiological associations.

Associations between exposures and outcomes reported in epidemiological studies are typically unadjusted for genetic confounding. We propose a two-stage approach for estimating the degree to which such observed associations can be explained by genetic confounding. First, we assess attenuation of exposure effects in regressions controlling for increasingly powerful polygenic scores. Second, we use structural equation models to estimate genetic confounding using heritability estimates derived from both SNP-based and twin-based studies. We examine associations between maternal education and three developmental outcomes - child educational achievement, Body Mass Index, and Attention Deficit Hyperactivity Disorder. Polygenic scores explain between 14.3% and 23.0% of the original associations, while analyses under SNP- and twin-based heritability scenarios indicate that observed associations could be almost entirely explained by genetic confounding. Thus, caution is needed when interpreting associations from non-genetically informed epidemiology studies. Our approach, akin to a genetically informed sensitivity analysis can be applied widely.

Comparing Within- and Between-Family Polygenic Score Prediction.

Polygenic scores are a popular tool for prediction of complex traits. However, prediction estimates in samples of unrelated participants can include effects of population stratification, assortative mating, and environmentally mediated parental genetic effects, a form of genotype-environment correlation (rGE). Comparing genome-wide polygenic score (GPS) predictions in unrelated individuals with predictions between siblings in a within-family design is a powerful approach to identify these different sources of prediction. Here, we compared within- to between-family GPS predictions of eight outcomes (anthropometric, cognitive, personality, and health) for eight corresponding GPSs. The outcomes were assessed in up to 2,366 dizygotic (DZ) twin pairs from the Twins Early Development Study from age 12 to age 21. To account for family clustering, we used mixed-effects modeling, simultaneously estimating within- and between-family effects for target- and cross-trait GPS prediction of the outcomes. There were three main findings: (1) DZ twin GPS differences predicted DZ differences in height, BMI, intelligence, educational achievement, and ADHD symptoms; (2) target and cross-trait analyses indicated that GPS prediction estimates for cognitive traits (intelligence and educational achievement) were on average 60% greater between families than within families, but this was not the case for non-cognitive traits; and (3) much of this within- and between-family difference for cognitive traits disappeared after controlling for family socio-economic status (SES), suggesting that SES is a major source of between-family prediction through rGE mechanisms. These results provide insights into the patterns by which rGE contributes to GPS prediction, while ruling out confounding due to population stratification and assortative mating.

Exploring the causal effect of maternal pregnancy adiposity on offspring adiposity: Mendelian randomisation using polygenic risk scores.

BACKGROUND: Greater maternal adiposity before or during pregnancy is associated with greater offspring adiposity throughout childhood, but the extent to which this is due to causal intrauterine or periconceptional mechanisms remains unclear. Here, we use Mendelian randomisation (MR) with polygenic risk scores (PRS) to investigate whether associations between maternal pre-/early pregnancy body mass index (BMI) and offspring adiposity from birth to adolescence are causal. METHODS: We undertook confounder adjusted multivariable (MV) regression and MR using mother-offspring pairs from two UK cohorts: Avon Longitudinal Study of Parents and Children (ALSPAC) and Born in Bradford (BiB). In ALSPAC and BiB, the outcomes were birthweight (BW; N = 9339) and BMI at age 1 and 4 years (N = 8659 to 7575). In ALSPAC only we investigated BMI at 10 and 15 years (N = 4476 to 4112) and dual-energy X-ray absorptiometry (DXA) determined fat mass index (FMI) from age 10-18 years (N = 2659 to 3855). We compared MR results from several PRS, calculated from maternal non-transmitted alleles at between 29 and 80,939 single nucleotide polymorphisms (SNPs). RESULTS: MV and MR consistently showed a positive association between maternal BMI and BW, supporting a moderate causal effect. For adiposity at most older ages, although MV estimates indicated a strong positive association, MR estimates did not support a causal effect. For the PRS with few SNPs, MR estimates were statistically consistent with the null, but had wide confidence intervals so were often also statistically consistent with the MV estimates. In contrast, the largest PRS yielded MR estimates with narrower confidence intervals, providing strong evidence that the true causal effect on adolescent adiposity is smaller than the MV estimates (Pdifference = 0.001 for 15-year BMI). This suggests that the MV estimates are affected by residual confounding, therefore do not provide an accurate indication of the causal effect size. CONCLUSIONS: Our results suggest that higher maternal pre-/early-pregnancy BMI is not a key driver of higher adiposity in the next generation. Thus, they support interventions that target the whole population for reducing overweight and obesity, rather than a specific focus on women of reproductive age.

Causal Associations Between Modifiable Risk Factors and the Alzheimer's Phenome.

OBJECTIVE: The purpose of this study was to infer causal relationships between 22 previously reported risk factors for Alzheimer's disease (AD) and the "AD phenome": AD, AD age of onset (AAOS), hippocampal volume, cortical surface area and thickness, cerebrospinal fluid (CSF) levels of amyloid-ß (Aß42 ), tau, and ptau181 , and the neuropathological burden of neuritic plaques, neurofibrillary tangles (NFTs), and vascular brain injury (VBI). METHODS: Polygenic risk scores (PRS) for the 22 risk factors were computed in 26,431 AD cases/controls and the association with AD was evaluated using logistic regression. Two-sample Mendelian randomization (MR) was used to infer the causal effect of risk factors on the AD phenome. RESULTS: PRS for increased education and diastolic blood pressure were associated with reduced risk for AD. MR indicated that only education was causally associated with reduced risk of AD, delayed AAOS, and increased cortical surface area and thickness. Total- and LDL-cholesterol levels were causally associated with increased neuritic plaque burden, although the effects were driven by single nucleotide polymorphisms (SNPs) within the APOE locus. Diastolic blood pressure and pulse pressure are causally associated with increased risk of VBI. Furthermore, total cholesterol was associated with decreased hippocampal volume; smoking initiation with decreased cortical thickness; type 2 diabetes with an earlier AAOS; and sleep duration with increased cortical thickness. INTERPRETATION: Our comprehensive examination of the genetic evidence for the causal relationships between previously reported risk factors in AD using PRS and MR supports a causal role for education, blood pressure, cholesterol levels, smoking, and diabetes with the AD phenome. ANN NEUROL 2021;89:54-65.

Dabigatran Reversal With Idarucizumab in Patients Requiring Urgent Surgery: A Subanalysis of the RE-VERSE AD Study.

OBJECTIVE: To further examine anticoagulation reversal and clinical outcomes in dabigatran treated patients requiring urgent surgery or procedural interventions. BACKGROUND: Idarucizumab, a humanized monoclonal antibody fragment, reverses dabigatran anticoagulation. METHODS: Data from surgical and procedural patients in RE-VERSE AD, a multicenter, open-label, single-arm, prospective cohort of dabigatran reversal were evaluated. A total of 202 patients in this group received 5 g of idarucizumab before surgery or procedures. RESULTS: The interventions included 49 abdominal, 45 orthopedic, 34 vascular, 8 neurologic, and 4 genitourinary surgical procedures, or 29 catheter-based cases, 20 cases for drainage, and 8 diagnostic procedures. Five patients did not undergo their intended intervention after receiving idarucizumab. Complete reversal of the dabigatran anticoagulant effect occurred within minutes in almost all patients, with normal hemostasis in more than 91% of patients. The median time from the first vial of idarucizumab to surgery or procedures was less than 2 hours in all groups except neurosurgery, where it was 3.3 hours. Fresh frozen plasma and packed red cells were the most frequently transfused blood products. Postreversal thromboembolic events occurred in 10 (5%) patients at 30 days, 5 of whom had restarted anticoagulation before the event. Overall 30-day mortality was 12.6%. There were no serious adverse safety signals due to idarucizumab dosing. CONCLUSIONS: Idarucizumab facilitates management of patients requiring urgent procedures by providing rapid dabigatran reversal, and is the only agent of its class studied in surgical patients.

Maternal health around pregnancy and autism risk: a diagnosis-wide, population-based study.

BACKGROUND: Many studies have reported an increased risk of autism spectrum disorder (ASD) associated with some maternal diagnoses in pregnancy. However, such associations have not been studied systematically, accounting for comorbidity between maternal disorders. Therefore our aim was to comprehensively test the associations between maternal diagnoses around pregnancy and ASD risk in offspring. METHODS: This exploratory case-cohort study included children born in Israel from 1997 to 2008, and followed up until 2015. We used information on all ICD-9 codes received by their mothers during pregnancy and the preceding year. ASD risk associated with each of those conditions was calculated using Cox proportional hazards regression, adjusted for the confounders (birth year, maternal age, socioeconomic status and number of ICD-9 diagnoses during the exposure period). RESULTS: The analytic sample consisted of 80 187 individuals (1132 cases, 79 055 controls), with 822 unique ICD-9 codes recorded in their mothers. After extensive quality control, 22 maternal diagnoses were nominally significantly associated with offspring ASD, with 16 of those surviving subsequent filtering steps (permutation testing, multiple testing correction, multiple regression). Among those, we recorded an increased risk of ASD associated with metabolic [e.g. hypertension; HR = 2.74 (1.92-3.90), p = 2.43 × 10-8], genitourinary [e.g. non-inflammatory disorders of cervix; HR = 1.88 (1.38-2.57), p = 7.06 × 10-5] and psychiatric [depressive disorder; HR = 2.11 (1.32-3.35), p = 1.70 × 10-3] diagnoses. Meanwhile, mothers of children with ASD were less likely to attend prenatal care appointment [HR = 0.62 (0.54-0.71), p = 1.80 × 10-11]. CONCLUSIONS: Sixteen maternal diagnoses were associated with ASD in the offspring, after rigorous filtering of potential false-positive associations. Replication in other cohorts and further research to understand the mechanisms underlying the observed associations with ASD are warranted.

Genome-wide gene-environment analyses of major depressive disorder and reported lifetime traumatic experiences in UK Biobank.

Depression is more frequent among individuals exposed to traumatic events. Both trauma exposure and depression are heritable. However, the relationship between these traits, including the role of genetic risk factors, is complex and poorly understood. When modelling trauma exposure as an environmental influence on depression, both gene-environment correlations and gene-environment interactions have been observed. The UK Biobank concurrently assessed Major Depressive Disorder (MDD) and self-reported lifetime exposure to traumatic events in 126,522 genotyped individuals of European ancestry. We contrasted genetic influences on MDD stratified by reported trauma exposure (final sample size range: 24,094-92,957). The SNP-based heritability of MDD with reported trauma exposure (24%) was greater than MDD without reported trauma exposure (12%). Simulations showed that this is not confounded by the strong, positive genetic correlation observed between MDD and reported trauma exposure. We also observed that the genetic correlation between MDD and waist circumference was only significant in individuals reporting trauma exposure (rg = 0.24, p = 1.8 × 10-7 versus rg = -0.05, p = 0.39 in individuals not reporting trauma exposure, difference p = 2.3 × 10-4). Our results suggest that the genetic contribution to MDD is greater when reported trauma is present, and that a complex relationship exists between reported trauma exposure, body composition, and MDD.

Repurposing FDA approved drugs as radiosensitizers for treating hypoxic prostate cancer.

BACKGROUND: The presence of hypoxia is a poor prognostic factor in prostate cancer and the hypoxic tumor microenvironment promotes radioresistance. There is potential for drug radiotherapy combinations to improve the therapeutic ratio. We aimed to investigate whether hypoxia-associated genes could be used to identify FDA approved drugs for repurposing for the treatment of hypoxic prostate cancer. METHODS: Hypoxia associated genes were identified and used in the connectivity mapping software QUADrATIC to identify FDA approved drugs as candidates for repurposing. Drugs identified were tested in vitro in prostate cancer cell lines (DU145, PC3, LNCAP). Cytotoxicity was investigated using the sulforhodamine B assay and radiosensitization using a clonogenic assay in normoxia and hypoxia. RESULTS: Menadione and gemcitabine had similar cytotoxicity in normoxia and hypoxia in all three cell lines. In DU145 cells, the radiation sensitizer enhancement ratio (SER) of menadione was 1.02 in normoxia and 1.15 in hypoxia. The SER of gemcitabine was 1.27 in normoxia and 1.09 in hypoxia. No radiosensitization was seen in PC3 cells. CONCLUSION: Connectivity mapping can identify FDA approved drugs for potential repurposing that are linked to a radiobiologically relevant phenotype. Gemcitabine and menadione could be further investigated as potential radiosensitizers in prostate cancer.

Evidence for gene-environment correlation in child feeding: Links between common genetic variation for BMI in children and parental feeding practices.

The parental feeding practices (PFPs) of excessive restriction of food intake ('restriction') and pressure to increase food consumption ('pressure') have been argued to causally influence child weight in opposite directions (high restriction causing overweight; high pressure causing underweight). However child weight could also 'elicit' PFPs. A novel approach is to investigate gene-environment correlation between child genetic influences on BMI and PFPs. Genome-wide polygenic scores (GPS) combining BMI-associated variants were created for 10,346 children (including 3,320 DZ twin pairs) from the Twins Early Development Study using results from an independent genome-wide association study meta-analysis. Parental 'restriction' and 'pressure' were assessed using the Child Feeding Questionnaire. Child BMI standard deviation scores (BMI-SDS) were calculated from children's height and weight at age 10. Linear regression and fixed family effect models were used to test between- (n = 4,445 individuals) and within-family (n = 2,164 DZ pairs) associations between the GPS and PFPs. In addition, we performed multivariate twin analyses (n = 4,375 twin pairs) to estimate the heritabilities of PFPs and the genetic correlations between BMI-SDS and PFPs. The GPS was correlated with BMI-SDS (ß = 0.20, p = 2.41x10-38). Consistent with the gene-environment correlation hypothesis, child BMI GPS was positively associated with 'restriction' (ß = 0.05, p = 4.19x10-4), and negatively associated with 'pressure' (ß = -0.08, p = 2.70x10-7). These results remained consistent after controlling for parental BMI, and after controlling for overall family contributions (within-family analyses). Heritabilities for 'restriction' (43% [40-47%]) and 'pressure' (54% [50-59%]) were moderate-to-high. Twin-based genetic correlations were moderate and positive between BMI-SDS and 'restriction' (rA = 0.28 [0.23-0.32]), and substantial and negative between BMI-SDS and 'pressure' (rA = -0.48 [-0.52 - -0.44]. Results suggest that the degree to which parents limit or encourage children's food intake is partly influenced by children's genetic predispositions to higher or lower BMI. These findings point to an evocative gene-environment correlation in which heritable characteristics in the child elicit parental feeding behaviour.

Idarucizumab for Dabigatran Reversal in the Management of Patients With Gastrointestinal Bleeding.

BACKGROUND: Although dabigatran has a favorable risk-benefit profile compared with vitamin K antagonist therapy for venous thromboembolism and nonvalvular atrial fibrillation, major bleeding events, including gastrointestinal (GI) bleeding, may occur. Therefore, our aim was to provide insights into the efficacy and safety of idarucizumab for urgent dabigatran reversal in patients with major GI bleeding. METHODS: Patients with uncontrollable GI bleeding requiring reversal were enrolled from June 2014 through July 2016 in the RE-VERSE AD study (Reversal of Dabigatran Anticoagulant Effect With Idarucizumab), a prospective, multicenter, open-label study of idarucizumab, and were followed up for 90 days for primary and secondary outcomes. Patients were to receive a 5-g dose of intravenous idarucizumab, administered as 2 bolus infusions of 2.5 g no more than 15 minutes apart. The primary end point was the maximum reversal of dabigatran anticoagulation within 4 hours after administration of idarucizumab as measured by the dabigatran-specific assays diluted thrombin time and ecarin clotting time. Further end points included investigator-reported bleeding cessation within the first 24 hours and incidence of rebleeding, thromboembolic events, or mortality. RESULTS: GI bleeding occurred in 137 patients enrolled in RE-VERSE AD, of which 84% was adjudicated as major or life-threatening, 48 (35.0%) was upper GI tract in origin, 43 (31.4%) was lower GI in origin, and 46 (33.6%) was either both or unknown. Complete reversal of dabigatran was observed in 118 of 121 patients (97.5%) with an elevated diluted thrombin time at presentation and 95 of 131 patients (72.5%) with an elevated ecarin clotting time and was similar for upper and lower GI bleeding. Bleeding cessation within 24 hours was reported in 92 of 134 evaluable patients (68.7%) after a median duration of 2.4 hours (interquartile range, 2.0-3.9 hours). During the 90-day follow-up, 6 patients (4.4%) had a postreversal thromboembolic event, and 20 patients (14.6%) died. CONCLUSIONS: Idarucizumab showed a rapid and complete reversal of dabigatran activity in nearly all patients presenting with GI bleeding, facilitating emergency patient care without the additional presence of anticoagulation. CLINICAL TRIAL REGISTRATION: URL: https://www.clinicaltrials.gov . Unique identifier: NCT02104947.

West Nile Virus in Farmed Crocodiles, Zambia, 2019.

We detected West Nile virus (WNV) nucleic acid in crocodiles (Crocodylus niloticus) in Zambia. Phylogenetically, the virus belonged to lineage 1a, which is predominant in the Northern Hemisphere. These data provide evidence that WNV is circulating in crocodiles in Africa and increases the risk for animal and human transmission.

NUQA: Estimating Cancer Spatial and Temporal Heterogeneity and Evolution through Alignment-Free Methods.

Longitudinal next-generation sequencing of cancer patient samples has enhanced our understanding of the evolution and progression of various cancers. As a result, and due to our increasing knowledge of heterogeneity, such sampling is becoming increasingly common in research and clinical trial sample collections. Traditionally, the evolutionary analysis of these cohorts involves the use of an aligner followed by subsequent stringent downstream analyses. However, this can lead to large levels of information loss due to the vast mutational landscape that characterizes tumor samples. Here, we propose an alignment-free approach for sequence comparison-a well-established approach in a range of biological applications including typical phylogenetic classification. Such methods could be used to compare information collated in raw sequence files to allow an unsupervised assessment of the evolutionary trajectory of patient genomic profiles. In order to highlight this utility in cancer research we have applied our alignment-free approach using a previously established metric, Jensen-Shannon divergence, and a metric novel to this area, Hellinger distance, to two longitudinal cancer patient cohorts in glioma and clear cell renal cell carcinoma using our software, NUQA. We hypothesize that this approach has the potential to reveal novel information about the heterogeneity and evolutionary trajectory of spatiotemporal tumor samples, potentially revealing early events in tumorigenesis and the origins of metastases and recurrences. Key words: alignment-free, Hellinger distance, exome-seq, evolution, phylogenetics, longitudinal.

Idarucizumab for Dabigatran Reversal - Full Cohort Analysis.

BACKGROUND: Idarucizumab, a monoclonal antibody fragment, was developed to reverse the anticoagulant effect of dabigatran. METHODS: We performed a multicenter, prospective, open-label study to determine whether 5 g of intravenous idarucizumab would be able to reverse the anticoagulant effect of dabigatran in patients who had uncontrolled bleeding (group A) or were about to undergo an urgent procedure (group B). The primary end point was the maximum percentage reversal of the anticoagulant effect of dabigatran within 4 hours after the administration of idarucizumab, on the basis of the diluted thrombin time or ecarin clotting time. Secondary end points included the restoration of hemostasis and safety measures. RESULTS: A total of 503 patients were enrolled: 301 in group A, and 202 in group B. The median maximum percentage reversal of dabigatran was 100% (95% confidence interval, 100 to 100), on the basis of either the diluted thrombin time or the ecarin clotting time. In group A, 137 patients (45.5%) presented with gastrointestinal bleeding and 98 (32.6%) presented with intracranial hemorrhage; among the patients who could be assessed, the median time to the cessation of bleeding was 2.5 hours. In group B, the median time to the initiation of the intended procedure was 1.6 hours; periprocedural hemostasis was assessed as normal in 93.4% of the patients, mildly abnormal in 5.1%, and moderately abnormal in 1.5%. At 90 days, thrombotic events had occurred in 6.3% of the patients in group A and in 7.4% in group B, and the mortality rate was 18.8% and 18.9%, respectively. There were no serious adverse safety signals. CONCLUSIONS: In emergency situations, idarucizumab rapidly, durably, and safely reversed the anticoagulant effect of dabigatran. (Funded by Boehringer Ingelheim; RE-VERSE AD ClinicalTrials.gov number, NCT02104947 .).

Machine Learning to Predict Mortality and Critical Events in a Cohort of Patients With COVID-19 in New York City: Model Development and Validation.

BACKGROUND: COVID-19 has infected millions of people worldwide and is responsible for several hundred thousand fatalities. The COVID-19 pandemic has necessitated thoughtful resource allocation and early identification of high-risk patients. However, effective methods to meet these needs are lacking. OBJECTIVE: The aims of this study were to analyze the electronic health records (EHRs) of patients who tested positive for COVID-19 and were admitted to hospitals in the Mount Sinai Health System in New York City; to develop machine learning models for making predictions about the hospital course of the patients over clinically meaningful time horizons based on patient characteristics at admission; and to assess the performance of these models at multiple hospitals and time points. METHODS: We used Extreme Gradient Boosting (XGBoost) and baseline comparator models to predict in-hospital mortality and critical events at time windows of 3, 5, 7, and 10 days from admission. Our study population included harmonized EHR data from five hospitals in New York City for 4098 COVID-19-positive patients admitted from March 15 to May 22, 2020. The models were first trained on patients from a single hospital (n=1514) before or on May 1, externally validated on patients from four other hospitals (n=2201) before or on May 1, and prospectively validated on all patients after May 1 (n=383). Finally, we established model interpretability to identify and rank variables that drive model predictions. RESULTS: Upon cross-validation, the XGBoost classifier outperformed baseline models, with an area under the receiver operating characteristic curve (AUC-ROC) for mortality of 0.89 at 3 days, 0.85 at 5 and 7 days, and 0.84 at 10 days. XGBoost also performed well for critical event prediction, with an AUC-ROC of 0.80 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. In external validation, XGBoost achieved an AUC-ROC of 0.88 at 3 days, 0.86 at 5 days, 0.86 at 7 days, and 0.84 at 10 days for mortality prediction. Similarly, the unimputed XGBoost model achieved an AUC-ROC of 0.78 at 3 days, 0.79 at 5 days, 0.80 at 7 days, and 0.81 at 10 days. Trends in performance on prospective validation sets were similar. At 7 days, acute kidney injury on admission, elevated LDH, tachypnea, and hyperglycemia were the strongest drivers of critical event prediction, while higher age, anion gap, and C-reactive protein were the strongest drivers of mortality prediction. CONCLUSIONS: We externally and prospectively trained and validated machine learning models for mortality and critical events for patients with COVID-19 at different time horizons. These models identified at-risk patients and uncovered underlying relationships that predicted outcomes.