Search

Background: Timely diagnosis of structural heart disease improves patient outcomes, yet many remain underdiagnosed. While population screening with echocardiography is impractical, ECG-based prediction models can help target high-risk patients. We developed a novel ECG-based machine learning approach to predict multiple structural heart conditions, hypothesizing that a composite model would yield higher prevalence and positive predictive values to facilitate meaningful recommendations for echocardiography. Methods: Using 2 232 130 ECGs linked to electronic health records and echocardiography reports from 484 765 adults between 1984 to 2021, we trained machine learning models to predict the presence or absence of any of 7 echocardiography-confirmed diseases within 1 year. This composite label included the following: moderate or severe valvular disease (aortic/mitral stenosis or regurgitation, tricuspid regurgitation), reduced ejection fraction 15 mm. We tested various combinations of input features (demographics, laboratory values, structured ECG data, ECG traces) and evaluated model performance using 5-fold cross-validation, multisite validation trained on 1 site and tested on 10 independent sites, and simulated retrospective deployment trained on pre-2010 data and deployed in 2010. Results: Our composite rECHOmmend model used age, sex, and ECG traces and had a 0.91 area under the receiver operating characteristic curve and a 42% positive predictive value at 90% sensitivity, with a composite label prevalence of 17.9%. Individual disease models had area under the receiver operating characteristic curves from 0.86 to 0.93 and lower positive predictive values from 1% to 31%. Area under the receiver operating characteristic curves for models using different input features ranged from 0.80 to 0.93, increasing with additional features. Multisite validation showed similar results to cross-validation, with an aggregate area under the receiver operating characteristic curve of 0.91 across our independent test set of 10 clinical sites after training on a separate site. Our simulated retrospective deployment showed that for ECGs acquired in patients without preexisting structural heart disease in the year 2010, 11% were classified as high risk and 41% (4.5% of total patients) developed true echocardiography-confirmed disease within 1 year. Conclusions: An ECG-based machine learning model using a composite end point can identify a high-risk population for having undiagnosed, clinically significant structural heart disease while outperforming single-disease models and improving practical utility with higher positive predictive values. This approach can facilitate targeted screening with echocardiography to improve underdiagnosis of structural heart disease.

Supplemental Digital Content is available in the text., Background: Atrial fibrillation (AF) is associated with substantial morbidity, especially when it goes undetected. If new-onset AF could be predicted, targeted screening could be used to find it early. We hypothesized that a deep neural network could predict new-onset AF from the resting 12-lead ECG and that this prediction may help identify those at risk of AF-related stroke. Methods: We used 1.6 M resting 12-lead digital ECG traces from 430 000 patients collected from 1984 to 2019. Deep neural networks were trained to predict new-onset AF (within 1 year) in patients without a history of AF. Performance was evaluated using areas under the receiver operating characteristic curve and precision-recall curve. We performed an incidence-free survival analysis for a period of 30 years following the ECG stratified by model predictions. To simulate real-world deployment, we trained a separate model using all ECGs before 2010 and evaluated model performance on a test set of ECGs from 2010 through 2014 that were linked to our stroke registry. We identified the patients at risk for AF-related stroke among those predicted to be high risk for AF by the model at different prediction thresholds. Results: The area under the receiver operating characteristic curve and area under the precision-recall curve were 0.85 and 0.22, respectively, for predicting new-onset AF within 1 year of an ECG. The hazard ratio for the predicted high- versus low-risk groups over a 30-year span was 7.2 (95% CI, 6.9–7.6). In a simulated deployment scenario, the model predicted new-onset AF at 1 year with a sensitivity of 69% and specificity of 81%. The number needed to screen to find 1 new case of AF was 9. This model predicted patients at high risk for new-onset AF in 62% of all patients who experienced an AF-related stroke within 3 years of the index ECG. Conclusions: Deep learning can predict new-onset AF from the 12-lead ECG in patients with no previous history of AF. This prediction may help identify patients at risk for AF-related strokes.

Several large trials have employed age or clinical features to select patients for atrial fibrillation (AF) screening to reduce strokes. We hypothesized that a machine learning (ML) model trained to predict AF risk from 12‑lead electrocardiogram (ECG) would be more efficient than criteria based on clinical variables in indicating a population for AF screening to potentially prevent AF-related stroke.We retrospectively included all patients with clinical encounters in Geisinger without a prior history of AF. Incidence of AF within 1 year and AF-related strokes within 3 years of the encounter were identified. AF-related stroke was defined as a stroke where AF was diagnosed at the time of stroke or within a year after the stroke. The efficiency of five methods was evaluated for selecting a cohort for AF screening. The methods were selected from four clinical trials (mSToPS, GUARD-AF, SCREEN-AF and STROKESTOP) and the ECG-based ML model. We simulated patient selection for the five methods between the years 2011 and 2014 and evaluated outcomes for 1 year intervals between 2012 and 2015, resulting in a total of twenty 1-year periods. Patients were considered eligible if they met the criteria before the start of the given 1-year period or within that period. The primary outcomes were numbers needed to screen (NNS) for AF and AF-associated stroke.The clinical trial models indicated large proportions of the population with a prior ECG for AF screening (up to 31%), coinciding with NNS ranging from 14 to 18 for AF and 249-359 for AF-associated stroke. At comparable sensitivity, the ECG ML model indicated a modest number of patients for screening (14%) and had the highest efficiency in NNS for AF (7.3; up to 60% reduction) and AF-associated stroke (223; up to 38% reduction).An ECG-based ML risk prediction model is more efficient than contemporary AF-screening criteria based on age alone or age and clinical features at indicating a population for AF screening to potentially prevent AF-related strokes.

The UK Biobank is a prospective study of 502,543 individuals, combining extensive phenotypic and genotypic data with streamlined access for researchers around the world1. Here we describe the release of exome-sequence data for the first 49,960 study participants, revealing approximately 4 million coding variants (of which around 98.6% have a frequency of less than 1%). The data include 198,269 autosomal predicted loss-of-function (LOF) variants, a more than 14-fold increase compared to the imputed sequence. Nearly all genes (more than 97%) had at least one carrier with a LOF variant, and most genes (more than 69%) had at least ten carriers with a LOF variant. We illustrate the power of characterizing LOF variants in this population through association analyses across 1,730 phenotypes. In addition to replicating established associations, we found novel LOF variants with large effects on disease traits, including PIEZO1 on varicose veins, COL6A1 on corneal resistance, MEPE on bone density, and IQGAP2 and GMPR on blood cell traits. We further demonstrate the value of exome sequencing by surveying the prevalence of pathogenic variants of clinical importance, and show that 2% of this population has a medically actionable variant. Furthermore, we characterize the penetrance of cancer in carriers of pathogenic BRCA1 and BRCA2 variants. Exome sequences from the first 49,960 participants highlight the promise of genome sequencing in large population-based studies and are now accessible to the scientific community., Exome sequences from the first 49,960 participants in the UK Biobank highlight the promise of genome sequencing in large population-based studies and are now accessible to the scientific community.

Background Heart failure is a prevalent, costly disease for which new value-based payment models demand optimized population management strategies. Objectives This study sought to generate a strategy for managing populations of patients with heart failure by leveraging large clinical datasets and machine learning. Methods Geisinger electronic health record data were used to train machine learning models to predict 1-year all-cause mortality in 26,971 patients with heart failure who underwent 276,819 clinical episodes. There were 26 clinical variables (demographics, laboratory test results, medications), 90 diagnostic codes, 41 electrocardiogram measurements and patterns, 44 echocardiographic measurements, and 8 evidence-based “care gaps”: flu vaccine, blood pressure of Results Machine learning models achieved areas under the receiver-operating characteristic curve (AUCs) of 0.74 to 0.77 in a split-by-year training/test scheme, with the nonlinear XGBoost model (AUC: 0.77) outperforming linear logistic regression (AUC: 0.74). Out of 13,238 currently living patients, 2,844 were predicted to die within a year, and closing all care gaps was predicted to save 231 of these lives. Prioritizing patients for intervention by using the predicted reduction in 1-year mortality risk outperformed all other priority rankings (e.g., random selection or Seattle Heart Failure risk score). Conclusions Machine learning can be used to priority-rank patients most likely to benefit from interventions to optimize evidence-based therapies. This approach may prove useful for optimizing heart failure population health management teams within value-based payment models.

BackgroundEarly diagnosis of structural heart disease improves patient outcomes, yet many remain underdiagnosed. While population screening with echocardiography is impractical, electrocardiogram (ECG)-based prediction models can help target high-risk patients. We developed a novel ECG-based machine learning approach to predict multiple structural heart conditions, hypothesizing that a composite model would yield higher prevalence and positive predictive values (PPVs) to facilitate meaningful recommendations for echocardiography.MethodsUsing 2,232,130 ECGs linked to electronic health records and echocardiography reports from 484,765 adults between 1984-2021, we trained machine learning models to predict the presence of any of seven echocardiography-confirmed diseases within one year. This composite label included: moderate or severe valvular disease (aortic/mitral stenosis or regurgitation, tricuspid regurgitation), reduced ejection fraction 15mm. We tested various combinations of input features (demographics, labs, structured ECG data, ECG traces) and evaluated model performance using 5-fold cross-validation, multi-site validation trained on one clinical site and tested on 11 other independent sites, and simulated retrospective deployment trained on pre-2010 data and deployed in 2010.FindingsOur composite “rECHOmmend” model using age, sex and ECG traces had an area under the receiver operating characteristic curve (AUROC) of 0.91 and a PPV of 42% at 90% sensitivity at a prevalence of 17.9% for our composite label. Individual disease models had AUROCs ranging from 0.86-0.93 and lower PPVs from 1%-31%. The AUROC for models using different input features ranged from 0.80-0.93, increasing with additional features. Multi-site validation showed similar results to the cross-validation, with an aggregate AUROC of 0.91 across our independent test set of 11 clinical sites after training on a separate site. Our simulated retrospective deployment showed that for ECGs acquired in patients without pre-existing known structural heart disease in a single year, 2010, 11% were classified as high-risk, of which 41% developed true, echocardiography-confirmed disease within one year.InterpretationAn ECG-based machine learning model using a composite endpoint can predict previously undiagnosed, clinically significant structural heart disease while outperforming single disease models and improving practical utility with higher PPVs. This approach can facilitate targeted screening with echocardiography to improve under-diagnosis of structural heart disease.

Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) causes coronavirus disease 2019 (COVID-19), a respiratory illness that can result in hospitalization or death. We used exome sequence data to investigate associations between rare genetic variants and seven COVID-19 outcomes in 586,157 individuals, including 20,952 with COVID-19. After accounting for multiple testing, we did not identify any clear associations with rare variants either exome wide or when specifically focusing on (1) 13 interferon pathway genes in which rare deleterious variants have been reported in individuals with severe COVID-19, (2) 281 genes located in susceptibility loci identified by the COVID-19 Host Genetics Initiative, or (3) 32 additional genes of immunologic relevance and/or therapeutic potential. Our analyses indicate there are no significant associations with rare protein-coding variants with detectable effect sizes at our current sample sizes. Analyses will be updated as additional data become available, and results are publicly available through the Regeneron Genetics Center COVID-19 Results Browser.

Variability in SARS-CoV-2 susceptibility and COVID-19 disease severity between individuals is partly due to genetic factors. Here, we identify 4 genomic loci with suggestive associations for SARS-CoV-2 susceptibility and 19 for COVID-19 disease severity. Four of these 23 loci likely have an ethnicity-specific component. Genome-wide association study (GWAS) signals in 11 loci colocalize with expression quantitative trait loci (eQTLs) associated with the expression of 20 genes in 62 tissues/cell types (range: 1:43 tissues/gene), including lung, brain, heart, muscle, and skin as well as the digestive system and immune system. We perform genetic fine mapping to compute 99% credible SNP sets, which identify 10 GWAS loci that have eight or fewer SNPs in the credible set, including three loci with one single likely causal SNP. Our study suggests that the diverse symptoms and disease severity of COVID-19 observed between individuals is associated with variants across the genome, affecting gene expression levels in a wide variety of tissue types., Graphical abstract, D’Antonio et al. characterize associations between GWAS signals for COVID-19 disease and eQTLs in 69 human tissues to identify causal variants and their underlying molecular mechanisms. They show that diverse symptoms and disease severity of COVID-19 are associated with variants affecting gene expression in a wide variety of tissues.

SARS-CoV-2 enters host cells by binding angiotensin-converting enzyme 2 (ACE2). Through a genome-wide association study, we show that a rare variant (MAF = 0.3%, odds ratio 0.60, P=4.5x10-13) that down-regulates ACE2 expression reduces risk of COVID-19 disease, providing human genetics support for the hypothesis that ACE2 levels influence COVID-19 risk. Further, we show that common genetic variants define a risk score that predicts severe disease among COVID-19 cases.

Exome and genome sequencing are increasingly utilized in research studies and clinical care and can provide clinically relevant information beyond the initial intent for sequencing, including medically actionable secondary findings. Despite ongoing debate about sharing this information with patients and participants, a growing number of clinical laboratories and research programs routinely report secondary findings that increase the risk for selected diseases. Recently, there has been a push to maximize the potential benefit of this practice by implementing proactive genomic screening at the population level irrespective of medical history, but the feasibility of deploying population-scale proactive genomic screening requires scaling key elements of the genomic data evaluation process. Herein, we describe the motivation, development, and implementation of a population-scale variant-first screening pipeline combining bioinformatics-based filtering with a manual review process to screen for clinically relevant findings in research exomes generated through the DiscovEHR collaboration within Geisinger's MyCode® research project. Consistent with other studies, this pipeline yields a screen-positive detection rate between 2.1 and 2.6% (depending on inclusion of those with prior indication-based testing) in 130,048 adult MyCode patient-participants screened for clinically relevant findings in 60 genes. Our variant-first pipeline affords cost and time savings by filtering out negative cases, thereby avoiding analysis of each exome one-by-one, as typically employed in the diagnostic setting. While research is still needed to fully appreciate the benefits of population genomic screening, MyCode provides the first demonstration of a program at scale to help shape how population genomic screening is integrated into routine clinical care.

Introduction: Healthcare cost has increased drastically in the last decade, and over 50% of the cost can be attributed to a small portion (5-10%) of the population. Certain clinical programs, such as home-based care, aim to reduce this utilization but need methods to identify the most appropriate patients to enroll. We hypothesized that machine learning can predict patients with high future utilization with good accuracy. Methods: 683,160 cardiology patients (defined broadly as those with an ECG, echocardiogram or cardiology visit) with ~17 million clinical episodes since 2004 were identified from Geisinger’s electronic health records. Utilization was estimated as total cost of care for outpatient, inpatient and emergency department visits. Patients with the highest 10% utilization in a given year were defined as high utilizers. Machine learning models were used to predict high utilization over the next 3, 6 and 12 months. Input variables (n=191) included age, sex, smoking, 5 vital signs, 21 labs, 18 medications (current and past), 40 ECG and 44 echocardiographic measurements, 43 comorbidities, 7 time / cyclical features, 6 past utilization metrics and 4 social metrics (e.g. distance to healthcare facilities). Results: XGBoost achieved the best performance with areas under the ROC curve (AUC) of 0.82, 0.81 and 0.78 for 3, 6, 12-month models, and average precision scores (AP) of 0.31, 0.36 and 0.37, respectively, while the commonly used Charlson Comorbidity Index had poor performance with AUCs of 0.63 - 0.64 and APs of 0.1 - 0.17. Past utilization was the best predictor of future utilization. Targeting patients with the top 5 and 10% highest risk for utilization achieved sensitivities of 26 and 40% and positive predictive values of 50 and 38% (12-month model, Figure). Conclusions: Machine learning can be used to predict which patients will have high future healthcare utilization. This may help target populations for intervention programs aimed at reducing utilization.

Introduction: An intriguing U-shape relationship between left ventricular ejection fraction (LVEF) and survival has recently been reported, with LVEFs of 60-65% associated with the lowest mortality risk. In heart failure, LVEF recovery has been linked to improved outcomes; however, the relationship between changes in LVEF (ΔLVEF) and survival in a general clinical population has not been studied. We hypothesized that ΔLVEF would have a non-linear relationship with all-cause mortality. Methods: A total of 194,599 echocardiograms from 57,823 patients with physician-reported LVEF were identified from Geisinger health records, along with dates of death or last living encounter, age, sex, smoking status, height, weight, and active diagnoses. ΔLVEF for a given echocardiogram was calculated for 136,776 studies as the difference between the most recent previous and current LVEF. Cox Proportional Hazards Regression was used to relate interaction between ΔLVEF and current LVEF to all-cause mortality while adjusting for confounders. Results: Death occurred in 15,419 patients who underwent 39,562 (29%) echocardiograms. Median follow up duration was 3.6 years (IQR, 1.3-7.2), and median time between tests was 1.0 year (IQR, 0.2-2.3). The interaction between LVEF and ΔLVEF ( P < 0.001) demonstrated that a stable LVEF of 60-65% was associated with the best survival (Figure). A decreased LVEF was universally associated with increased mortality, while LVEFs that had increased showed non-linear interactions. In general, an LVEF of 35-55% that had increased from the previous test was associated with a similar or slightly lower mortality than the same LVEF that was unchanged or had decreased, while an LVEF >55% that had increased was associated with an increase in mortality risk compared to those with a stable LVEF. Conclusions: Changes in LVEF have a non-linear relationship with all-cause mortality. In general, a constant, stable LVEF associates with the lowest risk of mortality.

Introduction: Managing heart failure patients relies on mortality prediction models such as the Seattle Heart Failure (SHF) model. We hypothesized that a deep learning model that leverages echocardiography video data could improve prediction performance. Methods: We trained deep neural networks (DNN) to predict 1-year all-cause mortality using echocardiography videos and 158 additional clinical variables (labs, vitals, diagnoses, etc.) acquired from 34,362 patients (812,278 videos). We tested both the DNN and the SHF model on a separate cohort of 2,404 patients with heart failure who underwent 3,384 echocardiograms. We computed the area under the receiver operating characteristic curve (AUC) and its bootstrapped 95% confidence interval (CI) in the test set. Results: The DNN model (AUC of 0.76, 95% CI [0.74, 0.77]) outperformed the SHF model (AUC of 0.70, 95% CI [0.68, 0.71]). This superior performance was observed for patients with both reduced (n=2,026, AUC of 0.76 [0.74, 0.78] vs 0.70 [0.67, 0.72]) and preserved left ventricular ejection fraction (n=1,356, AUC of 0.75 [0.72, 0.78] vs 0.69 [0.66, 0.72]). A Cox Proportional Hazard survival analysis showed that, despite the prediction being for 1-year all-cause mortality, the result held long-term predictive power over the next 9 years with a superior hazard ratio of 2.9 [2.6, 3.2] for the DNN model compared to 2.2 [2, 2.4] for the SHF model. At mid-range operating points, the DNN model also maintained a higher negative predictive value, predicting survival, compared to the SHF model (89% vs 83%, respectively), while maintaining the same positive predictive value of 40%. Conclusion: A deep neural network that automatically analyzes echocardiography videos can outperform traditional risk modeling approaches such as the Seattle Heart Failure Model. This provides additional evidence that deep learning can help improve our ability to make clinically relevant predictions by leveraging complex datasets.

Background: Atrial fibrillation (AF) is associated with stroke, especially when AF goes undetected. Deep neural networks (DNN) can predict incident AF from a 12-lead resting ECG. We hypothesize that use of a DNN to predict new onset AF from an ECG may identify patients at risk of sustaining a potentially preventable AF-related stroke. Methods: We trained a DNN model to predict new-onset AF using 382,604 ECGs prior to 2010. We then evaluated the model performance on a test set of ECGs from 2010 through 2014 linked to patients in an institutional stroke registry. There were 181,969 patients in the test set with at least one ECG and no prior history of AF. Of those patients 3,497 (1.9%) had a stroke following an ECG that did not show AF. Within the set of patients with stroke, 375 had the stroke within 3 years of the ECG and were diagnosed with new AF between -3 and 365 days of the stroke. We considered these potentially preventable AF-related strokes. We report the sensitivity and positive predictive value (PPV) of the model for appropriately risk stratifying these 375 patients who sustained a potentially preventable AF-related stroke. Results: We used F β scores to identify different risk prediction thresholds (operating points) for the model. Operating points chosen by F 0.5 , F 1 , and F 2 scores identified 4, 12, and 21% of the population as high risk for the development of AF within 1 year (Figure 1). Screening 1, 4, 12, and 21% of the overall population resulted in PPV of 28, 21, 15, and 12%, respectively, for identification of new onset AF in one year. Using those same thresholds yielded sensitivities of 4, 17, 45, and 62% for identifying potentially preventable AF-related strokes. The different risk prediction thresholds resulted in a low (120-162) number needed to screen to detect one potentially preventable AF-related stroke at 3 years. Conclusions: Use of a deep learning model to predict new onset AF may identify patients at high risk of sustaining a potentially preventable AF-related stroke.

Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) causes coronavirus disease-19 (COVID-19), a respiratory illness that can result in hospitalization or death. We investigated associations between rare genetic variants and seven COVID-19 outcomes in 543,213 individuals, including 8,248 with COVID-19. After accounting for multiple testing, we did not identify any clear associations with rare variants either exome-wide or when specifically focusing on (i) 14 interferon pathway genes in which rare deleterious variants have been reported in severe COVID-19 patients; (ii) 167 genes located in COVID-19 GWAS risk loci; or (iii) 32 additional genes of immunologic relevance and/or therapeutic potential. Our analyses indicate there are no significant associations with rare protein-coding variants with detectable effect sizes at our current sample sizes. Analyses will be updated as additional data become available, with results publicly browsable athttps://rgc-covid19.regeneron.com.

BackgroundAtrial fibrillation (AF) is associated with substantial morbidity, especially when it goes undetected. If new onset AF could be predicted, targeted population screening could be used to find it early. We hypothesized that a deep neural network could predict new onset AF from the resting 12-lead electrocardiogram (ECG) and that this prediction may help prevent AF-related stroke.MethodsWe used 1.6M resting 12-lead ECG voltage-time traces from 430k patients collected from 1984-2019 in this study. Deep neural networks were trained to predict new onset AF (within 1 year) in patients without a history of AF. Performance was evaluated using areas under the receiver operating characteristic curve (AUROC) and precision-recall curve (AUPRC). We performed an incidence-free survival analysis for a period of 30 years following the ECG stratified by model predictions. To simulate real-world deployment, we trained a separate model using all ECGs prior to 2010 and evaluated model performance on a test set of ECGs from 2010 through 2014 that were linked to our stroke registry. We used standard metrics to explore different prediction thresholds for the model and also calculated how many AF-related strokes might be potentially prevented.ResultsThe AUROC and AUPRC were 0.83 and 0.21, respectively, for predicting new onset AF within 1 year of an ECG. Adding age and sex improved the AUROC to 0.85 and the AUPRC to 0.23. The hazard ratio for the predicted high- vs. low-risk groups over a 30-year span was 7.2 [95% confidence interval: 6.9 – 7.6]. In a simulated deployment scenario, using the F2 score to select the risk prediction threshold, the model predicted new onset AF at 1 year with a sensitivity of 69%, specificity of 81%, and positive predictive value (PPV) of 12%. This model correctly predicted new onset AF in 62% of all patients that experienced an AF-related stroke within 3 years of the ECG.ConclusionsDeep learning can predict new onset AF from the 12-lead ECG in patients with no prior history of AF. This prediction may prove useful in preventing AF-related strokes.

Machine learning promises to assist physicians with predictions of mortality and of other future clinical events by learning complex patterns from historical data, such as longitudinal electronic health records. Here we show that a convolutional neural network trained on raw pixel data in 812,278 echocardiographic videos from 34,362 individuals provides superior predictions of one-year all-cause mortality. The model’s predictions outperformed the widely used pooled cohort equations, the Seattle Heart Failure score (measured in an independent dataset of 2,404 patients with heart failure who underwent 3,384 echocardiograms), and a machine learning model involving 58 human-derived variables from echocardiograms and 100 clinical variables derived from electronic health records. We also show that cardiologists assisted by the model substantially improved the sensitivity of their predictions of one-year all-cause mortality by 13% while maintaining prediction specificity. Large unstructured datasets may enable deep learning to improve a wide range of clinical prediction models. A deep learning model trained on raw pixel data in hundreds of thousands of echocardiographic videos for the prediction of one-year all-cause mortality outperforms clinical scores and improves predictions by cardiologists.

Aims We investigated the relationship between clinically assessed left ventricular ejection fraction (LVEF) and survival in a large, heterogeneous clinical cohort. Methods and results Physician-reported LVEF on 403 977 echocardiograms from 203 135 patients were linked to all-cause mortality using electronic health records (1998–2018) from US regional healthcare system. Cox proportional hazards regression was used for analyses while adjusting for many patient characteristics including age, sex, and relevant comorbidities. A dataset including 45 531 echocardiograms and 35 976 patients from New Zealand was used to provide independent validation of analyses. During follow-up of the US cohort, 46 258 (23%) patients who had undergone 108 578 (27%) echocardiograms died. Overall, adjusted hazard ratios (HR) for mortality showed a u-shaped relationship for LVEF with a nadir of risk at an LVEF of 60–65%, a HR of 1.71 [95% confidence interval (CI) 1.64–1.77] when ≥70% and a HR of 1.73 (95% CI 1.66–1.80) at LVEF of 35–40%. Similar relationships with a nadir at 60–65% were observed in the validation dataset as well as for each age group and both sexes. The results were similar after further adjustments for conditions associated with an elevated LVEF, including mitral regurgitation, increased wall thickness, and anaemia and when restricted to patients reported to have heart failure at the time of the echocardiogram. Conclusion Deviation of LVEF from 60% to 65% is associated with poorer survival regardless of age, sex, or other relevant comorbidities such as heart failure. These results may herald the recognition of a new phenotype characterized by supra-normal LVEF.

Angiopoietin-like 4 (ANGPTL4) is an endogenous inhibitor of lipoprotein lipase that modulates lipid levels, coronary atherosclerosis risk, and nutrient partitioning. We hypothesize that loss of ANGPTL4 function might improve glucose homeostasis and decrease risk of type 2 diabetes (T2D). We investigate protein-altering variants in ANGPTL4 among 58,124 participants in the DiscovEHR human genetics study, with follow-up studies in 82,766 T2D cases and 498,761 controls. Carriers of p.E40K, a variant that abolishes ANGPTL4 ability to inhibit lipoprotein lipase, have lower odds of T2D (odds ratio 0.89, 95% confidence interval 0.85–0.92, p = 6.3 × 10−10), lower fasting glucose, and greater insulin sensitivity. Predicted loss-of-function variants are associated with lower odds of T2D among 32,015 cases and 84,006 controls (odds ratio 0.71, 95% confidence interval 0.49–0.99, p = 0.041). Functional studies in Angptl4-deficient mice confirm improved insulin sensitivity and glucose homeostasis. In conclusion, genetic inactivation of ANGPTL4 is associated with improved glucose homeostasis and reduced risk of T2D., Genetic variation in ANGPTL4 is associated with lipid traits. Here, the authors find that predicted loss-of-function variants in ANGPTL4 are associated with glucose homeostasis and reduced risk of type 2 diabetes and that Angptl4−/− mice on a high-fat diet show improved insulin sensitivity.

Health care delivery is increasingly influenced by the emerging concepts of precision health and the learning health care system. Although not synonymous with precision health, genomics is a key enabler of individualized care. Delivering patient-centered, genomics-informed care based on individual-level data in the current national landscape of health care delivery is a daunting challenge. Problems to overcome include data generation, analysis, storage, and transfer; knowledge management and representation for patients and providers at the point of care; process management; and outcomes definition, collection, and analysis. Development, testing, and implementation of a genomics-informed program requires multidisciplinary collaboration and building the concepts of precision health into a multilevel implementation framework. Using the principles of a learning health care system provides a promising solution. This article describes the implementation of population-based genomic medicine in an integrated learning health care system-a working example of a precision health program.

Large-scale human genetics studies are ascertaining increasing proportions of populations as they continue growing in both number and scale. As a result, the amount of cryptic relatedness within these study cohorts is growing rapidly and has significant implications on downstream analyses. We demonstrate this growth empirically among the first 92,455 exomes from the DiscovEHR cohort and, via a custom simulation framework we developed called SimProgeny, show that these measures are in-line with expectations given the underlying population and ascertainment approach. For example, we identified ∼66,000 close (first- and second-degree) relationships within DiscovEHR involving 55.6% of study participants. Our simulation results project that >70% of the cohort will be involved in these close relationships as DiscovEHR scales to 250,000 recruited individuals. We reconstructed 12,574 pedigrees using these relationships (including 2,192 nuclear families) and leveraged them for multiple applications. The pedigrees substantially improved the phasing accuracy of 20,947 rare, deleterious compound heterozygous mutations. Reconstructed nuclear families were critical for identifying 3,415 de novo mutations in ∼1,783 genes. Finally, we demonstrate the segregation of known and suspected disease-causing mutations through reconstructed pedigrees, including a tandem duplication in LDLR causing familial hypercholesterolemia. In summary, this work highlights the prevalence of cryptic relatedness expected among large healthcare population genomic studies and demonstrates several analyses that are uniquely enabled by large amounts of cryptic relatedness.

Most phenome-wide association studies (PheWASs) to date have used a small to moderate number of SNPs for association with phenotypic data. We performed a large-scale single-cohort PheWAS, using electronic health record (EHR)-derived case-control status for 541 diagnoses using International Classification of Disease version 9 (ICD-9) codes and 25 median clinical laboratory measures. We calculated associations between these diagnoses and traits with ∼630,000 common frequency SNPs with minor allele frequency > 0.01 for 38,662 individuals. In this landscape PheWAS, we explored results within diseases and traits, comparing results to those previously reported in genome-wide association studies (GWASs), as well as previously published PheWASs. We further leveraged the context of functional impact from protein-coding to regulatory regions, providing a deeper interpretation of these associations. The comprehensive nature of this PheWAS allows for novel hypothesis generation, the identification of phenotypes for further study for future phenotypic algorithm development, and identification of cross-phenotype associations.

In a retrospective study of 19 171 mother-child dyads, elevated random plasma glucose values during early pregnancy were directly correlated with increased risk for congenital heart disease in offspring. Plasma glucose levels proximal to the period of cardiac development may represent a modifiable risk factor for congenital heart disease in expectant mothers without diabetes.

BackgroundLoss-of-function variants in the angiopoietin-like 3 gene (ANGPTL3) have been associated with decreased plasma levels of triglycerides, low-density lipoprotein (LDL) cholesterol, and high-density lipoprotein (HDL) cholesterol. It is not known whether such variants or therapeutic antagonism of ANGPTL3 are associated with a reduced risk of atherosclerotic cardiovascular disease. MethodsWe sequenced the exons of ANGPTL3 in 58,335 participants in the DiscovEHR human genetics study. We performed tests of association for loss-of-function variants in ANGPTL3 with lipid levels and with coronary artery disease in 13,102 case patients and 40,430 controls from the DiscovEHR study, with follow-up studies involving 23,317 case patients and 107,166 controls from four population studies. We also tested the effects of a human monoclonal antibody, evinacumab, against Angptl3 in dyslipidemic mice and against ANGPTL3 in healthy human volunteers with elevated levels of triglycerides or LDL cholesterol. ResultsIn th...

Background: Arrhythmogenic right ventricular cardiomyopathy (ARVC) is associated with variants in desmosome genes. Secondary findings of pathogenic/likely pathogenic variants, primarily loss-of-function (LOF) variants, are recommended for clinical reporting; however, their prevalence and associated phenotype in a general clinical population are not fully characterized. Methods: From whole-exome sequencing of 61 019 individuals in the DiscovEHR cohort, we screened for putative loss-of-function variants in PKP2 , DSC2 , DSG2 , and DSP . We evaluated measures from prior clinical ECG and echocardiograms, manually over-read to evaluate ARVC diagnostic criteria, and performed a PheWAS (phenome-wide association study). Finally, we estimated expected penetrance using Bayesian inference. Results: One hundred forty individuals (0.23%; 59±18 years old at last encounter; 33% male) had an ARVC variant (G + ). None had an existing diagnosis of ARVC in the electronic health record, nor significant differences in prior ECG or echocardiogram findings compared with matched controls without variants. Several G + individuals satisfied major repolarization (n=4) and ventricular function (n=5) criteria, but this prevalence matched controls. PheWAS showed no significant associations of other heart disease diagnoses. Combining our best genetic and disease prevalence estimates yields an estimated penetrance of 6.0%. Conclusions: The prevalence of ARVC loss-of-function variants is ≈1:435 in a general clinical population of predominantly European descent, but with limited electronic health record-based evidence of phenotypic association in our population, consistent with a low penetrance estimate. Prospective deep phenotyping and longitudinal follow-up of a large sequenced cohort is needed to determine the true clinical relevance of an incidentally identified ARVC loss-of-function variant.

Background: Truncating variants in the Titin gene (TTNtvs) are common in individuals with idiopathic dilated cardiomyopathy (DCM). However, a comprehensive genomics-first evaluation of the impact of TTNtvs in different clinical contexts, and the evaluation of modifiers such as genetic ancestry, has not been performed. Methods: We reviewed whole exome sequence data for >71 000 individuals (61 040 from the Geisinger MyCode Community Health Initiative (2007 to present) and 10 273 from the PennMedicine BioBank (2013 to present) to identify anyone with TTNtvs. We further selected individuals with TTNtvs in exons highly expressed in the heart (proportion spliced in [PSI] >0.9). Using linked electronic health records, we evaluated associations of TTNtvs with diagnoses and quantitative echocardiographic measures, including subanalyses for individuals with and without DCM diagnoses. We also reviewed data from the Jackson Heart Study to validate specific analyses for individuals of African ancestry. Results: Identified with a TTNtv in a highly expressed exon (hiPSI) were 1.2% individuals in PennMedicine BioBank and 0.6% at Geisinger. The presence of a hiPSI TTNtv was associated with increased odds of DCM in individuals of European ancestry (odds ratio [95% CI]: 18.7 [9.1–39.4] {PennMedicine BioBank} and 10.8 [7.0–16.0] {Geisinger}). hiPSI TTNtvs were not associated with DCM in individuals of African ancestry, despite a high DCM prevalence (odds ratio, 1.8 [0.2–13.7]; P =0.57). Among 244 individuals of European ancestry with DCM in PennMedicine BioBank, hiPSI TTNtv carriers had lower left ventricular ejection fraction (β=–12%, P =3×10 –7 ), and increased left ventricular diameter (β=0.65 cm, P =9×10 –3 ). In the Geisinger cohort, hiPSI TTNtv carriers without a cardiomyopathy diagnosis had more atrial fibrillation (odds ratio, 2.4 [1.6–3.6]) and heart failure (odds ratio, 3.8 [2.4–6.0]), and lower left ventricular ejection fraction (β=–3.4%, P =1×10 –7 ). Conclusions: Individuals of European ancestry with hiPSI TTNtv have an abnormal cardiac phenotype characterized by lower left ventricular ejection fraction, irrespective of the clinical manifestation of cardiomyopathy. Associations with arrhythmias, including atrial fibrillation, were observed even when controlling for cardiomyopathy diagnosis. In contrast, no association between hiPSI TTNtvs and DCM was discerned among individuals of African ancestry. Given these findings, clinical identification of hiPSI TTNtv carriers may alter clinical management strategies.

Maturity onset diabetes of the young (MODY) is a monogenic form of diabetes attributed to a highly penetrant variant in one of several genes, including GCK. GCK-MODY rarely requires pharmacotherapy, however, the clinical management of pregnant mothers depends on whether the fetus inherited the GCK variant. The classic GCK-MODY phenotype is young (diagnosed before 25 years), lean (BMI < 25 mg/kg2), with stable, mild hyperglycemia (Hba1c 5.8-7.6%). We hypothesized that strict application of these guidelines has led to under-reporting of GCK-MODY (prevalence 1.1 per 1000). In a cohort of 92,412 individuals from the DiscovEHR study, we identified 252 heterozygous carriers of 88 rare GCK variants (minor allele frequency < 0.02%). None had an existing diagnosis of monogenic diabetes. The median BMI of GCK carriers was 35 mg/kg2, compared to 31 mg/kg2 for nondiabetic noncarriers and 38 mg/kg2 for noncarriers with type 2 diabetes (T2D). Excluding variants with no evidence of hyperglycemia in 50% of the carriers resulted in 70 variants in 213 individuals. GCK-MODY variant carriers had mean HbA1c, fasting blood glucose (FBG), and random glucose measurements that were greater than nondiabetic non-carriers but less than those with T2D. Kaplan-Meier analysis adjusted for left-truncation bias showed that the age at which 50% of GCK-MODY subjects diagnosed with impaired fasting glucose and diabetes is 35 and 55 years, respectively, compared to 60 and 65 years in non-carriers with T2D (p< 2e-16). Within-individual variability of lifetime FBG measurements of GCK-MODY carriers was less than noncarriers with T2D (GCK-MODY, 7.9 ± 2.0 mg/dL; T2D, 17.6 ± 0.03 mg/dl, p Disclosure U.L. Mirshahi: None. J.M. Goehringer: None. Y. Hu: None. A.H. Wardeh: None. J. Williams: None. C.B. Manney: None. J.C. Staples: Employee; Self; Regeneron Pharmaceuticals. J.B. Leader: None. A.R. Shuldiner: Employee; Self; Regeneron Pharmaceuticals. Stock/Shareholder; Self; Rhythm Pharmaceuticals, Inc. T.I. Pollin: Other Relationship; Self; Regeneron Genetics Center. D.J. Carey: None.

Lipodystrophies are a group of disorders characterized by absence or loss of adipose tissue and abnormal fat distribution, commonly accompanied by metabolic dysregulation. Although considered rare disorders, their prevalence in the general population is not well understood. We aimed to evaluate the clinical and genetic prevalence of lipodystrophy disorders in a large clinical care cohort. We interrogated the electronic health record (EHR) information of >1.3 million adults from the Geisinger Health System for lipodystrophy diagnostic codes. We estimate a clinical prevalence of disease of 1 in 20,000 individuals. We performed genetic analyses in individuals with available genomic data to identify variants associated with inherited lipodystrophies and examined their EHR for comorbidities associated with lipodystrophy. We identified 16 individuals carrying the p.R482Q pathogenic variant in LMNA associated with Dunnigan familial partial lipodystrophy. Four had a clinical diagnosis of lipodystrophy, whereas the remaining had no documented clinical diagnosis despite having accompanying metabolic abnormalities. We observed a lipodystrophy-associated variant carrier frequency of 1 in 3,082 individuals in our cohort with substantial burden of metabolic dysregulation. We estimate a genetic prevalence of disease of ∼1 in 7,000 in the general population. Partial lipodystrophy is an underdiagnosed condition. and its prevalence, as defined molecularly, is higher than previously reported. Genetically guided stratification of patients with common metabolic disorders, like diabetes and dyslipidemia, is an important step toward precision medicine.

Background Familial hypobetalipoproteinemia is a genetic disorder caused by rare protein-truncating variants (PTV) in the gene encoding APOB (apolipoprotein B), the major protein component of LDL (low-density lipoprotein) and triglyceride-rich lipoprotein particles. Whether heterozygous APOB deficiency is associated with decreased risk for coronary heart disease (CHD) is uncertain. We combined family-based and large scale gene-sequencing to characterize the association of rare PTVs in APOB with circulating LDL-C (LDL cholesterol), triglycerides, and risk for CHD. Methods We sequenced the APOB gene in 29 Japanese hypobetalipoproteinemia families, as well as 57 973 individuals derived from 12 CHD case-control studies-18 442 with early-onset CHD and 39 531 controls. We defined PTVs as variants that lead to a premature stop, disrupt canonical splice-sites, or lead to insertions/deletions that shift reading frame. We tested the association of rare APOB PTV carrier status with blood lipid levels and CHD. Results Among 29 familial hypobetalipoproteinemia families, 8 families harbored APOB PTVs. Carrying 1 APOB PTV was associated with 55 mg/dL lower LDL-C ( P=3×10-5) and 53% lower triglyceride level ( P=2×10-4). Among 12 case-control studies, an APOB PTV was present in 0.038% of CHD cases as compared to 0.092% of controls. APOB PTV carrier status was associated with a 43 mg/dL lower LDL-C ( P=2×10-7), a 30% decrease in triglycerides ( P=5×10-4), and a 72% lower risk for CHD (odds ratio, 0.28; 95% CI, 0.12-0.64; P=0.002). Conclusions Rare PTV mutations in APOB which are associated with lower LDL-C and reduced triglycerides also confer protection against CHD. Dr Peloso is supported by the National Heart, Lung, and Blood Institute (NHLBI) of the National Institutes of Health under Award Number K01HL125751. Dr Nomura was supported by the Yoshida Scholarship Foundation. Dr Khera is supported by an institutional grant from the Broad Institute of Massachusetts Institute of Technology (MIT) and Harvard (BroadIgnite), a K08 from the National Human Genome Research Institute (K08HG010155), and a Junior Faculty Award from the National Lipid Association. Dr Kathiresan is supported by a research scholar award from the Massachusetts General Hospital, the Donovan Family Foundation, and grant R01 HL127564 from the NHLBI. Funding for the EOMI study (Exome Sequencing Project Early-Onset Myocardial Infarction) was provided by grants RC2 HL103010 (HeartGO, Heart Grand Opportunity), RC2 HL102923 (LungGO, Lung Grand Opportunity), and RC2 HL102924 (WHISP) from the NHLBI. Exome sequencing was performed through grants RC2 HL102925 (BroadGO, Broad Grand Opportunity) and RC2 HL102926 (SeattleGO, Seattle Grand Opportunity) from the NHLBI. Exome sequencing in ATVB (Italian Atherosclerosis, Thrombosis, and Vascular Biology), the PROCARDIS study (Precocious Coronary Artery Disease), the OHS (Ottawa Heart Study), PROMIS (Pakistan Risk of Myocardial Infarction Study), South German MI study (South German Myocardial Infarction), and the JHS (Jackson Heart Study) was supported by grant 5U54HG003067 from the National Institutes of Health. Fieldwork, genotyping, and standard clinical chemistry assays in PROMIS were principally supported by grants awarded to the University of Cambridge from the British Heart Foundation, UK Medical Research Council, Wellcome Trust, European Union (EU) Framework 6–funded Bloodomics Integrated Project, Pfizer, Novartis, and Merck. Additional support for PROMIS was provided by the UK Medical Research Council (MR/L003120/1), British Heart Foundation (RG/13/13/30194), UK National Institute for Health Research Cambridge Biomedical Research Centre, European Research Council (268834), and European Commission Framework Programme 7 (HEALTH-F2-2012–279233). The Jackson Heart Study is supported by contracts HHSN268201300046C, HHSN268201300047C, HHSN268201300048C, HHSN268201300049C, and HHSN268201300050C from the NHLBI and the National Institute on Minority Health and Health Disparities. Dr Wilson is supported by U54GM115428 from the National Institute of General Medical Sciences. REGICOR study (Registre Gironí del COR [Gerona Heart Registry]) was supported by the Spanish Ministry of Economy and Innovation through the Carlos III Health Institute (Red Investigación Cardiovascular RD12/0042, PI09/90506), European Funds for Development (ERDF-FEDER), and by the Catalan Research and Technology Innovation Interdepartmental Commission (2014SGR240). Samples for the Leicester (Leicester Myocardial Infarction) cohort were collected as part of projects funded by the British Heart Foundation (British Heart Foundation Family Heart Study, RG2000010; UK Aneurysm Growth Study, CS/14/2/30841) and the National Institute for Health Research (NIHR Leicester Cardiovascular Biomedical Research Unit Biomedical Research Informatics Centre for Cardiovascular Science, IS_BRU_0211_20033). The South MI Study is supported by the German Federal Ministry of Education and Research (BMBF) in the context of the e:Med program (e:AtheroSysMed) and the FP7 European Union project CVgenes@target (261123). Additional grants were received from the Fondation Leducq (CADgenomics: Understanding Coronary Artery Disease Genes, 12CVD02). This study was also supported through the Deutsche Forschungsgemeinschaft cluster of excellence Inflammation at Interfaces and SFB 1123. The ATVB study was supported by a grant from RFPS-2007-3-644382 and Programma di ricerca Regione-Università 2010–2012 Area 1–Strategic Programmes–Regione Emilia-Romagna. The authors would like to thank the MyCode Community Health Initiative participants for their permission to utilize their health and genomics information in the DiscovEHR (DiscovEHR partnership of the Regeneron Genetics Center and Geisinger Health System) collaboration. The DiscovEHR study was funded, in part, by the Regeneron Genetics Center.

The electrocardiogram (ECG) is a widely used medical test, consisting of voltage versus time traces collected from surface recordings over the heart1. Here we hypothesized that a deep neural network (DNN) can predict an important future clinical event, 1-year all-cause mortality, from ECG voltage–time traces. By using ECGs collected over a 34-year period in a large regional health system, we trained a DNN with 1,169,662 12-lead resting ECGs obtained from 253,397 patients, in which 99,371 events occurred. The model achieved an area under the curve (AUC) of 0.88 on a held-out test set of 168,914 patients, in which 14,207 events occurred. Even within the large subset of patients (n = 45,285) with ECGs interpreted as ‘normal’ by a physician, the performance of the model in predicting 1-year mortality remained high (AUC = 0.85). A blinded survey of cardiologists demonstrated that many of the discriminating features of these normal ECGs were not apparent to expert reviewers. Finally, a Cox proportional-hazard model revealed a hazard ratio of 9.5 (P

SUMMARYThe UK Biobank is a prospective study of 502,543 individuals, combining extensive phenotypic and genotypic data with streamlined access for researchers around the world. Here we describe the first tranche of large-scale exome sequence data for 49,960 study participants, revealing approximately 4 million coding variants (of which ~98.4% have frequency < 1%). The data includes 231,631 predicted loss of function variants, a >10-fold increase compared to imputed sequence for the same participants. Nearly all genes (>97%) had ≥1 predicted loss of function carrier, and most genes (>69%) had ≥10 loss of function carriers. We illustrate the power of characterizing loss of function variation in this large population through association analyses across 1,741 phenotypes. In addition to replicating a range of established associations, we discover novel loss of function variants with large effects on disease traits, including PIEZO1 on varicose veins, COL6A1 on corneal resistance, MEPE on bone density, and IQGAP2 and GMPR on blood cell traits. We further demonstrate the value of exome sequencing by surveying the prevalence of pathogenic variants of clinical significance in this population, finding that 2% of the population has a medically actionable variant. Additionally, we leverage the phenotypic data to characterize the relationship between rare BRCA1 and BRCA2 pathogenic variants and cancer risk. Exomes from the first 49,960 participants are now made accessible to the scientific community and highlight the promise offered by genomic sequencing in large-scale population-based studies.

Background Empirical data on conditions that increase risk of coronavirus disease 2019 (COVID-19) progression are needed to identify high risk individuals. We performed a comprehensive quantitative assessment of pre-existing clinical phenotypes associated with COVID-19-related hospitalization. Methods Phenome-wide association study (PheWAS) of SARS-CoV-2-positive patients from an integrated health system (Geisinger) with system-level outpatient/inpatient COVID-19 testing capacity and retrospective electronic health record (EHR) data to assess pre-COVID-19 pandemic clinical phenotypes associated with hospital admission (hospitalization). Results Of 12,971 individuals tested for SARS-CoV-2 with sufficient pre-COVID-19 pandemic EHR data at Geisinger, 1604 were SARS-CoV-2 positive and 354 required hospitalization. We identified 21 clinical phenotypes in 5 disease categories meeting phenome-wide significance (P-4), including: six kidney phenotypes, e.g. end stage renal disease or stage 5 CKD (OR = 11.07, p = 1.96x10-8), six cardiovascular phenotypes, e.g. congestive heart failure (OR = 3.8, p = 3.24x10-5), five respiratory phenotypes, e.g. chronic airway obstruction (OR = 2.54, p = 3.71x10-5), and three metabolic phenotypes, e.g. type 2 diabetes (OR = 1.80, p = 7.51x10-5). Additional analyses defining CKD based on estimated glomerular filtration rate, confirmed high risk of hospitalization associated with pre-existing stage 4 CKD (OR 2.90, 95% CI: 1.47, 5.74), stage 5 CKD/dialysis (OR 8.83, 95% CI: 2.76, 28.27), and kidney transplant (OR 14.98, 95% CI: 2.77, 80.8) but not stage 3 CKD (OR 1.03, 95% CI: 0.71, 1.48). Conclusions This study provides quantitative estimates of the contribution of pre-existing clinical phenotypes to COVID-19 hospitalization and highlights kidney disorders as the strongest factors associated with hospitalization in an integrated US healthcare system.

Familial hypercholesterolemia (FH) is an underdiagnosed dominant genetic condition affecting approximately 0.4% of the population and has up to a 20-fold increased risk of coronary artery disease if untreated. Simple screening strategies have false positive rates greater than 95%. As part of the FH Foundation′s FIND FH initiative, we developed a classifier to identify potential FH patients using electronic health record (EHR) data at Stanford Health Care. We trained a random forest classifier using data from known patients (n = 197) and matched non-cases (n = 6590). Our classifier obtained a positive predictive value (PPV) of 0.88 and sensitivity of 0.75 on a held-out test-set. We evaluated the accuracy of the classifier′s predictions by chart review of 100 patients at risk of FH not included in the original dataset. The classifier correctly flagged 84% of patients at the highest probability threshold, with decreasing performance as the threshold lowers. In external validation on 466 FH patients (236 with genetically proven FH) and 5000 matched non-cases from the Geisinger Healthcare System our FH classifier achieved a PPV of 0.85. Our EHR-derived FH classifier is effective in finding candidate patients for further FH screening. Such machine learning guided strategies can lead to effective identification of the highest risk patients for enhanced management strategies.

Protein-coding genetic variants that strongly affect disease risk can provide important clues into disease pathogenesis. Here we report an exome sequence analysis of 20,791 type 2 diabetes (T2D) cases and 24,440 controls from five ancestries. We identify rare (minor allele frequency30 SLC30A8 alleles, and (b) within 12 gene sets, including those corresponding to T2D drug targets (p=6.1×10−3) and candidate genes from knockout mice (p=5.2×10−3). Within our study, the strongest T2D rare variant gene-level signals explain at most 25% of the heritability of the strongest common single-variant signals, and the rare variant gene-level effect sizes we observe in established T2D drug targets will require 110K-180K sequenced cases to exceed exome-wide significance. To help prioritize genes using associations from current smaller sample sizes, we present a Bayesian framework to recalibrate association p-values as posterior probabilities of association, estimating that reaching ppwww.type2diabetesgenetics.org.

Rationale : Therapies that inhibit CETP (cholesteryl ester transfer protein) have failed to demonstrate a reduction in risk for coronary heart disease (CHD). Human DNA sequence variants that truncate the CETP gene may provide insight into the efficacy of CETP inhibition. Objective: To test whether protein-truncating variants (PTVs) at the CETP gene were associated with plasma lipid levels and CHD. Methods and Results: We sequenced the exons of the CETP gene in 58 469 participants from 12 case–control studies (18 817 CHD cases, 39 652 CHD-free controls). We defined PTV as those that lead to a premature stop, disrupt canonical splice sites, or lead to insertions/deletions that shift frame. We also genotyped 1 Japanese-specific PTV in 27561 participants from 3 case–control studies (14 286 CHD cases, 13 275 CHD-free controls). We tested association of CETP PTV carrier status with both plasma lipids and CHD. Among 58 469 participants with CETP gene-sequencing data available, average age was 51.5 years and 43% were women; 1 in 975 participants carried a PTV at the CETP gene. Compared with noncarriers, carriers of PTV at CETP had higher high-density lipoprotein cholesterol (effect size, 22.6 mg/dL; 95% confidence interval, 18–27; P −4 ), lower low-density lipoprotein cholesterol (−12.2 mg/dL; 95% confidence interval, −23 to −0.98; P =0.033), and lower triglycerides (−6.3%; 95% confidence interval, −12 to −0.22; P =0.043). CETP PTV carrier status was associated with reduced risk for CHD (summary odds ratio, 0.70; 95% confidence interval, 0.54–0.90; P =5.1×10 −3 ). Conclusions: Compared with noncarriers, carriers of PTV at CETP displayed higher high-density lipoprotein cholesterol, lower low-density lipoprotein cholesterol, lower triglycerides, and lower risk for CHD.

Copy number variants (CNVs) are a substantial source of genomic variation and contribute to a wide range of human disorders. Gene-disrupting exonic CNVs have important clinical implications as they can underlie variability in disease presentation and susceptibility. The relationship between exonic CNVs and clinical traits has not been broadly explored at the population level, primarily due to technical challenges. We surveyed common and rare CNVs in the exome sequences of 50,726 adult DiscovEHR study participants with linked electronic health records (EHRs). We evaluated the diagnostic yield and clinical expressivity of known pathogenic CNVs, and performed tests of association with EHR-derived serum lipids, thereby evaluating the relationship between CNVs and complex traits and phenotypes in an unbiased, real-world clinical context. We identified CNVs from megabase to exon-level resolution, demonstrating reliable, high-throughput detection of clinically relevant exonic CNVs. In doing so, we created a catalog of high-confidence common and rare CNVs and refined population frequency estimates of known and novel gene-disrupting CNVs. Our survey among an unselected clinical population provides further evidence that neuropathy-associated duplications and deletions in 17p12 have similar population prevalence but are clinically under-diagnosed. Similarly, adults who harbor 22q11.2 deletions frequently had EHR documentation of neurodevelopmental/neuropsychiatric disorders and congenital anomalies, but not a formal genetic diagnosis (i.e., deletion). In an exome-wide association study of lipid levels, we identified a novel five-exon duplication within LDLR segregating in a large kindred with features of familial hypercholesterolemia. Exonic CNVs provide new opportunities to understand and diagnose human disease.

Importance The activity of lipoprotein lipase (LPL) is the rate-determining step in clearing triglyceride-rich lipoproteins from the circulation. Mutations that damage the LPL gene ( LPL ) lead to lifelong deficiency in enzymatic activity and can provide insight into the relationship of LPL to human disease. Objective To determine whether rare and/or common variants in LPL are associated with early-onset coronary artery disease (CAD). Design, Setting, and Participants In a cross-sectional study, LPL was sequenced in 10 CAD case-control cohorts of the multinational Myocardial Infarction Genetics Consortium and a nested CAD case-control cohort of the Geisinger Health System DiscovEHR cohort between 2010 and 2015. Common variants were genotyped in up to 305 699 individuals of the Global Lipids Genetics Consortium and up to 120 600 individuals of the CARDIoGRAM Exome Consortium between 2012 and 2014. Study-specific estimates were pooled via meta-analysis. Exposures Rare damaging mutations in LPL included loss-of-function variants and missense variants annotated as pathogenic in a human genetics database or predicted to be damaging by computer prediction algorithms trained to identify mutations that impair protein function. Common variants in the LPL gene region included those independently associated with circulating triglyceride levels. Main Outcomes and Measures Circulating lipid levels and CAD. Results Among 46 891 individuals with LPL gene sequencing data available, the mean (SD) age was 50 (12.6) years and 51% were female. A total of 188 participants (0.40%; 95% CI, 0.35%-0.46%) carried a damaging mutation in LPL , including 105 of 32 646 control participants (0.32%) and 83 of 14 245 participants with early-onset CAD (0.58%). Compared with 46 703 noncarriers, the 188 heterozygous carriers of an LPL damaging mutation displayed higher plasma triglyceride levels (19.6 mg/dL; 95% CI, 4.6-34.6 mg/dL) and higher odds of CAD (odds ratio = 1.84; 95% CI, 1.35-2.51; P LPL variants resulted in an odds ratio for CAD of 1.51 (95% CI, 1.39-1.64; P = 1.1 × 10 −22 ) per 1-SD increase in triglycerides. Conclusions and Relevance The presence of rare damaging mutations in LPL was significantly associated with higher triglyceride levels and presence of coronary artery disease. However, further research is needed to assess whether there are causal mechanisms by which heterozygous lipoprotein lipase deficiency could lead to coronary artery disease.

BACKGROUND: Varenicline, a selective α4β2 nicotinic acetylcholine receptor partial agonist, is a pharmacotherapy indicated for smoking cessation treatment. To our knowledge, no studies have described varenicline treatment adherence and efficacy from real-world treatment patterns in a U.S. primary care setting. OBJECTIVE: To estimate adherence to varenicline prescription orders and subsequent quit rates among smokers in a primary care setting. METHODS: In this retrospective cohort study, eligible patients were enrolled with Geisinger Health Plan, had an initial varenicline prescription written by a Geisinger provider between January 1, 2006, and December 31, 2009, and had a follow-up clinic visit within the subsequent 12 months. Adherence was derived from linking electronic prescriptions with adjudicated pharmacy claims. Smoking status was collected at each health care encounter. RESULTS: Of the 1,477 eligible patients, 823 (55.7%) were primary nonadherent, having failed to initiate on the prescribed varenicline therapy. Of the remaining 654 patients, 359 (54.9%) were adherent, having completed a full 12-week course of therapy, and 295 (45.1%) were partially adherent, having initiated but not completed the full course of therapy. A total of 521 patients (35.3%) ceased smoking during the 12-month follow-up period: 182 (50.7%) of the adherent cohort, 82 (27.8%) of the partially adherent population, and 257 (31.2%) of the nonadherent cohort. No significant difference was found in quit rates between the partially adherent and nonadherent patient cohorts (adjusted HR 0.88 [95% CI = 0.69-1.13]). However, patients adherent to the varenicline regimen were almost twice as likely to succeed in quitting smoking compared with completely nonadherent patients (HR 1.93 [95% CI = 1.59-2.33]). CONCLUSION: Smoking cessation occurred more often among individuals adherent to varenicline therapy; however, medication nonadherence was common. After prescribing varenicline, clinicians and payers could consider active patient follow-up to maximize adherence and optimize treatment outcomes.

Unleashing the power of precision medicine Precision medicine promises the ability to identify risks and treat patients on the basis of pathogenic genetic variation. Two studies combined exome sequencing results for over 50,000 people with their electronic health records. Dewey et al. found that ∼3.5% of individuals in their cohort had clinically actionable genetic variants. Many of these variants affected blood lipid levels that could influence cardiovascular health. Abul-Husn et al. extended these findings to investigate the genetics and treatment of familial hypercholesterolemia, a risk factor for cardiovascular disease, within their patient pool. Genetic screening helped identify at-risk patients who could benefit from increased treatment. Science , this issue p. 10.1126/science.aaf6814 , p. 10.1126/science.aaf7000

PurposeArrhythmogenic right ventricular cardiomyopathy (ARVC) is an inherited heart disease. Clinical follow-up of incidental findings in ARVC-associated genes is recommended. We aimed to determine the prevalence of disease thus ascertained.MethodsIndividuals (n = 30,716) underwent exome sequencing. Variants in PKP2, DSG2, DSC2, DSP, JUP, TMEM43, or TGFβ3 that were database-listed as pathogenic or likely pathogenic were identified and evidence-reviewed. For subjects with putative loss-of-function (pLOF) variants or variants of uncertain significance (VUS), electronic health records (EHR) were reviewed for ARVC diagnosis, diagnostic criteria, and International Classification of Diseases (ICD-9) codes.ResultsEighteen subjects had pLOF variants; none of these had an EHR diagnosis of ARVC. Of 14 patients with an electrocardiogram, one had a minor diagnostic criterion; the rest were normal. A total of 184 subjects had VUS, none of whom had an ARVC diagnosis. The proportion of subjects with VUS with major (4%) or minor (13%) electrocardiogram diagnostic criteria did not differ from that of variant-negative controls. ICD-9 codes showed no difference in defibrillator use, electrophysiologic abnormalities or nonischemic cardiomyopathies in patients with pLOF or VUSs compared with controls.ConclusionpLOF variants in an unselected cohort were not associated with ARVC phenotypes based on EHR review. The negative predictive value of EHR review remains uncertain.

The past decade has seen exponential growth in the numbers of sequenced and genotyped individuals and a corresponding increase in our ability of collect and catalogue phenotypic data for use in the clinic. We now face the challenge of integrating these diverse data in new ways new that can provide useful diagnostics and precise medical interventions for individual patients. One of the first steps in this process is to accurately map the phenotypic consequences of the genetic variation in human populations. The most common approach for this is the genome wide association study (GWAS). While this technique is relatively simple to implement for a given phenotype, the choice of how to define a phenotype is critical. It is becoming increasingly common for each individual in a GWAS cohort to have a large profile of quantitative measures. The standard approach is to test for associations with one measure at a time; however, there are many justifiable ways to define a set of phenotypes, and the genetic associations that are revealed will vary based on these definitions. Some phenotypes may only show a significant genetic association signal when considered together, such as through principle components analysis (PCA). Combining correlated measures may increase the power to detect association by reducing the noise present in individual variables and reduce the multiple hypothesis testing burden. Here we show that PCA and k-means clustering are two complimentary methods for identifying novel genotype-phenotype relationships within a set of quantitative human traits derived from the Geisinger Health System electronic health record (EHR). Using a diverse set of approaches for defining phenotype may yield more insights into the genetic architecture of complex traits and the findings presented here highlight a clear need for further investigation into other methods for defining the most relevant phenotypes in a set of variables. As the data of EHR continue to grow, addressing these issues will become increasingly important in our efforts to use genomic data effectively in medicine.

Unleashing the power of precision medicine Precision medicine promises the ability to identify risks and treat patients on the basis of pathogenic genetic variation. Two studies combined exome sequencing results for over 50,000 people with their electronic health records. Dewey et al. found that ∼3.5% of individuals in their cohort had clinically actionable genetic variants. Many of these variants affected blood lipid levels that could influence cardiovascular health. Abul-Husn et al. extended these findings to investigate the genetics and treatment of familial hypercholesterolemia, a risk factor for cardiovascular disease, within their patient pool. Genetic screening helped identify at-risk patients who could benefit from increased treatment. Science , this issue p. 10.1126/science.aaf6814 , p. 10.1126/science.aaf7000