The power of representation: Statistical analysis of diversity in US Alzheimer's disease genetics data.

INTRODUCTION: Alzheimer's disease (AD) is a complex disease influenced by genetics and environment. More than 75 susceptibility loci have been linked to late-onset AD, but most of these loci were discovered in genome-wide association studies (GWAS) exclusive to non-Hispanic White individuals. There are wide disparities in AD risk across racially stratified groups, and while these disparities are not due to genetic differences, underrepresentation in genetic research can further exacerbate and contribute to their persistence. We investigated the racial/ethnic representation of participants in United States (US)-based AD genetics and the statistical implications of current representation. METHODS: We compared racial/ethnic data of participants from array and sequencing studies in US AD genetics databases, including National Institute on Aging Genetics of Alzheimer's Disease Data Storage Site (NIAGADS) and NIAGADS Data Sharing Service (dssNIAGADS), to AD and related dementia (ADRD) prevalence and mortality. We then simulated the statistical power of these datasets to identify risk variants from non-White populations. RESULTS: There is insufficient statistical power (probability <80%) to detect single nucleotide polymorphisms (SNPs) with low to moderate effect sizes (odds ratio [OR]<1.5) using array data from Black and Hispanic participants; studies of Asian participants are not powered to detect variants OR <= 2. Using available and projected sequencing data from Black and Hispanic participants, risk variants with OR = 1.2 are detectable at high allele frequencies. Sample sizes remain insufficiently powered to detect these variants in Asian populations. DISCUSSION: AD genetics datasets are largely representative of US ADRD burden. However, there is a wide discrepancy between proportional representation and statistically meaningful representation. Most variation identified in GWAS of non-Hispanic White individuals have low to moderate effects. Comparable risk variants in non-White populations are not detectable given current sample sizes, which could lead to disparities in future studies and drug development. We urge AD genetics researchers and institutions to continue investing in recruiting diverse participants and use community-based participatory research practices.

Development and External Validation of Models to Predict Persistent Hypoxemic Respiratory Failure for Clinical Trial Enrichment.

OBJECTIVES: Improving the efficiency of clinical trials in acute hypoxemic respiratory failure (HRF) depends on enrichment strategies that minimize enrollment of patients who quickly resolve with existing care and focus on patients at high risk for persistent HRF. We aimed to develop parsimonious models predicting risk of persistent HRF using routine data from ICU admission and select research immune biomarkers. DESIGN: Prospective cohorts for derivation ( n = 630) and external validation ( n = 511). SETTING: Medical and surgical ICUs at two U.S. medical centers. PATIENTS: Adults with acute HRF defined as new invasive mechanical ventilation (IMV) and hypoxemia on the first calendar day after ICU admission. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We evaluated discrimination, calibration, and practical utility of models predicting persistent HRF risk (defined as ongoing IMV and hypoxemia on the third calendar day after admission): 1) a clinical model with least absolute shrinkage and selection operator (LASSO) selecting Pa o2 /F io2 , vasopressors, mean arterial pressure, bicarbonate, and acute respiratory distress syndrome as predictors; 2) a model adding interleukin-6 (IL-6) to clinical predictors; and 3) a comparator model with Pa o2 /F io2 alone, representing an existing strategy for enrichment. Forty-nine percent and 69% of patients had persistent HRF in derivation and validation sets, respectively. In validation, both LASSO (area under the receiver operating characteristic curve, 0.68; 95% CI, 0.64-0.73) and LASSO + IL-6 (0.71; 95% CI, 0.66-0.76) models had better discrimination than Pa o2 /F io2 (0.64; 95% CI, 0.59-0.69). Both models underestimated risk in lower risk deciles, but exhibited better calibration at relevant risk thresholds. Evaluating practical utility, both LASSO and LASSO + IL-6 models exhibited greater net benefit in decision curve analysis, and greater sample size savings in enrichment analysis, compared with Pa o2 /F io2 . The added utility of LASSO + IL-6 model over LASSO was modest. CONCLUSIONS: Parsimonious, interpretable models that predict persistent HRF may improve enrichment of trials testing HRF-targeted therapies and warrant future validation.

The Association of Upper Airway Anatomy with Brain Structure: The Multi-Ethnic Study of Atherosclerosis.

Sleep apnea, affecting an estimated 1 in 4 American adults, has been reported to be associated with both brain structural abnormality and impaired cognitive function. Obstructive sleep apnea is known to be affected by upper airway anatomy. To better understand the contribution of upper airway anatomy to pathways linking sleep apnea with impaired cognitive function, we investigated the association of upper airway anatomy with structural brain abnormalities. Based in the Multi-Ethnic Study of Atherosclerosis, a longitudinal cohort study of community-dwelling adults, a comprehensive sleep study and an MRI of the upper airway and brain were performed on 578 participants. Machine learning models were used to select from 74 upper airway measures those measures most associated with selected regional brain volumes and white matter hyperintensity volume. Linear regression assessed associations between the selected upper airway measures, sleep measures, and brain structure. Maxillary divergence was positively associated with hippocampus volume, and mandible length was negatively associated with total white and gray matter volume. Both coefficients were small (coefficients per standard deviation 0.063 mL, p = 0.04, and - 7.0 mL, p < 0.001 respectively), and not affected by adjustment for sleep study measures. Self-reported snoring >2 times per week was associated with larger hippocampus volume (coefficient 0.164 mL, p = 0.007), and higher percentage of time in the N3 sleep stage was associated with larger total white and gray matter volume (4.8 mL, p = 0.004). Despite associations of two upper airway anatomy measures with brain volume, the evidence did not suggest that these upper airway and brain structure associations were acting primarily through the pathway of sleep disturbance.

Proteome Network Analysis Identifies Potential Biomarkers for Brain Aging.

BACKGROUND: Alzheimer's disease and related dementias (ADRD) involve biological processes that begin years to decades before onset of clinical symptoms. The plasma proteome can offer insight into brain aging and risk of incident dementia among cognitively healthy adults. OBJECTIVE: To identify biomarkers and biological pathways associated with neuroimaging measures and incident dementia in two large community-based cohorts by applying a correlation-based network analysis to the plasma proteome. METHODS: Weighted co-expression network analysis of 1,305 plasma proteins identified four modules of co-expressed proteins, which were related to MRI brain volumes and risk of incident dementia over a median 20-year follow-up in Framingham Heart Study (FHS) Offspring cohort participants (n = 1,861). Analyses were replicated in the Cardiovascular Health Study (CHS) (n = 2,117, mean 6-year follow-up). RESULTS: Two proteomic modules, one related to protein clearance and synaptic maintenance (M2) and a second to inflammation (M4), were associated with total brain volume in FHS (M2: p = 0.014; M4: p = 4.2×10-5). These modules were not significantly associated with hippocampal volume, white matter hyperintensities, or incident all-cause or AD dementia. Associations with TCBV did not replicate in CHS, an older cohort with a greater burden of comorbidities. CONCLUSIONS: Proteome networks implicate an early role for biological pathways involving inflammation and synaptic function in preclinical brain atrophy, with implications for clinical dementia.

The association of upper airway anatomy with cognitive test performance: the Multi-Ethnic Study of Atherosclerosis.

BACKGROUND: Numerous upper airway anatomy characteristics are risk factors for sleep apnea, which affects 26% of older Americans, and more severe sleep apnea is associated with cognitive impairment. This study explores the pathophysiology and links between upper airway anatomy, sleep, and cognition. METHODS: Participants in the Multi-Ethnic Study of Atherosclerosis underwent an upper airway MRI, polysomnography to assess sleep measures including the apnea-hypopnea index (AHI) and completed the Cognitive Abilities Screening Instrument (CASI). Two model selection techniques selected from among 67 upper airway measures those that are most strongly associated with CASI score. The associations of selected upper airway measures with AHI, AHI with CASI score, and selected upper airway anatomy measures with CASI score, both alone and after adjustment for AHI, were assessed using linear regression. RESULTS: Soft palate volume, maxillary divergence, and upper facial height were significantly positively associated with higher CASI score, indicating better cognition. The coefficients were small, with a 1 standard deviation (SD) increase in these variables being associated with a 0.83, 0.75, and 0.70 point higher CASI score, respectively. Additional adjustment for AHI very slightly attenuated these associations. Larger soft palate volume was significantly associated with higher AHI (15% higher AHI (95% CI 2%,28%) per SD). Higher AHI was marginally associated with higher CASI score (0.43 (95% CI 0.01,0.85) per AHI doubling). CONCLUSIONS: Three upper airway measures were weakly but significantly associated with higher global cognitive test performance. Sleep apnea did not appear to be the mechanism through which these upper airway and cognition associations were acting. Further research on the selected upper airway measures is recommended.

Pharmacologic and Genetic Downregulation of Proprotein Convertase Subtilisin/Kexin Type 9 and Survival From Sepsis.

OBJECTIVES: Treatments that prevent sepsis complications are needed. Circulating lipid and protein assemblies-lipoproteins play critical roles in clearing pathogens from the bloodstream. We investigated whether early inhibition of proprotein convertase subtilisin/kexin type 9 (PCSK9) may accelerate bloodstream clearance of immunogenic bacterial lipids and improve sepsis outcomes. DESIGN: Genetic and clinical epidemiology, and experimental models. SETTING: Human genetics cohorts, secondary analysis of a phase 3 randomized clinical trial enrolling patients with cardiovascular disease (Evaluation of Cardiovascular Outcomes After an Acute Coronary Syndrome During Treatment With Alirocumab [ODYSSEY OUTCOMES]; NCT01663402), and experimental murine models of sepsis. PATIENTS OR SUBJECTS: Nine human cohorts with sepsis (total n = 12,514) were assessed for an association between sepsis mortality and PCSK9 loss-of-function (LOF) variants. Incident or fatal sepsis rates were evaluated among 18,884 participants in a post hoc analysis of ODYSSEY OUTCOMES. C57BI/6J mice were used in Pseudomonas aeruginosa and Staphylococcus aureus bacteremia sepsis models, and in lipopolysaccharide-induced animal models. INTERVENTIONS: Observational human cohort studies used genetic PCSK9 LOF variants as instrumental variables. ODYSSEY OUTCOMES participants were randomized to alirocumab or placebo. Mice were administered alirocumab, a PCSK9 inhibitor, at 5 mg/kg or 25 mg/kg subcutaneously, or isotype-matched control, 48 hours prior to the induction of bacterial sepsis. Mice did not receive other treatments for sepsis. MEASUREMENTS AND MAIN RESULTS: Across human cohort studies, the effect estimate for 28-day mortality after sepsis diagnosis associated with genetic PCSK9 LOF was odds ratio = 0.86 (95% CI, 0.67-1.10; p = 0.24). A significant association was present in antibiotic-treated patients. In ODYSSEY OUTCOMES, sepsis frequency and mortality were infrequent and did not significantly differ by group, although both were numerically lower with alirocumab vs. placebo (relative risk of death from sepsis for alirocumab vs. placebo, 0.62; 95% CI, 0.32-1.20; p = 0.15). Mice treated with alirocumab had lower endotoxin levels and improved survival. CONCLUSIONS: PCSK9 inhibition may improve clinical outcomes in sepsis in preventive, pretreatment settings.

Identification of circulating proteins associated with general cognitive function among middle-aged and older adults.

Identifying circulating proteins associated with cognitive function may point to biomarkers and molecular process of cognitive impairment. Few studies have investigated the association between circulating proteins and cognitive function. We identify 246 protein measures quantified by the SomaScan assay as associated with cognitive function (p < 4.9E-5, n up to 7289). Of these, 45 were replicated using SomaScan data, and three were replicated using Olink data at Bonferroni-corrected significance. Enrichment analysis linked the proteins associated with general cognitive function to cell signaling pathways and synapse architecture. Mendelian randomization analysis implicated higher levels of NECTIN2, a protein mediating viral entry into neuronal cells, with higher Alzheimer's disease (AD) risk (p = 2.5E-26). Levels of 14 other protein measures were implicated as consequences of AD susceptibility (p < 2.0E-4). Proteins implicated as causes or consequences of AD susceptibility may provide new insight into the potential relationship between immunity and AD susceptibility as well as potential therapeutic targets.

WHOLE GENOME SEQUENCING ANALYSIS OF BODY MASS INDEX IDENTIFIES NOVEL AFRICAN ANCESTRY-SPECIFIC RISK ALLELE.

Obesity is a major public health crisis associated with high mortality rates. Previous genome-wide association studies (GWAS) investigating body mass index (BMI) have largely relied on imputed data from European individuals. This study leveraged whole-genome sequencing (WGS) data from 88,873 participants from the Trans-Omics for Precision Medicine (TOPMed) Program, of which 51% were of non-European population groups. We discovered 18 BMI-associated signals (P < 5 × 10-9). Notably, we identified and replicated a novel low frequency single nucleotide polymorphism (SNP) in MTMR3 that was common in individuals of African descent. Using a diverse study population, we further identified two novel secondary signals in known BMI loci and pinpointed two likely causal variants in the POC5 and DMD loci. Our work demonstrates the benefits of combining WGS and diverse cohorts in expanding current catalog of variants and genes confer risk for obesity, bringing us one step closer to personalized medicine.

Serum NfL and GFAP are associated with incident dementia and dementia mortality in older adults: The cardiovascular health study.

INTRODUCTION: Circulating neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP) have been independently associated with dementia risk. Their additive association, and their associations with dementia-specific mortality, have not been investigated. METHODS: We associated serum NfL, GFAP, total tau ,and ubiquitin carboxyl-terminal hydrolase-L1, measured in 1712 dementia-free adults, with 19-year incident dementia and dementia-specific mortality risk, and with 3-year cognitive decline. RESULTS: In adjusted models, being in the highest versus lowest tertile of NfL or GFAP associated with a hazard ratio (HR) of 1.49 (1.20-1.84) and 1.38 (1.15-1.66) for incident dementia, and 2.87 (1.79-4.61) and 2.76 (1.73-4.40) for dementia-specific mortality. Joint third versus first tertile exposure further increased risk; HR = 2.06 (1.60-2.67) and 9.22 (4.48-18.9). NfL was independently associated with accelerated cognitive decline. DISCUSSION: Circulating NfL and GFAP may, independently and jointly, provide useful clinical insight regarding dementia risk and prognosis.

Whole-exome sequence analysis of anthropometric traits illustrates challenges in identifying effects of rare genetic variants.

Anthropometric traits, measuring body size and shape, are highly heritable and significant clinical risk factors for cardiometabolic disorders. These traits have been extensively studied in genome-wide association studies (GWASs), with hundreds of genome-wide significant loci identified. We performed a whole-exome sequence analysis of the genetics of height, body mass index (BMI) and waist/hip ratio (WHR). We meta-analyzed single-variant and gene-based associations of whole-exome sequence variation with height, BMI, and WHR in up to 22,004 individuals, and we assessed replication of our findings in up to 16,418 individuals from 10 independent cohorts from Trans-Omics for Precision Medicine (TOPMed). We identified four trait associations with single-nucleotide variants (SNVs; two for height and two for BMI) and replicated the LECT2 gene association with height. Our expression quantitative trait locus (eQTL) analysis within previously reported GWAS loci implicated CEP63 and RFT1 as potential functional genes for known height loci. We further assessed enrichment of SNVs, which were monogenic or syndromic variants within loci associated with our three traits. This led to the significant enrichment results for height, whereas we observed no Bonferroni-corrected significance for all SNVs. With a sample size of ∼20,000 whole-exome sequences in our discovery dataset, our findings demonstrate the importance of genomic sequencing in genetic association studies, yet they also illustrate the challenges in identifying effects of rare genetic variants.

Beyond the Individual: Community-Centric Approaches to Increase Diversity in Biomedical Research.

Community-centric approaches to engage underrepresented populations-including community engagement, community-level consent practices, and capacity development for research-are means to enhance diversity in biomedical research populations in a more ethical way. Low diversity is a known problem in biomedical research that presents challenges in translating the benefits of research to the global population. Through long-term partnerships built on trust and collaboration, communities who would otherwise avoid research may be more willing to participate. When communities are engaged in research as partners and research questions are motivated by community health priorities, research is more meaningful and research methods are more respectful. Conversely, a lack of consultation throughout the research process can further alienate the very communities that these efforts are designed to engage. A number of underserved populations-for example American Indian and Alaska Native peoples-may value the benefits of research to a community equally or more than individual benefits. A community's autonomy must be considered, particularly when that community is not adequately protected by traditional informed consent processes. Opportunities for capacity development to support collaborative partnerships between communities and researchers are required to support engagement and understanding of the research process. Changes to research processes and infrastructure that encourage a higher level of research oversight within the community should be supported. In this paper, we present approaches that may improve diversity and equitable access to research and the delivery of health innovations for people that have historically been left out of biomedical research.

Association of immune cell subsets with incident heart failure in two population-based cohorts.

AIMS: Circulating inflammatory markers are associated with incident heart failure (HF), but prospective data on associations of immune cell subsets with incident HF are lacking. We determined the associations of immune cell subsets with incident HF as well as HF subtypes [with reduced ejection fraction (HFrEF) and preserved ejection fraction (HFpEF)]. METHODS AND RESULTS: Peripheral blood immune cell subsets were measured in adults from the Multi-Ethnic Study of Atherosclerosis (MESA) and Cardiovascular Health Study (CHS). Cox proportional hazard models adjusted for demographics, HF risk factors, and cytomegalovirus serostatus were used to evaluate the association of the immune cell subsets with incident HF. The average age of the MESA cohort at the time of immune cell measurements was 63.0 ± 10.4 years with 51% women, and in the CHS cohort, it was 79.6 ± 4.4 years with 62% women. In the meta-analysis of CHS and MESA, a higher proportion of CD4+ T helper (Th) 1 cells (per one standard deviation) was associated with a lower risk of incident HF [hazard ratio (HR) 0.91, (95% CI 0.83-0.99), P = 0.03]. Specifically, higher proportion of CD4+ Th1 cells was significantly associated with a lower risk of HFrEF [HR 0.73, (95% CI 0.62-0.85), <0.001] after correction for multiple testing. No association was observed with HFpEF. No other cell subsets were associated with incident HF. CONCLUSIONS: We observed that higher proportions of CD4+ Th1 cells were associated with a lower risk of incident HFrEF in two distinct population-based cohorts, with similar effect sizes in both cohorts demonstrating replicability. Although unexpected, the consistency of this finding across cohorts merits further investigation.

Genome-wide meta-analyses reveal novel loci for verbal short-term memory and learning.

Understanding the genomic basis of memory processes may help in combating neurodegenerative disorders. Hence, we examined the associations of common genetic variants with verbal short-term memory and verbal learning in adults without dementia or stroke (N = 53,637). We identified novel loci in the intronic region of CDH18, and at 13q21 and 3p21.1, as well as an expected signal in the APOE/APOC1/TOMM40 region. These results replicated in an independent sample. Functional and bioinformatic analyses supported many of these loci and further implicated POC1. We showed that polygenic score for verbal learning associated with brain activation in right parieto-occipital region during working memory task. Finally, we showed genetic correlations of these memory traits with several neurocognitive and health outcomes. Our findings suggest a role of several genomic loci in verbal memory processes.

Proteomics and Population Biology in the Cardiovascular Health Study (CHS): design of a study with mentored access and active data sharing.

BACKGROUND: In the last decade, genomic studies have identified and replicated thousands of genetic associations with measures of health and disease and contributed to the understanding of the etiology of a variety of health conditions. Proteins are key biomarkers in clinical medicine and often drug-therapy targets. Like genomics, proteomics can advance our understanding of biology. METHODS AND RESULTS: In the setting of the Cardiovascular Health Study (CHS), a cohort study of older adults, an aptamer-based method that has high sensitivity for low-abundance proteins was used to assay 4979 proteins in frozen, stored plasma from 3188 participants (61% women, mean age 74 years). CHS provides active support, including central analysis, for seven phenotype-specific working groups (WGs). Each CHS WG is led by one or two senior investigators and includes 10 to 20 early or mid-career scientists. In this setting of mentored access, the proteomic data and analytic methods are widely shared with the WGs and investigators so that they may evaluate associations between baseline levels of circulating proteins and the incidence of a variety of health outcomes in prospective cohort analyses. We describe the design of CHS, the CHS Proteomics Study, characteristics of participants, quality control measures, and structural characteristics of the data provided to CHS WGs. We additionally highlight plans for validation and replication of novel proteomic associations. CONCLUSION: The CHS Proteomics Study offers an opportunity for collaborative data sharing to improve our understanding of the etiology of a variety of health conditions in older adults.

Monocyte subsets, T cell activation profiles, and stroke in men and women: The Multi-Ethnic Study of Atherosclerosis and Cardiovascular Health Study.

BACKGROUND AND AIMS: Despite mechanistic data implicating unresolving inflammation in stroke pathogenesis, data regarding circulating immune cell phenotypes - key determinants of inflammation propagation versus resolution - and incident stroke are lacking. Therefore, we aimed to comprehensively define associations of circulating immune phenotypes and activation profiles with incident stroke. METHODS: We investigated circulating leukocyte phenotypes and activation profiles with incident adjudicated stroke in 2104 diverse adults from the Multi-Ethnic Study of Atherosclerosis (MESA) followed over a median of 16.6 years. Cryopreserved cells from the MESA baseline examination were thawed and myeloid and lymphoid lineage cell subsets were measured using polychromatic flow cytometry and intracellular cytokine activation staining. We analyzed multivariable-adjusted associations of cell phenotypes, as a proportion of parent cell subsets, with incident stroke (overall) and ischemic stroke using Cox regression models. RESULTS: We observed associations of intermediate monocytes, early-activated CD4+ T cells, and both CD4+ and CD8+ T cells producing interleukin-4 after cytokine stimulation (Th2 and Tc2, respectively) with higher risk for incident stroke; effect sizes ranged from 35% to 62% relative increases in risk for stroke. Meanwhile, differentiated and memory T cell phenotypes were associated with lower risk for incident stroke. In sex-stratified analyses, positive and negative associations were especially strong among men but null among women. CONCLUSIONS: Circulating IL-4 producing T cells and intermediate monocytes were significantly associated with incident stroke over nearly two decades of follow-up. These associations were stronger among men and not among women. Further translational studies are warranted to define more precise targets for prognosis and intervention.

Meta-analysis of genome-wide association studies identifies ancestry-specific associations underlying circulating total tau levels.

Circulating total-tau levels can be used as an endophenotype to identify genetic risk factors for tauopathies and related neurological disorders. Here, we confirmed and better characterized the association of the 17q21 MAPT locus with circulating total-tau in 14,721 European participants and identified three novel loci in 953 African American participants (4q31, 5p13, and 6q25) at P < 5 × 10-8. We additionally detected 14 novel loci at P < 5 × 10-7, specific to either Europeans or African Americans. Using whole-exome sequence data in 2,279 European participants, we identified ten genes associated with circulating total-tau when aggregating rare variants. Our genetic study sheds light on genes reported to be associated with neurological diseases including stroke, Alzheimer's, and Parkinson's (F5, MAP1B, and BCAS3), with Alzheimer's pathological hallmarks (ADAMTS12, IL15, and FHIT), or with an important function in the brain (PARD3, ELFN2, UBASH3B, SLIT3, and NSD3), and suggests that the genetic architecture of circulating total-tau may differ according to ancestry.

Identifying risk of adverse outcomes in COVID-19 patients via artificial intelligence-powered analysis of 12-lead intake electrocardiogram.

BACKGROUND: Adverse events in COVID-19 are difficult to predict. Risk stratification is encumbered by the need to protect healthcare workers. We hypothesize that artificial intelligence (AI) can help identify subtle signs of myocardial involvement in the 12-lead electrocardiogram (ECG), which could help predict complications. OBJECTIVE: Use intake ECGs from COVID-19 patients to train AI models to predict risk of mortality or major adverse cardiovascular events (MACE). METHODS: We studied intake ECGs from 1448 COVID-19 patients (60.5% male, aged 63.4 ± 16.9 years). Records were labeled by mortality (death vs discharge) or MACE (no events vs arrhythmic, heart failure [HF], or thromboembolic [TE] events), then used to train AI models; these were compared to conventional regression models developed using demographic and comorbidity data. RESULTS: A total of 245 (17.7%) patients died (67.3% male, aged 74.5 ± 14.4 years); 352 (24.4%) experienced at least 1 MACE (119 arrhythmic, 107 HF, 130 TE). AI models predicted mortality and MACE with area under the curve (AUC) values of 0.60 ± 0.05 and 0.55 ± 0.07, respectively; these were comparable to AUC values for conventional models (0.73 ± 0.07 and 0.65 ± 0.10). There were no prominent temporal trends in mortality rate or MACE incidence in our cohort; holdout testing with data from after a cutoff date (June 9, 2020) did not degrade model performance. CONCLUSION: Using intake ECGs alone, our AI models had limited ability to predict hospitalized COVID-19 patients' risk of mortality or MACE. Our models' accuracy was comparable to that of conventional models built using more in-depth information, but translation to clinical use would require higher sensitivity and positive predictive value. In the future, we hope that mixed-input AI models utilizing both ECG and clinical data may be developed to enhance predictive accuracy.

Single Nucleotide Variant in FAS Associates With Organ Failure and Soluble Fas Cell Surface Death Receptor in Critical Illness.

OBJECTIVES: Multiple organ failure in critically ill patients is associated with poor prognosis, but biomarkers contributory to pathogenesis are unknown. Previous studies support a role for Fas cell surface death receptor (Fas)-mediated apoptosis in organ dysfunction. Our objectives were to test for associations between soluble Fas and multiple organ failure, identify protein quantitative trait loci, and determine associations between genetic variants and multiple organ failure. DESIGN: Retrospective observational cohort study. SETTING: Four academic ICUs at U.S. hospitals. PATIENTS: Genetic analyses were completed in a discovery (n = 1,589) and validation set (n = 863). Fas gene expression and flow cytometry studies were completed in outpatient research participants (n = 250). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: In discovery and validation sets of critically ill patients, we tested for associations between enrollment plasma soluble Fas concentrations and Sequential Organ Failure Assessment score on day 3. We conducted a genome-wide association study of plasma soluble Fas (discovery n = 1,042) and carried forward a single nucleotide variant in the FAS gene, rs982764, for validation (n = 863). We further tested whether the single nucleotide variant in FAS (rs982764) was associated with Sequential Organ Failure Assessment score, FAS transcriptional isoforms, and Fas cell surface expression. Higher plasma soluble Fas was associated with higher day 3 Sequential Organ Failure Assessment scores in both the discovery (ß = 4.07; p < 0.001) and validation (ß = 6.96; p < 0.001) sets. A single nucleotide variant in FAS (rs982764G) was associated with lower plasma soluble Fas concentrations and lower day 3 Sequential Organ Failure Assessment score in meta-analysis (-0.21; p = 0.02). Single nucleotide variant rs982764G was also associated with a lower relative expression of the transcript for soluble as opposed to transmembrane Fas and higher cell surface expression of Fas on CD4+ T cells. CONCLUSIONS: We found that single nucleotide variant rs982764G was associated with lower plasma soluble Fas concentrations in a discovery and validation population, and single nucleotide variant rs982764G was also associated with lower organ dysfunction on day 3. These findings support further study of the Fas pathway as a potential mediator of organ dysfunction in critically ill patients.