No abstract

Background and aims The prevalence and significance of digestive manifestations in COVID-19 remain uncertain. We aimed to assess the prevalence, spectrum, severity, and significance of digestive manifestations in patients hospitalized with COVID-19. Methods Consecutive patients hospitalized with COVID-19 were identified across a geographically diverse alliance of medical centers in North America. Data pertaining to baseline characteristics, symptomatology, laboratory assessment, imaging, and endoscopic findings from the time of symptom onset until discharge or death were manually abstracted from electronic health records to characterize the prevalence, spectrum, and severity of digestive manifestations. Regression analyses were performed to evaluate the association between digestive manifestations and severe outcomes related to COVID-19. Findings A total of 1992 patients across 36 centers met eligibility criteria and were included. Overall, 53% of patients experienced at least one gastrointestinal symptom at any time during their illness, most commonly diarrhea (34%), nausea (27%), vomiting (16%), and abdominal pain (11%). In 74% of cases, gastrointestinal symptoms were judged to be mild. In total, 35% of patients developed an abnormal alanine aminotransferase or total bilirubin level; these were elevated to less than 5 times the upper limit of normal in 77% of cases. After adjusting for potential confounders, the presence of gastrointestinal symptoms at any time (odds ratio 0.93, 95% confidence interval 0.76-1.15) or liver test abnormalities on admission (odds ratio 1.31, 95% confidence interval 0.80-2.12) were not independently associated with mechanical ventilation or death. Conclusion Among patients hospitalized with COVID-19, gastrointestinal symptoms and liver test abnormalities were common but the majority were mild and their presence was not associated with a more severe clinical course.

Key Points Question Can deep learning recurrent neural network (RNN) models using raw longitudinal data extracted directly from electronic health records outperform conventional regression models in predicting the risk of developing hepatocellular carcinoma (HCC)? Findings This prognostic study included 48 151 patients with hepatitis C virus (HCV)–related cirrhosis in the national Veterans Health Administration who had at least 3 years of follow-up after the diagnosis of cirrhosis. Deep learning RNN models outperformed conventional linear regression models and could be used to identify patients with HCV-related cirrhosis at high risk of developing HCC. Meaning The findings of this study suggest that RNN models could have multiple applications in clinical practice and could be applied to HCC outreach and surveillance strategies.

Primary age-related tauopathy (PART) is a neurodegenerative pathology with features distinct from but also overlapping with Alzheimer disease (AD). While both exhibit Alzheimer-type temporal lobe neurofibrillary degeneration alongside amnestic cognitive impairment, PART develops independently of amyloid-β (Aβ) plaques. The pathogenesis of PART is not known, but evidence suggests an association with genes that promote tau pathology and others that protect from Aβ toxicity. Here, we performed a genetic association study in an autopsy cohort of individuals with PART (n = 647) using Braak neurofibrillary tangle stage as a quantitative trait. We found some significant associations with candidate loci associated with AD (SLC24A4, MS4A6A, HS3ST1) and progressive supranuclear palsy (MAPT and EIF2AK3). Genome-wide association analysis revealed a novel significant association with a single nucleotide polymorphism on chromosome 4 (rs56405341) in a locus containing three genes, including JADE1 which was significantly upregulated in tangle-bearing neurons by single-soma RNA-seq. Immunohistochemical studies using antisera targeting JADE1 protein revealed localization within tau aggregates in autopsy brains with four microtubule-binding domain repeats (4R) isoforms and mixed 3R/4R, but not with 3R exclusively. Coimmunoprecipitation in post-mortem human PART brain tissue revealed a specific binding of JADE1 protein to four repeat tau lacking N-terminal inserts (0N4R). Finally, knockdown of the Drosophila JADE1 homolog rhinoceros (rno) enhanced tau-induced toxicity and apoptosis in vivo in a humanized 0N4R mutant tau knock-in model, as quantified by rough eye phenotype and terminal deoxynucleotidyl transferase dUTP nick end-labeling (TUNEL) in the fly brain. Together, these findings indicate that PART has a genetic architecture that partially overlaps with AD and other tauopathies and suggests a novel role for JADE1 as a modifier of neurofibrillary degeneration.

No abstract