Artificial intelligence and big data processing for the development of predictive models to support Covid-19 hospital patients

Coronavirus disease 2019 (COVID-19) has caused more than 6 million deaths. Higher values of the FIB-4 index have been shown to be associated with disease severity. Although vaccination has helped to improve clinical outcomes and overall mortality, it remains important to identify clinical parameters that can predict a likely worse prognosis. Artificial intelligence and big data processing were used to retrieve data from patients with Covid-19 admitted during the period March 2020-January 2022 at the Fondazione Policlinico Gemelli IRCCS. Patients and clinical characteristics of patients with available FIB 4 data derived from the Gemelli Generator Real World Data (G2 RWD) were used to develop predictive models of mortality during the 4 waves of the Covid-19 pandemic. A logistic regression model was applied to the training and test set. The performance of the model was assessed by means of the ROC curve. After the pre-processing steps, 1143 patients and 35 variables were included in the final dataset. The FIB-4 discretization algorithm identified a cut-off of 2.54. After fitting the model for multiple mortality regression analysis: FIB-4>= 2.53 (OR=4.53, [CI: 2.83 - 7.25]), wave 3 (OR=0.34, [CI: 0.15 - 0.75], wave 4 (OR=0.40 [CI: 0.24 - 0.66]) and LDH (OR=1.001, [CI: 1.000 - 1.002]) were considered. The machine learning approach identified a cut-off of 2.53 for FIB-4 above which the risk of death increases significantly. These data may be useful in the clinical management of patients with Covid-19, as they can be calculated from the blood test after hospital admission.

Adenosine-based cancer immunotherapy for COVID-19

Cancer immunotherapy has shaped the way in which we design cancer treatments, introducing the paradigm of taking advantage of our immune system to fight cancer. We propose that the same concept could be applied to infectious diseases and, especially, to those that hamper the immune system like COVID-19. It is well known that an adaptive immune response is able to eradicate viral infections and CD8+ T-cells are key in such anti-viral response. Furthermore, several studies reported that the number of CD8+ T-cells is reduced in COVID-19 patients since the beginning of SARS-CoV-2 infection and further decreased in severe cases. CD8+ T-cells often show signs of exhaustion, loss of T-cell functions and suppression in COVID-19 patients, suggesting that hampering CD8+ T-cells could be a way by which SARS-CoV-2 infection progresses. Importantly, the number of CD8+ T cells appears to re-increase in patients that are recovering from COVID-19, suggesting that CD8+ T-cells could be a key factor that determines whether our body can recover from the disease. We propose that therapies that boost CD8+ T cells could be effective in clearing SARS-CoV-2 infection in COVID-19. To test this, we will take advantage of the adenosine-mediated immunomodulation. Adenosine, an ATP metabolite that is produced during inflammation, hypoxia and in the tumor microenvironment, was found to suppress the immune response through activation of adenosine receptors present on immune cells. Small molecules antagonists that block one of these receptors, adenosine A2A receptor (A2AR) antagonists, are currently being studied to boost anti-cancer T-cell mediated immune responses. Our data show that treatment with an A2AR antagonist restores and stabilizes Notch1, a key pathway for T-cell functions, along with production of INF-gamma and Granzyme B and proliferation in CD8+ T-cells. As a proof of concept, our data indicates that treatment with an A2AR antagonist increases CD8+ T-cells anti-cancer response in tumor-derived organoids, suggesting that the treatment boosts CD8+ T-cells-mediated immune response. Ongoing work aims to test whether the A2AR antagonist treatment restores several parameters of T-cell function that are specifically modified in COVID-19. This analysis will help to predict the action of the A2AR antagonist on T-cells in vivo and we ultimately aim to test this treatment in a mouse model for COVID-19. Overall, our work could introduce a new paradigm and new therapies for COVID-19 and other infectious diseases.

Evaluation of Rural-Urban Differences in Hospitalization and Mortality Rates for US COVID-19 Patients in the United States

Background. Rural communities are among the most vulnerable and resourcescarce populations in the United States. Rural data is rarely centralized, precluding comparability across regions, and no significant studies have studied this population at scale. The purpose of this study is to present findings from the National COVID Cohort Collaborative (N3C) to provide insight into future research and highlight the urgent need to address health disparities in rural populations. N3C Patient Distribution This figure shows the geospatial distribution of the N3C COVID-19 positive population. N3C contains data from 55 data contributors from across the United States, 40 of whom include sufficient location information to map by ZIP Code centroid spatially. Of those sites, we selected 27 whose data met our minimum robustness qualifications for inclusion in our study. This bubble map is to scale with larger bubbles representing more patients. A. shows all N3C patients. B. shows only urban N3C distribution. C. shows the urban-adjacent rural patient distribution. D. shows the nonurban-adjacent rural patient distribution, representing the most isolated patients in N3C. Methods. This retrospective cohort of 573,018 patients from 27 hospital systems presenting with COVID-19 between January 2020 and March 2021, of whom 117,897 were admitted (see Data Analysis Plan diagram for inclusion/exclusion criteria), analyzes outcomes and 30-day survival for the hospitalized population by the degree of rurality. Multivariate Cox regression analysis and mixed-effects models were used to estimate the association between rurality, hospitalization, and all-cause mortality, controlling for major risk factors associated with rural-urban health discrepancies and differences in health system outcomes. The difference in distribution by rurality is described as well as supplemented by population-level statistics to confirm representativeness. Data Analysis Plan This data analysis plan includes an overview of study inclusion and exclusion criteria, the matrix for data robustness to determine potential sites to include, and our covariate selection, model building, and residual testing strategy. Results. This study demonstrates a significant difference between hospital admissions and outcomes in urban versus urban-adjacent rural (UAR) and nonurban-adjacent rural (NAR) lines. Hospital admissions for UAR (OR 1.41, p< 0.001, 95% CI: 1.37 - 1.45) and NAR (OR 1.42, p< 0.001, 95% CI: 1.35 - 1.50) were significantly higher than their urban counterparts. Similar distributions were present for all-cause mortality for UAR (OR 1.39, p< 0.001, 95% CI: 1.30 - 1.49) and NAR (OR 1.38, p< 0.001, 95% CI: 1.22 - 1.55) compared to urban populations. These associations persisted despite adjustments for significant differences in BMI, Charlson Comorbidity index Score, gender, age, and the quarter of diagnosis for COVID-19. Baseline Characteristics Hospitalized COVID-19 Positive Population by Rurality Category, January 2020 - March 2021 Survival Curves in Hospitalized Patients Over 30 Days from Day of Admission This figure shows a survival plot of COVID-19 positive hospitalized patients in N3C by rural category (A), Charlson Comorbidity Index (B), Quarter of Diagnosis (C), and Age Group (D) from hospital admission through day 30. Events were censored at day 30 based on the incidence of death or transfer to hospice care. These four factors had the highest predictive power of the covariates evaluated in this study. Unadjusted and Adjusted Odds Ratios for Hospitalization and All-Cause Mortality by Rural Category, January 2020 - March 2021 This figure shows the adjusted and unadjusted odds ratios for being hospitalized or dying after hospitalization for the COVID-19 positive population in N3C. Risk is similar between adjusted and unadjusted models, suggesting a real impact of rurality on all-cause mortality. A shows the unadjusted odds ratios for admission to the hospital after a positive COVID-19 diagnosis for all N3C patients. B shows the unadjusted odds ratios for all-cause mortalit at any point after hospitalization for COVID-19 positive patients. C shows the adjusted odds ratios for being admitted to the hospital after a positive COVID-19 diagnosis for all N3C patients. D shows the adjusted odds ratios for all-cause mortality for all-cause mortality at any point after hospitalization for COVID-19 positive patients. Adjusted models include adjustments for gender, race, ethnicity, BMI, age, Charlson Comorbidity Index (CCI) composite score, rurality, and quarter of diagnosis. The data provider is included as a random effect in all models. Conclusion. In N3C, we found that hospitalizations and all-cause mortality were greater among rural populations when compared to urban populations after adjustment for several factors, including age and co-morbidities. This study also identified key demographic and clinical disparities among rural patients that require further investigation.

Clinical features of patients with new onset of autoimmune hepatitis following SARS-CoV-2 vaccination

Introduction: Autoimmune hepatitis (AIH) is a relatively rare chronic immune-mediated liver disease, which develops in genetically predisposed individuals following an environmental trigger. A few cases of AIH have been recently reported after the SARS-CoV-2 vaccination. Aims: The aim of this study was to describe clinical-epidemiological profile of a series of adult patients who experienced AIH onset following SARS-CoV-2 vaccination. Materials and Methods: This multicentric observational study collected clinical data of adult patients who had received SARS-CoV-2 vaccination and thereafter were diagnosed with AIH between 03/2021 and 10/2021 in Italy, using an online survey among members of the Italian Association for the study of the Liver (AISF). Results: Among the 12 patients included: 50% were females, median age 62 years (range 32-80), 6 (50%) had preexisting extrahepatic autoimmune disease (3 thyroiditis, 2 rheumatoid arthritis, 1 systemic lupus erythematosus), 7 patients have received Comirnaty (BioNTech/Pfizer) vaccine, 2 Spikevax (Moderna Biotech) and 3 Vaxzevria (AstraZeneca). Ten patients (83%) had acute onset of AIH with transaminase levels ≥10 times the upper limit of normal (ULN, range 13-77 x ULN), 8 (67%) with jaundice (total bilirubin 3.5-18.6 x ULN). At AIH diagnosis (median time from first and second vaccine dose: 48 and 10 days, respectively) median AST was 18 x ULN (range 5-85), ALT 23.8 x ULN (range 7-83), total bilirubin 3.8 x ULN (range 0.6-18.6), alkaline phosphatase 1.3 x ULN (range 0.8-7.1), immunoglobulin G 1.2 x ULN (median 0.8-1.5). Eight (67%) patients had autoantibodies: 6 ANA, 1 SMA, 1 LKM-1. Liver biopsy was typical for AIH in 8 and compatible in 3 patients. After 3 months 58% achieved complete biochemical response to standard immunosuppressive treatment. Conclusion: While intensive vaccination against SARS-CoV-2 continues, the diagnosis of AIH secondary to vaccines should be included in the differential diagnosis in cases of acute hepatitis of unexplained aetiology.

Gastrointestinal symptoms and digestive comorbidities in an Italian cohort of patients with COVID-19.

OBJECTIVE: The Coronavirus Disease 2019 (COVID-19) pandemic mainly involves respiratory symptoms, though gastrointestinal (GI) symptoms are increasingly being recognized. In this context, the presence of comorbidities appears to be associated with adverse outcomes. However, the role of digestive manifestations is not yet well defined. The primary aim of this study was to assess the prevalence of GI symptoms and digestive comorbidities in a cohort of patients with COVID-19 compared to controls. The secondary aim was to determine the association of GI-symptoms and digestive comorbidities with clinical outcomes. PATIENTS AND METHODS: Inpatients with COVID-19 and controls with similar symptoms and/or radiological findings were enrolled. Symptoms at admission and throughout hospitalization were collected as they were comorbidities. The measured clinical outcomes were mortality, intensive care unit admission and cumulative endpoint. RESULTS: A total of 105 patients were included: 34 with COVID-19 and 71 controls. At admission, the prevalence of GI symptoms among COVID-19 patients was 8.8%. During hospitalization, the frequency of GI symptoms was higher in patients with COVID-19 than in controls (p=0.004). Among patients with COVID-19, the mortality and a cumulative endpoint rates of those with GI symptoms were both lower than for those without GI symptoms (p=0.016 and p=0.000, respectively). Finally, we found digestive comorbidities to be associated with a milder course of COVID-19 (p=0.039 for cumulative endpoint). CONCLUSIONS: Our results highlighted the non-negligible frequency of GI symptoms in patients with COVID-19, partly attributable to the therapies implemented. In addition, the presence of GI symptoms and digestive comorbidities is associated with better outcomes. Most likely, digestive comorbidities do not hinder the host's immune response against SARS-COV-2, and the occurrence of GI symptoms might be linked to a faster reduction of the viral load via the faecal route.

SARS-CoV2 RNA detection in a pancreatic pseudocyst sample.

The involvement of gastrointestinal system in SARS-CoV2 related disease, COVID-19, is increasingly recognized. COVID-19 associated pancreatic injury has been suggested, but its correlation with pancreatic disease is still unclear. In this case report, we describe the detection of SARS-CoV2 RNA in a pancreatic pseudocyst fluid sample collected from a patient with SARS-CoV2 associated pneumonia and a pancreatic pseudocyst developed as a complication of an acute edematous pancreatitis. The detection of SARS-CoV2 within the pancreatic collection arise the question of whether this virus has a tropism for pancreatic tissue and whether it plays a role in pancreatic diseases occurrence.