An ensemble prediction model for COVID-19 mortality risk.

Background: It's critical to identify COVID-19 patients with a higher death risk at early stage to give them better hospitalization or intensive care. However, thus far, none of the machine learning models has been shown to be successful in an independent cohort. We aim to develop a machine learning model which could accurately predict death risk of COVID-19 patients at an early stage in other independent cohorts. Methods: We used a cohort containing 4711 patients whose clinical features associated with patient physiological conditions or lab test data associated with inflammation, hepatorenal function, cardiovascular function, and so on to identify key features. To do so, we first developed a novel data preprocessing approach to clean up clinical features and then developed an ensemble machine learning method to identify key features. Results: Finally, we identified 14 key clinical features whose combination reached a good predictive performance of area under the receiver operating characteristic curve 0.907. Most importantly, we successfully validated these key features in a large independent cohort containing 15 790 patients. Conclusions: Our study shows that 14 key features are robust and useful in predicting the risk of death in patients confirmed SARS-CoV-2 infection at an early stage, and potentially useful in clinical settings to help in making clinical decisions.

Integrative omics provide biological and clinical insights into acute respiratory distress syndrome.

PURPOSE: Acute respiratory distress syndrome (ARDS) is accompanied by a dysfunctional immune-inflammatory response following lung injury, including during coronavirus disease 2019 (COVID-19). Limited causal biomarkers exist for ARDS development. We sought to identify novel genetic susceptibility targets for ARDS to focus further investigation on their biological mechanism and therapeutic potential. METHODS: Meta-analyses of ARDS genome-wide association studies were performed with 1250 cases and 1583 controls in Europeans, and 387 cases and 387 controls in African Americans. The functionality of novel loci was determined in silico using multiple omics approaches. The causality of 114 factors potentially involved in ARDS development was assessed using Mendelian Randomization analysis. RESULTS: There was distinct genetic heterogeneity in ARDS between Europeans and African Americans. rs7967111 at 12p13.2 was functionally associated with ARDS susceptibility in Europeans (odds ratio = 1.38; P = 2.15 × 10-8). Expression of two genes annotated at this locus, BORCS5 and DUSP16, was dynamic but ultimately decreased during ARDS development, as well as downregulated in immune cells alongside COVID-19 severity. Causal inference implied that comorbidity of inflammatory bowel disease and elevated levels of C-reactive protein and interleukin-10 causally increased ARDS risk, while vitamin D supplementation and vasodilator use ameliorated risk. CONCLUSION: Our findings suggest a novel susceptibility locus in ARDS pathophysiology that implicates BORCS5 and DUSP16 as potentially acting in immune-inflammatory processes. This locus warrants further investigation to inform the development of therapeutic targets and clinical care strategies for ARDS, including those induced by COVID-19.

Outcome of Hospitalization for COVID-19 in Patients with Interstitial Lung Disease. An International Multicenter Study.

Rationale: The impact of coronavirus disease (COVID-19) on patients with interstitial lung disease (ILD) has not been established.Objectives: To assess outcomes in patients with ILD hospitalized for COVID-19 versus those without ILD in a contemporaneous age-, sex-, and comorbidity-matched population.Methods: An international multicenter audit of patients with a prior diagnosis of ILD admitted to the hospital with COVID-19 between March 1 and May 1, 2020, was undertaken and compared with patients without ILD, obtained from the ISARIC4C (International Severe Acute Respiratory and Emerging Infection Consortium Coronavirus Clinical Characterisation Consortium) cohort, admitted with COVID-19 over the same period. The primary outcome was survival. Secondary analysis distinguished idiopathic pulmonary fibrosis from non-idiopathic pulmonary fibrosis ILD and used lung function to determine the greatest risks of death.Measurements and Main Results: Data from 349 patients with ILD across Europe were included, of whom 161 were admitted to the hospital with laboratory or clinical evidence of COVID-19 and eligible for propensity score matching. Overall mortality was 49% (79/161) in patients with ILD with COVID-19. After matching, patients with ILD with COVID-19 had significantly poorer survival (hazard ratio [HR], 1.60; confidence interval, 1.17-2.18; P = 0.003) than age-, sex-, and comorbidity-matched controls without ILD. Patients with an FVC of <80% had an increased risk of death versus patients with FVC ≥80% (HR, 1.72; 1.05-2.83). Furthermore, obese patients with ILD had an elevated risk of death (HR, 2.27; 1.39-3.71).Conclusions: Patients with ILD are at increased risk of death from COVID-19, particularly those with poor lung function and obesity. Stringent precautions should be taken to avoid COVID-19 in patients with ILD.