Machine learning for prediction of in-hospital mortality in coronavirus disease 2019 patients: results from an Italian multicenter study.

BACKGROUND: Several risk factors have been identified to predict worse outcomes in patients affected by SARS-CoV-2 infection. Machine learning algorithms represent a novel approach to identifying a prediction model with a good discriminatory capacity to be easily used in clinical practice. The aim of this study was to obtain a risk score for in-hospital mortality in patients with coronavirus disease infection (COVID-19) based on a limited number of features collected at hospital admission. METHODS AND RESULTS: We studied an Italian cohort of consecutive adult Caucasian patients with laboratory-confirmed COVID-19 who were hospitalized in 13 cardiology units during Spring 2020. The Lasso procedure was used to select the most relevant covariates. The dataset was randomly divided into a training set containing 80% of the data, used for estimating the model, and a test set with the remaining 20%. A Random Forest modeled in-hospital mortality with the selected set of covariates: its accuracy was measured by means of the ROC curve, obtaining AUC, sensitivity, specificity and related 95% confidence interval (CI). This model was then compared with the one obtained by the Gradient Boosting Machine (GBM) and with logistic regression. Finally, to understand if each model has the same performance in the training and test set, the two AUCs were compared using the DeLong's test. Among 701 patients enrolled (mean age 67.2 ±â€Š13.2 years, 69.5% male individuals), 165 (23.5%) died during a median hospitalization of 15 (IQR, 9-24) days. Variables selected by the Lasso procedure were: age, oxygen saturation, PaO2/FiO2, creatinine clearance and elevated troponin. Compared with those who survived, deceased patients were older, had a lower blood oxygenation, lower creatinine clearance levels and higher prevalence of elevated troponin (all P < 0.001). The best performance out of the samples was provided by Random Forest with an AUC of 0.78 (95% CI: 0.68-0.88) and a sensitivity of 0.88 (95% CI: 0.58-1.00). Moreover, Random Forest was the unique model that provided similar performance in sample and out of sample (DeLong test P = 0.78). CONCLUSION: In a large COVID-19 population, we showed that a customizable machine learning-based score derived from clinical variables is feasible and effective for the prediction of in-hospital mortality.

Implications of atrial fibrillation on the clinical course and outcomes of hospitalized COVID-19 patients: results of the Cardio-COVID-Italy multicentre study.

AIMS: To assess the clinical relevance of a history of atrial fibrillation (AF) in hospitalized patients with coronavirus disease 2019 (COVID-19). METHODS AND RESULTS: We enrolled 696 consecutive patients (mean age 67.4 ± 13.2 years, 69.7% males) admitted for COVID-19 in 13 Italian cardiology centres between 1 March and 9 April 2020. One hundred and six patients (15%) had a history of AF and the median hospitalization length was 14 days (interquartile range 9-24). Patients with a history of AF were older and with a higher burden of cardiovascular risk factors. Compared to patients without AF, they showed a higher rate of in-hospital death (38.7% vs. 20.8%; P < 0.001). History of AF was associated with an increased risk of death after adjustment for clinical confounders related to COVID-19 severity and cardiovascular comorbidities, including history of heart failure (HF) and increased plasma troponin [adjusted hazard ratio (HR): 1.73; 95% confidence interval (CI) 1.06-2.84; P = 0.029]. Patients with a history of AF also had more in-hospital clinical events including new-onset AF (36.8% vs. 7.9%; P < 0.001), acute HF (25.3% vs. 6.3%; P < 0.001), and multiorgan failure (13.9% vs. 5.8%; P = 0.010). The association between AF and worse outcome was not modified by previous or concomitant use of anticoagulants or steroid therapy (P for interaction >0.05 for both) and was not related to stroke or bleeding events. CONCLUSION: Among hospitalized patients with COVID-19, a history of AF contributes to worse clinical course with a higher mortality and in-hospital events including new-onset AF, acute HF, and multiorgan failure. The mortality risk remains significant after adjustment for variables associated with COVID-19 severity and comorbidities.

Acute pulmonary embolism in COVID-19 disease: Preliminary report on seven patients.

There is some evidence that Covid 19 pneumonia is associated with prothrombotic status and increased risk of venous thromboembolic events (deep venous thrombosis and pulmonary embolism). Over a two-week period we admitted in our Unit 25 patients with Covid-19 pneumonia, of these pulmonary embolism was diagnosed using computed tomography angiography in 7. We report on clinical and biochemical features of these patients. They were all males, with a mean age of 70.3 years (range 58-84); traditional risk factors for venous thromboembolism were identified in the majority of patients with pulmonary embolism, however not differently from those without pulmonary embolism. Clinical presentation of pulmonary embolism patients was usually characterized by persistence or worsening of respiratory symptoms, with increasing oxygen requirement. D-dimer levels were several fold higher than the upper threshold of normal; in patients in whom PE was recognized during hospital stay, a rapid and relevant increase of D-dimer levels was observed. Computed tomographic findings ranged from massive acute pulmonary embolism to a segmental or sub-segmental pattern; furthermore, thrombosis of sub-segmental pulmonary arteries within lung infiltrates were occasionally seen, suggesting local mechanisms. Six out of 7 patients were treated with unfractionated or low molecular weight heparin with clinical benefit within few days; one patient needed systemic thrombolysis (death from hemorrhagic complication).