Improving prediction of COVID-19 mortality using machine learning in the Spanish SEMI-COVID-19 registry.

COVID-19 is responsible for high mortality, but robust machine learning-based predictors of mortality are lacking. To generate a model for predicting mortality in patients hospitalized with COVID-19 using Gradient Boosting Decision Trees (GBDT). The Spanish SEMI-COVID-19 registry includes 24,514 pseudo-anonymized cases of patients hospitalized with COVID-19 from 1 February 2020 to 5 December 2021. This registry was used as a GBDT machine learning model, employing the CatBoost and BorutaShap classifier to select the most relevant indicators and generate a mortality prediction model by risk level, ranging from 0 to 1. The model was validated by separating patients according to admission date, using the period 1 February to 31 December 2020 (first and second waves, pre-vaccination period) for training, and 1 January to 30 November 2021 (vaccination period) for the test group. An ensemble of ten models with different random seeds was constructed, separating 80% of the patients for training and 20% from the end of the training period for cross-validation. The area under the receiver operating characteristics curve (AUC) was used as a performance metric. Clinical and laboratory data from 23,983 patients were analyzed. CatBoost mortality prediction models achieved an AUC performance of 84.76 (standard deviation 0.45) for patients in the test group (potentially vaccinated patients not included in model training) using 16 features. The performance of the 16-parameter GBDT model for predicting COVID-19 hospital mortality, although requiring a relatively large number of predictors, shows a high predictive capacity.

Differences in clinical features and mortality in very old unvaccinated patients (≥ 80 years) hospitalized with COVID-19 during the first and successive waves from the multicenter SEMI-COVID-19 Registry (Spain).

BACKGROUND: Old age is one of the most important risk factors for severe COVID-19. Few studies have analyzed changes in the clinical characteristics and prognosis of COVID-19 among older adults before the availability of vaccines. This work analyzes differences in clinical features and mortality in unvaccinated very old adults during the first and successive COVID-19 waves in Spain. METHODS: This nationwide, multicenter, retrospective cohort study analyzes unvaccinated patients ≥ 80 years hospitalized for COVID-19 in 150 Spanish hospitals (SEMI-COVID-19 Registry). Patients were classified according to whether they were admitted in the first wave (March 1-June 30, 2020) or successive waves (July 1-December 31, 2020). The endpoint was all-cause in-hospital mortality, expressed as the case fatality rate (CFR). RESULTS: Of the 21,461 patients hospitalized with COVID-19, 5,953 (27.7%) were ≥ 80 years (mean age [IQR]: 85.6 [82.3-89.2] years). Of them, 4,545 (76.3%) were admitted during the first wave and 1,408 (23.7%) during successive waves. Patients hospitalized in successive waves were older, had a greater Charlson Comorbidity Index and dependency, less cough and fever, and met fewer severity criteria at admission (qSOFA index, PO2/FiO2 ratio, inflammatory parameters). Significant differences were observed in treatments used in the first (greater use of antimalarials, lopinavir, and macrolides) and successive waves (greater use of corticosteroids, tocilizumab and remdesivir). In-hospital complications, especially acute respiratory distress syndrome and pneumonia, were less frequent in patients hospitalized in successive waves, except for heart failure. The CFR was significantly higher in the first wave (44.1% vs. 33.3%; -10.8%; p < 0.001) and was higher among patients ≥ 95 years (54.4% vs. 38.5%; -15.9%; p < 0.001). After adjustments to the model, the probability of death was 33% lower in successive waves (OR: 0.67; 95% CI: 0.57-0.79). CONCLUSIONS: Mortality declined significantly between the first and successive waves in very old unvaccinated patients hospitalized with COVID-19 in Spain. This decline could be explained by a greater availability of hospital resources and more effective treatments as the pandemic progressed, although other factors such as changes in SARS-CoV-2 virulence cannot be ruled out.

Effectiveness of glucocorticoids in patients hospitalized for severe SARS-CoV-2 pneumonia.

BACKGROUND: Several studies have reported the beneficial effect of glucocorticoids in the treatment of cytokine storm that occurs in patients with severe COVID-19. Various glucocorticoids regimens have been proposed. METHODS: Retrospective observational study that includes patients with severe SARS-CoV-2 pneumonia and compares admission to an Intensive Care Unit (ICU) or death during hospitalization in three groups of patients: no glucocorticoids treatment, use of glucocorticoids doses equivalent to less than 250 mg of prednisone daily and use of equivalent doses greater than or equal to 250 mg of prednisone daily. Multivariate analysis was performed using logistic regression, using the propensity index as a covariant. RESULTS: Of the 259 patients enrolled in the study, 67 (25.9%) had an unfavorable evolution, dying or requiring ICU admission. Comparative analyzes between different glucocorticoids treatments and the association with ICU admission or death were: glucocorticoids treatment (any dose) versus no glucocorticoids treatment (OR: 0.71 [0.30-1.66]), treatment with glucocorticoids (≥250 mg prednisone daily) versus no glucocorticoids treatment (OR: 0.35 [0.11-1.08]) and glucocorticoids treatment (≥250 mg prednisone daily) versus patients with glucocorticoids doses <250 mg prednisone daily or without glucocorticoids treatment (OR: 0.30 [0.10-0.88]). CONCLUSION: The results of this study show that patients with severe SARS-CoV-2 pneumonia treated with glucocorticoids pulses with equivalent doses of prednisone greater than or equal to 250 mg have a more favorable evolution (less mortality and less admission to ICU).

INTRODUCCIÓN: Se han comunicado varios trabajos donde se ha demostrado un efecto beneficioso de los glucocorticoides como tratamiento de la tormenta de citocinas que se asocia a los cuadros graves por SARS-CoV-2, plateándose diferentes pautas de glucocorticoides. MÉTODOS: Estudio observacional retrospectivo que incluye pacientes con neumonía grave por SARS-CoV-2 y compara el ingreso en una unidad de cuidados intensivos (UCI) o fallecimiento durante la hospitalización en 3 grupos de pacientes: sin tratamiento con glucocorticoides, uso de dosis diarias de glucocorticoides equivalentes menores a 250 mg de prednisona y dosis diarias equivalentes mayores o iguales a 250 mg de prednisona. Se realizó un análisis multivariante mediante regresión logística, utilizando el índice de propensión como covariante. RESULTADOS: De los 259 pacientes incorporados al estudio 67 (25,9%) tuvieron una evolución desfavorable, falleciendo o precisando ingreso en UCI. Los análisis comparativos entre diferentes tratamientos con glucocorticoides, y la asociación con ingreso en UCI o fallecimiento fueron: tratamiento con glucocorticoides (cualquier dosis) versus sin tratamiento con glucocorticoides (OR: 0,71 [0,30­1,66]), tratamiento con glucocorticoides (≥250 mg de prednisona al día) versus sin tratamiento con glucocorticoides (OR: 0,35 [0,11­1,08]) y tratamiento con glucocorticoides (≥250 mg de prednisona al día) versus pacientes con dosis de glucocorticoides < 250 mg de prednisona o sin tratamiento con glucocorticoides (OR: 0,30 [0,10­0,88]). CONCLUSIÓN: Los resultados de este estudio muestran que los paciente con neumonía grave por SARS-CoV-2 tratados con pulsos con glucocorticoides con dosis equivalentes de prednisona mayor o igual de 250 mg tienen una evolución más favorable (menos mortalidad e ingreso en UCI).

[Effectiveness of glucocorticoids in patients hospitalized for severe SARS-CoV-2 pneumonia]. / Efectividad de los glucocorticoides en pacientes hospitalizados por neumonía grave por SARS-CoV-2.

INTRODUCTION: Several studies have reported the beneficial effect of glucocorticoids in the treatment of cytokine storm that occurs in patients with severe COVID-19. Various glucocorticoids regimens have been proposed. METHODS: Retrospective observational study that includes patients with severe SARS-CoV-2 pneumonia and compares admission to an Intensive Care Unit (ICU) or death during hospitalization in three groups of patients: no glucocorticoids treatment, use of glucocorticoids doses equivalent to less than 250mg of prednisone daily and use of equivalent doses greater than or equal to 250mg of prednisone daily. Multivariate analysis was performed using logistic regression, using the propensity index as a covariant. RESULTS: Of the 259 patients enrolled in the study, 67 (25.9%) had an unfavorable evolution, dying or requiring ICU admission. Comparative analyzes between different glucocorticoids treatments and the association with ICU admission or death were: glucocorticoids treatment (any dose) versus no glucocorticoids treatment (OR: 0.71 [0.30-1.66]), treatment with glucocorticoids (≥250mg prednisone daily) versus no glucocorticoids treatment (OR: 0.35 [0.11-1.08]) and glucocorticoids treatment (≥250mg prednisone daily) versus patients with glucocorticoids doses <250mg prednisone daily or without glucocorticoids treatment (OR: 0.30 [0.10-0.88]). CONCLUSION: The results of this study show that patients with severe SARS-CoV-2 pneumonia treated with glucocorticoids pulses with equivalent doses of prednisone greater than or equal to 250mg have a more favorable evolution (less mortality and less admission to ICU).