P-1981. MMCD score for predicting kidney replacement therapy in hospitalized patients with COVID-19 from 2020 to 2022

Le résumé fourni par la source

Abstract Background Acute kidney injury (AKI) requiring kidney replacement therapy (KRT) in its most severe forms is a significant complication of patients with COVID-19. The development of a risk score to predict KRT requirement could optimize health resource allocation. Thus, this study aimed to develop and validate a score to predict the KRT requirement in hospitalized patients with COVID-19 from 2020 to 2022. Area under the receiver operating characteristic curve for the 2021-2022 validation of MMCD score Methods This is a multicenter retrospective cohort of consecutive patients hospitalized for COVID-19, confirmed by laboratory tests, in 40 Brazilian hospitals, from March 2020 to July 2022. Patients under 18 years old, pregnant, in palliative care, or in dialysis therapy at admission were excluded from the study. Predictor variables were selected using generalized additive models (GAM) and the least absolute shrinkage and selection operator (LASSO) regression was used for score derivation. The score was developed in the period from March to July 2020, with temporal and geographic validation from July to September 2020 and new temporal validation from March 2021 to July 2022. The performance of the score was evaluated by the area under the receiver operating characteristic curve (AUROC, with a 95% confidence interval), graphical analysis with intercept and slope test, and Brier score. Calibration for the validation of MMCD score Results The study included 3,680 patients in the development sample, 1,532 in the 2020 temporal validation, 1,378 in the geographic validation, and 9,473 in the 2021-2022 temporal validation. Four predictors were identified: mechanical ventilation at any time during hospitalization, male sex, admission creatinine, and diabetes mellitus. The score named MMCD showed excellent performance in the derivation and validations cohorts (development AUROC 0.929, CI 0.918–0.939, Brier 0.057; 2020 temporal validation AUROC 0.927, CI 0.911–0.941, Brier 0.056; 2020 geographic validation AUROC 0.819, CI 0.792–0.845, Brier 0.122; 2021/2022 temporal validation AUROC 0.916, CI 0.909-0.924, Brier 0.057). Conclusion The MMCD showed excellent predictive ability for KRT in different phases of the pandemic, which can contribute to support more assertive decisions in the allocation of care resources, in addition to efficient clinical management. Disclosures All Authors: No reported disclosures

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.