Prognostic model and risk factors for hospital mortality in patients with diffuse large B-cell lymphoma associated with coronavirus infection: a single-center cohort study

• Published in: Onkogematologii͡a Vol. 18; no. 4; pp. 74 - 85
• Main Authors: Polyakov, Yu. Yu, Baryakh, E. A., Misyurina, E. N., Zhelnova, E. I., Yatskov, K. V., Makeshova, A. B., Mingalimov, M. A., Tolstykh, T. N., Chudnova, T. S., Ivanova, D. D., Koneva, A. I., Kochneva, O. L., Zotina, E. N., Gagloeva, D. E., Grishina, E. Yu, Shimanovskaya, L. T., Yakimets, V. N.
• Format: Journal Article
• Language: English
• Published: 15.01.2024
• ISSN: 1818-8346; 2413-4023
• DOI: 10.17650/1818-8346-2023-18-4(Suppl)-74-85

Background . Coronavirus disease (COVID-19), caused by SARS-CoV-2, presents new challenges to hematologists, highlighting the vulnerability of patients with hematological malignancies, in particular with diffuse large B-cell lymphoma (DLBCL). Identification of hospital mortality risk factors is necessary for subsequent stratification of patients into risk groups, which will allow further risk-based therapy. Aim . To develop a prognostic model and identify risk factors for hospital mortality in patients with DLBCL associated with COVID-19. Materials and methods . The interim retrospective study included 112 patients with an immunohistochemically confirmed diagnosis of DLBCL, coronavirus infection verified based on polymerase chain reaction (PCR) for SARS-CoV-2, and viral pneumonia associated with COVID-19. To determine the risk factors for hospital mortality, a multivariate (logistic regression) statistical analysis was performed. The study end point was a binary variable - the patient vital status (discharged alive or died). Results and conclusion . Of the 112 patients, 24 died. Due to the limited number of patients compared to the number of predictors and to avoid overfitting, a two-stage approach to constructing a predictive model was used. In univariate analysis, statistically significant during hospitalization were the hematological disease status (complete remission/partial remission, progression/relapse, de novo ), positive PCR result, C-reactive protein level >6 mg/L, platelets <100 thousand/pL, hemoglobin <120 g/L, albumin <35 g/L, lactate dehydrogenase >248 U/L, D-dimer >500 ng/mL and the degree of lung tissue damage according to computed tomography >50 % (grade II and above), respiratory failure I degrees and higher. The final model was constructed by minimizing the Akaike information criterion. The final model included a positive PCR result, stage II respiratory failure, hematologic disease status (relapse/progression), and albumin level at the time of hospital admission.