Prediction of COVID-19 course in patients with stage 3 chronic kidney disease and type 2 diabetes mellitus

Abstract

BACKGROUND. Patients with Diabetes Mellitus 2 (DM2) and Chronic Kidney Disease (CKD) are at a high risk for severe clinical course of COVID-19. The high mortality rate due to COVID-19 and widespread distribution of DM2 and CKD all over the world make it necessary to determine the predictors of adverse outcome of novel coronavirus infection (NCI).AIM. The identification of predictors of NCI adverse outcome in patients with DM2 and CKD stage 3 due to diabetic kidney disease.Patients and Methods. The patients with NCI and CKD stage 3 were included in observational retrospective uncontrolled study during the follow-up period from 04.01. to 10.30.2020. The study endpoints were the outcome of NCI (survivors/nonsurvivors). Data were collected from electronic versions of case records. Demographic, DM2-related, CKD-related and NCI-related baseline parameters/signs were studied as independent variables.RESULTS. 90 patients with DM2 and CKD stages 3 (Me GFR 43[37; 49] ml/ min/1,73m2) were included, mean age 70 [69; 78] y, females – 56 %, the mortality rate – 21 %. The independent predictors of NCI adverse outcome were detected using a single factor analysis (odds ratio). Among them are: initial prandial glycemia ≥ 10 mmol/l (ОR 11,8; 95 % CI 3,13–44,9; р <0,001), albuminemia at admission ≤ 35 g/l (ОR 5,52; 95 % CI 1,85–16,55; р = 0,012), initial proteinuria ≥ 1 g/л (ОR 6,69; 95 % CI 1,95–23,00; р = 0,002), News2 ≥ 5 at admission (ОR 14,7; 95 % CI 3,15–48,8; р <0,001), lung damage CT 3–4 at admission (ОR 31,7; 95 % CI 6,59–52,85; р = 0,04). A prognostic model was constructed to determine the risk of lethal outcome using logistic regression method. The detected risk factors were used as variables. The predictive value of the model was 93 % according to ROC-analyses data.CONCLUSION. The detected predictors of adverse outcome are the part of routine screening available in pre-hospital setting and at hospital admission. Early identification of predictors allows optimizing patient routing and selecting the best treatment strategy for each patient.