ABSTRACT
Background and Aim: Coronavirus is an infectious disease that is now known as an epidemic, early and accurate diagnosis helps the patient receive more care. The aim of this study is to investigate Covid-19 using blood tests and multilayer perceptron neural network and affective factors in improving and preventing Covid-19. Methods: This cross-sectional study was performed on 200 patients referred to Sina Hospital, Tehran, Iran, who were confirmed cases of Covid-19 by computerized tomography-scan analysis between 2 March 2020 to 5 April 2020. After verification of lung involvement, blood sampling was done to separate the sera for C-reactive protein (CRP), magnesium (Mg), lymphocyte percentage, and vitamin D analysis in healthy and unhealthy people. Blood samples from healthy and sick people were applied to the multilayer perceptron network for 70% of the data for training and 30% for testing. Result: By examining the features, it was found that in patients with Covid-19, there was a significant relationship between increased CRP and decreased lymphocyte levels, and increased Mg (p < 0.01). In these patients, the amount of CRP and Mg in women and the number of lymphocytes and vitamin D in men were significantly higher (p < 0.01). Conclusion: The important advantage of using a multilayer perceptron neural network is to speed up the diagnosis and treatment.
ABSTRACT
BACKGROUND: Without an adequate immune response, SARS-CoV2 virus can simply spread throughout the body of the host. Two of the well-known immunonutrients are selenium (Se) and zinc (Zn). Se and Zn deficiency might lead to inflammation, oxidative stress, and viral entry into the cells by decreasing ACE-2 expression; three factors that are proposed to be involved in COVID-19 pathogenesis. Thus, in the current study we aimed at evaluating the correlation between serum Se and Zn status and COVID-19 severity. METHODS: Eighty-four COVID-19 patients were enrolled in this observational study. Patients were diagnosed based on an infectious disease specialist diagnosis, using WHO interim guidance and the recommendations of the Iranian National Committee of Covid-19. The patients with acute respiratory tract infection symptoms were checked for compatibility of chest computed tomography (CT) scan results with that of Covid-19 and Real-time polymerase chain reaction (RT-PCR) for corona virus infection. The severity of Covid-19 was categorized into three groups (mild, moderate, and severe) using CDC criteria. Serum Zn and Se level of all subjects was measured. The severity of the disease was determined only once at the onset of disease. RESULTS: According to the results of linear regression test, there was a significant association between Zn and Se level and COVID-19 severity (ß = - 0.28, P-value = 0.01 for Se; ß = - 0.26, P-value = 0.02). However the significance disappeared after adjusting for confounding factors. Spearman correlation analysis showed a significant negative association between serum Zn, Se and CRP level (r = - 0.35, P-value = 0.001 for Se; r = - 0.41, P-value < 0.001 for Zn). CONCLUSION: Results suggest that increasing levels of Se and Zn were accompanied by a decrease in serum CRP level. However, the significant association between Se, Zn, and disease severity was lost after adjusting for confounding factors.