RESUMEN
BackgroundSince the end of 2019, physicians became more and more familiar with SARS-CoV-2 infection and the variety of forms in which it may present and evolve. There have been a lot of studies trying to understand and predict why some patients develop a dysregulation of the immune response, with an exaggerated release of pro-inflammatory cytokines, called cytokine storm (1–4). There is scarce evidence in Romania regarding this aspect.ObjectivesThis study aims to verify the correlation between some laboratory parameters and the development of cytokine storm in SARS-CoV-2 infection in a cohort of over 200 patients admitted in a tertiary hospital from Romania, hoping that early identification of these risk factors of progression to a severe form of the disease can bring considerable benefit to patient care.MethodsThis is an analytical, observational, case-control study which includes 219 patients (all COVID-19 hospitalized patients on the Internal Medicine 3 department of Colentina Clinical Hospital, Bucharest, from 01 March 2020 to 1 April 2021). A series of data were collected, the laboratory parameters being the most important, including: albumin, lymphocyte (percentage), neutrophil (absolute value), aspartate aminotransferase, alanine aminotransferase, D-dimers, lactate dehydrogenase (LDH), anionic gap, chloremia, potassium and the BUN:creatinine ratio (BUN - blood urea nitrogen). The laboratory parameters used for the statistical analysis represent the average values of the first 7 days of hospitalization for those who did not develop cytokine storm, respectively until the day of its development, for the others. Patients were classified into these groups, those who developed cytokine storm, respectively those who did not have this complication taking into account the clinical and paraclinical criteria (impairment of respiratory function, elevations of certain markers 2-3 times above the upper limit of normal, those who died as a result of SARS-CoV-2 infection). Then Binary Univariate Logistic Regression was applied in order to verify the individual impact of every laboratory parameter on cytokine storm development. Furthermore, all laboratory parameters were subsequently included in the multivariate analysis, using the backward selection technique to achieve a model as predictive as possible.ResultsWe mention that the analysis of demographic data was previously performed, showing no statistically significant relationship between patient gender, age or comorbidities (history of neoplasm, lung diseases, cardiac pathology, obesity, type II diabetes and hypertension) and their evolution to cytokine storm. After performing binary univariate logistic regression we concluded that 8 of the 13 laboratory analyzes have had a significant change between groups (ferritin, PCR, albumin, Lymphocyte, Neutrophils, TGO, LDH, BUN:creatinine ratio). Only 150 patients were then included in the multivariate analysis. After the analysis, some of the variables lost their statistical significance, the final model including C-reactive protein, neutrophilia, LDH, ferritin and the BUN:creatinine ratio. This model correctly predicts the development of cytokine storm in 88% of cases.ConclusionHigh C-reactive protein, neutrophilia, LDH, ferritin and the BUN:creatinine ratio are risk factors for cytokine storm development and should be monitored in all COVID-19 patients in order to predict their evolution.References[1]Pedersen SF et all. SARS-CoV-2: A storm is raging[2]Mehta P et al. COVID-19: consider cytokine storm syndromes and immunosuppression[3]Hu B et al. The cytokine storm and COVID-19.[4]Caricchio R et al. Preliminary predictive criteria for COVID-19 cytokine stormAcknowledgements:NIL.Disclosure of InterestsNone Declared.