ABSTRACT
In this study, we used administrative data of six health insurance companies in the Czech Republic, to analyse the incidence of hospitalization in COVID-19 patients early treated with corticosteroids (CS). Cohort selection was based on index date defined by the date of the PCR test reported with COVID-19-related ICD-10 diagnoses found within a period of March 2020 - November 2020. The CS treated cohort was defined patients treated with CS within the 7 days after index date with no history of antibiotics nor corticosteroids within 30 days before index date. The control groups were defined either with or without a use of codeine within the 7 days after index date but always with no use of antibiotics nor corticosteroids within 30 days before and within the 7 days after index date. In total 590 274 records fulfilled the inclusion criteria, of whom 3 670 (0.62%) were hospitalized and 2 467 (0.42%) used CS within 7 days after ID. The contrast between a studied medication and a control was analysed with the logistic regression analysis with adjustment on age. Treatment with CS as compared with a cohort of patients with no curative treatment was associated with significantly (Chi square test p<.0001) increased rate of hospitalization in those treated with CS with odds ratio calculated as 3.1. Compared with a cohort of patients taking codeine the CS was associated with significantly (Chi square test p=0.001) lower rate of hospitalization, giving a moderate odds ratio 0.6. No beneficial effect of CS can be concluded based on this study. The strong influence of the composition of the control group is also evident from this study, as is evident from the reverse effect of the use of codeine in the control group. © 2022 IEEE.