The ongoing COVID-19 pandemic calls for extensive research on various medical topics. Since the beginning of the pandemic, multiple studies investigated the impact of SARS CoV-2 on thyroid function. However, crucial data, such as trend progression over time or influence of commonly used drugs, might still be missing. We checked the thyroid function in 174 patients with PCR-confirmed COVID-19. Our research covered three separate time points of hospitalization (days 1, 4, and 10). We did not exclude patients treated with glucocorticoids but, instead, compared them with patients not treated with steroids. We correlated the results of thyroid function tests with markers of systemic inflammation. We checked if abnormal thyroid function can predict unfavorable outcomes defined as combined primary endpoint and/or secondary endpoints; the combined primary endpoint was the occurrence of death, mechanical ventilation, non-invasive ventilation, vasopressor infusion, or prolonged hospital stay, and the secondary endpoint was any of the listed events. In general, 80.46% of evaluated patients displayed abnormalities in thyroid function tests over at least one time point throughout the observation. We noticed a high prevalence of features typical for thyroid dysfunction in non-thyroidal illness (NTI). Free triiodothyronine (fT3) concentration was significantly lower in the group requiring glucocorticoids. Patients displaying abnormal thyroid function were statistically more likely to meet the predefined combined primary endpoint. We found that fT3 measured at admission could be perceived as an independent predictor of endpoint completion for all analyzed groups. Thyroid involvement is common in COVID-19. Our study supports the idea of thyroid function abnormalities being important clinical tools and allowing early recognition of possible detrimental outcomes of the disease.