Background : The objective of this cohort study was to determine whether elevated CRP in early COVID-19 was associated with 14-day mortality in geriatric patients. Methods : Plasma CRP levels at hospital admission and 14-day all-cause mortality were assessed in geriatric inpatients hospitalized for COVID-19. Potential confounders were age, sex, functional abilities, history of malignancies, hypertension, cardiomyopathy, albuminemia, number of acute health issues, use of antibiotics and respiratory treatments. Results : Ninety-five participants (mean±SD 88.0±5.5years; 49.5%women; mean CRP, 76.7±77.5mg/L; mean albuminemia, 32.9±6.0g/L) were included. Sixteen participants who did not survive at day 14 exhibited higher CRP level at baseline than the others (120.3±71.2 versus 67.9±76.1 mg/L, P = 0.002). There was no difference in albuminemia (P = 0.329). Plasma CRP level was directly associated with 14-day mortality (fully adjusted HR = 1.11, P = 0.025). The cut-off for CRP associated with 14-day mortality was set at 35mg/L (sensitivity = 0.88; specificity = 0.56). Those with CRP<35mg/L had longer survival time than the others (log-rank P<0.001). Conclusions : Elevated CRP levels were associated with poorer 14-day survival in hospitalized geriatric COVID-19 patients.