Abstract
Systemic inflammation and hypercoagulopathy are known pathophysiological processes of coronavirus disease 2019 (COVID-19), particularly in patients with known cardiovascular disease or its risk factors (CVD). However, whether a cumulative assessment of these biomarkers at admission could contribute to the prediction of in-hospital outcomes remains unknown. The CLAVIS-COVID registry was a Japanese nationwide retrospective multicenter observational study, supported by the Japanese Circulation Society. Consecutive hospitalized patients with pre-existing CVD and COVID-19 were enrolled. Patients were stratified by the tertiles of CRP and D-dimer values at the time of admission. Multivariable Cox proportional hazard models were constructed. In 461 patients (65.5% male; median age, 70.0), the median baseline CRP and D-dimer was 58.3 (interquartile range, 18.2-116.0) mg/L and 1.5 (interquartile range, 0.8-3.0) mg/L, respectively. Overall, the in-hospital mortality rate was 16.5%, and the rates steadily increased in concordance with both CRP (5.0%, 15.0%, and 28.2%, respectively p < 0.001) and D-dimer values (6.8%, 19.6%, and 22.5%, respectively p = 0.001). Patients with the lowest tertiles of both biomarkers (CRP, 29.0 mg/L; D-dimer, 1.00 mg/L) were at extremely low risk of in-hospital mortality (0% until day 50, and 1.4% overall). Conversely, the elevation of both CRP and D-dimer levels was a significant predictor of in-hospital mortality (Hazard ratio, 2.97; 95% confidence interval, 1.57-5.60). A similar trend was observed when the biomarker threshold was set at a clinically relevant threshold. In conclusion, the combination of these abnormalities may provide a framework for rapid risk estimation for in-hospital COVID-19 patients with CVD.