Scoring system for identifying Japanese patients with COVID-19 at risk of requiring oxygen supply: A retrospective single-center study

Kitajima, Heita; Hirashima, Tomonori; Suzuki, Hidekazu; Arai, Tsuyoshi; Tamura, Yoshitaka; Hashimoto, Shoji; Morishita, Hiroshi; Matsuoka, Hiroto; Kashiwa, Yozo; Han, Yuki; Minamoto, Seijiro; Tanaka, Toshio; Nagai, Takayuki

Introduction Japan is facing the threat of medical system collapse due to the rapid spread of coronavirus disease 2019 (COVID-19). The present scoring system may help assess disease severity and oxygen supply requirements in COVID-19 patients. Methods Data on patient characteristics at baseline and throughout hospitalization for COVID-19 were extracted from medical records. Disease severity was dichotomized into two categories without or with oxygen supply as asymptomatic, mild, and moderate illness (AMMI), and severe and critical illness (CSI). The AMMI and CSI groups were compared. Predictors of disease severity, previously identified in the outpatient setting, were included in multivariable logistic regression analysis; the obtained coefficients were converted to integers and assigned a score. Results A total of 206 patients diagnosed with COVID-19 were included in this study. Correlation between COVID-19 severity and medical information was examined by comparing AMMI and CSI. Age, hemodialysis, and C-reactive protein (CRP) levels were candidate predictors of the need for oxygen supply in patients with COVID-19. Coefficients associated with age, hemodialysis, and CRP were as follows: 1×age (in years, coded as 0 for values of <50, and as 1 for values of ≥50) + 1× hemodialysis (coded as 0 for “no”, and as 1 for “yes”) + 1×CRP (in mg/dL, coded as 0 for values of <1.0, and as 1 for values of ≥1.0). Patients with scores of ≥2 points required oxygen supply (sensitivity, 68.4%; specificity, 79.0%) Conclusion The present model can help predict disease severity and oxygen requirements in COVID-19 patients in Japan.