Clinical, radiological, and laboratory characteristics and risk factors for severity and mortality of 289 hospitalized COVID-19 patients

Abstract

Background: The coronavirus disease 2019 (COVID-19) has become a global pandemic, with 10%-20% of severe cases and over 508 000 deaths worldwide.

Objective: This study aims to address the risk factors associated with the severity of COVID-19 patients and the mortality of severe patients.

Methods: 289 hospitalized laboratory-confirmed COVID-19 patients were included in this study. Electronic medical records, including patient demographics, clinical manifestation, comorbidities, laboratory tests results, and radiological materials, were collected and analyzed. According to the severity and outcomes of the patients, they were divided into three groups: nonsurvived (n = 49), survived severe (n = 78), and nonsevere (n = 162) groups. Clinical, laboratory, and radiological data were compared among these groups. Principal component analysis (PCA) was applied to reduce the dimensionality and visualize the patients on a low-dimensional space. Correlations between clinical, radiological, and laboratory parameters were investigated. Univariate and multivariate logistic regression methods were used to determine the risk factors associated with mortality in severe patients. Longitudinal changes of laboratory findings of survived severe cases and nonsurvived cases during hospital stay were also collected.

Results: Of the 289 patients, the median age was 57 years (range, 22-88) and 155 (53.4%) patients were male. As of the final follow-up date of this study, 240 (83.0%) patients were discharged from the hospital and 49 (17.0%) patients died. Elder age, underlying comorbidities, and increased laboratory variables, such as leukocyte count, neutrophil count, neutrophil-to-lymphocyte ratio (NLR), C-reactive protein (CRP), procalcitonin (PCT), D-dimer, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and blood urea nitrogen (BUN) on admission, were found in survived severe cases compared to nonsevere cases. According to the multivariate logistic regression analysis, elder age, a higher number of affected lobes, elevated CRP levels on admission, increased prevalence of chest tightness/dyspnea, and smoking history were independent risk factors for death of severe patients. A trajectory in PCA was observed from "nonsevere" toward "nonsurvived" via "severe and survived" patients. Strong correlations between the age of patients, the affected lobe numbers, and laboratory variables were identified. Dynamic changes of laboratory findings of survived severe cases and nonsurvived cases during hospital stay showed that continuing increase of leukocytes and neutrophil count, sustained lymphopenia and eosinopenia, progressing decrease in platelet count, as well as high levels of NLR, CRP, PCT, AST, BUN, and serum creatinine were associated with in-hospital death.

Conclusions: Survived severe and nonsurvived COVID-19 patients had distinct clinical and laboratory characteristics, which were separated by principle component analysis. Elder age, increased number of affected lobes, higher levels of serum CRP, chest tightness/dyspnea, and smoking history were risk factors for mortality of severe COVID-19 patients. Longitudinal changes of laboratory findings may be helpful in predicting disease progression and clinical outcome of severe patients.

Keywords: clinical characteristics; coronavirus disease 2019; mortality; risk factors; severity.

FIGURE 1

Difference in risk factors of death among three groups of COVID‐19 patients with different severity and outcomes. Demographic parameters and variables of laboratory tests (data on admission) with significant differences among the three groups of patients are illustrated, including continuous variables (A‐E) and categorical variables (F‐K). Elder age (A), increased number of affected lobes (B), increased leukocyte count (C), elevated levels of CRP (D) and BUN (E), higher prevalence of patients with smoking history (F), dyspnea (G), a larger proportion of patients with increased NLR (H), PCT (I), BUN (J), and serum creatinine (K) were identified in the nonsurvived group compared to the survived severe group. Percentages in the bars of F‐K represent the percentages of patients with specific demographic/abnormal laboratory findings in each subgroup. Continuous variables of the three groups (A‐E) were compared using one‐way ANOVA test or Kruskal‐Wallis test, as appropriate. Categorical variables of the nonsurvived and survived severe groups (F‐K) were compared via chi‐square test or Fisher's exact test, as appropriate. * denotes a P value of ≤.05, ** denotes P ≤ .01, *** denotes P ≤ .001. BUN, blood urea nitrogen; CRP, C‐reactive protein; NLR, neutrophil‐to‐lymphocyte ratio; PCT, procalcitonin

FIGURE 2

Principal Component Analysis (PCA). Principal Component Analysis (PCA) was used for dimensionality reduction and visualization of the patients. All patients were included in the analysis; parameters including laboratory results on admission, age, and affected lobe numbers were used in the analysis; results are represented by colored dots separated by three groups of severity. Despite no clear separation between the three groups, there was a clear trajectory from "nonsevere" toward "nonsurvived" via " survived severe"

FIGURE 3

Heatmap of Spearman correlations among laboratory results, as well as with age and affected lobe numbers. Spearman correlation heatmap with correlation coefficient and significance levels based on the laboratory results on admission, as well as patients’ age and affected lobe number. Positive correlations are marked in red and negative ones in blue (color scale on the right side). * denotes P ≤ .05, ** denotes P ≤ .01, *** denotes P ≤ .001. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; CK, serum creatinine kinase; CK‐MB, creatinine kinase‐MB; CRP, C‐reactive protein; D‐D, D‐dimer; NLR, neutrophil‐to‐lymphocyte ratio; PCT, procalcitonin; SAA, serum amyloid A

FIGURE 4

Selected Spearman correlations between the number of affected lobe number(s), age of patients and laboratory parameters in COVID‐19 patients. Scatter plots showing the correlations between affected lobe numbers, age of patients, and laboratory variables (data on admission). Strong positive correlations were observed in all plots. Spearman's test was used to evaluate the correlations. BUN, blood urea nitrogen; CRP, C‐reactive protein; D‐D, D‐dimer; NLR, neutrophil‐to‐lymphocyte ratio; PCT, procalcitonin

FIGURE 5

Differences in longitudinal course of laboratory findings between nonsurvived and survived severe cases. Severe COVID‐19 patients were divided into nonsurvived and survived severe groups according to the clinical outcomes as of March 28, 2020. Data from patients with available laboratory results on admission, 3‐7 days after admission, and 8‐14 days after admission are shown; “n” represents the number of patients with available follow‐up data for each parameter. The red lines represent the values of nonsurvived patients of each parameter, and the blue lines show the values of survived severe patients; * denotes P ≤ .05, ** denotes P ≤ .01, *** denotes P ≤ .001. AST, aspartate aminotransferase; BUN, blood urea nitrogen; CRP, C‐reactive protein; NLR, neutrophil‐to‐lymphocyte ratio; PCT, procalcitonin