key: cord-0855161-6veeygzp
authors: Yu, Caizheng; Lei, Qing; Li, Wenkai; Wang, Xiong; Liu, Wei; Fan, Xionglin; Li, Wengang
title: Clinical Characteristics, Associated Factors, and Predicting COVID-19 Mortality Risk: A Retrospective Study in Wuhan, China
date: 2020-05-27
journal: Am J Prev Med
DOI: 10.1016/j.amepre.2020.05.002
sha: 98be6b54750ba939dff9a90c7f476b21f353cbfc
doc_id: 855161
cord_uid: 6veeygzp

Introduction Coronavirus disease 2019 (COVID-19) has become a serious global pandemic. This study investigates the clinical characteristics and risk factors for COVID-19 mortality, and establishes a novel scoring system to predict mortality risk in COVID-19 patients. Methods A cohort of 1,663 hospitalized COVID-19 patients in Wuhan, China, of whom 212 died and 1,252 recovered, were included in the present study. Demographic, clinical, and laboratory data on admission were collected from electronic medical records between January 14, 2020, and February 28, 2020. Clinical outcomes were collected until March 26, 2020. Multivariable logistic regression was used to explore the association between potential risk factors and COVID-19 mortality. The receiver operating characteristic curve was used to predict COVID-19 mortality risk. All analyses were conducted in April 2020. Results Multivariable regression showed increased odds of COVID-19 mortality associated with older age (OR=2.15, 95% CI=1.35, 3.43), male sex (OR=1.97, 95% CI=1.29, 2.99), history of diabetes (OR=2.34, 95% CI=1.45, 3.76), lymphopenia (OR=1.59, 95% CI=1.03, 2.46), and increased procalcitonin (OR=3.91, 95% CI=2.22, 6.91, per SD increase) on admission. Spline regression analysis indicated that the correlation between procalcitonin levels and COVID-19 mortality was nonlinear (p=0.0004 for nonlinearity). The area under the receiver operating curve of the COVID-19 mortality risk was 0.765 (95% CI=0.725, 0.805). Conclusions The independent risk factors for COVID-19 mortality included older age, male sex, history of diabetes, lymphopenia, and increased procalcitonin, which could help clinicians to identify COVID-19 patients with poor prognosis at an earlier stage. The COVID-19 mortality risk score model may assist clinicians in reducing COVID-19-related mortality by implementing better strategies for more effective use of limited medical resources.

This study investigates the clinical characteristics and risk factors for COVID-19 mortality, and establishes a novel scoring system to predict mortality risk in COVID-19 patients.

Methods: A cohort of 1,663 hospitalized COVID-19 patients in Wuhan, China, of whom 212 died and 1,252 recovered, were included in the present study. Demographic, clinical, and laboratory data on admission were collected from electronic medical records between January 14, 2020, and February 28, 2020. Clinical outcomes were collected until March 26, 2020.

Multivariable logistic regression was used to explore the association between potential risk factors and COVID-19 mortality. The receiver operating characteristic curve was used to predict COVID-19 mortality risk. All analyses were conducted in April 2020. 

The independent risk factors for COVID-19 mortality included older age, male sex, history of diabetes, lymphopenia, and increased procalcitonin, which could help clinicians to identify COVID-19 patients with poor prognosis at an earlier stage. The COVID-19 mortality risk score model may assist clinicians in reducing COVID-19-related mortality by implementing better strategies for more effective use of limited medical resources.

In December 2019, several pneumonia cases of unknown origin were identified in Wuhan, Hubei, China. 1, 2 The pathogen has been identified as a novel coronavirus, belonging to the βcoronavirus genus, and has been renamed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2, previously named 2019-nCoV). 3 This novel virus shared 87.99% sequence identity to Bat SARS-like coronavirus, and 79.5% of its sequence with severe acute respiratory syndrome coronavirus (SARS-CoV). 4,5 SARS-CoV-2 has a strong affinity for angiotensinconverting enzyme 2 receptors, which was an early indicator of its potential for becoming a pandemic threat. 6 The clinical characteristics of COVID-19 have been well described, 1, [8] [9] [10] [11] but there are few published analyses focused specifically on COVID-19 mortality. 12 In addition, there have been limited studies exploring the potential risk factors for COVID-19 mortality. Therefore, the present study examines potential risk factors for COVID-19 mortality and aims to establish a COVID-19 mortality risk prediction model at a single-center hospital.

The authors obtained the medical records of 1,663 hospitalized patients with laboratoryconfirmed diagnosed COVID-19 from Tongji Hospital between January 14, 2020, and February 28, 2020. As of March 26, the clinical outcomes of the total hospitalized patient population were collected. After exclusion of patients who were still hospitalized (n=196) or transferred to other hospitals (n=3), a total of 1,464 eligible patients were included in the final analysis. Patients missing procalcitonin (PCT; n=324) and lymphocyte count (LY; n=117) data were further excluded leaving 1,140 and 1,347 patients included in the analyses of PCT and LY with COVID-19 mortality, respectively. The study population selection is shown in Appendix Figure 1 .

A laboratory-confirmed case of COVID-19 was defined as a positive real-time reverse transcriptase polymerase chain reaction (RT-PCR) test result assay obtained through oral pharyngeal swab specimens. Investigators collected demographic information, exposure history, medical history, comorbidities, signs and symptoms, chest computed tomography, laboratory findings on admission, and clinical outcomes from electronic medical records. Laboratory results (blood count, chemical analysis, and coagulation testing) were included in laboratory testing.

The date of disease onset, SARS-CoV-2 laboratory confirmation, hospital admission, discharge, and death were also recorded. The study was approved by Tongji Hospital Ethics Committee.

Oral pharyngeal swab samples (stored in 5-mL virus preservation solution) were collected for SARS-CoV-2 viral nucleic acid detection. Virus RNA was extracted within 24 hours by Tianlong PANA9600 automatic nucleic acid extraction system (Tianlong, China Referring to previous studies, [12] [13] [14] [15] [16] age (<65 years, ≥65 years), sex (female, male), history of hypertension (yes/no) and diabetes (yes/no), lymphopenia (<1.1×10 9 /L, ≥1.1×10 9 /L), increased alanine aminotransferase (<40 U/L, ≥41 U/L), increased lactate dehydrogenase (<214 U/L, ≥214 U/L), increased D-dimer (<0.5 mg/L, ≥0.5 mg/L), and increased PCT (<0.05 ng/mL, ≥0.05 ng/mL) were included in multivariable logistic regression model. In the analysis of PCT and LY with COVID-19 mortality, PCT and LY were categorized into three groups according to the tertile of distribution. The p-value for trend was calculated from group medians. The association of PCT concentration with risk of COVID-19 mortality was also evaluated using restricted cubic splines, with 3 konts defined at the 5th, 50th, and 90th percentiles of the PCT concentrations; the reference value (OR=1) was 0.05 ng/mL for PCT concentrations; data from the <5th and >95th percentiles were deleted. Variables that were at a statistically significant level (p<0.05) in the multivariable logistic regression were included in the prediction model. The receiver operating characteristic curve was used for prediction of COVID-19 mortality, and the Youden index was used to identify the optimal cut off point. 17 The novel scoring model was established, and the mortality risk scores were determined by multivariate logistic regression to reflect their weights of impact on the COVID-19 mortality. The mortality risk score was calculated according to the ORs and rounded to the nearest integer. 18 The total risk score was the sum of the scores of each variable (age, sex, history of diabetes, lymphopenia, and increased PCT). SPSS version, 13.0 and SAS, version 9.4 were used to conduct all statistical analyses. All analyses were conducted in April 2020. Two-sided statistical tests were considered significant at p<0.05.

Baseline characteristics of the recovered patients (n=1,252) and patients who died from COVID-19 (n=212) are shown in Table 1 Compared with recovered patients, those who died from COVID-19 were more likely to be male, older, and tended to have a shorter time from onset of symptoms to death and shorter time of hospital stay (all p<0.01). In addition, patients who died from COVID-19 had a higher proportion of comorbidities including hypertension, diabetes, and coronary heart disease (all p<0.01) and the presence of clinical symptoms such as fever, cough, and fatigue (all p<0.01).

More than 50% of patients had decreased LY (50.5%), and increased levels of lactate dehydrogenase (75.8%), increased C-reactive protein (79.9%), and increased D-dimer (62.9%).

Increased PCT (49%), mononucleosis (26.2%), increased alanine aminotransferase (20.3%), and increased aspartate aminotransferase (19.9%) were observed. Additionally, lymphopenia, neutrophilia, thrombocytopenia, leukocytosis, increased aspartate aminotransferase, increased creatinine, increased lactate dehydrogenase, increased D-dimer, prolonged thrombin time, and increased PCT were significantly different between recovered patients and those who died from COVID-19 (all p<0.05). Table 2 Table 1 ). Further adjustment for lymphopenia, increased alanine aminotransferase, increased lactate dehydrogenase, and increased D-dimer did not substantially change the association. A 1-SD (SD=0.6 ng/mL) increase in PCT concentration was associated with a 3.91-fold increased risk of COVID-19 mortality after adjustment for potential confounders (OR=3.91, 95% CI=2.22, 6.91).

Further, spline regression analysis indicated that the association between PCT concentrations and COVID-19 mortality was nonlinear (p=0.0004 for nonlinearity) (Appendix Figure 2 ).

The association of LY levels with risk of COVID-19 mortality is presented in Appendix Table 1 Table 1 ). After further adjustment for increased aspartate aminotransferase, increased creatinine, increased D-dimer, and increased PCT, LY T1 had marginally higher risk of COVID-19 mortality, when compared with T3 (OR=1.64, 95% CI=0.95, 2.84; p=0.08 for trend) (Appendix Table 1 ). A 1-SD (SD=0.52 ×10 9 /L) decrease in LY concentration was associated with a 31% increased risk of COVID-19 mortality after adjustment for potential confounders (OR=1.31, 95% CI=1.03, 1.67) (Appendix Table 1 ). The categorical variable model for COVID-19 mortality prediction is shown in Table 3 . Age (<65 years, ≥65 years), sex (female, male), history of diabetes (yes/no), and increased PCT (<0.05 ng/mL, ≥0.05 ng/mL) were significantly associated with COVID-19 mortality, and lymphopenia (<1.1×10 9 /L, ≥1.1×10 9 /L) had marginal association with COVID-19 mortality. In order to more fully inform clinical utilization, the authors developed a novel scoring system for COVID-19 mortality risk ( Table 4 ). The optimal cut off point for COVID-19 mortality risk was 3, and the area under the receiver operating curve of the COVID-19 mortality risk score was 0.765 (95% CI=0.725, 0.805) (Appendix Figure 3 ).

In this study of hospitalized COVID-19 patients in Wuhan, China conducted between mid-January to late March 2020, the authors found patients who were male, elderly (>65 years), and had a history of diabetes, lymphopenia, and increased PCT tended to have higher odds of mortality. After further adjustment for potential confounders, significant independent associations were observed between older age, male sex, history of diabetes, lymphopenia, and increased PCT and higher risk of COVID-19 mortality.

The age (median=64.0 years, IQR=51.0-71.0 years) of the overall population in the present study was higher than that of individuals in other studies, which might be related to the fact that more serious patients were admitted to Tongji hospital. Consistent with a previous study, 13 the current study found that increased age was positively correlated with risk of COVID-19 mortality. Prior studies report older age was an independent predictor of mortality in SARS and MERS. 19 ,20 A macaque model found that older macaques tended to have stronger host innate responses to SARS-CoV infection compared with younger macaque. 21 Additionally, with increased age, T-cell and B-cell function become potentially more defective with overproduction of type 2 cytokines, which might be implicated in the poor clinical prognosis with COVID-19 infection. 13, 22 These findings might help explain the relationship between older age and COVID-19 mortality, as was observed in this and other studies.

Compared with female patients, male patients had higher odds of COVID-19 mortality after adjustment for potential risk factors, which was inconsistent with the findings from another study based on 191 patients from two different hospitals. 13 This might due in part to difference in the size of the study sample and different sociodemographic composition of the study populations.

However, other studies have also found that male patients tended to have higher risk of COVID-19 mortality, 12,23 consistent with the present study. Moreover, previous studies have reported that more men than women were affected by SARS and MERS infection. 19, 24 Compared with men, women may tend to have healthier lifestyles and behaviors combined with gender differences in immune response, which might explain the potential mechanism behind this observed sex difference. 25 Findings from the present study indicated that patients with a history of diabetes had higher odds of COVID-19 mortality after adjustment for potential risk factors. Previous studies found that the presence of diabetes increased morbidity and mortality in patients with COVID-19, which was consistent with the present findings. 12, 26 In addition, a prior study found that plasma glucose levels and diabetes were independent predictors for mortality and morbidity in patients with SARS. 27 Patients with diabetes tended to have higher affinity cellular binding and efficient virus entry, decreased viral clearance, diminished T cell function, and increased susceptibility to hyper-inflammation and cytokine storm syndrome, which could all be contributing factors to greater susceptibility to COVID-19 among diabetics and their generally poorer prognosis. 28 No studies have yet investigated whether PCT is an independent risk factor for COVID-19

mortality. This study showed that PCT concentrations were positively correlated with COVID-19 mortality after adjustment for potential risk factors. Although the inflammatory mediator PCT is an established marker of bacterial infection and antibiotic stewardship, 29, 30 PCT has been reported to be associated with clinical prognosis in myocardial infarction, cancer, sepsis patients, and ventilator-associated pneumonia. 15, 16, 31 In addition, previous studies have found that a high PCT concentration was an independent prognostic biomarker of mortality risk in both healthy populations and critically ill patients. 15, 16, 32 In sepsis, PCT promotes inflammation and immunosuppression, and can play a dual role as a biomarker of diagnosis and prognosis as well as a disease mediator. 33 In vitro, Liappis et al. 34 Therefore, PCT-increased COVID-19 mortality might be implicated in the induction of proinflammatory cytokines although the exact mechanism behind this requires further investigation.

Previous studies have shown that deceased patients with COVID-19 tended to have lower lymphocyte count, which is consistent with the findings of the present study. 12, 13 However, Zhou and colleagues 13 did not find a significant association between lymphopenia and COVID-19 mortality after adjustment for potential risk factors. Different sample size of the study population and their sociodemographic composition again might explain some of these observed differences. The present study revealed that patients with lymphopenia had higher risk of COVID-19 mortality whereas a prior study found the percentage of lymphocytes in the blood was negatively correlated with the severity and prognosis of COVID-19. 36 SARS-CoV-2 might contribute to the destruction of lymphatic organs, cause lymphocytic dysfunction, induce apoptosis or necrosis of lymphocytes, and suppress lymphocytes via disordered metabolic molecules, which might work collectively to result in lymphopenia. 36 Further studies are needed to clarify the underlying mechanism.

The strengths of present study include the relatively large sample size, and the ability to investigate the associations between potential risk factors and COVID-19 mortality with moderate statistical power. This study is the first that the authors are aware to report that male sex, increased PCT levels, and lymphopenia are independent risk factors for COVID-19 mortality. Additionally, a novel scoring system was established to predict mortality risk in COVID-19 patients in the present study.

Nonetheless, some limitations should be taken into consideration. First, the present study was performed in single medical center; thus, the findings may not be representative of the general population. Second, the authors have not yet collected information on treatments in the present study. The mechanism between risk factors and COVID-19 mortality still requires further study.

This study of hospitalized COVID-19 patients in Wuhan, China, found many patients had at least one comorbidity with hypertension, diabetes, and coronary heart disease as the most common pre-existing conditions. Older age, male sex, history of diabetes, lymphopenia, and increased PCT on admission had significant associations with COVID-19 mortality. These independent risk factors can assist clinicians in identifying patients who are likely to have a poorer prognosis at an early stage in the clinical course of disease. In addition, the COVID-19 mortality risk score model developed in this study is intended to help clinicians reduce the COVID-19 related mortality by implementing better strategies for more effective use of limited medical resources.

This work was supported by the grants from the HUST COVID-19 Rapid Response Call (No. 2020kfyXGYJ040).

The authors would like to thank all study subjects for participating in the present study.

CY and QL contributed equally to this work. CY, XF, and WL were co-corresponding authors.

CY, QL, XF, and WL designed the study, interpreted data, and wrote the first draft of the paper.

CY, QL, WL, and XW take responsibility for the accuracy of the data analysis. CY, QL, WL, and WL performed data collection and designed the study's analytic strategy. All authors have read and approved the final manuscript.

No financial disclosures were reported by the authors of this paper. 

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

Outbreak of pneumonia of unknown etiology in Wuhan, China: the mystery and the miracle

A novel coronavirus from patients with pneumonia in China

Notes from the field: a novel coronavirus genome identified in a cluster of pneumonia cases -Wuhan

A pneumonia outbreak associated with a new coronavirus of probable bat origin

Evolution of the novel coronavirus from the ongoing Wuhan outbreak and modeling of its spike protein for risk of human transmission

Coronavirus disease (COVID-2019) situation reports

Characteristics of and important lessons from the Coronavirus Disease 2019 (COVID-19) outbreak in China: summary of a report of 72314 cases from the Chinese Center for Disease Control and Prevention

Clinical characteristics of 140 patients infected with SARS-CoV-2 in Wuhan, China. Allergy. In press

Clinical features of patients infected with 2019 novel coronavirus in Wuhan

Clinical characteristics of 138 hospitalized patients with 2019 novel coronavirus-infected pneumonia in Wuhan

Clinical characteristics of 113 deceased patients with coronavirus disease 2019: retrospective study

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Prediction for progression risk in patients with COVID-19 pneumonia: the CALL Score

Serum procalcitonin level and mortality risk in critically ill patients with ventilator-associated pneumonia

Plasma procalcitonin is associated with all-cause and cancer mortality in apparently healthy men: a prospective population-based study

Nonlaboratory-based risk assessment algorithm for undiagnosed type 2 diabetes developed on a nation-wide diabetes survey

Studies reporting ROC curves of diagnostic and prediction data can be incorporated into meta-analyses using corresponding odds ratios

Outcomes and prognostic factors in 267 patients with severe acute respiratory syndrome in Hong Kong

Predictors of mortality in Middle East respiratory syndrome (MERS)

Exacerbated innate host response to SARS-CoV in aged non-human primates

The immunopathogenesis of sepsis in elderly patients

Clinical features of 85 fatal cases of COVID-19 from Wuhan: a retrospective observational study

Middle East respiratory syndrome coronavirus (MERS-CoV) outbreak in South Korea, 2015: epidemiology, characteristics and public health implications

Sex differences in immune responses

Case-fatality rate and characteristics of patients dying in relation to COVID-19 in Italy

Plasma glucose levels and diabetes are independent predictors for mortality and morbidity in patients with SARS

COVID-19 pandemic, corona viruses, and diabetes mellitus

Procalcitonin-guided diagnosis and antibiotic stewardship revisited

Procalcitonin to initiate or discontinue antibiotics in acute respiratory tract infections

Serial procalcitonin predicts mortality in severe sepsis patients: results from the Multicenter Procalcitonin MOnitoring SEpsis (MOSES) Study

Procalcitonin increase in early identification of critically ill patients at high risk of mortality

Procalcitonin in sepsis and systemic inflammation: a harmful biomarker and a therapeutic target

Exogenous procalcitonin evokes a proinflammatory cytokine response

COVID-19: immunopathology and its implications for therapy

Lymphopenia predicts disease severity of COVID-19: a descriptive and predictive study