Background
The outbreak of COVID-19 has aroused global concerns. Few studies focused on characteristics of COVID-19 nonsurvivors.
Objective In this study, we aimed to describe the clinical characteristics of 101 nonsurvivors with COVID-19 and analyze risk factors associated with the rapid disease progress to death after admission.
Design Retrospective study using electronic medical records.
Participants 101 hospitalized patients with confirmed COVID-19 died in Renmin Hospital of Wuhan University before February 15, 2020, were included in this study.
Main measures We obtained epidemiological, demographic, clinical, laboratory, and outcome data from electronic medical records. Univariate logistic regression analyses were performed to evaluate risk factors associated with rapid disease progress to death after admission.
Key results Among included nonsurvivors, the median age was 71.0 years (IQR, 59.0-80.0), 60 (59.4%) were men. Eighty-two (79.2%) had one or more comorbidities including hypertension (58.4%), diabetes (21.8%) etc. Respiratory failure (99.0%), acute cardiac injury (52.5%), sepsis (40.6%) and acute kidney injury (22.8%) were most common complications. Patients died within 3 days of admission were more likely to develop sepsis compared with those survived 3 days of admission. Higher heart rate and respiration rate, increased white blood cell count and neutrophil count, elevated myoglobin level and depressed oxygen saturation on admission were associated with rapid disease progress to death within 3 days of admission. The AUCs for white blood cell count and neutrophil count were 0.71 (95% CI, 0.60-0.80) and 0.70 (95% CI, 0.59-0.79), respectively.
Conclusions Older patients with underlying comorbidities suffering COVID-19 were at increased risks of death. Respiratory failure, acute cardiac injury and acute kidney injury played crucial roles in the death of COVID-19 patients. Of importance, early development of sepsis, increased white blood cell count and neutrophil count on admission were associated with the rapid disease progress to death.
Introduction
Since December 2019, the outbreak of coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) swept across China and emerged in many other countries have caused global concerns. 1 As of March 24, 2020, a total of 372,755 COVID-19 cases around the world have been confirmed and 16,231 lost their lives. 2 In the early outbreak of COVID-19, respiratory infection symptoms occurred, with some patients rapidly developed acute respiratory distress syndrome (ARDS), acute respiratory failure and even death. Compared with severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS), the concealed onset, similar or stronger contagion and lower mortality could be defined as the clinical features of the disease. 3, 4, 5 However, many patients suffered rapid disease progress to death after admission in the early breakout in Wuhan, China.
Previous studies focused on clinical characteristics and outcomes of COVID-19 patients varied from mild, severe, even deceased cases. 6, 7 But the causes and risk factors associated with rapid disease progress to death need to be addressed further. In this study, we aimed to describe the clinical characteristics and outcomes of nonsurvivors confirmed with SARS-CoV-2 infection in the early outbreak, and to analyze the possible risk factors associated with rapid disease progress to death after admission in patients with COVID-19.
Methods
Study Design and Participants
This study was approved by the Ethics Committee of Renmin Hospital of Wuhan University (No. WDRY2020-K059). 101 patients with confirmed COVID-19 admitted to Renmin Hospital of Wuhan University (including Shouyi campus and Eastern campus) and died before February 15, 2020 in Wuhan, China, were enrolled. Renmin Hospital of Wuhan University is one of the major tertiary hospitals in Wuhan, China and is a government designated hospital for COVID-19 treatment. All patients with COVID-19 enrolled in this study were diagnosed according to World Health Organization interim guidance 8. Large proportion of patients died within 3 days of admission in the early outbreak of COVID-19, to further investigate the causes and risk factors associated with the rapid disease progress to death, we also compared demographic, clinical, laboratory and radiological characteristics of patients died within 3 days with those survived 3 days of admission.
Procedures
We obtained epidemiological, demographic, clinical, laboratory, and outcome data from electronic medical records. Missing data were collected through direct communication with involved health-care providers or family member of patients. All data were checked by two sophisticated physicians.
The date of disease onset was defined as the day when the onset of symptoms were noticed. Acute respiratory distress syndrome (ARDS) and acute respiratory failure were defined according to the Berlin definition and previously reported. 9, 10 Acute kidney injury (AKI) was identified based on serum creatinine. 11 Acute liver injury was identified according to previous report 12. Acute cardiac injury was defined if the serum ultra-sensitive troponin I was above the 99th percentile upper reference limit 6. Sepsis was defined according to the Third International Consensus Definitions for Sepsis and Septic Shock. 13 Characteristic chest computer tomography (CT) imaging of these nonsurvivors were evaluated by previously described method. 14
The primary outcome was organ injury or failure (including lung, heart, kidney and liver etc.) and risk factors for the rapid disease progress to death. Secondary outcomes were to present epidemiological and demographic data, comorbidities, signs and symptoms on admission, laboratory indicators, CT findings and treatments, and to compare the differences of patients died within 3 days with those survived 3 days of admission.
Statistical analysis
We expressed descriptive data as median (IQR) for continuous variables and number (%) for categorical variables. We used Pearson’s chi-square test, the Mann-Whitney test and Fisher’s exact test for comparisons between patients died within 3 days and those died after 3 days of admission, as appropriate. Univariate logistic regression analyses were performed to evaluate risk factors associated with rapid death within 3 days of admission. Receiver operating characteristic (ROC) curves were used to analyze the performance of associated risk factors for rapid disease progress to death. The area under the curve (AUC) > 0.70 was defined as good performance. All statistical analyses were performed using SPSS version 22.0 software (SPSS Inc). P<0.05 was considered statistically significant.
Results
By February 15, 2020, 101 hospitalized patients with confirmed COVID-19 died in Renmin Hospital of Wuhan University (including Shouyi campus and Eastern campus) were included in this study. The median age was 71.0 (IQR, 59.0-80.0) years. 60 (59.41%) were males. There were 57 (56.4%) cases whose age over 70 years (Table 1). The most common symptoms were fever (76.2%), cough (58.4%), dyspnea (54.5%), fatigue (37.6%) and diarrhea (9.9%). Other less common symptoms were anorexia (5.9%), nausea and vomiting (2.0%) and abdominal pain (1.0%) (Table 1). The median duration from the onset of symptoms to admission and from admission to death were 9.0 (IQR, 5.0-13.0) days and 4.0 days (IQR, 2.0-7.0), respectively.
Forty-eight cases (47.52%) died within 3 days of admission, the median age was 72.0 (IQR, 59.0-78.0) years, 28 (58.3%) were male. Fifty-three cases (52.5%) died after 3 days of admission, the median age was 71.0 (IQR, 59.0-81.0) years, 32 (60.4%) were male. Compared with patients who survived 3 days, those died within 3 days of admission exhibited significantly higher respiration rate (25 vs 20 rpm, P=0.005) and heart rate (97 vs 84 bpm, P=0.006) on admission. There were no significant differences of self-report symptoms between two groups included fever, dyspnea, cough, fatigue etc. (Table 1). Among all the nonsurvivors, 81 (79.2%) patients had one or more underling medical conditions. The most common comorbidities included hypertension (58.4%), cardiovascular disease (22.8%), diabetes (20.8%), chronic pulmonary disease (14.9%), cerebrovascular disease (12.9%) (Table 1). There is no significant difference on above mentioned comorbidities between who died within 3 days of admission and those survived 3 days of admission (Table 1).
We observed substantial abnormalities (below or over normal range) in laboratory findings of these nonsurvivors on admission. Inflammation was obvious and bacterial infection was suspected as shown by elevated neutrophil count, C-reactive protein and procalcitonin levels. Lymphpenia was obvious on admission as revealed by decreased lymphocyte count, especially decreased CD3+ T cells below normal range. Elevated levels of ultra-sensitive troponin I, myoglobin, D-dimer, aspartate aminotransferase, lactate, decreased albumin level, depressed PaO2 and SpO2 demonstrated impaired organ functions of lung, heart, liver as well as coagulation disruption on admission (Table 2). There were several differences of laboratory parameters between patients who died within 3 days and those survived 3 days of admission. Patients died within 3 days of admission had significantly higher white blood cell count (10.8 vs 6.7×109/L, P= 0.001), neutrophil count (8.9 vs 5.5×109/L, P=0.001). For other biochemical indexes, creatine kinase-MB (3.3 vs 2.0 ng/ml, P=0.036), myoglobin (186.6 vs 118.7 ug/L, P=0.034) were significantly higher in patients died within 3 days of admission (Table 2). Compared with patients survived 3 days of admission, those died within 3 days had longer prothrombin time (13.2 vs 12.5 s, P=0.040) and higher D-dimer concentration (7.64 vs 2.82, P=0.040). Blood gas analyses showed patients died within 3 days presented with significantly lower oxygen saturation (85.0% vs 91.0%, P=0.008) and higher lactate level (2.9 vs 2.2 mmol/L, P=0.042) compared with those died after 3 days which reflected poor oxygenation in these patients (Table 2). Chest CT scans were used to evaluate the features of lung injury. Patients suffered severe lung damage and their lung imaging features presented vast or merged glass shadow, but quantification of lung injury revealed no significant difference between patients died within 3 days and those died after 3 days of admission (Table 2 and Figure 1).
Among treatment strategies, most patients received antiviral therapy (96.0%) and antibacterial therapy (98.0%). Many received glucocorticoid therapy (43.6%) and immunoglobulin therapy (35.6%). Seventy-eight (77.2%) patients received high flow oxygen inhalation, 41 (40.6%) received noninvasive mechanical ventilation while 14 (13.9%) received invasive mechanical ventilation and 5 (5.0%) received continuous renal replacement therapy (Table 3).
Regarding to organ functions, 100 (99.0%) patients developed respiratory failure, 53 (52.5%) had acute cardiac injury, 41 (40.6%) developed sepsis, acute kidney injury were developed in 24 cases (23.8%), acute liver dysfunction were found in 18 cases (17.8%) and 15 (14.9%) suffered coagulation disorders among which one patient died of massive upper gastrointestinal bleeding. Compared with patients died after 3 days of admission, patients died within 3 days of admission were more likely to develop sepsis. (52.1% vs 30.2%, P=0.025) (Table 3).
As summarized in Table 4, higher heart rate and respiration rate, increased white blood cell count and neutrophil count, elevated myoglobin concentration and depressed oxygen saturation were associated with rapid disease progress to death within 3 days of admission. The AUCs for white blood cell count and neutrophil count were 0.71 (95% CI 0.60-0.80) and 0.70 (95% CI 0.59-0.79), respectively. The best cut-off values for white blood cell count and neutrophil count were 8.13×109/L and 7.28×109/L, respectively (Figure 2). The AUCs of other indicators were with either poor sensitivity or specificity.
Discussion
As of February 15, 2020, 101 patients died because of COVID-19 were included in this study. In the early outbreak of COVID-19, major gaps in our knowledge of the epidemiology, clinical features and treatment measures of COVID-19 posed huge challenge for our clinical practice. Although previous study reported the mortality of patients with COVID-19 was 1.4%, 15 some special population including the aged and individuals with underlying comorbidities may be vulnerable to SARS-CoV-2 infection. In our study, the median age of nonsurvivors was 71.0 years, 81.2% of them combined with one or more underlying medical conditions including hypertension, cardiovascular disease, diabetes, chronic pulmonary disease etc. This suggested that aged individuals with chronic medical conditions were more likely to suffer a severe attack and death.
Common symptoms at onset of COVID-19 were fever, dry cough, myalgia, fatigue etc. 6, 16 In our study, we observed that besides fever and cough, dyspnea (54.5%) presented early in most nonsurvivors with confirmed COVID-19, similar finding has also been observed in critically ill patients reported by Yang etc. 17 Recent publication suggested the median time from onset of symptoms to hospital admission of deceased patients was 10.0 (IQR, 7.0-13.0) days, which tended to be longer than that of recovered patients (9.0 (6.0-12.0) days). 18 In our study, the median time from symptom onset to admission of patients died within 3 days and after 3 days of admission was 10.0 (6.0-13.0) days and 8.0 (4.0-14.0) days, respectively, but with no significantly difference between two groups. This indicated that although longer duration before admission might be associated with poor prognosis of COVID-19 patients, it was not contributed to the rapid disease progress to death within 3 days of admission.
Abnormal laboratory findings on admission suggested impaired pulmonary, cardiac, liver and coagulation functions as well as disturbed immune function of these nonsurvivors even before hospitalization. ARDS was the most common complication in nonsurviving patients with COVID-19 and oxygen saturation was an important manifestation of ARDS. In our study, SpO2 of patients died within 3 days of admission was 85.0% on admission, which was significantly lower compared with those died after 3 days of admission. These suggested that PaO2 and SpO2 on admission might serve as early warning signals of poor prognosis and should attract more attentions. In terms of SARS, severe respiratory failure was the leading cause of mortality, with little other organ failure. 19 In our study, it was noticed that 52.5% and 23.8% nonsurvivors suffered acute cardiac injury and AKI, respectively. Angiotensin-converting enzyme 2 (ACE2), a crucial SARS-CoV-2 receptor, distributed widely in heart, kidney and intestine etc. 20, 21, 22 Whether SARS-CoV-2 attack heart and kidney or other vital organs directly needs to be confirmed further.
Usually, neutrophils do not elevated significantly during viral infections. However, viral infection might impair the immune system and facilitate opportunistic infection and neutrophils are crucial in protection against bacterial infections. 23 Of interest, the occurrence of sepsis and significantly elevated neutrophils were noticed in patients died with COVID-19, especially in those died within 3 days of admission. This reminded us that secondary infection could be related to the rapid disease progress to death. Frequent detection of bacterial infection indicators like procalcitonin and early implementation of antibiotics might be beneficial for patients suspected of bacterial infections. Compared with patients died after 3 days of admission, prolonged prothrombin time and remarkably increased D-dimer were noticed in patients died within 3 days of admission, which together with depressed platelet count indicated coagulation disturbance and tendency of disseminated intravascular coagulation in these patients.
Up to now, no antiviral treatment has been recommended for coronavirus treatment except supportive care and organ support. 6, 24 In current clinical practice, arbidol was used as an antiviral treatment for SARS-CoV-2 infected patients, but the efficacy remains to be proved. Other antiviral medications including remdesivir, lopinavir and traditional Chinese medicine are still being verified in clinical trials. Application of glucocorticoid depended on the severity of inflammatory response. In our study, most nonsurvivors received high flow oxygen inhalation, but quite a low proportion of patients received machine ventilation partly due to limited medical resources against the steeply increasing number of patients in early outbreak.
Univariate logistic regression analyses and ROC were conducted to assess risk factors and their performance associated with the rapid disease progress to death within 3 days of admission in COVID-19 patients. Previous study has demonstrated that old age was an independent risk factor for death of COVID-19 patients. 25 In our study, the median age of nonsurvivors was 71.0 year, this also suggested that old age was a critical factor for the death of COVID-19 patients, but it was not associated with the rapid disease progress to death. The AUCs for white blood cell count and neutrophil count were 0.71 (95% CI, 0.60-0.80) and 0.70 (95% CI, 0.59-0.79), respectively. Other indicators including respiration rate, heart rate, oxygen saturation and myoglobin, to some extent, also related to the rapid disease progress to death of COVID-19 patients, but the performance of these indictors is not good due to either poor sensitivity or specificity.
Several limitations existed in this study. Firstly, not all laboratory tests were performed in all patients, the missing data might lead to bias and the roles of some indicators might be underestimated. Secondly, in the early outbreak of COVID-19, relatively inadequate medical resources might affect the observation of the natural course of disease. Finally, only 101 patients died with COVID-19 in one hospital were included in this study, the interpretation of our results might be limited, so further studies including multi-centers are still needed.
In conclusion, the findings of this study suggested that older patients with comorbidities had a steeply increased risk of death with COVID-19. Respiratory failure, acute cardiac and kidney injury played crucial roles in the death of patients. Of importance, the early development of sepsis, increased white blood cell count and neutrophil count were associated with rapid disease progress to death.
Data Availability
all original data were saved in Renmin Hospital of Wuhan University, Wuhan, China
Author Contributions
WW and CC conceptualized the paper. QS, JY and FJ drafted the manuscript. QS, KZ, JY, JF, XZ, XC, PH, YH and ML collected the epidemiological and clinical data. KZ, FJ, KZ contributed to the statistical analyses. JY and FL contributed to radiological analyses. CC and WW were the manuscript’s guarantors and supervised the overall study. All authors reviewed and approved the final version of the manuscript.
Conflict of Interests
The authors declare that there are no conflicts of interest.
Funding statement
This study was funded by the National Natural Science Foundation of China (No. 81800574; No. 81870442) and Natural Science Foundation of Hubei Province (No. 2018CFB649). Funding sources for this study had no roles in the study design; data collection, analyses, or interpretation; or writing of the manuscript. The corresponding authors had full access to all the data in the study and were responsible for the final decision to submit for publication.
Acknowledgment
We thank all the patients and health providers involved in this study.