Coronavirus disease 2019 (COVID-19) may lead to the severe acute respiratory syndrome, and results in a major pandemic worldwide. This has been a global health crisis since the flu pandemic in 1918 [1]. There are many risk factors related to severe COVID-19 that were mentioned, such as older age, diabetes, obesity, cardiovascular disease, chronic lung disease, cancer, chronic kidney disease, pregnancy, smoking, organ transplantation, and using immunosuppressive drugs [2]. Although COVID-19 is more and more understood, the mortality rate of admitted COVID-19 patients is obviously high, with several studies providing a mortality rate of up to 61.51% [3]. Therefore, understanding the risk factors of death in relation to COVID-19 patients in each region and each country is always necessary for clinicians to devise treatment strategies, monitor, and prognosticate. During the pandemic outbreak in Ho Chi Minh City of Vietnam, many hospitals received a large amount of COVID-19 patients, including the Hospital for Rehabilitation-Professional diseases. We have conducted this research to understand the risk factors related to death of patients admitted to our hospital during the peak of the epidemic from August 2021 to October 2021, aiming two objectives: (1) Describe clinical, subclinical characteristics and mortality rate of COVID-19 inpatients, (2) Identify mortality risk factors of COVID-19 inpatients.

During the pandemic period in 2021, we recruited 104 patients who were admitted to our hospital in emergency conditions. Most of the hospitals in the city were overloaded so COVID-19 patients were often treated at home and hospitalized in severe conditions. Therefore, the average age of the study population was high, approximately 62 years old and the oldest was 92 years old. The mean age of the group of deaths (65.9 ± 11.2 years) was significantly higher (p=0.042) than that of the non-death group (59.9 ± 14.4 years). There was no gender difference between the 2 groups (p > 0.05) while the mean BMI of the non-death group was lower statistical significance than the death group (p=0.027).

The median illness day was 7, IQR of 4 to 8 days. The time of disease onset in the death and non-death groups was not statistically significant (p = 0.615). This time of admission is similar to the record of Davide Ippolito [4] in that most patients are admitted to the hospital after 5 days of disease onset. According to the literature, the time of severe progression requiring hospitalization is usually around 5-10 days [5].

The common chief complaints that led to admission were dyspnea, fever, lethargy, dry cough, and productive cough. Davide Ippolito [4] reported that the main symptoms were fever, cough, and dyspnea. The difference in reasons for hospitalization is the time of admission in our study, which was at the overload period in Ho Chi Minh City; patients were often admitted in severe (32.7%) or critical conditions (40.4%).

The most common symptoms were fever, shortness of breath and fatigue. Less common were anorexia, and cough. The least common clinical symptoms are abdominal pain, urticaria, maculopapular, and erythematous papules. These findings were similar to those of Davide Ippolito [4].

Over half of the study population was vaccinated, and the mortality rate in the vaccinated group was significantly lower than in the non-vaccinated group (p=0.043). The most common condition was hypertension, diabetes, gastritis, and ischemic heart disease. The number of comorbidities in the death group was higher than in the non-death group; this difference was statistically significant (p=0.013). Similarly, Morgan Spencer Gold [6] analyzed 33 studies and recorded 74.4% of deaths with comorbidities and commonly hypertension, diabetes and respiratory disease.

The death group compared with the non-death group had a higher respiratory rate and a lower SpO2; the difference was statistically significant (p<0.001). These results were similar to a study on 601 COVID-19 patients hospitalized at Lombardy hospital in Italy [7], which recorded oxygen saturation and breathing rate to help identify severe patients requiring early non-invasive mechanical ventilation.

There was no difference in the number of white blood cell, neutrophil, and neutrophil percentage between the non-death and death COVID-19 group (p>0.05). Additionally, in the death group, the mean lymphocytes and median percentage of lymphocytes were lower than those of the recovered group (p<0.05). These were similar to the study of Zachary Illg [8].

There were also no significant differences between the two groups who recovered from the disease and died (p>0.05). This result as well as the study of Jianguo Zhang [9] did not find this association.

The mean platelet count in the COVID-19 death group was significantly lower than in the non-death group (p=0.005). The same results were reported by Davood Bashash [3] who analyzed 19 studies, 3383 COVID-19 patients and concluded that decreased platelet count is often associated with severe COVID-19 (Table 1)

In comparing the 2 groups, we found that urea, creatinine, eGFR, AST, D Dimer, Ferritin, CRP of the mortality group were significantly higher than that of the recovered group (p<0.05). This was like the study of Fen Lan [10]. In contrast, the median sodium concentration in the mortality group was lower than in the recovered group (p=0.008) (Table 2). Andrea Berni [11] also reported similarly and concluded that hyponatremia was an independent predictor for death on admission (2.7 times higher than normal sodium). Decreased sodium could be correlated with a 14.4% increased risk of death. Elevated liver enzymes were also reported in a retrospective study of 373 COVID-19 patients at 5 hospitals in Madrid [12]. Accordingly, 33.1% of hospitalized patients had an increase in AST. Rise-to-base AST is an independent factor of severe COVID-19 disease

The mean D-Dimer, Ferritin, CRP increased, and the mortality group was significantly higher (p<0.001) in comparison to the recovered group (Table 2). These results matched with the study’s results of Fei Zhou and Luo [7, 13].

The average value of the cut-off CT of real-time RT-PCR SARS-CoV-2 was no difference between the two groups (p=0.071) (Table 2). The results of Abdulkarim Abdulrahman’s study2 had the same record that CT had no correlation with COVID-19 disease status.

This finding in ECG disorders was similar to that reported by Brit Long [14]: Sinus tachycardia is the most common abnormality; other abnormalities include supraventricular tachycardia, ventricular arrhythmias, bradycardia, ST segment changes and T wave.

Median brixia radiographic score in the non-death group was significantly lower than the death group (p<0.001). The results of our study are similar to those of Davide Ippolito4. The median Brixia score of our study is lower than that of Agrawal [15] in which the Brixia score (>12 points) is associated with increased mortality from COVID-19. The difference was due to the severity of Agrawal's study population [15] and there was upper double deaths (60.7%) rate than our study (29.8%).

Our study collected patients during the peak period of the pandemic in Ho Chi Minh City; hospitals were overloaded and did not have enough medicine, especially COVID-19 antiviral medicine and lack of medical equipment. Only serious patients were admitted to the hospital, mild and moderate cases were usually recommended to self-monitor at home under the consultants of telemedicine teams. Therefore, the stratification of severity in our study was quite high and most of the patients at admission required respiratory support. The mortality rate in our study is similar to the study of the author Anant Dinesh [16] with the hospitalization rate of 21.26%, the mortality rate in Brooklyn public hospitals is 29.9%, Queens 28.1%, and Manhattan 20.4%.

Based on the cut-off points in Table 3, we conducted univariate and multivariate logistic regression analyses of factors related to the death of COVID-19 patients. The results received 6 risk factors such as age ≥ 60, history of no vaccination for SARS-CoV-2, SpO2 ≤ 87%, Urea > 7 mmol/L, Ferritin > 578 ng/ml, Brixia score > 9, which were independently related factors to mortality of the study population (p<0.05) (Table 4). Our results were like that reported by Zelalem G. Dessie [17], who analyzed 42 studies with 423,117 patients with comorbidities, gender, age, smoking status, obesity, and lesions. Acute renal failure and D-dimer are risk factors for admission to the ICU. According to the U.S. Centers for Disease Control (CDC), of 1,228,664 people who completed the first dose of immunization between December 2020 and October 2021, serious outcomes associated with COVID-19 (0.015%) or death (0.0033%) were rare . As such, vaccines are a protective factor against COVID-19, which matches our univariate analysis. Similarly, for respiratory rate variables, SpO2, CRP, Ferritin are the factors associated with COVID-19 mortality. These findings were also reported by Ahmed Mukhtar et al. [18] in a retrospective observational study analyzing SpO2 on breathing air within the first 6 hours after hospital admission of COVID-19 patients. 19. The lower SpO2 cut-off point in Ahmed Mukhtar study (78% vs 87%) may be due to the different sample population which was more severe in Ahmed’s research. Huang et al. [19] studied 5350 patients who were aggregated from 25 studies with increased CRP (RR 1.84, 95% CI 1.45-2.33, p < 0.00), increased D-dimer (RR 2.93, 95% CI 2.14, 4.01, p < 0.001) was associated with an increased risk of poor outcomes. Besides, according to the results of the retrospective cohort study by Alonso Soto [20], mortality corresponded to elderly, history of surgery, low oxygen saturation, presence of comorbidities, unconsciousness and the number of white blood cells. The model of Alonso Soto [20] has 2 similar factors to our research model: elder and low oxygen saturation are independent predictors of mortality. However, the author's cut-off point for oxygen saturation was lower than ours because the author's study had more severe patients and a higher mortality rate than our study (46% versus 29.8%).

Coronavirus disease 2019 (COVID-19) with diverse clinical symptoms on multi-organs can lead to acute respiratory failure and death. During the outbreak of the pandemic in Vietnam, especially in the year of BUFFALO, most of the COVID-19 patients who were admitted to the hospital were severe and critical indicating ICU admission and needing respiratory support. One-third had to intubate and use mechanical ventilation. Independent risk factors related to the death of COVID-19 included: Brixia score > 9, Urea > 7 mmol/L, Ferritin > 578 ng/ ml, Failure to get vaccinated, Age > 60, Low Oxygen SpO2 < 87% (BUFFALO).