Epidemiologic Characteristics of and Prognostic Factors for COVID-19 Among Hospitalized Patients: Updated Implications From Hubei Province, China

Introduction: The roles of some indicators in the prognosis of patients with coronavirus disease-19 (COVID-19) remain unclear and controversial. This study aimed to explore the epidemiologic characteristics of and prognostic factors for COVID-19 to provide updated recommendations for its prevention, diagnosis, and treatment. Methods: For this retrospective study, demographic, epidemiologic, and clinical data were extracted from the medical records of patients admitted to the Maternal and Child Hospital of Hubei Province (Optical Valley) with COVID-19 between February 19, 2020, and March 19, 2020. The primary outcome was the prognosis that was determined at discharge as mentioned in the medical records. Descriptive statistics, univariate analyses, and stepwise logistic regression analysis were used for data analysis. Results: Of the 1,765 patients included, 93.1% were cured and the mortality was 1.8%. Univariate analyses identified 63 factors significantly associated with COVID-19 prognosis. Logistic regression analysis revealed that a poorer prognosis was associated with undergoing resuscitation, complex disease manifestations, consultation with outside specialists, elevated basophil or lymphocyte counts, an albumin (ALB)/globulin (A/G) ratio > 2.4, and elevated levels of serum aspartate aminotransferase (AST) or creatinine. Patients had a better prognosis if the following conditions were met: dry cough reported as an initial symptom, fatigue as a clinical manifestation, and a diagnosis based on laboratory testing. Conclusion: To prevent clinical deterioration, clinicians should provide special care to patients who underwent resuscitation, with a critical disease, or requiring consultation with outside specialists. Extra attention should be paid to patients with high basophil or lymphocyte counts, a high A/G ratio, and elevated AST or creatinine levels.


INTRODUCTION
In early December 2019, several pneumonia cases of unknown origin emerged in Wuhan, China, later attributed to a novel coronavirus (1,2). On February 11, 2020, the WHO officially named the disease coronavirus disease-19 (COVID-19) (3). Since then, COVID-19 has spread globally, and on March 11, 2020, the WHO declared COVID-19 a pandemic (3). By July 30, 2021, there had been 196,553,009 confirmed cases of COVID-19, of which 4,200,412 deaths were reported to the WHO, of which 102,553 confirmed cases and 5,635 deaths were from China (4). Although great progress has been achieved in epidemiological experience, clinical management, and vaccination, the pandemic continues to progress globally.
Because China was the first country to report the outbreak of COVID-19, Chinese clinicians and researchers have continued to explore and update the assessment of epidemic characteristics and COVID-19 prognostic factors, providing evidence for its prevention, diagnosis, and treatment. The prognostic factors include demographic characteristics (5), underlying diseases (6), laboratory parameters (7,8), clinical manifestations (9,10), radiological features (11), and treatment (12). In addition, vaccination becomes a positive prognostic factor recently, with more than 10 vaccines of different strategies being reported to reduce the severity of the disease (13)(14)(15)(16)(17). Similar findings regarding the aforementioned prognostic factors have been reported from various countries, namely, France (18), Saudi Arabia (19), Spain (20), Italy (21), and others (22).
Despite overall consistency in findings between different studies, data on several prognostic factors, such as sex (23,24), clinical manifestations (25), and underlying diseases (25,26), remain inconsistent or variable. For example, the African COVID-19 Critical Care Outcomes Study revealed that sex was not independently associated with outcome (24), while a study in Detroit revealed that male sex was significantly associated with mortality (23). Though having chronic diseases has been reported to be a risk factor for worse prognosis in some studies, other studies have found that they had no influence on prognosis (25)(26)(27)(28). In addition, many studies had a small sample size, which may cause biases. Some prognostic factors, such as some laboratory test results, radiological features, and treatments, have not been explored owing to limited data availability.
Because there remains much about COVID-19 that is unknown or poorly understood and the disease continues to be prevalent and evolve as new variants emerge, it is important to understand demographic, epidemiologic characteristics, clinical features, laboratory test results, radiological features, and treatment, especially the controversial indicators or those not explored in previous studies. Therefore, to gain insight into Abbreviations: AST, Aspartate aminotransferase; COVID-19, Coronavirus disease; CRP, C-reactive protein; CT, Computed tomography; D-Bil, Direct bilirubin; DBP, Diastolic blood pressure; GGO, Ground-glass opacity; MCH, Maternal and Child Hospital of Hubei; MCHC, Mean corpuscular hemoglobin concentration; PLT, Platelet, RBC, Red blood cell, RDW, Red blood cell distribution width, SBP, Systolic blood pressure, TP, Total protein, UN, Urea nitrogen, WBC, White blood cell, WHO, World Health Organization. these inconsistencies and expand the understanding of additional indicators, this study aimed to explore the epidemiologic characteristics and factors influencing COVID-19 prognosis with a relatively large sample size. The findings of the potential prognostic indicators from this study will provide an updated assessment of their implications for the prevention, diagnosis, and treatment of COVID-19.

Study Design and Participants
The study was a retrospective review of patients with COVID- 19

Data Collection
Medical records of patients at MCH-Optical Valley were obtained from the electronic medical record system. The inclusion criteria were (1) hospitalization owing to confirmed or suspected COVID-19; (2) hospitalization at MCH-Optical Valley between February 19, 2020, andMarch 19, 2020;and (3) complete medical records covering the seven dimensions mentioned below. The exclusion criteria were (1) outpatient consultation; (2) hospitalized during other time periods; and (3) in complete medical records. Finally, a total of 1,765 patients with detailed records were included in this study. Data in this medical record system could be divided into seven dimensions: (1) demographic characteristics, (2) epidemiological exposure history, (3) admission characteristics, (4) hospitalization and treatment characteristics, (5) imaging features, (6) laboratory findings, and (7) diagnosis and prognosis.
Demographic characteristics included sex, marital status, occupation, and age. Epidemiological exposure history referred to exposure history (contact with someone with confirmed or suspected COVID-19 during the 2 weeks preceding the onset of illness), the relationship between patients and people with COVID-19 to whom they were exposed, source of exposure, time period of exposure, and possible location of exposure. Admission characteristics included admission route, illness condition on admission, critical degree on admission, and vital signs on admission (body temperature, pulse rate, respiratory rate, systolic blood pressure [SBP], and diastolic blood pressure [DBP]). Hospitalization and treatment characteristics included whether patients underwent surgery, underwent resuscitation (life-threatening patients needing rescue), developed critical illness, required consultation, had complex cases, reported initial symptoms, current symptoms during hospitalization, clinical manifestations, previous medical history, self-reported underlying diseases, smoking habits, treatment, and length of stay. Imaging features on CT or chest radiography included ground-glass opacity (GGO) characteristics, location, and other observations. Laboratory findings were classified into two categories (according to the sample), with a total of 44 indicators. The first category comprised the following indicators tested using whole blood: C-reactive protein (CRP), white blood cell (WBC) count, percentage of lymphocytes, percentage of eosinophils, neutrophil count, monocyte count, basophil count, hemoglobin level, mean corpuscular volume, mean corpuscular hemoglobin concentration (MCHC), platelet (PLT) count, percentage of neutrophils, percentage of monocytes, percentage of basophils, lymphocyte count, eosinophil count, red blood cell count, packed cell volume, mean corpuscular hemoglobin, red blood cell distribution width (RDW), and mean platelet volume. The second category comprised the following indicators tested using serum: blood glucose, albumin (ALB), ALB/globulin (A/G) ratio, direct bilirubin (D-BiL), alanine aminotransferase, alkaline phosphatase (ALP), total bile acid, sodium, calcium, serum magnesium, urea nitrogen (UN), uric acid, total protein (TP), globulin, total bilirubin, indirect bilirubin (I-BiL), aspartate aminotransferase (AST), γ-glutamyl transpeptidase, kalium, chlorine, phosphorus, total carbon dioxide, and creatinine. Diagnosis and prognosis considered outpatient and emergency diagnosis, admission diagnosis, discharge diagnosis, a subtype of COVID-19 at discharge (mild, normal, severe, critical, suspected, and clinical diagnosis subtypes), consistency between discharge and outpatient diagnoses, consistency between admission and discharge diagnoses, consistency between preoperative and postoperative diagnoses, consistency between clinical and pathological diagnoses, consistency between radiological and pathological diagnoses, the highest evidence of diagnosis (clinical diagnosis; radiography, CT, ultrasound, and endoscopy; biochemical and immunological test results; and cytological blood smear test), and the prognosis at discharge (dead, unhealed, improved, and cured).

Data Analysis
Statistical analyses, such as descriptive statistics, univariate analyses, and multivariate analysis, were conducted using SAS 8.2 (SAS Institute Inc., Cary, NC, USA) and SPSS 18.0 (SPSS Inc., Chicago, IL, USA). First, descriptive statistics were used to describe the frequency and percentage of categorical variables. Continuous variables are presented as the median and interquartile ranges because all continuous variables had a skewed distribution. Second, univariate analyses were conducted to determine the association between prognosis and each variable. Specifically, the Wilcoxon rank test was used to test the relationship between prognosis and dichotomous variables (e.g., sex). The Kruskal-Wallis H test was used to test the relationship between prognosis and multivariate variables (e.g., marital status). The Spearman rank correlation test was used to test the relationship between prognosis and continuous variables (e.g., age). Third, because the prognosis (the dependent variable) was an ordinal multi-categorical variable, stepwise logistic regression analysis was used to explore the factors influencing COVID-19 prognosis. Factors having a significant association with prognosis in the univariate analysis were included in the logistic model, with the significance level for entry set at 0.10 and the significance level for selection set at 0.15. To explain the results more explicitly, continuous variables were transformed into categorical variables according to their meanings and clinical reference ranges in the logistic regression analysis. All tests were two-sided, and P < 0.05 were considered statistically significant.

Ethical Approval and Patient Consent
This study was approved by the 903rd Hospital of PLA ethics committee (approval reference number: 20210224/02/01/002). The requirement for informed consent was waived by the ethics committee because of the retrospective nature of the study and the urgent nature of the pandemic.

Demographic and Epidemic Characteristics of Patients With COVID-19
As shown in Table 1, 58.4% of the 1,765 patients with COVID-19 were women, 89.6% were married, and 43.3% were retired. The average age was 58.7 ± 15.1 years (Supplementary Table 1).
Of the total number of patients, 27.0% had an exposure history of contact with someone with suspected or confirmed COVID-19 during the 2 weeks preceding the onset of their illness. Most contacts (75.8%) were with people other than family members, colleagues, and people with social interactions and shared transportation. The most common source of exposure was through treatment and care. Of the patients, 74.8% had an uncertain time period of exposure and 53.5% were exposed in locations other than their home, workplace, dormitories, hospitals, or indoor public places ( Table 1).
Most patients were admitted to the hospital through other routes (rather than through the emergency department, outpatient department, or referral from other hospitals); 98.1% had symptoms at the time of admission, and 92.2% had moderately severe disease on admission ( Table 1). The vital signs on admission, such as body temperature, pulse rate, respiratory rate, SBP, and DBP, are shown in Supplementary Table 1.
The average length of stay was 13.3 ± 6.1 days (Supplementary Table 1). During hospitalization, most did not undergo surgery (99.5%) or resuscitation (96.7%), did not develop critical illness (88.3%) or require consultation with outside specialists (68.2%), and did not have complex cases. The top three initial symptoms reported were fever, cough, and fatigue. Over 60% of the patients experienced cough (61.4%) and fever (60.0%) during hospitalization. The top three clinical manifestations were consistent with the reported initial symptoms. Most patients had no previous medical history of note. The top three self-reported underlying diseases were hypertension, diabetes, and coronary heart disease. Of the patients, 78.6% were non-smokers. Most received antiviral therapy during hospitalization. Chest CT or radiography showed that most had GGO, observed at multiple     Most patients were diagnosed with confirmed COVID-19 at the time of outpatient or emergency department consultation, admission, and discharge. Over 50% of the patients were diagnosed with a normal subtype of COVID-19 at discharge. For most patients, the discharge and outpatient diagnoses and the admission and discharge diagnoses were consistent. However, the preoperative and postoperative diagnoses, clinical and pathological diagnoses, and radiological and pathological diagnoses could not be compared because of insufficient data. In terms of the diagnoses mentioned above, 52.6% of patients were diagnosed based on the highest level of evidence of biochemical and immunological tests, indicating that most (93.1%) patients were cured at discharge and the case fatality rate was 1.8% ( Table 1).

Univariate Analysis of Factors Influencing COVID-19 Prognosis
Univariate analyses (Supplementary Tables 1, 2) suggested the following 63 factors significant to the prognosis: marital status; occupation; age; exposure history (contact with someone with confirmed or suspected COVID-19 during the 2 weeks preceding the onset of illness); source of exposure; time period of exposure; critical degree on admission; undergoing surgery; undergoing resuscitation; critical disease; need for consulting outside specialists; complex disease manifestations; reported initial symptom of dry cough, fatigue, dyspnea, disturbance of consciousness and apathy, and unknown; current symptom of dyspnea; clinical manifestation of fever, cough, and fatigue; medical history; self-reported underlying hypertension, diabetes, atrial fibrillation, and Alzheimer's disease; smoking habits; treatment method of oxygen therapy measures, mechanical ventilation, and immunotherapy with tocilizumab; imaging feature of pleural effusion; lymphocyte count; percentage of lymphocytes; percentage of eosinophils; neutrophil count; basophil count; MCHC; PLT count; percentage of neutrophils; percentage of monocytes; percentage of basophils; lymphocyte count; eosinophil count; RDW; blood glucose; ALB; A/G ratio; D-BiL; ALP; UN; TP; AST; chlorine; creatinine; CRP; outpatient and emergency diagnosis; admission diagnosis; discharge diagnosis; subtype of COVID-19 at discharge; consistency between discharge and outpatient diagnoses; consistency between admission and discharge diagnoses; consistency between preoperative and postoperative diagnoses; and the highest level of evidence of diagnosis.

Logistic Regression Analysis of COVID-19 Prognosis
Logistic regression analysis (Table 2) revealed that the prognosis of patients with COVID-19 was influenced by the following factors: undergoing resuscitation, developing a critical illness, requiring outside specialist consultation, having dry cough as an initial symptom, the clinical manifestation of fatigue, the highest level of evidence of diagnosis, basophil count, lymphocyte count, A/G ratio, and AST and creatinine levels.
A poorer prognosis was associated with undergoing resuscitation (odds ratio

DISCUSSION
In this study, 93.1% of the 1,765 patients with COVID-19 were cured, 4.8% improved, 0.2% were unhealed, and 1.8% died. Univariate analyses identified 63 significant factors for COVID-19 prognosis, while logistic regression analysis identified factors related to the severity of illness, symptoms and manifestations, diagnosis, and laboratory findings as factors independently associated with COVID-19 prognosis.
Patients with more severe diseases had a poorer prognosis. Undergoing resuscitation, consultation with outside specialists, and critical disease during hospitalization were all associated with more severe diseases. More critical patients were at a higher risk of death, which was consistent with the findings of a previous study, conducted in Wuhan, of 109 patients with COVID-19 who were serious enough to be admitted to the intensive care unit; all died owing to the rapid progress of the disease (29). Our finding was also indirectly supported by findings from a study from Detroit (23) and a study from Saudi Arabia (30) that emphasized admission to the intensive care unit would increase the incidence of death.
Patients who reported initial symptoms of dry cough and who had clinical manifestations of fatigue had a better prognosis. Our findings were consistent with those of other studies that found that though fatigue and dry cough are common symptoms and clinical manifestations, especially at the onset of illness (1,31), among patients with a fatal disease, the most common symptom was dyspnea, followed by fatigue (29). Therefore, it was understandable that patients with a dry cough had a 6-fold better survival than those without this symptom and that those with the clinical manifestation of fatigue had double the survival of those without fatigue. However, some other studies have found the opposite association with survival. A study of 47 patients with COVID-19 in Xinyu, China (32) and a systematic review of 207 studies (33) found that fatigue was associated with a poorer prognosis. We were unable to find any studies that assessed dry cough as a prognostic indicator, and only a few studies have mentioned cough when describing the clinical characteristics of patients with COVID-19 (2,34,35).
Those diagnosed based on the highest level of evidence of biochemical and immunological tests were more than twice as likely to have a favorable outcome than patients with a diagnosis based on clinical findings. Considering that the admission period of patients in this study was from February 19, 2020, to March 19, 2020, the diagnosis and treatment were based on the Diagnosis and Treatment Plan for COVID-19, Version 6.0 (36) and Version 7.0 (37). The diagnostic criteria indicated that clinical diagnoses, along with radiological examinations (i.e., radiography and CT), should be used to identify suspected cases, while biochemical and immunological tests should be used for confirmation. Patients with confirmed COVID-19 were more likely to receive appropriate treatment, which could explain the better prognosis among patients diagnosed with COVID-19 based on the highest level of evidence of biochemical and immunological tests.
In terms of laboratory parameters, basophil count, lymphocyte count, A/G ratio, AST, and creatinine levels were associated with prognosis. Several previous studies have found that lymphocyte count is one of the most significant factors for COVID-19 prognosis and is a predictor of death (38). Most results of previous studies conducted in China and other countries found that lymphopenia was associated with a poorer prognosis (33, 35,[39][40][41][42], and one study found that a lymphocyte count <0.8 × 10 9 cells/L was associated with an increased risk of severe COVID-19 (43). However, in our study, a lymphocyte count exceeding the upper limit of normal (>3.2 × 10 9 cells/L) was associated with a poorer prognosis. Similarly, a study from Saudi Arabia found that the lymphocyte count of patients with moderate disease was higher than that of patients with mild disease (19), which indicated that the severity of the disease might be a confounding factor. Therefore, more evidence from larger samples in other countries or other regions of China is needed to explore the contradictory results of the lymphocyte count.
Concerning AST, a study from Libya (42) reached a similar conclusion as that of our study. The increased level of AST would increase the possibility of death.
Creatinine was another sensitive prognostic laboratory indicator. A higher level of creatinine was associated with a higher probability of death or severe disease. This finding was consistent with a study on 113 patients in China with fatal disease (7), a study on 1,207 patients in Libya (42), and a systematic review of 207 studies from multiple countries (33). Furthermore, to identify and care for patients with increasing creatinine levels, attention should be focused on male patients with creatinine levels >104 µmol/L and female patients with creatinine levels > 90 µmol/L.
Furthermore, our study found that basophil count and A/G ratio were associated with COVID-19 prognosis, though few other studies have investigated these parameters. Extra attention should be paid to patients with a basophil count >0.06 × 10 9 cells/L and A/G ratio > 2.4 to prevent disease progression. However, some controversial factors in previous studies were not significant to the prognosis in our study; these included sex and comorbidities (23,24,30).
There are some limitations to this study. First, it was a singlecenter study, which might have led to some biases. Further, the generalization of our findings is limited to some extent. Second, many patients were transferred from other hospitals and were not first diagnosed at MCH-Optical Valley; however, all laboratory indicators were tested on admission, implying that the laboratory findings of some patients were not their first test results-this might have introduced some biases. Third, data in this study were from February 19, 2020, to March 19, 2020, i.e., relatively old, though analyses were conducted on some inconsistent or previously unexplored indicators. Follow-up after discharge should be considered to compare the first wave with the following waves (e.g., the emergence of new variants) in future studies to provide further evidence for the prevention, diagnosis, and treatment of COVID-19. Fourth, some important factors (e.g., lactate dehydrogenase, D-dimer, and procalcitonin) investigated in other studies were not explored in this study owing to a lack of data. Fifth, this study had no control group, and the heterogeneity of the population might influence the generalization of the findings to other populations. Sixth, because some characteristics were self-reported, there existed some unreasonable classifications, such as viral pneumonia and tuberculosis in the reported initial symptoms. We did not make any corrections about these self-reported characteristics due to the lack of follow-up, which might bring some biases.

CONCLUSIONS
The cure rate of 1,765 patients with COVID-19 in this study was 93.1%, and the mortality rate was 1.8%. To prevent the deterioration of the condition of the patient, clinicians should provide special care to patients developing critical illness during hospitalization, undergoing resuscitation, or needing consultation with outside specialists. In addition, patients with a basophil count >0.06 × 10 9 cells/L, lymphocyte count > 3.2 × 10 9 cells/L, an A/G ratio > 2.4, AST level > 34 U/L, and male sex with creatinine levels >104 µmol/L, or female sex with creatinine levels > 90 µmol/L were at a higher risk of having a poorer prognosis. Dry cough as an initial symptom, the clinical manifestation of fatigue, and being diagnosed with COVID-19 based on the highest level of evidence of biochemical and immunological tests were protective factors and was associated with a more favorable prognosis.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the 903rd Hospital of PLA Ethics Committee. Written informed consent from the participants' legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
All the authors contributed to the conception and design of the study. Material preparation, data collection, and analysis were performed by XiaL, LZ, TL, XibL, WY, and QC.