Comparison and Clinical Characteristics of COVID-19 in Different Time Period in Wuhan, China

Background SARS-CoV-2 is making deadly impact on the human lives all over the world. Therefore, we aimed to analyze the changes involved in clinical characteristics of COVID-19 patients over time. Methods We conducted a retrospective study to compare the patients whose onset of illness in January with the patients whose onset of illness in February in Wuhan, China. Results Among enrolled 896 patients, the median age was 60 years (47-69 years), 685 (76.5%) were categorized into group A (patients with illness onset in January), and 211 (23.5%) were categorized into group B (patients with illness onset in February). Compared with group B, group A had a higher rate of fever (p<0.001), the lower rate of asymptomatic (p<0.001). Group A had a higher incidence of neutrophilia (p=0.043), elevated D-miner (p<0.001), increased LDH (p=0.002), but lower incidence of normal CT scan (p=0.001). CD3 cells (p=0.015) and CD4 cells p=0.04) count signicantly reduced in group A. Critical patients decreased (p=0.005) and mild patients increased (p=0.001) in group B. The fatality rate signicantly decreased in group B (p=0.028). Conclusions The condition of patients with onset of illness in January were more serious than patients with onset of illness in February. It indicates that virulence showed reducing effect, but more basic research is required to support the hypothesis.

4, 2020, no new cases were diagnosed with the COVID-19 in Wuhan city [10]. Previous studies have shown that the SARS-CoV and MERS-CoV showed reduced pathogenicity overtime. However, it was unclear whether the pathogenicity of SARS-CoV-2 would also degrade over time. Therefore, we compared clinical characteristics and outcome of COVID-19 patients in the different period for analyzing the change of pathogenicity of SARS-CoV-2.

Study population
We administered 1347 patients who admitted to the Renmin Hospital of Wuhan University from 30th January 2020 to until they got discharged before 31st March 2020. However, we excluded the patients as follows: 167 patients who were transferred to Fangcang shelter hospital around 9th February for missing outcome information, whereas, 156 patients who were transferred from Fangcang shelter hospital in early March for without initial data, 44 patients with constant negative SARS-CoV-2 test results in pharyngeal swab samples, and 84 patients for missing essential data. Therefore, we nally considered 896 patients diagnosed with COVID-19 ( Fig. 1). Oral informed consent was obtained from all the patients. This

Procedures
Trained physicians and medical students reviewed demographics, clinical characteristics, radiologic and laboratory ndings, treatment, and clinical outcomes from electronic medical records. Symptoms appearing before admission were also recorded. Vital signs and disease severity status were recorded during the admissions. The results of the blood test and radiologic ndings after admission were analyzed. The most intense level of oxygen support during hospitalization was recorded. During the whole study period, patients were protected for con dentiality. The nasopharyngeal swab samples of patients were collected during admission and hospital stay if needed. The double nucleic acid detection kit (Bio-Germ, Shanghai, China) was used to detect SARS-CoV-2 in the open reading frame 1ab (ORF1ab) / nucleocapsid protein (N) gene. Nearly all patients received routine blood test, coagulation, biochemical tests and computed tomography.

De nitions
The COVID-19 was de ned based on the positive results of nasopharyngeal swab samples at any one time during clinical course plus any one point including: fever or respiratory symptoms and bilateral or multiple ground-glass opacity in CT scan. The severity of COVID-19 was de ned based on the Chinese National for COVID-19 Diagnosis and Treatment Guidelines version-7 [11]: Mild types were showed with only slight clinical symptoms. Common types were presented with both slight clinical symptoms and an abnormal CT scan. Severe types were exhibited with dyspnea, respiratory rate ≥ 30 times/min, SpO 2 (percutaneous oxygen saturation) ≤ 93%, and oxygenation index (PaO 2 /FiO 2 ) ≤ 300 mmHg. The critical type patients were presented with respiratory failure, shock or multiple organ failure which ultimately required mechanical ventilation and intensive care. Fever was de ned with not less than 37.3℃. The ARDS (acute respiratory distress syndrome) and the AKI (acute kidney injury) were de ned according to the Berlin De nition [12] and the KDIGO clinical practice guidelines [13], respectively. The sepsis was de ned according to 2016 Third International Consensus De nition [14]. The AHF (acute heart failure) and the ALF (acute liver failure) were de ned based on Recommendations on pre-hospital and early hospital management of acute heart failure [15] and American Gastroenterological Association Institute Guidelines for the Diagnosis and Management of Acute Liver Failure [16], respectively. The hypoalbuminemia was de ned as the blood albumin less than 35 g/L. The shock was de ned as systolic pressure lower than 90 mmHg or mean arterial pressure lower than 65 mmHg.

Statistical analysis
Categorical data such as sex, symptoms, comorbidity, vital signs, disease severity status, abnormal blood test, imaging features, treatment and outcomes were compared using the χ2 test. Non-normal distributed continuous data were compared using the Mann-Whitney-Wilcoxon test, including age, time from illness onset to hospital admission, blood test, time from illness onset to discharge or death. Categorical variables were represented by frequency and percentage, and abnormally distributed continuous variables were described by the median and interquartile range (IQR). Survival data were analyzed by the Kaplan-Meier plot. All statistical data were calculated using SPSS (v23) software and plotted gures by GRAPHPAD PRISM (v8.0). P-value less than 0.05 was considered as statistically signi cant.

Demographics and clinical characteristics on admission
A total of 896 patients with a con rmed diagnosed with the COVID-19 were enrolled. The median age was 60 years (IQR, 47-69 years), 431(48.1%) were male; the most common symptoms were fever (80.4%) and cough (50.4%). Some people showed symptoms related to digestive tract, including diarrhea (17.7%), nausea or vomiting (5.8%), abdominal pain and distention (2.3%) and inappetence (13.1%). Few patients had symptoms such as itchiness in the eyes (0.8%) ( Table 1). According to time of illness onset, 896 patients were divided into group A (n = 685, 76.5%; patients with onset of illness in January) and group B (n = 211, 23.5%; patients with onset of illness in February). The median age was 60 (IQR 47-69 years) and 59 (IQR 47-69 years) respectively. Approximately 1:1 ratio of male and female were present in both group A and group B. Compared with group B, group A had a higher incidence of fever (p < 0.001), chills (p = 0.016), myalgia (p = 0.037) but a lower rate of chest congestion (p = 0.019) and chest pain (p = 0.013). The patients with nonspeci c symptoms, including inappetence (p = 0.001) and fatigue (p = 0.006), reduced in group B. Signi cant decrease in the patients with dyspnea (p = 0.02), and increased asymptomatic patients (p < 0.001) in group B were observed (Table 1).
Compared with group B, group A had a higher rate of mortality (p = 0.028), the higher rate of critical patients (p = 0.005), but a lower rate of mild patients (p = 0.001). The number of common patients and severe patients had no signi cant difference between the two groups. There was signi cant difference in time from the onset of illness to the admission (p = 0.001). The median time was eleven days (IQR 8-15 days) in group A and seven days (IQR 5-13days) in group B. The number of patients with a respiratory rate higher than 24 breaths per min during admission were observed to be increased signi cantly in group A (p = 0.031) ( Table 1).

Laboratory ndings and image feature
Laboratory results were collected during the admission. Among all the patients, the patients with most common abnormal blood test showed increased D-miner (63.3%), about half of patients demonstrated elevated CRP (60%), brinogen (53.6%), LDH (50.1%) and decreased lymphocyte count (48.7%). Whereas, few patients showed kidney injury with increased level of creatinine (7.7%) and urea (13.6%) ( Table 2). Compared with group B, group A had a higher rate of increased D-miner (p < 0.001) and PCT (p = 0.018), as well as neutrophilia (p = 0.043) and elevated levels of LDH (p = 0.002). Group B showed signi cant increase in CD3 cells count (p = 0.015) and CD4 cells count (p = 0.04). Whereas, in group A, liver injury indices including ALT and AST, were signi cantly higher than that in group B. More hypoalbuminemia occurred in group A (p < 0.001). However, there was no signi cantly difference between two groups in white blood cell count, lymphocyte count and CRP (Table 2).
Compared with group B, more CT scan results in group A showed bilateral pulmonary in ltration (p = 0.003) and ground-glass opacity (p < 0.001), however, the group B exhibited increased normal CT scan results (p = 0.001) ( Table 2).

Treatments and outcomes
Among 896 patients, most patients received antibiotics (77%), Lianhua Qingwen capsule (71.8%) and antiviral treatment (93.4%) which including Oseltamivir, Arbidol and Ganciclovir. Compared to group B, group A received more treatment of antibiotics (p = 0.009), corticosteroids (p = 0.036), intravenous immunoglobin (p < 0.001) and a-interferon (p = 0.008). We recorded the most intense level of oxygen support, and 701(78.24%) patients showed the need of oxygen support. More treatments of high-ow nasal cannula oxygen therapy (p = 0.028) and non-invasive mechanical ventilation (p = 0.027) were administrated in patients of group A (Table 3). patients. The Shock was observed in 21(2.3%) patients. Among the patients with shock, 9 (42.9%) patients were caused by sepsis, 10 (47.6%) patients were contributed to heart failure and 3 (14.3%) patients were caused by acute gastrointestinal bleeding. None of the patients complicated with shock were survived. Compared with group B, group A had a higher rate of mortality (p = 0.028) (Fig. 2). The incidence of developing to ARDS (p = 0.003) was higher in group A. Among all the patients, the median time from illness onset to death or discharge was 33 days (IQR 21-44 days) and there was no signi cant difference observed between the two groups (Table 3).

Discussion
In our study, we found distinction of severity of COVID-19 in different time period signi cantly. Compared with patients whose onset of illness began in January, the illness condition was improved in patients with onset of illness in February, which represented with lower mortality, lower rate of critical patients, lower level of neutrophil, D-miner, LDH, PCT, ALT, AST and a higher level of CD3 and CD4 cells count. Also, we observed signi cant increase in mild types of patients, asymptomatic patients and patients with normal CT scan whose illness onset began in February.
Increased neutrophil was positively related to poor outcome in coronavirus diseases. The study carried out by Wu et al demonstrated that neutrophilia is a risk factor related to the progression of ARDS and development from ARDS to death in COVID-19 patients [17]. The neutrophilia was also a risk factor for intubation in SARS patients [18]. The severity of lung damage positively correlated with neutrophils counts in the peripheral blood in patients with MERS [19,20]. The neutrophil was able to produce chemokines and cytokines which can resist against the virus but at the same time it led to an excess in ammatory response that resulted in multiple organ injury. Whereas, D-miner was an independent risk factor for mortality in COVID-19 patients [17,21]. Moreover, coagulation disorders was a displayed in the COVID-19 patients [22] and non-survivor had a higher level of D-miner, FDP and prolonged PT, APTT in COVID-19 patients [23]. Anticoagulant therapy might play the key role in relieving hypoxemia due to existence of micro-thrombosis in the lung of COVID-19 patients [24].The possible mechanism might be that the in ammatory cascade triggered by a virus and secondary infection led to injury of microvascular wall and release of procoagulant factors which resulted in ischemia and thrombosis. Evidence revealed that higher CD3 and CD4 T-cell counts protected patients from developing ARDS [17]. A recent study showed using electron microscopy that SARS-CoV-2 were in CD4 cells of COVID-19 patients, however, SARS-CoV-2 showed no productive replication [25]. This indicate that CD4 might control the replication of SARS-CoV-2. The incidence of secondary liver injury was positively related to poor outcome in COVID-19 patients [26]. A recent study showed that the liver injury was caused by activated cytotoxic T lymphocyte triggered by SARS-CoV-2 [27], other than drug-induced liver injury. Increased LDH was also associated with poor outcome in COVID-19 patients [17,21]. Mild patients and asymptomatic patients increased with onset of illness in February, a recent study showed that COVID-19 patients who had mild or no symptoms, were dominant with virus in the upper respiratory tract [28,29].
The mortality was reduced, symptoms and results of laboratory test were improved in patients with onset of illness in February compared with patients with onset of illness in January, however, there was no evident difference in age, sex, underlying diseases and environmental factors (district and seasonality) between them. Therefore, we speculate that the improved outcome might contribute to the diminishing virulence.
Recently, variants [30], mutations [31] and genetic diversity [32,33] is found in SARS-CoV-2, which might be an explanations for diminishing virulence. A 382-nt deletion appeared during SARS-CoV-2 evolution [34], and a 29-nt deletion appeared during SARS-CoV evolution. Both of the deletions covered open reading frame 8 (ORF8). Evidence showed that the 29-nt deletion resulted in the decreased replicative ability of SARS-CoV in-vitro independent of the cell system [6]. The 382-nt deletion might also reduce virulence of SARS-CoV-2. An 81-nucleotide deletion had been found in ORF7 of SARS-CoV-2, leading to a 27 amino-acid in-frame deletion [35], which might also change pathogenicity of SARS-CoV-2. In the United States, researchers found discrepancy in mortality of COVID-19 patients between the East Coast and West Coast of America. They speculated that the discrepancy probably existed due to the change in amino-acid of viral spike protein in SARS-CoV-2 [36]. The study carried out by Li et al similarly found distinct pathogenicity of SARS-CoV-2 in vitro. The Vero-E6 cells were used to be infected with 11 SARS-CoV-2 viral isolates from COVID-19 patients. It showed huge discrepancy in pathogenicity on the Vero-E6 cells [37]. In addition, reduced virulence probably contributed to generation of antibody. A report showed that the improvement of clinical outcome in ve COVID-19 patients who received an infusion of plasma from ve patients who recovered from the COVID-19 [38].
In addition to diminishing virulence, some other factors which might have caused a difference in severity of COVID-19 in a different time period. At the early stage of the outbreak, the time from illness onset to hospital admission was longer, probably because of a shortage of medical resources and lack of awareness for COVID-19. It might aggravate illness after a few days of delay. Moreover, in the later time of outbreaks, more strict managements were employed [39], more number of Fangfang shelter hospital were established and the healthcare environment was improved. Furthermore, experiences and researches accumulated in early time played a crucial role in the improvement of clinical outcomes. (Fig. 3).
However, there were few limitations in our study. First, in the early stage of outbreak, many patients with clinical improvement were transferred to other isolated sites for further observation and waited for viral clearance, thus we did not obtain the time of viral clearance. Second, due to the retrospective study, some data in the analysis is missing. Third, some patients with onset of illness in February were not included as they were not discharged from the hospital yet. Moreover, our hospital was designated for to enroll the patients with severe and critical type patients with COVID-19, therefore the disease severity status and mortality in our research could not re ect the actual situation in Wuhan city. Nonetheless, our study was conducted at a single-center hospital with limited sample size.

Conclusion
We analyzed clinical characteristics and outcomes of COVID-19 during the different time period in Wuhan. Our research suggested that the patients with onset of illness in February showed improved clinical outcome compared to the patients with onset of illness in January. This probably indicate reduced virulence from the aspect of clinical perspective, but more factors should be taken into consideration. Further, virus sequence and population immunity surveillance are required to con rm our hypothesis.

Consent for publication
Not applicable.

Competing interests
No con icts of interest to be declared.

Funding
This work was supported by the Science and technology key project on novel coronavirus pneumonia, Hubei province, China (grant number: 2020FCA002).

Authors' contributions
Ke Hu and Ruhao Yang contribute to study design, data analysis, and protocol writing. Weihua Hu contribute to protocol writing, collection of data, and data analysis. Jie Wei contributes to study design, protocol writing and is responsible for the statistical analyses. Jun Xiong and Menglin Liu contribute to data collection. The authors read and approved the nal manuscript. Survival curve in COVID-19 patients with illness onset in January or in February Figure 3