Adult inpatients (age ≥ 18 years old) diagnosed with COVID-19 pneumonia in Xi’an People’s Hospital (Xi’an Fourth Hospital) from December 8, 2022, to January 31, 2023, were included in this cross-sectional study. According to literature research and clinical experience, exceptions included pregnant women and patients with incomplete electronic medical records.

In this study, the sample size was calculated by using the following formula: n = z²p(1–p)/d², where n referred to the sample size, z referred to coefficient of confidence interval (1.96), p represented prevalence rate, and d indicated type I error level of 0.05. The severity rate of COVID-19 patients was assumed to be 15.7% based on previous studies (12). In China, the overall death rate from COVID-19 was 11% (13). Therefore, based on the above assumptions, the minimum sample size was 203 patients. Finally, 223 inpatients were included in this study.

According to the tenth edition of “Diagnosis and Treatment Protocol for Novel Coronavirus Infection” (6), the clinical types of inpatients with COVID-19 infection were as follows: 1. Mild (mild clinical symptoms with no sign of pneumonia on imaging); 2. Moderate (fever, respiratory symptoms, and imaging manifestations of pneumonia); 3. Severe (patients met one of the following criteria: respiratory distress and respiratory rate (RR) ≥30 breaths per minute; arterial oxygen saturation (SaO₂) ≤93% at rest; arterial partial pressure of oxygen /fraction of inspired oxygen (PaO₂/FiO₂) ≤300 mmHg (1 mmHg = 0.133 kPa); lung infiltration >50% within 24 ~ 48 h); and 4. Critical (patients met any of the following criteria: respiratory failure occurs and mechanical ventilation is required; shock occurs; concomitant failure of other organs occurs and intensive care unit monitoring and treatment is required). The severity of the disease was evaluated within 48 h of hospital admission. To better understand the clinical features, this study classified moderate cases into the non-severe group (n = 150) and the severe and critical cases into the severe group (n = 73).

Data on the patients’ demographic and clinical characteristics, laboratory tests, radiological findings at admission, treatments and outcomes were extracted from electronic medical records. All data were collected by two pharmacists independently and verified by two additional clinicians. To get the laboratory results, the indicators that could reflect the blood routine, inflammatory status, cardiac function, coagulation function, hepatorenal function, and blood glucose level, were collected. Radiologic evaluation was performed using chest X-rays or chest computed tomography (CT) scans. Besides, the information concerning drug treatments accepted by inpatients, including antiviral drugs, antibiotics, corticosteroids, intravenous immunoglobulin, anticoagulant and Chinese herbs, was also collected. Further, patients would receive supplemental oxygen inhalation including nasal catheter for oxygen, face mask oxygen inhalation, high-flow oxygen, noninvasive ventilation and tracheal intubation if necessary. Two outcomes were evaluated: hospital discharge and in-hospital death. When patients’ condition got improved obviously (demonstrated by the stable vital signs, the temperature had returned to normal for more than 24 h, the acute exudative disease on the lung image was significantly improved, the patients could be converted to oral drug treatment, and there were no complications that need further treatment), the patients could get discharged from the hospital. An in-hospital death is defined as a death that occurred during hospitalization.

The study endpoint was the risk factors associated with the severity and mortality of COVID-19 pneumonia during the first wave of the pandemic after the dynamic zero-COVID policy was retracted in Xi’an, China. Differences in the severity and mortality risk factors for patients hospitalized for COVID-19 pneumonia were assessed between the early wave and the very late stage of the pandemic.

Descriptive statistics were presented using frequencies (percentages) for categorical variables and median (interquartile range) for abnormal continuous variables. Continuous variables of all laboratory tests were converted into categorical variables according to their reference range. Candidate variables of the patients’ demographic and clinical characteristics, laboratory tests, and radiological findings at admission were included initially. Then, 4 laboratory variables were excluded for a missing rate > 20%, including interleukin (IL)-6, brain natriuretic peptide (BNP), myoglobin, and troponin I. Differences in the candidate variables between non-severe and severe inpatients, as well as the candidate variables between the patients discharged from hospital and those died in hospital, were evaluated using the Chi-square test for categorical variables, the Mann–Whitney test for continuous variables. Univariable and multivariable ordinal logistic regression analyses were performed to identify the independent factors associated with the severity and mortality of COVID-19 pneumonia. Variables found to be significant at p value <0.05 from the univariable logistic regression, along with age and sex, were included in the multivariable logistic regression model. As for vital signs, part of the criterion for distinguishing the disease severity, they were not included in the regression analyses of factors associated with the severity of the disease. All statistical analyses were performed using SPSS V26.0 Statistical Software Package for Windows. A p value <0.05 was considered statistically significant.

Of these 223 patients, 150 (67.3%) patients were categorized into the non-severe group, while 73 (32.7%) patients were categorized into the severe group. A total of 174 (78.0%) patients were discharged from the hospital, while 49 (22.0%) patients died in hospital. The demographic, clinical characteristics, laboratory tests, radiological findings at admission, treatments and outcomes of 223 patients were shown in Table 1. The median age of the 223 participants was 75 (IQR, 60.0–85.0) years old, and the majority of severe (58, 79.5%) and death (44, 89.8%) cases occurred in patients aged 65 or above. There were significant differences in the age-grading between the non-severe and severe groups, as well as the discharged and the death groups (p < 0.05). Males accounted for 67.3%. Most of the patients (183, 82.1%) suffered from at least one of the comorbidities. Hypertension and cardiovascular disease were the most common comorbidity, with 118 (52.9%) and 100 (44.8%) patients, respectively, whilst chronic liver disease and dementia were the rarer, with 11 (4.9%) and 7 (3.1%) patients, respectively. The proportions of cardiovascular disease in the death group were higher than those in the discharged group, and the difference was significant (p < 0.05). In terms of clinical manifestation, the incidences of cough (202, 90.6%) in COVID-19 patients were higher than fever (177, 79.4%). The incidences of chest tightness were significantly different between the severe patients (46.6%) and the non-severe (22.0%) patients (p < 0.05). The incidences of consciousness disorders were significantly different between the death group (12.2%) and the discharged group (2.9%) (p < 0.05). Pulse velocity in the death group was significantly higher than that in the discharged group (p < 0.05). There were significant differences in the RR, SaO₂ and PaO₂/FiO₂ between the non-severe and severe groups, as well as the discharged and the death groups (p < 0.05).

The following parameters had statistical difference between the non-severe and severe groups: white blood cell count (WBC), neutrophil count (N), N%, lymphocyte count (L), L%, neutrophil-to-lymphocyte ratio (NLR), C-reactive protein (CRP), procalcitonin (PCT), D-dimer, creatine kinase (CK), lactate dehydrogenase (LDH), α-Hydroxybutyrate dehydrogenase (α-HBDH), aspartate aminotransferase (AST), albumin (ALB), urea nitrogen (BUN) and fasting blood glucose (FBG). Between the discharged and death groups, the following parameters had statistical difference: WBC, N, N%, L, L%, NLR, platelet count (PLT), CRP, PCT, D-dimer, CK, LDH, α-HBDH, AST, alanine aminotransferase (ALT), ALB, BUN, creatinine (CR) and FBG.

Diffuse plaques shadow (DPS) (59.2%) and ground glass opacification (GGO) (33.2%) were typical manifestations of radiological findings in COVID-19 patients. GGO was significantly more frequently observed in the severe group than the non-severe group (p < 0.05), while DPS was significantly more frequently observed in the non-severe group than the severe group (p < 0.05).

Effective SARS-CoV-2 antivirals would alleviate severe cases and reduce mortality. As shown in Table 1, a total of 129 patients (57.8%) received antiviral treatments, including nirmatrelvir/ritonavir and azvudine, indicating widespread use of antivirals in patients with COVID-19 during the first wave of the pandemic after the dynamic zero-COVID policy was retracted. Antibiotics were used by 94.2%, glucocorticoids by 56.5% and anticoagulants by 31.8% of the patients. Compared with the non-severe group, azvudine, glucocorticoids and anticoagulants treatment were more frequently administered in the severe group (p < 0.05). Compared with the discharged group, azvudine and glucocorticoids treatment were more frequently administered in the death group (p < 0.05). In addition, oxygen therapy was administered in 94.6% of the inpatients, and there were significant differences in the oxygen mode between the non-severe group and severe group, as well as the discharged group and the death group (p < 0.001).

Univariable and multivariable logistic regression analyses of demographic and clinical factors associated with the severity of COVID-19 were shown in Table 2, and laboratory and radiological factors associated with the severity of COVID-19 were shown in Table 3. In the univariable analyses, nineteen factors were significantly associated with increasing risks of the severity of COVID-19 pneumonia: age ≥ 65 years, chest tightness, WBC ≤10⨯10⁹/L, N > 7⨯10⁹/L, N% >70, L < 0.8⨯10⁹/L, L% <20, NLR >4.4, PLT <100⨯10⁹/L, CRP >10 mg/L, PCT >0.25 ng/mL, D-dimer >0.55 mg/L, CK >190 U/L, LDH >220 U/L, α-HBDH >182 U/L, ALB <35 g/L, FBG >6.1 mmol/L, DPS and GGO in chest imaging examination.

Multivariable logistic regression analyses revealed that aged 65 years or above (adjusted odds ratio [OR] and 95% confidence interval [CI], 2.171 [1.096, 4.297]; p = 0.029), chest tightness (adjusted OR 3.095 [1.682, 5.694]; p < 0.001), NLR >4.4 (adjusted OR 2.683 [1.172, 6.141]; p = 0.020), α-HBDH >182 U/L (adjusted OR 5.465 [2.556, 11.684]; p < 0.001), albumin <35 g/L (adjusted OR 2.270 [1.053, 4.896]; p = 0.037), and GGO (adjusted OR 2.417 [1.158, 5.047]; p = 0.010) in radiological finding were independent risk factors associated with the severity of COVID-19 pneumonia.

Univariable and multivariable logistic regression analyses of demographic and clinical factors associated with in-hospital death of COVID-19 were shown in Table 4, and laboratory and radiological factors associated with in-hospital death of COVID-19 were shown in Table 5. In the univariable analyses, 28 factors were significantly associated with increasing risks of the mortality of COVID-19 pneumonia: aged 65 years or above, live in urban areas, cardiovascular disease, consciousness disorders, RR ≥30 breaths per minute, SaO₂ ≤ 93%, PaO₂/FiO₂ ≤ 300 mmHg, WBC ≤10⨯10⁹/L, N > 7⨯10⁹/L, N% >70, L < 0.8⨯10⁹/L, L% <20, NLR >4.4, PLT <100⨯10⁹/L, CRP >10 mg/L, PCT >0.25 ng/mL, APTT >40 s, D-dimer >0.55 mg/L, CK >190 U/L, LDH >220 U/L, α-HBDH >182 U/L, total bilirubin (TBIL) >20.5 μmol/L, AST >34 U/L, ALT >55 U/L, ALB <35 g/L, BUN >7.4 mmol/L, CR >110.5 μmol/L and FBG >6.1 mmol/L.

Multivariable logistic regression analyses revealed that cardiovascular disease (adjusted OR 2.747 [1.214, 6.220]; p = 0.015), PaO₂/FiO₂ ≤ 300 mmHg (adjusted OR 4.716 [2.115, 10.518]; p < 0.001), PLT <100⨯10⁹/L (adjusted OR 15.149 [3.255, 70.508]; p = 0.001), D-dimer>0.55 mg/L (adjusted OR 9.483 [1.773, 50.728]; p = 0.009), α-HBDH>182 U/L (adjusted OR 8.709 [2.787, 27.217]; p = 0.001), TBIL >20.5 μmol/L (adjusted OR 4.588 [1.479, 14.225]; p = 0.008), ALT >55 U/L (adjusted OR 5.438 [1.022, 28.920]; p = 0.047), BUN >7.4 mmol/L (adjusted OR 4.320 [1.676, 11.137]; p = 0.002), CR >110.5 μmol/L (adjusted OR 6.430 [2.155, 19.186]; p = 0.001), and FBG >6.1 mmol/L (adjusted OR 5.892 [1.646, 21.084]; p = 0.006) were independent risk factors associated with the mortality of COVID-19 pneumonia.

This study was the first to characterize differences in the severity and mortality risk factors for patients hospitalized for COVID-19 pneumonia between the early wave and the very late stage of the pandemic. The investigation of these differences could help healthcare providers monitor susceptible population at an early stage and offer theoretical assistance for future management of this disease. This study had several limitations. Firstly, this study was conducted in a single center, which wasn’t representative of the general situation in China. Further larger and more representative studies are needed to explore how these factors affect disease severity and mortality. Secondly, candidate predictors were collected from the electronic medical records in this retrospective study. Examinations and tests were carried out based on individual specific condition. Missing data of some variables from hospitalized patients, such as detailed information of blood pressure and laboratory data including interleukin (IL)-6, brain natriuretic peptide (BNP), myoglobin, and troponin I, made it impossible to characterize differences in new-onset disorders (such as diabetes, hypertension, gastrointestinal and cardiac symptoms) for COVID-19 patients between different stages of the pandemic. Early identification of risk factors for new-onset disorders could help prevent long-term complications. However, long-term follow-up for liver, cardiac, neurological, pulmonary and endocrine/genitourinary systems complications were not conducted in our study. Thirdly, patients were categorized as non-severe and severe groups within 48 h of hospital admission based on initial clinical presentation. Patients experienced clinical deterioration (admitted as moderate cases but developed into severe cases) during hospitalization were not discriminated in our study, which might cause bias. Finally, effective treatment with antivirals for COVID-19 has been recommend in the Chinese guidelines (6). Nirmatrelvir/ritonavir and azvudine (the first homegrown anti-COVID-19 drug by China) are available during the Omicron predominant period (10), which certainly play a major role in improving patients’ survival (63, 64). Due to the diverse treatment schemes for COVID-19 among different patients and clinical departments (some patients had already received antiviral treatment before admission), the effect of treatments was not considered as candidate predictor of disease mortality.