According to the COVID-19 treatment guidelines from China and the WHO[8, 9], the severity of COVID-19 pneumonia is categorized into four levels according to the condition at the time of admission: mild (no radiological evidence of pneumonia and mild clinical symptoms), moderate (pneumonia on chest radiograph with fever and evidence of respiratory symptoms), severe (pneumonia with any of the following indications: PaO2/FiO2≤300mmHg, oxygen saturation≤93% at rest, tachypnoea at RR≥30/min or respiratory distress), and critical (patients who either developed organ failure requiring ICU monitoring or respiratory failure requiring mechanical ventilation).
Forty-six patients with COVID-19 pneumonia treated at the First Affiliated Hospital, Harbin Medical University since February 2020 were enrolled in this study for follow-up in July 2020. These patients were categorized into three groups: Group A (n=24) who were diagnosed with moderate pneumonia in April 2020 and were followed up at three months after diagnosis; Group B (n=11) who were diagnosed with severe pneumonia in April 2020 and were followed up at three months after diagnosis; Group C (n=11) who were diagnosed with severe pneumonia in February 2020 and were followed up at six months after diagnosis. During hospitalization, all patients were treated in accordance with the COVID-19 treatment guidelines of China. Patients were given oxygen therapy, high-flow nasal cannula, non-invasive ventilator, or invasive ventilator according to their pulmonary conditions. They were discharged from our hospital when the following criteria were met: oropharyngeal swab SARS-CoV-2 nucleic acid was negative twice from tests done at least 24 hours apart; body temperature has returned to normal for more than 3 days; respiratory symptoms have improved significantly; lung radiograph shows significant improvement in acute exudative lesion.
These patients were all followed up in the outpatient clinic in July 2020. Figure 1 shows a flowchart of the study. They were assessed for their pulmonary function, chest high-resolution CT (HRCT), arterial blood gas analysis, blood tests, modified Medical Research Council (mMRC) dyspnea score, and health-related quality of life (HRQoL). The clinical research ethics committee of the First Affiliated Hospital, Harbin Medical University approved this study (Protocol No. IRB-AF/SC-04/01.0).
Blood analyses included blood cell counts, renal and liver function tests, coagulation profile, and immunoglobulin test for SARS-CoV-2. Serum SAR-CoV-2 IgG and IgM antibody titers (AU/mL) were analyzed with a chemiluminescent immunoassay (Shenzhen Yahuilong Biotechnology Co., Ltd, Shenzhen, China), with a reference level of 10 AU/mL. A blood gas analyzer (GEM Premier 3000; Instrumentation Laboratory, New York, USA) was used to quantify arterial oxygen partial pressure and the alveolar-arterial oxygen pressure gradient (PO2 (A-a)). Arterial blood gas analysis was not performed for two patients because of their disconsent.
All patients first were assessed with an oropharyngeal swab SARS-CoV-2 nucleic acid test to exclude active viral infection. Standard single-breath pulmonary function testing (MasterScreen Body/Diff, Jaeger Co., Germany) was then carried out to determine total lung capacity (TLC), forced vital capacity (FVC), vital capacity (VC), forced expiratory volume at first second (FEV1), and diffusing lung capacity for carbon monoxide single-breath (DLCO SB). DLCO SB values were corrected for individual hemoglobin levels. A DLCO SB of less than 80% was interpreted as diffusion deficit, while all other results were depicted in terms of percentages of predicted normal values. Pulmonary function test was not done for one patient because of positive SARS-CoV-2 IgM.
Chest high-resolution CT Scan (HRCT)
All COVID-19 patients underwent chest HRCT with a 256-slice multi-detector CT scanner (Brilliance iCT, Philips Healthcare, Holland) when hospitalized, and chest HRCT at the follow-up visit was performed with a 40-row multi-detector CT scanner (uCT 528, Shanghai United Imaging Healthcare, Shanghai, China). Imaging parameters were set as follows: slice thickness, 1 mm; tube voltage, 100 kV; 125 mas; 0.30-second gantry rotation time, automatic. All images were analyzed with the Extended Brilliance Workspace V 18.104.22.168 (Philips Healthcare, Cleveland, OH, USA). For each patient, chest HRCT was performed at four stages: when the patient was admitted (Admission), when clinical condition reached its worst (Progression), at the time of hospital discharge (Discharge), and at the time of follow-up (Follow-up).
Two experienced radiologists without prior knowledge of the clinical profiles reviewed and graded CT images independently. A consensus had to be reached between these two radiologists about the abnormalities. When there was a disagreement, the final decision would be made by a third senior radiologist with more than 10 years of experience. A scoring system was adapted for this study. Each image was assigned a score ranging from 0 to 5 based on the presence of air trapping, fibrosis, consolidation, and ground-glass opacity (GGO). Score 0 indicated a normal lung; Score 1 was scored if <5% of a lobe presented GGO; Score 2 if 6-25% was involved; Score 3 if 26-50% was involved; 4 points if 51-75% was involved; Score 5 if more than 75% was involved. All 5 lung lobes were scored and an abnormal CT score (range 0-25) was generated by adding scores of individual lobes.
Health-related quality of life (HRQoL) assessment and mMRC dyspnea score
The HRQoL of patients was determined using the St. George's Respiratory Questionnaire (SGRQ). This questionnaire encompassed information regarding symptom severity, activity tolerance, and impact on daily life. A higher score was indicative of worse overall functional status. The mMRC dyspnea scale was used to assess dyspnea (score 0–4, with 4 indicating the worst dyspnea).
Statistical analysis was performed using SPSS 25.0, For normally distributed variables, the data were expressed as mean ± standard deviation (SD); the differences among these three groups were analyzed with one-way ANOVA and then Fisher’s LSD tests. For variables that are not normally distributed, data were presented as medians (interquartile range) and analyzed with the Kruskal-Wallis H test and then Nemenyi tests. Categorical variables were presented as frequencies or percentages and statistically analyzed with the Chi-square test or Fisher's exact test. A p-value of <0.05 was taken to indicate statistical significance.