Locations and patients
In this prospective cross-sectional study, we recruited patients who had been treated for moderate or severe COVID-19 pneumonia as inpatients and discharged from Rohani hospital in Babol, northern Iran, during March 2020. The SARS-CoV-2 infection was confirmed by real-time polymerase chain reaction (RT-PCR) on nasopharyngeal swab samples collected from cases initially presented with suspicious symptoms (e.g., fever, cough, dyspnea, sputum discharge, etc.). These patients also underwent chest CT scan at admission. The clinical severity of COVID-19 pneumonia was classified as moderate (evidence of lower respiratory disease with oxygen saturation ≥ 94%) and severe (oxygen saturation < 94%, a ratio of arterial partial pressure of oxygen to fraction of inspired oxygen < 300, respiratory rate > 30, or lung infiltrates > 50%) as per the definition by World Health Organization (WHO) [8]. Patients with mild disease were not admitted as per the national COVID-19 protocol.
After assessing the patients’ medical records, the required data were extracted, including demographic information (such as sex and age), symptoms, comorbidities (such as cardiovascular diseases [CVDs], asthma, chronic obstructive pulmonary disease [COPD], and diabetes). The criteria for discharging were based on the flowchart of diagnosis and treatment of COVID-19 disease in Iran (improved general conditions, increased oxygen saturation without respiratory distress, suppressed fever for at least three days, improved laboratory results) [9]. Within the 3 months of follow-up (in June 2020), the next chest CT scan was performed on the patients who had residual symptoms and/or would like to monitor the changes of their CT images. The exclusion criteria were incomplete information about RT-PCR results or comorbidities, as well as unwillingness to participate in the study. We also did not include patients with history of smoking to prevent its potential confounding effects on the study outcomes. As the secondary outcome, the patients with fibrotic abnormalities in their second CT scan were followed up in the next 3 months (in September 2020) to monitor their imaging changes.
Imaging procedures
Non-enhanced 16-detector-row CT scans were conducted on the patients in the supine position during deep inspiration breath-hold from the thoracic inlet to the diaphragm (siemens SOMATOM Emotion 16, Siemens Healthcare, Med Image Systems, Germany). The following scanning parameters were used: tube voltage, 100 kV for patients with BMI ≤ 30 and 120 kV for patients with BMI > 30; tube current, 50–100 mAs; pitch, 0.8–1.5; thickness, 1–3 mm; Matrix, 512. No additional image reconstructions were necessary. The CT scans (at both admission and follow-up) were conducted using the same scanners and assessed by two senior radiologists with experience of more than 15 years (R.M. and M.N.), who was not aware of the patients’ status. CT imaging features, including traction bronchiectasis, honeycombing, parenchymal bands, and interlobar septal thickening (IST), were considered as the fibrotic-like changes. Also, parenchymal bands and interlobar septal thickening were considered as mild/moderate fibrosis, and traction bronchiectasis and honeycombing were considered as severe fibrosis.
We also calculated the CT severity score (CSS) for all patients at admission, on the basis of the involvement of each five lung zones, which was as follows [10]: score 0, no involvement; score 1, < 5% involvement; score 2, 5–25% involvement; score 3, 26–50% involvement; score 4, 51–75% involvement; and score 5, > 75% involvement. Finally, the total CSS was calculated by summing the scores, ranging from 0 to 25.
Data analysis
We used descriptive analysis to calculate frequencies, percentages, mean, and standard deviations. Kolmogorov–Smirnov test was used to evaluate the normality of the data. Independent t-test and Mann–Whitney test were used for comparing parametric and nonparametric continuous data between the groups, respectively. We conducted chi-squared test and logistic regression analysis to investigate the association of baseline information and imaging findings of the patients with post-COVID-19 lung fibrosis. The factors with significant association (consolidation and severe disease) were entered into the multivariable analysis. Concerning CSS, we presented it as both of continuous variable and median with interquartile range (IQR). The results were presented as odds ratio (OR) as well as 95% confidence interval (CI). We also estimated the area under ROC curve (AUC) for predictive ability of the CT scan features. A p value less than 0.05 was considered statistically significant. All statistical analyses were performed by SPSS software.