This was a multicenter, case-control study with a prospective follow-up of 12 months after hospital discharge and visits at 1, 3, 6, and 12 months after discharge. Cases (n = 153) were defined as patients hospitalized due to COVID-19 infection who tested positive on nasal swabs from March to April 2020 in three university hospitals in Catalonia, Spain (Clínic Barcelona, Bellvitge Hospital, and Vic Hospital) that developed OP pattern in CT after 14 days of symptoms onset. We excluded patients with interstitial pathology that could be due to other causes such as coinfection by other germs demonstrated in microbiological cultures, although since the study was carried out in the first wave of the pandemic it was not possible, due to the risk involved, performing fibrobronchoscopy or taking biopsies. We also excluded immunocompromised patients (HIV, neoplastic disease or immunosuppressive drugs).

To investigate the risk factors for OP during hospitalization, we generated a control population (n = 140) from a random sample of COVID-19 patients admitted to the same hospitals in the same period who, at day 14 after the onset of symptoms, had improvement or disappearance of pulmonary opacities on chest radiography and followed a favorable evolutionary course without requiring in any case ventilatory support or admission to the ICU. The study protocol was approved by the Ethical Review Board of the coordinating hospital (HCB/ 2020/0410) and all patients provided informed consent.

The demographic, clinical, imaging and biological characteristics of the patients and controls were obtained from their electronic medical records. These data included age, sex, smoking status, body temperature, comorbidities (arterial hypertension, any chronic cardiovascular disease, any chronic respiratory disease, diabetes), inflammatory and metabolic parameters (ferritin, D-dimer, LDH, procalcitonin, lymphocyte count, C-reactive protein, and lactate dehydrogenase), and arterial blood gases (SaO2, PaO2 (FIO2), and PaCO2) at hospital admission and discharge. We also collected information on the following events: time from onset of symptoms to hospital admission, length of hospital stays, treatment (dose and duration) with azithromycin, antivirals (lopinavir/ritonavir, remdesivir), anti-inflammatory drugs (tocilizumab, anakinra, hydroxychloroquine), and/or oral corticosteroids. After discharge, FEV1, FVC, and DLco were measured only in OP-like pattern cases at three, six and 12 months, and CT thorax scans after one and 12 months (see below), following international standards. The reference values were those of Roca et al. [7, 8].

For optimal diagnostic performance of computed tomography, images were obtained in deep inspiration, with thin sections (1 mm) and high spatial resolution reconstruction (HRCT). The images were taken on different equipment all of them Siemens which were renewed after pandemic. All chest CT images were independently reviewed by two thoracic radiologists (MB and MS with 7 and 17 years of experience respectively) and a pneumologist (SC). Before scoring, each observer underwent supervised training to score the HRCT patterns described below. The observers had no knowledge of the pulmonary function or other clinical indicators of disease severity. When there was a discrepancy, the final result was reached through consensus. Radiological findings included ground glass, opacities, consolidation, reticulation, traction bronchiectasis, subpleural bands, honeycombing, and pleural retraction [9, 10]. An OP-like pattern was defined as peri-bronchovascular consolidations with perilobular distribution and/or the reverse halo sign [10]. Fibrotic-like changes were defined as the presence of traction bronchiectasis, pleural retraction, parenchymal bands and/or honeycombing [11]. A CT score [12] was used to quantify parenchymal abnormalities when OP was suspected, one month and one year after discharge. Both lungs were divided into their respective lobes. Involvement in each lobe was scored based on the following criteria: 0 (no involvement); 1 (< 10% involvement), 2 (10–25% involvement), 3 (25–50% involvement), 4 (50–75% involvement) and 5 (> 75% involvement). CT pulmonary angiography was performed when the blood D-dimer level was > 500 ng/ml, following the COVID-19 protocol from each hospital.

For continuous variables, descriptive statistics included n, proportion, mean ± SD, or median [25th-75th quartile]. We also compared the evolution of biological parameters between admission and hospital discharge in both cases and controls using Wilcoxon or McNemar paired tests. To investigate the risk factors for the development of OP-like pattern during hospitalization, we first compared clinical, biological, and radiological characteristics at admission between cases and controls using Student’s, Mann-Whitney U, chi-square, or Fisher’s exact tests as appropriate. We then built a multivariable logistic regression model that included all factors significantly different in this bivariate analysis. We used forward/backward strategies to obtain the most parsimonious model that explains the observations. Medical treatments were not included as potential risk factors, as they occurred during admission but not at baseline. To study the longitudinal evolution of OP patients, we assessed the distribution of radiological (at 1 and 12 months) and functional (at 3, 6, and 12 months) characteristics using univariate statistics and their changes over time using paired tests. Finally, we determined the factors at hospital discharge that predicted the persistence of fibrotic changes at one-year follow-up in OP patients using multivariable logistic regression (conditional stepwise forward model (Pin < 0.10, Pout < 0.05). Receiver operating characteristic (ROC) curves were used to determine the optimal cutoff values of daily corticosteroid dose in relation to the persistence of fibrotic changes. These cutoff values were included in the multivariate analysis. All analyses were performed using Stata 16.0 (StataCorp, College Station, TX, USA) and SPSS 26.0 (SPSS Inc., Chicago, IL, USA).