Pulmonary clinical, functional, and imaging follow-up of COVID-19 patients hospitalized in the University Hospital of Pisa during the first year of the pandemic

RATIONALE: COVID-19 pneumonia has affected millions of people in the world, with consequent many pulmonary sequelae. The aim of this thesis is to describe imaging and clinical-functional pulmonary data at 3 (T3) and 12 months (T12) after discharge in patients hospitalized for COVID-19. METHODS: In this single-center prospective observational study, COVID-19 patients discharged from the University Hospital of Pisa from March to September 2020 were evaluated (T0). No exclusion criteria were applied. We retrospectively collected data about the acute disease severity. Expert radiologists qualitatively assessed the evolution of COVID-19 pneumonia CT signs (PS) at three months (T0 vs. T3) by using an ad hoc coding system. Chest CT at T12 was performed only in patients with persistent PS at T3. Both at T3 and T12, all the patients underwent PFTs (spirometry, plethysmography, DLCO), and pulmonary visit. A subgroup of patients with persistent PS at T12 underwent to chest CT and PFTs at 24 months (T24). Through an LTA approach, we detected the cross-sectional phenotypes at T3 and T12 (based on imaging and PFTs data) and the longitudinal patterns of the evolution of the acute COVID-19 disease, with multiple logistic regressions adjusted for patients’ risk factors. RESULTS: Among 307 discharged patients, 57% and 44.3% of patients were followed up at T3 and T12, respectively, while 12.4% died within T3. Followed patients’ characteristics were: 62.8% men, mean age 61.1 (14.9) years, mean BMI 28.1 (4.8) kg/m2, 11.4% smokers, 43.4% ex-smokers, 59% had 1+ comorbidities [mean score of Charlson comorbidity index 2.3 (2.1)]. Median hospitalization was 15 (IQR 14) days; 15.4% of patients stayed 3+ days in ICU. According to the WHO Severity Index, 62.6%, 22.9%, and 14.5% of patients had mild, moderate, or severe COVID-19. Most patients showed a resolution of respiratory symptoms at T3, but at T12 36.1% of patients still referred dyspnea. Impaired DLCO was the most common abnormality, at both T3 (24.1%) and T12 (21.6%), with no statistically significant decrease. Patients with normal spirometry were 88.6% and 91.2% at T3 and T12, respectively, while 7% and 4.4% at T3 and T12, respectively, had a restrictive pattern. Mean measured values of TLC, FVC, and VC were observed to increase during follow-up. 52.1% of patients showed resolution of PS already at T3, while 47.9% had persistent PS. At T12, 35.8% of patients showed a complete resolution, while the remaining 64.2% showed persistence of PS. At T24, all the patients who had PS at T12 continued to have persistence of PS. LTA model found three cross-sectional phenotypes both at T3 and T12 (patients with no PS and normal PFTs, patients with PS and normal PFTs, and patients with PS and impaired PFTs) and four longitudinal patterns for the evolution of acute COVID-19 disease. 1) patients with normal cCT and PFTs at T3 (47.9%); 2) patients with resolution of PS and PFTs between T3 and T12 (14.4%); 3) patients who still showed persistence of PS at T12 (36.7%), with (11%) or without (25.7%) PFTs impairment. The regression analysis confirmed that patients of the last two patterns had higher age, more comorbidities, and severe COVID-19. CONCLUSIONS: About one-third of patients showed persistence of PS, and about one-fifth had DLCO abnormalities at T12. The LTA model identified different phenotypes and patterns of evolution of the COVID-19 acute disease, based on the persistence/absence of imaging and/or functional pulmonary abnormalities up to 12 months follow-up. A very long-term follow-up is needed to further study the natural history of COVID-19 and its impacts on the lungs