Pulmonary phenotypes and COVID-19 evolution patterns during the first year of the pandemic

Aim: To characterize clinical, functional, and imaging pulmonary phenotypes, and the evolution of COVID-19 in hospitalized patients. Methods: COVID-19 patients were evaluated at 3 (T3) and 12 months (T12) after discharge from the Pisa University Hospital (March-September 2020). Spirometry, lung volumes, DLCO, and the evolution of COVID-19 pneumonia CT signs (PS) were assessed. Through a latent transition analysis (LTA), based on PS, DLCO, and lung function (PFTs), we assessed COVID-19 cross-sectional phenotypes at T3 and T12 and their longitudinal patterns. Risk factors for longitudinal patterns were evaluated by multinomial logistic regression. Results: Among 307 discharged patients, 57% at T3 and 44.3% at T12 were followed up (62.8% males; mean age 61.1 years). Among patients with complete T12-follow-up, 24.1% at T3 and 21.6% at T12 had impaired DLCO, and 4.4% had a restrictive pattern at both T3 and T12. 48.9% and 18.3% had no PS at T3 and T12, respectively, but 32.8% had persistent PS at T12. LTA found three cross-sectional phenotypes at T3 and T12 (i.e., no PS and normal PFTs; PS and normal PFTs; PS and impaired PFTs), and four longitudinal patterns: 1) persistent no PS and normal PFTs (47.9%); 2) resolution of both PS and PFTs at T12 (15.4%); persistence of PS at T12 (36.7%), with 3) (11%) or without 4) (25.7%) PFTs impairment. The last two patterns were associated with higher age, more comorbidities, and severe acute COVID-19. Conclusions: In the first year of the pandemic, older patients, with more comorbidities and severe acute COVID-19, were those with the worst COVID-19 evolution patterns over 12 months and are those who currently need a very long-term follow-up