LAUSR

COVID-19 survivors are associated with acute respiratory failure (ARF) and show a high prevalence of impairment in physical performance. The present studied aimed to assess whether we may cluster these individuals according to an exercise test. The presented study is a retrospective analysis of 154 survivors who were admitted to two hospitals of Istituti Clinici Scientifici Maugeri network, Italy. Clinical characteristics, walked distance, heart rate (HR), pulse oximetry (SpO2), dyspnoea, and leg fatigue (Borg scale: Borg-D and Borg-F, respectively) while performing the six-minute walking test (6MWT) were entered into unsupervised clustering analysis. Multivariate linear regression identified variables that were informative for the set of variables used for cluster definition. Cluster 1 (C1: 86.4% of participants) and Cluster 2 (C2: 13.6%) were identified. Compared to C1, the individuals in C2 were significantly older, showed significantly higher increase in fatigue and in dyspnoea, greater reduction in SpO2, and a lower HRpeak during the test. The need of walking aids, time from admission to acute care hospitals, age, body mass index, endotracheal intubation, baseline HR and baseline Borg-D, and exercise-induced SpO2 change were significantly associated with the variables that were used for cluster definition. Different characteristics and physiological parameters during the 6MWT characterise survivors of COVID-19-associated ARF. These results may help in the management of the long-term effects of the disease.