COVID-19 may have a heterogeneous onset, especially in older age. However, whether and how COVID-19 signs and symptoms may present and aggregate together according to sociodemographic and health factors is unclear, as well as their prognostic value. This study included 981 COVID-19 inpatients who participated in the GeroCovid Observational study. Signs/symptoms at disease onset, sociodemographic, health, cognitive status, and mobility were systematically recorded. Clusters of signs/symptoms were identified through agglomerative hierarchical clustering. The associations of single signs/symptoms and symptom clusters with longer hospitalization (≥16 days) and in-hospital mortality were explored through logistic and Cox regressions. The signs/symptoms most reported in our sample (age 78.3 ± 9.39 years; 49.4% women) were fever (62.5%), cough (45.5%), and dyspnea (62.7%). Atypical symptoms were reported by up to one-third of patients, and delirium by 9.1%. Atypical symptoms were more frequent with advancing age and with lower pre-COVID-19 cognitive and mobility levels. Older men more likely reported respiratory symptoms than women. Dyspnea (hazard ratio [HR] = 1.47, 95% confidence interval [CI]: 1.02-2.12), tachypnea (HR = 1.53, 95% CI: 1.14-2.07), low oxygen saturation (HR = 1.95, 95% CI: 1.32-2.88) and delirium (HR = 1.60, 95% CI: 1.13-2.28) were associated with higher in-hospital mortality. Four symptom clusters were identified. Compared with the mild respiratory symptoms cluster, the severe clinical impairment cluster was associated with higher mortality (HR = 2.57, 95% CI: 1.58-4.18). The severe clinical impairment and aspecific symptoms clusters were associated with longer hospitalization (odds ratio [OR] = 2.38, 95% CI: 1.56-3.63, and OR = 1.75, 95% CI: 1.08-2.83, respectively). Multiple health aspects influence COVID-19 clinical presentation. A symptom clusters approach may help predict adverse health outcomes in older patients. In addition to respiratory symptoms, delirium is independently associated with mortality risk. ClinicalTrials.gov (NCT04379440).
Keywords: COVID-19; aged; cluster analysis; mortality; signs and symptoms.