To investigate the relationship between obesity and 30-day mortality in a cohort of older hospitalized COVID-19 inpatients.
Included patients were aged 70 years or more; hospitalized in acute geriatric wards between March and December 2020; with a positive PCR for COVID-19; not candidate to intensive care unit admission. Clinical data were collected from patients electronic medical records. Data on 30-day mortality were retrieved from the hospital administrative database.
Patients included (N = 294) were on average 83.4 ± 6.7 years old, 50.7% were women, and 21.7% were obese (BMI > 30 kg/m ² ). At 30-day, 85 (28.9%) patients were deceased. Compared to survivors in bivariable analysis, deceased patients were older (84.6 ± 7.6 vs 83.0 ± 6.3 years), more frequently with very complex health status (63.5% vs 39.7%, P < .001), but less frequently obese (13.4% vs 24.9%, P = .033) at admission. Over their stay, deceased patients more frequently (all P < .001) developed radiologic signs of COVID-19 (84.7% vs 58.9%), anorexia (84.7% vs 59.8%), hypernatremia (40.0% vs 10.5%), delirium (74.1% vs 30.1%), and need for oxygen (87.1% vs 46.4%) compared to survivors. In multivariable analysis that controlled for all markers of poor prognosis identified in bivariable analysis, obese patients remain with 64% (adjOR 0.36, 95%CI 0.14-0.95, P = .038) lower odds to be deceased at 30-day than non-obese patients.
In this population of older COVID-19 inpatients, an inverse association between obesity and 30-day mortality was observed even after adjusting for all already-known markers of poor prognosis. This result challenges previous observations in younger cohorts and would need to be replicated.