Hospitalized patients with respiratory failure due to SARS-CoV-2 pneumonia are at increased risk of malnutrition and related mortality. The predictive value of the Mini-Nutritional Assessment short form (MNA-sf), hand-grip strength (HGS), and bioelectrical impedance analysis (BIA) was determined with respect to in-hospital mortality or endotracheal intubation. The study included 101 patients admitted to a sub-intensive care unit from November 2021 to April 2022. The discriminative capacity of MNA-sf, HGS, and body composition parameters (skeletal mass index and phase angle) was assessed computing the area under the receiver operating characteristic curves (AUC). Analyses were stratified by age groups (<70/70+ years). The MNA-sf alone or in combination with HGS or BIA was not able to reliably predict our outcome. In younger participants, HGS showed a sensitivity of 0.87 and a specificity of 0.54 (AUC: 0.77). In older participants, phase angle (AUC: 0.72) was the best predictor and MNA-sf in combination with HGS had an AUC of 0.66. In our sample, MNA- sf alone, or in combination with HGS and BIA was not useful to predict our outcome in patients with COVID-19 pneumonia. Phase angle and HGS may be useful tools to predict worse outcomes in older and younger patients, respectively.