Guidelines for bioelectrical impedance analysis (BIA) recommend fasting before measurements, but how it affects the outcomes is unclear. This descriptive, before-and-after study examines the effect of fasting on BIA results and its impact on the diagnosis of malnutrition in hospitalized cancer patients. Fifty-three oncology patients (admitted in January-March 2020) were consecutively enrolled regardless of their nutritional status. Patients were assessed by the same dietician 24-48 h after admission, following the usual clinical practice. The measurements were taken after 12-h fasting (fasting state) and 60-90 min after breakfast (non-fasting state). Bioimpedance parameters (resistance [R], reactance [Xc], phase angle [PA]) and body composition indices (free-fat mass index [FFMI] and appendicular skeletal muscle index [ASMI]) were calculated. On average, R values did not significantly differ between fasting and non-fasting states (mean difference: +1.82 Ω; p = 0.64). The non-fasting Xc and PA were reduced (mean differences: -1.55 Ω, p = 0.93 and 0.09°, p = 0.82, respectively). Fasting and non-fasting FFMI and ASMI were similar (mean differences: -0.13 kg/m2 (p = 0.5) and -0.10 kg/m2 (p not calculated)). These results suggest that BIA does not require fasting, facilitating its routine use in hospitalized cancer patients.