A relationship between weight loss and inflammation has been described in patients with cancer. In the present study, the relationship between subjective global assessment (SGA) and the severity of inflammation, as defined by Glasgow prognostic score (GPS), as well as the relationship of both of these measures with the presence of complications and survival time, was assessed. In addition, we compared the diagnosis given by SGA with parameters of nutritional assessment, such as body mass index, triceps skinfold, midarm circumference (MAC), midarm muscle circumference (MAMC), phase angle (PA), adductor pollicis muscle thickness (APMT), and handgrip strength (HGS). According to the SGA, the nutritional status was associated with the GPS (P < 0.05), and both the SGA and GPS were associated with the presence of complications. However, the GPS [area under the curve (AUC): 0.77, P < 0.05, confidence interval (CI) = 0.580, 0.956] seems to be more accurate in identifying complications than the SGA (AUC: 0.679, P < 0.05, CI = 0.426, 0.931). Only GPS was associated with survival time. Comparing the different nutritional assessment methods with the SGA suggested that the MAC, MAMC, APMT, PA, and HGS parameters may be helpful in differentiating between nourished and malnourished patients, if new cutoffs are adopted.