Abstract
Malnourished patients with gastrointestinal tumours are at risk for postoperative complications and death. The aim of this study was to determine which nutritional assessment method better predicts outcome. Seventy-four patients, 45 men and 29 women; mean (SD) age of 63 (102) yr (range = 34 to 83), undergoing surgical resections for esophageal (n = 19) gastric (n = 43) and pancreatic (n = 12) tumors were preoperatively assessed by Patient Generated Subjective Global Assessment, anthropometry, and by laboratory sampling. Forty-three (58%) of them were unnourished; 25 Subjective Global Assessment (SGA)-A, 34 SGA-B, and 15 SGA-C cases. Mean (SD) of dominant hand adductor pollicis muscle thickness (DAPM) was 13 (3.5) mm and mean (SD) serum albumin was 3.8 (0.5) g/dL. Mean (SD) hospital staying for patients who complicated and died was 34 (29) days and 23 (13) days for survivors (not significant); SGA-B cases were significantly associated with higher mortality (n = 12, P<0.001). Patients with a mean (SD) DAPM below 10.8 (3.7) mm died more frequently than those with a mean (SD) greater than 14 (3) mm (P < 0.001). None of the methods was significantly related to hospital stay, but receiver operating characteristic curves (95% confidence interval) for PG-SGA and DAPM thickness (0.75 and 0.74) reliably predicted mortality (P<0.001) and these methods may be used as preoperative parameter.
ACKNOWLEDGMENTS
The authors are in debt with Mr. Marcos Martau Poziomyck, MSc for his critical review of the manuscript, to Dr. AC Weston, MD, Phd, Leader of the Porto Alegre Gastric Cancer Study Gorup for including their cases in the study and to Brazilian Postgraduate Program Council (CAPES) for partially supporting this study.