ABSTRACT
<p><b>OBJECTIVE</b>To investigate the nutritional status, and provide evidence for nutritional treatment option.</p><p><b>METHODS</b>A total of 452 patients with gastrointestinal cancer were selected, including 156 gastric cancer,117 colon cancer, and 180 rectal cancer. The nutritional risk screening 2002(NRS2002) was applied to grade the nutritional risk. A multi-frequency bioelectrical impedance analysis was used to measure the patients' body composition. Albumin (Alb), prealbumin(PA), transferring(Tf), retinol binding protein(RBP), red blood cell(RBC), hemoglobin (Hb), haematocrit(Hct) were measured after fasting.</p><p><b>RESULTS</b>The rate of patients with NRS2002 score more than 3 was 70.5%(110/156) for gastric cancer, 53.8%(63/117) for colon cancer, and 46.7%(86/180) for rectal cancer. The score for impaired nutritional status more than 1 for gastric cancer was higher than that for colorectal cancer(P<0.05), while patients with disease score more than 2 was less for gastric cancer(P<0.05). Body mass index(BMI), obesity degree, fat content, fat percentage, and arm circumference were lower in gastric cancer patients as compared to colorectal cancer patients(P<0.05); but protein percentage, muscle percentage, ratio of muscles of arm, and cell mass percentage were higher in gastric cancer patients(P<0.05). The proportions of patients with low Alb, PA, Tf, BC, Hb, Hct were higher for gastric cancer and colon cancer(P<0.05).</p><p><b>CONCLUSION</b>Patients with gastric cancer are prone to fat loss and therefore have a higher nutritional risk and malnutrition than those with colorectal cancer. Combination of body composition analysis and laboratory examination may achieve comprehensive evaluation of the nutritional status of patients, and provide the evidence of nutritional therapy by being combined with NRS2002 score.</p>