Objective To investigate the clinical values of the Patient-Generated Subjective Global Assessment (PG-SGA) and the Nutritional Risk Screening 2002 (NRS 2002) in identifying sarcopenia in patients with gastric cancer.
Methods A total of 136 patients with gastric cancer were selected for retrospective study, and their preoperative nutritional conditions were evaluated by PG-SGA and NRS 2002. Sarcopenia was defined according to the skeletal muscle index at the level of the third lumbar vertebrae (L3-SMI) on CT scans, the patients were divided into sarcopenia group and non-sarcopenia group, and the baseline characteristics were compared between the two groups. The identification abilities of PG-SGA and NRS 2002 for sarcopenia were evaluated by receiver operating characteristics (ROC) curve. Taking the presence of sarcopenia as the main independent variable, the multivariate Logistic regression analysis was used to evaluate the covariates with independent correlation.
Results According to the L3-SMI definition of sarcopenia, 43 patients (31.6%) with gastric cancer were diagnosed as sarcopenia. Compared with the non-sarcopenia patients, the average age of sarcopenia patients was significantly larger, the body mass index (BMI), L3-SMI, hemoglobin value, peripheral blood lymphocyte count, serum albumin, total cholesterol and triglyceride levels were significantly lower (P < 0.05 or P < 0.01). Sarcopenia was also significantly associated with NRS 2002 score≥3 and severe malnutrition defined by PG-SGA (P < 0.01). The ROC curve showed that the optimal cut-off score of PG-SGA in identifying sarcopenia in patients with gastric cancer was 5, and the area under the curve (AUC) was 0.714 (95%CI, 0.621 to 0.807); the AUC of NRS 2002 score≥3 for identifying sarcopenia was 0.658 (95%CI, 0.560 to 0.757). Multivariate Logistic regression analysis showed that when PG-SGA score≥5, the risk of sarcopenia increased by nearly 4.484 times (OR=4.484, 95%CI, 1.805 to 11.136, P=0.001). Other factors independently associated with sarcopenia in patients with gastric cancer were age (OR=1.046, 95%CI, 1.007 to 1.087, P=0.019), BMI (OR=0.734, 95%CI, 0.627 to 0.859, P < 0.001) and hemoglobin value (OR=0.805, 95%CI, 0.658 to 0.986, P=0.016).
Conclusion Sarcopenia defined by L3-SMI is significantly correlated with NRS 2002 and PG-SGA scores, and the PG-SGA score may be used as an effective tool to identify sarcopenia.