Associations between nutrition risk scores and sarcopenia in gastrointestinal cancer patients: a cross-sectional study

Purpose: Sarcopenia is an independent risk factor for poor prognosis of cancers. The nutritional risk screening 2002 (NRS2002) and patient-generated subjective global assessment (PG-SGA) tools are widely used tools for nutrition risk screening and assessing. The purpose of this study was to investigate whether NRS2002 and PG-SGA scores are associated with sarcopenia in gastrointestinal cancers.

Methods: A consecutive cohort comprised of 432 gastrointestinal cancer patients was conducted. We used NRS2002 and PG-SGA to assess their nutrition status. Sarcopenia was diagnosed with CT scan at the third lumber vertebra level. The correlations of nutritional scores with SMI, nutritional categories with sarcopenia were assessed by Spearman's correlation test and point biserial correlation. The cut-off value of nutritional scores for identifying sarcopenia was obtained by maximum Youden index. Logistic regression was used to confirm the associations.

Results: Sarcopenia patients had higher NRS2002 (2.63 ± 1.16 vs. 2.15 ± 1.20, p < 0.001) and PG-SGA (8.69 ± 1.16 vs. 5.56 ± 3.28, p < 0.001) scores. The NRS2002 (r = -0.198, p < 0.001) and PG-SGA (r = -0.409, p < 0.001) scores were significantly and negatively correlated with skeletal muscle mass index. The cut-off value of PG-SGA score for predicting sarcopenia was 7. In multivariate logistic regression, the PG-SGA exceeded 7 score (OR = 7.489, 95% CI: 4.122-13.608, p < 0.001) was significantly associated with increased risk of sarcopenia, while NRS2002 score showed no significant association with sarcopenia.

Conclusions: PG-SGA ≥ 7 was associated with increased risk of sarcopenia and could serve as a useful criterion for capturing sarcopenia in gastrointestinal cancers. Routine PG-SGA evaluation for patient with gastrointestinal cancers is important.