Prognostic significance of the modified Glasgow prognostic score in elderly patients with gastric cancer

The Glasgow prognostic score (GPS) is a preoperatively determined inflammation-based score. Reports suggest a significant correlation between the GPS and prognosis in several cancer types. We aimed to clarify the prognostic significance of the modified GPS (mGPS) in patients undergoing gastrectomy for gastric cancer. Two hundred and ninety-four patients with gastric cancer, 195 aged < 75 years (group NE) and 99 aged > 75 years (group E), who underwent gastrectomy from March 2005 to March 2011 were enrolled. Patients with an elevated C-reactive protein level (> 0.5 mg/dL) and hypoalbuminemia (< 3.8 g/dL) were assigned a mGPS of 2, those with either 1 abnormality were assigned a mGPS of 1, and those with neither abnormality were assigned a mGPS of 0. Cox proportional hazard models and Kaplan-Meier analysis were used to evaluate the usefulness of mGPS as a prognostic indicator. In the NE group, the prognosis of the 3 groups stratified by mGPS did not differ significantly. In multivariate Cox regression analysis, the type of gastrectomy, peritoneal metastasis, and stage were independently associated with poor prognosis. However, group E patients with a mGPS of 2 had significantly poorer prognosis than those with a mGPS of 0 or 1. In this age group, stage and mGPS were independently associated with poor prognosis. In patients aged > 75 years undergoing potentially curative gastrectomy, the preoperative mGPS was an independent predictor of survival. Therefore, mGPS can be a useful prognostic indicator in elderly patients with gastric cancer.