Objective: To perform a meta-analysis of prospective and retrospective studies exploring the association of the modified Glasgow prognostic score (mGPS) with overall survival (OS) in patients with lung cancer.
Methods: Relevant studies were identified by searching the Cochrane Library, Web of Science, Embase and PubMed until April 16, 2017. We combined hazard ratios (HRs) and 95% confidence intervals (CIs) to assess the correlation between mGPS and OS in patients with lung cancer.
Results: Eleven studies involving 5817 participants from several countries were included in the meta-analysis. In a pooled analysis of all studies, elevated mGPS predicted poorer OS (HR = 1.77; 95% CI: 1.35-2.31; P<0.05). Subgroup analyses stratified by mGPS showed that mGPS of 1 or 2 and mGPS≥1 were predictive of poorer OS and that the HR for mGPS of 2 (HR = 5.82; 95% CI: 1.85-18.22; P = 0.003) was significantly greater than that for mGPS of 1 (HR = 1.74; 95% CI: 1.24-2.45; P = 0.001) and mGPS≥1 (HR = 1.42; 95% CI: 1.14-1.76; P = 0.002). Among patients undergoing surgery, elevated mGPS had a non-significant correlation with reduced OS (HR = 2.48; 95% CI: 0.90-6.85; P = 0.079), whereas the correlation was significant for patients receiving chemotherapy or other palliative treatment (HR = 1.74; 95% CI: 1.31-2.30; P<0.05).
Conclusions: Our findings indicate that mGPS may have prognostic value in lung cancer, as we detected a significant association between elevated mGPS and poorer OS. The association between mGPS and poorer OS was non-significant among patients undergoing surgery, which may be attributable to lower tumor load. However, further studies are warranted to draw firm conclusions.