Surgical interventions for gastric cancer: a review of systematic reviews

Aim: To evaluate methodological quality and the extent of concordance among meta-analysis and/or systematic reviews on surgical interventions for gastric cancer (GC). Methods: A comprehensive search of PubMed, Medline, EMBASE, the Cochrane library and the DARE database was conducted to identify the reviews comparing different surgical interventions for GC prior to April 2014. After applying included criteria, available data were summarized and appraised by the Oxman and Guyatt scale. Results: Fifty six reviews were included. Forty five reviews (80.4%) were well conducted, with scores of adapted Oxman and Guyatt scale >= 14. The reviews differed in criteria for avoiding bias and assessing the validity of the primary studies. Many primary studies displayed major methodological flaws, such as randomization, allocation concealment, and dropouts and withdrawals. According to the concordance assessment, laparoscopy-assisted gastrectomy (LAG) was superior to open gastrectomy, and laparoscopy-assisted distal gastrectomy was superior to open distal gastrectomy in short-term outcomes. However, the concordance regarding other surgical interventions, such as D1 vs. D2 lymphadenectomy, and robotic gastrectomy vs. LAG were absent. Conclusion: Systematic reviews on surgical interventions for GC displayed relatively high methodological quality. The improvement of methodological quality and reporting was necessary for primary studies. The superiority of laparoscopic over open surgery was demonstrated. But concordance on other surgical interventions was rare, which needed more well-designed RCTs and systematic reviews.