[Comparison of safety and efficacy between proximal gastrectomy and total gastrectomy for upper third gastric cancer: a Meta-analysis].

To compare the safety and efficacy between proximal gastrectomy and total gastrectomy and to ascertain the optimized procedure for patients with upper third gastric cancer through meta-analysis. The English literatures about proximal gastrectomy and total gastrectomy for upper third gastric cancer were searched from PubMed, EMBASE, the Cochrane Library and the Web of Science database and then collected. The quality of enrolled studies was independently assessed by two researchers according to the Newcastle-Ottawa Scale for retrospective studies and Jadad scale for RCT studies. The basic information of the literature and related clinical indicators were extracted. The primary endpoints were 5-year overall survival rate and recurrence rate. The secondary endpoints were operative time, intraoperative blood loss, morbidity of postoperative complication, incidence of anastomotic stenosis and incidence of reflux esophagitis. Considering the influence of tumor staging on postoperative clinicopathological features and prognosis, a subgroup analysis was performed on the literatures including cases of early gastric cancer and those including cases of tumor stage I to IV. Statistical analyses were carried out by the "metafor" and "meta" software packages from RevMan 5.3 software and R software (V3.2.4). Twenty-five literatures involving 3667 patients (proximal gastrectomy for 1483, total gastrectomy for 2184) were finally enrolled for analysis, including 24 retrospective studies with ≥ 5 points and 1 RCT with 3 points, and all the literatures were of high quality. A total of 2516 cases of early gastric cancer were enrolled in 18 articles, including 1027 with proximal gastrectomy and 1489 with total gastrectomy. A total of 1151 cases with stage I to IV were enrolled in 7 articles, including 456 in proximal gastrectomy group and 695 in total gastrectomy group. Five-year survival rate was not significantly different for patients with early gastric cancer between the proximal gastrectomy group and total gastrectomy group (=1.16, 95% : 0.72 to 1.86, =0.54). Similarly, there was no significant difference for patients with stage I to IV between the proximal gastrectomy group and the total gastrectomy group (=1.19, 95% : 0.92 to 1.53, =0.18). Recurrence rate of early gastric cancer patients was not significantly different between the proximal gastrectomy group and the total gastrectomy group (=0.40, 95% : 0.05 to 3.16, =0.39).However, the recurrence rate of the proximal gastrectomy group was higher than that of the total gastrectomy group in patients with stage I to IV (=1.55, 95% : 1.09 to 2.19, <0.01), whose difference was statistically significant. There was no significant differences in postoperative complication between the groups, both in patients with early gastric cancer, and in those with stage I to IV (both >0.05). The incidences of postoperative anastomotic stenosis (=3.57, 95% : 1.82 to 6.99, <0.01) and reflux esophagitis (=2.83, 95% : 1.23 to 6.54, =0.01) in the proximal gastrectomy group were significantly higher than those in the total gastrectomy group in patients with early gastric cancer. There is no significant difference in long-term survival outcomes between total gastrectomy and proximal gastrectomy for upper gastric tumors. However,incidence of anastomotic stenosis and reflux esophagitis, and tumor recurrence rate after total gastrectomy are significantly lower. The total gastrectomy is recommended as the first choice for advanced upper gastric tumor.