Study Selection
A total of 2151 related clinical studies were initially searched. Among them, 2052 articles were excluded because of duplication or the title/abstract did not meet the inclusion criteria. After reviewing the full text of the remaining 99 articles, 4 non-original researches were excluded, and 47 were excluded because of lack of comparison between GJ and ES. In addition, 14 articles with invalid data to be extracted were excluded, and three articles were excluded due to duplicate reports. Finally, 31 studies [5, 9, 14-22, 30-49] were included according to inclusion and exclusion criteria, including 3 RCTs and 28 observational studies (Fig. 1). Details regarding the quality assessment of the included studies are presented in Supplementary Table 1.
Study Characteristics
The basic characteristics of the included studies are shown in Table 1. The 31 studies that met the inclusion criteria had 2444 GOO patients in GJ (1076) and ES (1368) groups. Most patients are malignant GOO. All included studies were conducted from 2002 to 2019, with the smallest sample size of 18 and the largest sample size of 347. In the included studies, the main etiology of malignant GOO is gastric cancer (n=1241) and pancreatic cancer (n=753), and others included biliary cancer, ampullary carcinoma, and metastatic cancer. Gastric cancer was the main etiology of malignancy in 13 studies, and pancreatic cancer was the main etiology of malignancy in 12 studies.
Surgery-related procedure outcomes
A total of 14 studies reported technical success. Compared with the ES group, the GJ group has a higher rate of technical success (OR, 3.79, 95%CI, 1.58-9.09, P=0.003) (Fig. 2a). A total of 17 studies reported clinical success, without a significant difference in the clinical success rate between GJ and ES groups (OR, 1.25, 95%CI, 0.65-2.38, P=0.50) (Fig. 2b). Six studies [14, 18, 21, 30, 34, 37] reported the hospital stay, and pooled results demonstrated that the hospitalization of ES group was shorter than that of GJ group (WMD, 7.34, 95%CI, 2.95-11.73, P=0.001) (Fig. 2c).
Complications
A total of 23 out of 31 included studies provided data on overall complications, and no significant difference was found (OR, 1.12, 95%CI, 0.68-1.86, P=0.65) (Fig. 3a). Nine studies reported minor complications, without significant difference between GJ and ES groups (OR, 1.38, 95%CI, 0.82-2.33, P=0.22) (Fig. 3b). Major complications were reported in 10 studies, however, no significant differences in the rates of major complications were observed between the two groups (OR, 0.57, 95%CI, 0.25-1.27, P=0.17) (Fig. 3c).
Short-term outcomes
Eight studies reported 30-day mortality in detail, without significant difference between the two groups (OR, 1.14, 95%CI, 0.43-3.01, P=0.17) (Fig. 4a). Nine studies compared the rate of postoperative chemotherapy between the two groups, and no significant difference was found (OR, 1.04, 95%CI, 0.63-1.70, P=0.89) (Fig. 4b). Six studies compared the re-obstruction rate between the two groups, and the results demonstrated that the GJ group had a significantly lower re-obstruction rate (OR, 0.41, 95%CI, 0.22-0.78, P=0.006) (Fig. 4c). Furthermore, the rate of reintervention was reported in 12 studies and was found to be significantly higher in the ES group than that in the GJ group (OR, 0.30, 95%CI, 0.22-0.41, P<0.001) (Fig. 4d).
In addition, since the leading malignant causes were divided into gastric cancer and pancreatic cancer, we conducted a subgroup analysis of the above indicators according to the malignant cause. The subgroup analysis results were similar to the overall analysis, and the detailed results are presented in Supplementary Table 2.
Survival
In this meta-analysis, seven studies [15, 17, 35, 37, 39, 45, 47] measured the overall survival of 427 patients in the GJ group and 601 patients in the ES group. The overall survival of GJ group was significantly better compared with that of the ES group (HR, 0.41, 95%CI, 0.24-0.72, P<0.001) (Supplementary Fig. 1), however, the heterogeneity of these studies could not be ignored (I2=82.1%), and the random-effect model was used. To further investigate the overall survival for malignant GOO patients, a subgroup analysis was performed according to etiology of malignancy (gastric cancer or pancreatic cancer). The results indicated that the overall survival of GJ was better than ES in gastric cancer group (HR, 0.33, 95%CI, 0.15-0.76, P=0.009). However, no significant statistical difference between GJ and ES was observed in the pancreatic cancer group (HR, 0.55, 95%CI, 0.24-1.26, P=0.159) (Fig. 5).