Comparison of the post-operative outcomes of surgical and chemical interventions on the gastric pylorus during oesophagectomy: A systematic review and meta-analysis

Abstract Background Oesophagectomy can lead to gastric dysmotility and pyloric dysfunction, causing delayed gastric emptying (DGE). DGE is associated with significant postoperative morbidity, namely aspiration pneumonia and anastomotic leak. Surgical drainage and botox injections (BI) on pylorus were described to combat DGE. Pyloric interventions however can increase risk of bile reflux, oesophagitis, dumping syndrome and pyloric leaks. Current literature showed equivocal short- and long-term outcomes of pyloric drainage procedures, with no consensus on the necessity of such procedures. We aim to evaluate the latest literature, comparing effects of different pylorus surgical drainage methods, BI and watch-and-wait approach on the pylorus during oesophagectomy Methods The OVID database was searched systematically. Articles were screened with the Rayyan software. Inclusion Criteria of articles are as follows: • RCT and comparative cohort studies (retrospective and prospective) which included any drainage methods, performed at the same time as oesophagectomy • Oesophagectomy for benign or malignant conditions • Studies of adult human subjects ≥ age of 18 years old • Studies published in English language • Studies published within the last 25 years Primary outcomes evaluated were DGE and pulmonary complication rates. A traditional pairwise meta-analysis was performed with Revman 5.4, utilizing a random effects model for analysis of the pooled outcome measures. Results Data of 3925 patients from 27 publications evaluating up to ten pyloric interventions were analyzed. Most studies were retrospective in nature (24 studies). Due to heterogeneity of reported outcomes, pooled meta-analyses were only performed on limited studies. Pooled analyses showed pyloric interventions reduced DGE rates in 19 studies and pulmonary complications in 11 studies. Relative risks (95% CI; p value) were 0.95 (0.7,1.29; p=0.74) and 0.88 (0.72,1.08; p=0.22), respectively.Figure 1:Forrest plot comparing DGE rates of Pyloric Interventions vs No Intervention Figure 2:Forrest plot comparing rates of Pulmonary Morbidity of Pyloric Interventions vs No Intervention Conclusions Our systematic review and meta-analysis did not observe statistically significant and conclusive evidence favouring pyloric drainage procedures to none. High quality comparative studies and inconsistent scrutinization of post operative outcomes were lacking. Prospectively designed RCT or propensity matched cohort studies comparing surgical pyloric drainage, BI and no intervention should be conducted to produce standardized comparison. We suggest analyzing post operative outcomes by measuring data which could be objectively recorded (DGE utilizing the latest definition, pulmonary morbidity, anastomotic leak, biliary reflux rates, radiographic or endoscopic findings etc.) combined with subjective measurement of patient symptoms, where a patient-centred questionnaire could be performed.