Extended versus limited lymph nodes dissection technique for adenocarcinoma of the stomach.

backgroundSurgeons disagree about the merits and risks of radical lymph node clearance during gastrectomy for cancer.

objectivesTo evaluate survival and peri-operative mortality after limited or extended lymph node removal during gastrectomy for cancer. SEARCH STRATEGY: We searched MEDLINE, EMBASE, CancerLit, LILACS, Central Medical Journal Japanese Database and the Cochrane register, references from relevant articles and conference proceedings. We contacted known workers in the field. SELECTION CRITERIA: Studies published after 1970 which reported 5 year survival or postoperative mortality rates, and clearly defined the node dissection performed, were considered. We excluded studies which overtly included patients receiving perioperative chemotherapy, and comparisons with clear systematic treatment allocation bias. Randomised controlled trials (RCTs), non-randomised comparisons and observational studies were considered separately. DATA COLLECTION AND ANALYSIS: Three reviewers selected trials for inclusion. Quality assessment and data extraction were performed independently by two reviewers. Results of trials of similar design were pooled. Meta-analysis was performed separately for randomised and non-randomised comparisons. MAIN

resultsTwo randomised and two non-randomised comparisons of limited (D1) versus extended (D2) node dissection and 11 cohort studies of either D1 or D2 resection were analysed. Meta-analysis of randomised trials did not reveal any survival benefit for extended lymph node dissection (Risk ratio = 0.95 (95% CI 0.83 - 1.09), but showed increased postoperative mortality (RR 2.23, 95% CI 1.45 - 3.45). Pre-specified subgroup analysis suggested a possible benefit in stage T3+ tumours (RR = 0.68, 95% CI 0.42-1.10). Non-randomised comparisons showed no significant survival benefit for extended dissection (RR 0.92, 95% CI 0.83 -1.02), but decreased mortality (RR 0.65, 95% CI 0.45-0.93). Subgroup analysis showed apparent benefit in UICC stage II and IIIa. Observational studies of D2 resection reported much better mortality and survival than those of D1 surgery, but the settings were strikingly different. REVIEWERS'

conclusionsD2 dissection carries increased mortality risks associated with spleen and pancreas resection, and probably with inexperience and low case volumes. Randomised studies show no evidence of overall survival benefit, but possible benefit in T3+ tumours. These results may be confounded by surgical learning curves and poor surgeon compliance. Non-randomised comparisons suggest a possible survival benefit for D2 in intermediate UICC stages. Observational studies show high 5 year survival and low operative mortality after D2 dissection in experienced units, and poor results after D1 dissection in non-specialist units. Further studies, with precautions to eliminate learning curve effects, contamination and non-compliance, are needed to evaluate D2 dissection in intermediate stage gastric cancer.