The Surgeon as a Prognostic Factor; A Systematic Review and Meta-analysis of Outcomes after Pancreatic Surgery

Abstract

Purpose: Centralising pancreatic surgery to fewer centres aims to develop and concentrate expertise thereby improving patient outcomes. The causal relationship between hospital volume and the outcomes from pancreatic surgery however, is complex and likely to be the result of a culmination of factors, including the ability of the surgeon. It was hypothesised that high volume surgeons (HVS) would therefore have better patient outcomes than low volume surgeons (LVS). Methods: A systematic literature search of published literature was conducted to identify studies that compared the outcomes after pancreatic surgery performed by HVS and LVS. A meta-analysis of studies meeting the inclusion criteria was performed. Results: The search strategy identified 1707 studies, 18 of which fulfilled the inclusion criteria and contributed 66,695 patients. Studies originated from 6 countries in total, with most (13/18) located in the USA. Most studies specified that the procedures included were pancreaticoduodenectomies, however 2 studies included left pancreatectomies and 6 did not specify the procedure other than as a ‘pancreatic resection’ The definition of HVS ranged from ≥4 to ≥55 procedures per annum. Sixteen studies calculated a post-operative mortality rate with. LVS mortality rates ranged from 1.1% to 14.7%, and for HVS from 0.3% to 4.8%. The pooled effect of pancreatic surgeon volume on inpatient mortality favoured HVS (OR 0.41, 95% CI 0.33-0.51, P<0.01). Due to the variation in SV cut-off criteria, a subgroup analysis of inpatient mortality was performed grouping studies annual SV; <10, 10-20, >20 resections/year. Studies that used lower end cut-offs for HVS (<10 per year) showed significantly lower inpatient mortality rates, OR 0.42 (95% CI 0.32,0.55, p < 0.001). HVS was also associated with a lower length of stay OR 0.75 (95% CI 0.55,1.02, p = 0.07) and lower hospital costs OR 0.95 (95% CI 0.92,0.99, p = 0.008). No difference was identified in blood transfusion requirements, OR 1.09 (0.40, 3.01, p = 0.87). Conclusion: Surgeons who perform a high volume of pancreatic surgeries have a lower risk of post-operative mortality and re-admissions with lower overall hospital costs. The lower operative mortality rate was most significant in surgeons with the highest annual volume. This analysis further supports the centralisation of pancreatic cancer surgery to specialised, units with high volume surgeons.