P14 Evaluation of clinical prognostic variables on short-term outcomes for colorectal cancer surgery: an overview and minimum dataset

Abstract Introduction Colorectal cancer resection remains a cause for morbidity and mortality despite the potential curative intent. Multiple systematic reviews have reported on individual factors affecting short term outcome following surgical resection. This umbrella review aims to synthesize the available evidence on host and other factors associated with short-term post-operative complications. Methods A systematic search identified systematic reviews reporting on short-term outcomes following colorectal cancer surgery. PubMed, Cochrane Database of Systematic Reviews and Web of Science were search from inception to 8 September 2020. All reported clinicopathological variables were extracted from included systematic reviews. Results This overview has identified multiple validated factors affecting short-term outcomes in patients undergoing colorectal cancer resection. There are multiple prognostic factors that affect the short term outcome following colorectal cancer resection although the association differs dependent on the type of complication; infective; non-infective or mortality. The presence of diabetes and measurements of body composition were both factors consistently associated across the short term outcome measures. Despite the large body of literature available, the heterogeneous nature of included prognostic factors and reported outcomes create a challenge for comparing outcomes. Conclusion A minimum dataset of mandatory, recommended, and potential baseline variables to be included in studies of patients undergoing colorectal cancer resection is proposed. This will maximize the benefit of such study databases. Take-home message The reported effect of prognostic risk factors associated with colorectal cancer resection vary dependent on outcome of interest. This highlights the shortcomings of the current literature and understand of outcomes in colorectal cancer which could be improved with a minimum dataset of baseline variables in future studies.