Risk assessment using a novel scoring system (NUn score) to predict major complications after oesophageal resection

Noble, Fergus, Curtis, Nathan, Sreekumar, Rahul, Uduko, C., Chande, S., Kelly, Jamie, Bailey, Ian, Byrne, James and Underwood, Tim (2011) Risk assessment using a novel scoring system (NUn score) to predict major complications after oesophageal resection. British Journal of Surgery, 98 (S7), 7.

Abstract

Background: the aim of this study was to establish a numerical scoring system that categorises a patients’ risk of developing major post-operative complications after oesophageal resection based on routine blood tests.

Methods: a prospective database of all upperGI resectionswith an oesophageal anastomosis between 2005 and 2010 was reviewed. C-reactive protein (CRP), white cell count (WCC) and albumin were recorded once pre-operatively and post-operatively daily until discharge or day 14. All post-operative complications were recorded using the Clavien-Dindo (CD) classification. The diagnostic accuracy of CRP,WCCand albumin levels were analysed by receiver operating characteristic (ROC) curve analysis with anastomotic leak and major complication or death (CD 3–5) as outcome measures.

Results: a total of 258 patients were identified (Median age 67 (37–85) years, Male 78%, Female 22%). A minimally invasive procedure was performed in 101 (40%) cases. A total of 63 (25%) patients developed a major complication and there were 7 (2•7%) deaths. 27 (10•5%) patients were diagnosed with an anastomotic leak at median post-operative day (POD) 7 (Range: 5–15). On univariable analysis there were no pre-operative patient or tumour characteristics that could predict post-operative complications. CRP (p=0•08), WCC (p=0•08) and albumin (p=0•003) were independent predictors of a major complication or death at POD 5. After multivariable analysis these factors were combined to create a novel scoring system (NUn score). On POD 4 the NUn score was highly predictive of an anastomotic leak (NUn score >0•97: sensitivity 100%, specificity 66%, diagnostic accuracy 0•78 (95%CI 0•655–0•905, p<0•0001)) and a major complication or death (NUn >0•9: sensitivity 73%, specificity 79%, diagnostic accuracy 0•71 (95%CI 0•609–0•81, p<0•0001)).

Conclusion: we describe the development of a novel (NUn) scoring system that accurately categorises patients at risk of anastomotic leak and major complications following oesophageal resection. In this cohort the NUn score accurately identified patients at risk of anastomotic leak 3 days prior to diagnosis. The sensitivity of the POD4NUnscore for anastomotic leak is such that patients with a score <0•97 can be confidently fast-tracked to enteral feeding and early discharge

This record has no associated files available for download.

Contributors

Author: Tim Underwood

Download statistics