Early prediction of adverse events in enhanced recovery based upon the host systemic inflammatory response.

AIM: Early identification of patients experiencing postoperative complications is imperative for successful management. C-reactive protein (CRP) is a nonspecific marker of inflammation used in many specialties to monitor patient condition. The role of CRP measurement early in the elective postoperative colorectal patient is unclear, particularly in the context of enhanced recovery (ERAS). METHODS: Five hundred and thirty-three consecutive patients who underwent elective colorectal surgery between October 2008 and October 2010 within an established ERAS programme were studied. Patients were separated into a development group of 265 patients and a validation group of 268 patients by chronological order. CRP and white cell count were added to a prospectively maintained ERAS database. The primary outcome of the study was all adverse events (including infective complications, postoperative organ dysfunction and prolonged length of stay) during the initial hospital admission. Significant predictors for adverse events on univariate analysis were submitted to multivariate regression analysis and the resulting model applied to the validation group. The validity and predictive accuracy of the regression model was assessed using receiver operating characteristic curve/area under the curve (AUC) analysis. RESULTS: CRP levels >150 mg/l on postoperative day 2 and a rising CRP on day 3 were independently associated with all adverse events during the hospital admission. A weighted model was applied to the validation group yielding an AUC of 0.65 (95% CI 0.58-0.73) indicating, at best, modest discrimination and predictive accuracy for adverse events. CONCLUSION: Measurement of CRP in patients after elective colorectal surgery in the first few days after surgery within ERAS can assist in identifying those at risk of adverse events and a prolonged hospital stay. A CRP value of >150 mg/l on day 2 and a rising CRP on day 3 should alert the surgeon to an increased likelihood of such events.