PWE-320 CRP can predict anastomotic leak in rectal cancer resection

Introduction

Post-operative colorectal cancer resections are high risk due to associated morbidity and mortality (30% and 6.7%, respectively). Anastomosis leak has been associated to a negative impact upon survival, independent of tumour staging. C-Reactive Protein (CRP; 125–190 mg/L) has a positive predictive value in predicting anastomotic leak in open colorectal resections.

Aims

To study the predictive value of CRP for anastomotic leak in all rectal resections. Secondly, to review if CRP values vary with laparoscopic and open rectal resections.

Method

Consecutive rectal cancer resection patients with an anastomosis performed at 2 high volume units for the period 2010–2014 were included. Clinical data was collected including demographics, CRP results for day 1 to day 7 and complications. Anastomotic leak was confirmed via CT imaging. Outcomes were statistically assessed with receiver operating curves (ROC) analysis.

Results

There were a total of 379 (131 Open; 248 laparoscopic) patients included, with a total of 49 (12.9%) anastomotic leaks: 26 in the open group and 23 in laparoscopic group. In laparoscopic resections, ROC analysis suggests day 3 CRP of 157 (ROC area: 0.713, sensitivity: 0.700, specificity: 0.753) is predictive of anastomotic leak. Comparatively in open resections, ROC analysis suggests day 3 CRP of 207 (ROC area: 0.765, sensitivity: 0.765, specificity: 0.700) is predictive.

Conclusion

CRP, in both laparoscopic and open rectal resections, has positive predictive value for an anastomotic leak; best seen at day 3. Laparoscopic resections have a much lower predictive threshold than open resections in the observed period.

Disclosure of interest

None Declared.