The role of C-reactive protein in predicting anastomotic leaks

One of the dilemmas in the employment of fast track protocols is the fear of discharging patient’s home “too soon” before postoperative complications manifest themselves. One of the most dreaded and challenging complications is anastomotic leak. While anastomotic leak may be diagnosed by clinical, radiographic, or endoscopic means many authors have suggested that C-reactive protein may be predictive of leak. To test the hypothesis of C-reactive protein on the third postoperative day, Pantel and colleagues at the Lahey Clinic evaluated 752. Their aim was to correlate the predictive value of C-reactive protein on the third postoperative day relative to leak or readmission within 30 days. The authors found that 73 of their 752 patients were readmitted within 30 days of surgery, 17 of whom had an anastomotic leak. Furthermore they found that the mean C-reactive protein levels in patients who had a leak was 228 mg/l, significantly higher than was the level in patients who were neither readmitted (127 mg/l), or readmitted without a leak (157 mg/l) (p=0.0000002). Using a receiver operating characteristic curve, the authors determined that employing a value of 145 mg/l a 93% predictive value was obtained. Moreover, using a cutoff value of 147 mg/l a 99% negative predictive value was obtained. The authors are to be congratulated for having shown this important correlation between a C-reactive protein level < 145 mg/l on the third postoperative day after colorectal resection as a correlation with several correlations. These correlations included a “low likelihood of readmission within 30 days and a very low likelihood of anastomotic leak”. It will be interesting to see whether this outstanding study prompts more surgeons in North America to routinely obtain C-reactive protein on the third postoperative day.