Unreasonable expectations in emergency colorectal cancer surgery

The recently published ACPGBI colorectal cancer (ACPGBI CRC) scoring system for predicting operative mortality has been suggested as an instrument to improve patient consent procedures and to compare results between centres. This study compares the results of a surgical unit against the standards set by the ACPGBI colorectal cancer model and for emergency surgery, against the p-POSSUM instrument.

Data for the ACPGBI CRC model were collected prospectively through 2003 at the Leicester Royal Infirmary. Additional data needed for the p-POSSUM was retrospectively collected from case records. The actual mortality was compared with that predicted by the models.

Seventy-two colorectal cancer operations were performed during the study period. The observed operative mortality in elective cases was lower, and in emergency cases higher, than predicted by the ACPGBI CRC model. With emergency cases the predicted mortality using P-POSSUM was significantly higher than that using the ACPGBI CRC model, particularly in the presence of faecal contamination.

The ACPGBI CRC model may be accurate for elective cases, but appears to significantly underestimate predicted mortality in the emergency setting, both actual and predicted by p-POSSUM. This may be due to a failure to incorporate adequate weighting for faecal peritonitis and the associated systemic insult into the ACPGBI model.