Predicting conversion to open surgery in laparoscopic colorectal resections. A simple clinical model.

OBJECTIVE

The objective of this study was to develop a simple model for clinical use in predicting the individual risk of conversion to open surgery in patients undergoing laparoscopic colorectal resections.

METHODS

A multiple logistic regression analysis of 367 laparoscopic colorectal resections completed between 1991 and 1998 was performed. The following 13 factors were considered: patient-specific factors (age, gender, weight levels less than 60 kg 60-90 kg, 90 kg or more), disease-specific factors (Crohn's disease, diverticulitis, malignancy, fistula), and procedure-specific factors (resection of the hepatic flexure, splenic flexure, sigmoid, rectum, perineum, experience with less than 50 cases). A scoring system was developed on the basis of the three factors found to be predictive of the risk for conversion to open surgery: diagnosis of malignancy (odds ratio 3.23; p = 0.0037; one point), surgeon experience with 50 cases or less (odds ratio 2.26; p = 0.0363; one point), and weight level (odds ratio 3.42; p = 0.005; 60 to 90 kg, one point, 90 kg or more, two points).

RESULTS

The predicted conversion rates for the cumulative scores of 0 to 4 points were 1.1%, 3.3%, 9.8%, 25.4%, and 49.7%, respectively. No significant difference was found between predicted and actual conversion rates, indicating a good fit of the model (chi square = 1.774; p > 0.5).

CONCLUSIONS

This novel scoring system is simple, accurate, and readily applicable in an office setting. It represents the large experience of one surgical group and remains to be validated by other centers.