Predictive model of heterotopy in Meckel's diverticulum in children.

Abstract

The presence of heterotopic tissue in Meckel's diverticulum (MD) increases the incidence of future complications. Based on this concept, this study aimed to detect clinical and morphological parameters able to predict the probability of heterotopy in incidentally discovered uncomplicated MD and to develop a risk-based classification tool. The clinical records of 146 patients who underwent MD resection were reviewed. MD complications (bleeding, inflammation, perforation or involvement in pathological processes such as intussusception) and MD dimensions (width, length and the product width × length) were evaluated in association with heterotopy. Regression analysis was used to explore the discrete effect of clinically relevant parameters, which were subsequently combined in a Classification and Regression Tree. The probability of heterotopy was higher in complicated MDs. The width of the base of the MD and the product width × length were also significant predictors of heterotopy. A stepwise combination (Classification and Regression Tree model) of (i) uncomplicated macroscopic presentation, (ii) width <2.5 cm, (iii) age <10 years and (iv) product width × length <4 cm2 resulted in a null probability of heterotopy in our study group. The combination of MD-related clinical and morphological parameters using advanced statistical tools may assist in estimating the probability of heterotopy. This may be applied in incidentally discovered uncomplicated MDs during abdominal surgery for other morbidities. Prospective studies are required to confirm our findings and test the practical usefulness of the classification tool developed here.