A Scoring System to Predict the Severity of Hirschsprung Disease at Diagnosis and its Correlation with Molecular Genetics.

Objectives: Hirschsprung disease (HSCR) has a wide range of severity. There are nonsevere forms treated conservatively until surgery and severe forms that require an early stoma and prolonged hospitalization. Our objective was to establish a clinical scoring system to predict the severity of HSCR and to evaluate the possible existence of a clinical-genetic correlation. Methods: We carried out a retrospective observational study including all HSCR cases treated in our hospital. The sample was divided into severe and nonsevere disease according to the number of surgical procedures, hospitalization time, and episodes of enterocolitis. The proposed score was applied at diagnosis, and the sensitivity, specificity, and optimal cut-point were determined. We conducted a prospective molecular study of RET, EDNRB, and EDN3 on all patients, as well as SOX10 in Waardenburg Syndrome type 4 forms. Results: Among the 42 patients treated between 1983 and 2013, 15 met the severe disease criteria. This group had a higher mean score (13.15 +/- 2.36) than the nonsevere group (8.15 +/- 2.13; p < 0.001). A score >= 11 had a sensitivity of 87% and a specificity of 81% in detecting the severe cases. Causative mutations were identified in 12 patients, 8 of them in the severe group (p = 0.015). Most of these mutations (75%) were located in the RET proto-oncogene. Conclusion: The proposed scoring system enables the early selection of patients with severe behavior of HSCR. A value >= 11 showed good sensitivity and specificity for this purpose. Causative mutations were identified in more than 50% of patients who met the criteria for severe disease.