A new prognostic classification for esophageal atresia.

BACKGROUND

Waterston's risk categories have been extensively used in the past for the classification and management of neonates with esophageal atresia. Advances in neonatal care have affected the prognostic usefulness of the Waterston classification. The purpose of this study is to formulate a new classification of risk factors that would more accurately predict outcome.

METHODS

We retrospectively reviewed 95 consecutive cases of esophageal atresia and/or tracheoesophageal fistula treated at the Montreal Children's Hospital. Fifteen patients were in the Waterston class A, with a 6.7% mortality; 35 patients, in class B with a 5.7% mortality; and 45 patients, in class C with a 26.7% mortality. Logistic regression analysis of the influence of each risk factor (weight, pulmonary status, and severity of associated anomalies) on mortality was performed.

RESULTS

Birth weight was not found to independently influence mortality. Only severe pulmonary dysfunction with preoperative ventilator dependence and severe associated anomalies had a prognostic influence. We therefore revised the classification to include only significant factors. The new high-risk class II consists of patients with either life-threatening anomalies or both major anomalies and ventilator dependence; the low-risk class I includes all other patients. These criteria radically changed the stratification of both the number of cases and the mortality among classes: the 82 patients in class I had a 7.3% mortality; the 13 patients in class II had a 69.2% mortality. Logistic regression analysis confirmed the validity of this new classification by showing displacement of individual variables by the revised classification but not by Waterston's.

CONCLUSIONS

This new "Montreal classification" can simplify and improve the stratification of patients with esophageal atresia. It also reflects the more favorable outcome of low birth weight neonates.