Prognostic factors for mortality of oesophageal atresia patients: Waterston revived.

Abstract

This study shows that the effect of medical progress on mortality is considerable and that the three factors of the Waterston classification still had a high impact on mortality in the most recent years. In 1962, Waterston et al. [5] proposed a prognostic classification for patients born with oesophageal atresia (OA) based on low birth weight, presence of pneumonia, and associated congenital malformations. Over the last decades, much progress has been made and several authors have questioned the validity of Waterston’s classification [1, 3, 4]. This study analyses the influence of several clinical characteristics on mortality of OA patients, and investigates the value of the Waterston classification within the past and current time frame. We studied all patients who were treated at the Paediatric Surgical Centre of Amsterdam for OA and/or tracheo-oesophageal fistula between 1947 and 2000 (n=371) [2]. Data were collected from patient charts, operative reports, and office notes, and were entered into a database. Medical progress was modelled by dividing the time interval arbitrarily into five time periods with approximately equal numbers of patients. Disease specific mortality was defined as mortality during the 1st year of life. To determine the independent impact of factors on disease specific mortality over the total time period (1947–2000), a multivariate logistic regression model was performed. Effect sizes were expressed as odds ratios (OR’s) and their 95% confidence limits; an OR >1 refers to an increased risk for mortality. To study the five separate time periods (and specifically the most recent time period), separate models were made for each of these time periods. Considering the total time period, the type of OA, associated malformations, low birth weight and medical progress modelled by the five time periods, all had an independent impact on mortality (Fig. 1). The OR depicting the medical progress were not significantly different for the two most recent time periods, suggesting a similar improvement between 1994 and 2000, compared to the reference time period. In the most recent time period (1994–2000), ‘birth weight <1500 g’ had a strong impact on mortality (OR 54.7). Factors with a less strong, but still clinically significant impact on mortality were ‘other malformations’ and ‘pre-operative