[Radiological index as a predictive factor of kyphosis in spina bifida]

The aim of this work was to validate the pertinence of a radiological index (P/A) used as a predictive factor for risk of kyphosis in children with spina bifida. P/A is defined as the ratio between P, the distance between the posterior vertebral wall and the remnant of the posterior arch, over A, the distance between the anterior wall and the posterior vertebral wall, measured on the lateral view of the spine in the most dysplastic vertebra before one year of age.

The files of 163 children were reviewed. The value of the P/A ratio was measured on lateral views of the spine obtained before the age of one year. The bone level on the anteroposterior film was also noted. The presence or absence of kyphosis was determined from the lateral views obtained at ten years of age. Paired variable analysis was used to compare the mean differences in P/A at one year and at ten years in 31 patients with a bone level of L1 or higher. The exact Fisher test was used to determine the difference in the kyphosis distributions in patients with a bone level of L1 or higher between those with P/A < or = 1/2 and the others. In addition, the specificity, sensitivity, and positive predictive value of a positive test (bone level L1 or higher and P/A < or = 1/2) for development of kyphosis were calculated. Results There was no statistically significant difference in the P/A mean value before the age of one year and after ten years. The distribution of patients with kyphosis was different between patients with a bone level of L1 or higher and P/A < /2 and the other patients (p < 0.01). Considering having a bone level of L1 or higher and a P/A < or = 1/2 as a prognostic test, the specificity was 100%, the sensitivity 87.56%, and the positive predictive value 100%. Many patients with spina bifida will develop kyphosis in the course of life. This spinal deformation raises many specific problems for patient management. Use of a radiological index as presented here allows early detection of patients at risk and can also predict which patients will remain free of deformation. This allows a selection of patients who can benefit most from more regular surveillance of the spinal static and avoid unnecessary radiographic examinations in others.