Prediction model for long-term seizure and developmental outcomes among children with infantile epileptic spasms syndrome

Yuto Arai, Tohru Okanishi, Sotaro Kanai, Yuko Nakamura, Hiroshi Sunada, Hinata Murakami, Kazuki Yamauchi, Hisashi Noma, Yoshihiro MaegakiFrontiers in Neurology 2023 · DOI

Abstract

Our prediction model may be useful for predicting the long-term prognosis of seizures and their development after 3 years. Understanding the long-term prognosis during the initial treatment may facilitate the selection of appropriate treatment.

Introduction

Infantile Epileptic Spasms Syndrome (IESS), historically known as West Syndrome, is a severe form of childhood epilepsy characterized by clusters of spasms and a specific "chaotic" brainwave pattern called hypsarrhythmia. While treatments like ACTH and vigabatrin are standard, the long-term prognosis remains challenging: over half of these children experience significant developmental delays, and two-thirds suffer from long-term seizures. Early identification of patients at high risk for poor outcomes is critical for optimizing treatment strategies, such as considering early epilepsy surgery or ketogenic diets.

Methods

Researchers conducted a retrospective cohort study of 44 patients with IESS treated at Tottori University Hospital between 1996 and 2022. The study aimed to develop a simple clinical scoring system to predict outcomes at 3 and 7 years post-onset. Three primary risk factors were identified and assigned one point each:

  1. Persistence of spasms/tonic seizures: Continuing more than 90 days after onset.
  2. Developmental delay before onset: Defined as a Developmental Quotient (DQ) < 70.
  3. MRI abnormalities: Presence of moderate or severe abnormalities on neuroimaging.

Results

The scoring model demonstrated strong predictive power, particularly for 3-year outcomes. For seizure prognosis, a cutoff score of 2 yielded a sensitivity of 86.7% and a specificity of 64.3% (AUC = 0.825). For developmental prognosis, the model was even more robust, achieving 88.9% sensitivity and 100% specificity (AUC = 0.974) for predicting severe delay (IQ/DQ < 35).

Figure 1: Prediction model for seizure and developmental prognosis after 3 years using clinical data. The table outlines the scoring system, while ROC curves illustrate the model's performance for seizure (left) and developmental (right) outcomes.

At the 7-year follow-up, pre-existing developmental delay remained the most significant predictor of long-term cognitive outcomes, suggesting that the underlying etiology plays a dominant role in cognitive trajectory. Notably, every patient who had a poor outcome at 7 years had already been identified as having a poor outcome at the 3-year mark, suggesting the 3-year milestone is a reliable indicator of permanent trajectory.

Discussion and Conclusion

The study highlights a "golden window" for treatment; failure to control spasms within 90 days is a major red flag for long-term seizure persistence. Furthermore, the high specificity of the developmental model at 3 years suggests it can reliably identify children who will face severe long-term challenges. While limited by a small, retrospective sample from a single center, this clinical scoring system provides a practical tool for early prognostic counseling and treatment planning in IESS. Future multi-center studies are needed to validate these findings across broader populations.

Summary updated

Abstract

Our prediction model may be useful for predicting the long-term prognosis of seizures and their development after 3 years. Understanding the long-term prognosis during the initial treatment may facilitate the selection of appropriate treatment.
BibTeX
@inproceedings{arai2023prediction,
  author = {Yuto Arai and Tohru Okanishi and Sotaro Kanai and Yuko Nakamura and Hiroshi Sunada and Hinata Murakami and Kazuki Yamauchi and Hisashi Noma and Yoshihiro Maegaki},
  title = {{Prediction model for long-term seizure and developmental outcomes among children with infantile epileptic spasms syndrome}},
  booktitle = {Frontiers in Neurology},
  year = {2023},
  doi = {10.3389/fneur.2023.1195252},
}
Figure