MDPI Children 2024 — CSWS Today: An Update

Source

• Children (MDPI) 11(2):169, February 2024. Open access.
• URL: https://www.mdpi.com/2227-9067/11/2/169

Why this paper is in the corpus

A recent, comprehensive update to the CSWS/ESES/DEE-SWAS treatment-and-outcome literature. It provides the modern framing of how behavioral and cognitive outcomes are expected to trajectory after each class of intervention (steroids, benzodiazepines, ASMs, surgery), and it anchors quantitative benchmarks that appear repeatedly in downstream reviews. Because this review cites Hempel 2019's ~59% pulse-prednisone language/behavior improvement figure and adds trajectory data from other methylprednisolone and topiramate cohorts, it is the single most useful review for interpreting "what should I expect 1-2 weeks after an apparently successful pulse?"

Key findings relevant to the recovery-trajectory question

• Core features of CSWS explicitly include behavioral disorders (attention deficit, hyperactivity, aggressiveness, difficulty in social interaction, occasionally psychotic-like or autistic-like presentations) alongside cognitive and language regression. These are not accidental co-features — they are part of the syndromic definition.
• Cognitive changes may reflect a lack of developmental progress, global loss of skills, or emerging specific deficits depending on the localization of ESES discharges. Younger age at onset is associated with worse cognitive and behavioral outcomes. This anchors the prediction that recovery should correspondingly show catch-up on a trajectory shaped by both the discharge topography and the child's chronological age at recovery.
• Behavioral change most often is characterized by hyperactivity and increased aggression and may persist even with remission of ESES. This is critical: the review explicitly warns that behavioral features can persist or re-emerge even when the EEG looks better, consistent with Hempel 2019's finding that behavioral/language response does not tightly track EEG response.

Quantitative response benchmarks

• Pulse methylprednisolone: global response rate 73% on seizures and 70% on EEG spike-wave index, with significant improvements in verbal performance, full IQ, learning and behavior. Recurrence at 1 year: ~29%.
• Hempel pulse-dose prednisone (cited): ~59% improvement in language or behavior in 17 pediatric ESES patients; improvement was not significantly related to ESES resolution on EEG or to history of regression; lower-IQ patients were more likely to improve.
• Topiramate: at 3 months, sleep EEG improved in 14/18 and normalized in 4; 16/18 showed cognitive or behavioral improvement. At 1 year, relapse occurred in nearly half.
• Perampanel: mostly retrospective evidence of behavioral/cognitive gains; specific quantitative effect sizes less robust than for pulse steroids.

Mixed-improvement trajectory

• The review acknowledges heterogeneity in the timing and pattern of recovery across domains. It does not identify a specific "worsening before improvement" U-shape as a stereotyped CSWS recovery phenomenon, but it repeatedly warns against equating EEG change with clinical change.
• Parent-rated outcomes are acknowledged as a real source of perceived benefit that may not match objective neuropsychological improvement, and vice versa. This matches caregiver accounts (e.g., the Mayo Clinic Connect parent who saw clean EEG with returning impulsivity during steroid weaning) and is consistent with a post-pulse window where behavior and EEG have partly dissociated.

Limitations relevant to Levi

• The review is a synthesis rather than a prospective cohort; it does not produce new trajectory data.
• Behavioral outcomes are grouped broadly and reported qualitatively; there are no standardized weekly-trajectory data the review could cite.
• Does not specifically address the question "do children typically show new negative behaviors alongside new positive developmental gains as DEE-SWAS suppresses?" — the review's description of recovery is higher-level and does not force a specific prediction on mixed-valence emergence.

Levi-relevant takeaways

• Confirms that the ~70% EEG/clinical response rate to pulse methylprednisolone, with a ~29% 1-year relapse rate, is the appropriate expected-value frame for Levi's own trajectory.
• Explicitly recognizes dissociation between EEG and behavior in both directions, supporting that Levi's clean-looking UCSF EEG plus mixed-valence behavior is a well-precedented pattern, not a novel puzzle.
• Supports the current post-pulse plan: watch the behavioral trajectory, expect some heterogeneity, do not over-interpret either a clean EEG or a bumpy behavioral window in isolation.

Cross-references in this corpus

• content/research/papers/2019-hempel-pulse-prednisone-eses-language-behavior.md — the primary ~59% figure source.
• content/research/papers/2017-kotagal-csws-treatment-review.md — the 575-case pooled treatment-efficacy reference.
• content/research/papers/2024-frontiers-dee-swas-high-dose-corticosteroid.md — relapse-prediction model (frontal discharges + age at seizure onset).
• content/research/papers/2019-gong-csws-hfo-relapse-biomarker.md — HFO-level relapse biomarker; supports that a clean SWI EEG can still predict relapse.
• content/research/notes/2026-04-18-steroid-pulse-rebound-behavior.md — companion memo for the post-pulse / GWS-like rebound hypothesis.
• content/research/notes/2026-04-19-eses-recovery-mixed-valence.md — synthesis memo tying this paper into the mixed-valence recovery framework for Levi.