Highlights
- People who later died of SUDEP exhibit a pathological lack of overnight decline in slow-wave activity (SWA), suggesting impaired sleep homeostasis.
- Increased variability in inter-breath intervals during non-rapid eye movement (NREM) sleep serves as a potent differentiator for SUDEP risk.
- The coefficient of variation of the inter-breath interval demonstrated high discriminative performance (AUC 0.80) in predicting SUDEP.
- Gender-specific differences were observed, with males showing a more pronounced increase in SWA slope compared to females.
The Silent Threat: Understanding SUDEP and the Need for Biomarkers
Sudden Unexpected Death in Epilepsy (SUDEP) remains the most devastating complication of chronic epilepsy, accounting for the majority of epilepsy-related mortality. Despite its clinical significance, the pathophysiology of SUDEP is complex and multifactorial, often involving a terminal cascade of seizure-induced respiratory arrest, cardiac arrhythmia, and cerebral autonomic failure. A critical observation in SUDEP research is its temporal predilection: the vast majority of cases occur during nighttime sleep, often following a generalized tonic-clonic seizure (GTCS).
Current clinical risk stratification relies heavily on the frequency of GTCS, which is the strongest known clinical risk factor. However, this metric is often subjective or reliant on patient self-reporting, which is notoriously unreliable in epilepsy. There is an urgent unmet medical need for objective, physiological biomarkers that can identify high-risk individuals before a fatal event occurs. This multicenter case-control study explores the intersection of sleep macroarchitecture, microarchitecture, and respiratory dynamics to uncover these elusive markers.
Study Design and Methodology: A Multicenter Comparative Approach
Researchers conducted a rigorous case-control study using prospectively collected data from a National Institutes of Health (NIH) multicenter study. The study population included 41 participants who later died of SUDEP (the SUDEP group) and 123 matched controls. These controls were categorized into three distinct cohorts: 41 individuals at high risk (≥1 GTCS/year), 41 at low risk (no history of GTCS), and 41 non-epilepsy controls. This design allowed for a nuanced comparison across the spectrum of seizure severity and mortality risk.
The primary focus was on seizure-free nights to identify interictal (between-seizure) biomarkers. The team analyzed sleep macroarchitecture (sleep stages) and microarchitecture, specifically focusing on sleep homeostasis. This was measured by calculating the overnight change in slow-wave activity (SWA; 0.5–4.0 Hz) during NREM sleep. Additionally, respiratory metrics were scrutinized, with a particular emphasis on the regularity of breathing patterns, quantified as the inter-breath interval variability. To evaluate the clinical utility of these markers, receiver operating characteristic (ROC) analysis was employed to determine their discriminative performance.
Key Findings: Disrupted Sleep Homeostasis
The study revealed a significant departure from normal sleep physiology in the SUDEP group. In healthy individuals and low-risk epilepsy patients, SWA typically follows a declining trajectory throughout the night, reflecting the dissipation of sleep pressure—a process fundamental to synaptic scaling and neural recovery. However, the SUDEP group showed an abnormal lack of this overnight decline. Instead, there was an increase in the slope of SWA power (SUDEP group mean 0.005 vs. non-epilepsy controls -0.007; p=0.017).
Interestingly, this impairment in sleep homeostasis exhibited sexual dimorphism. The increase in the SWA slope was significantly more pronounced in males than in females (p=0.005). Given that SUDEP often shows a slightly higher prevalence in males, this finding suggests that male patients may experience a more profound disruption in the neurophysiological mechanisms governing sleep-wake regulation, potentially contributing to their vulnerability.
Respiratory Biomarkers: The Predictive Power of Inter-breath Variability
While EEG markers provided insight into cortical health, respiratory data yielded even stronger predictive value. The variability of the inter-breath interval during NREM sleep was significantly higher in both the SUDEP group and the high-risk group compared to low-risk patients and non-epilepsy controls (p<0.0001).
The coefficient of variation (CV) of the inter-breath interval emerged as the most robust individual predictor. With an Area Under the Curve (AUC) of 0.80 (95% CI 0.70–0.90), this metric outperformed standard clinical variables. The higher CV indicates a fragmented or unstable respiratory rhythm during sleep, which may reflect underlying brainstem dysfunction. In the context of SUDEP, an unstable respiratory drive interictally may predispose a patient to terminal apnea when challenged by the physiological stress of a generalized seizure.
Expert Commentary and Mechanistic Insights
The findings of this study align with the ‘medullary dysfunction hypothesis’ of SUDEP. The brainstem, particularly the pre-Bötzinger complex and the raphe nuclei, is responsible for both respiratory rhythmogenesis and the modulation of sleep-wake states. The observed increase in inter-breath interval variability suggests that in patients at high risk of SUDEP, these homeostatic centers are already compromised.
The lack of SWA decline further suggests a failure in the glymphatic system or synaptic homeostasis. If the brain fails to ‘reset’ during sleep, it may be less resilient to the massive metabolic and electrical surge of a seizure. This interictal instability may create a ‘perfect storm’ when a seizure occurs, where the brain and lungs lack the compensatory reserves to prevent a fatal outcome.
Clinical Implications and Future Directions
For the clinician, these results suggest that polysomnography (PSG) may be a vital tool in the epilepsy clinic, not just for diagnosing obstructive sleep apnea, but for quantifying SUDEP risk. Specifically, monitoring respiratory variability and EEG slow-wave progression could provide a quantitative ‘risk score.’
However, several steps remain before these biomarkers can be integrated into standard care. First, multiday PSG studies are necessary to validate the stability of these markers over time. Second, the impact of anti-seizure medications (ASMs) on these sleep metrics needs further clarification, as some medications are known to alter sleep architecture. Finally, prospective intervention trials are needed to determine if improving sleep homeostasis (e.g., through optimized medication or CPAP therapy) can actually reduce the incidence of SUDEP.
Conclusion
This landmark study moves the field closer to a proactive model of SUDEP prevention. By identifying impaired sleep homeostasis and respiratory instability as hallmark features of the SUDEP-prone brain, it provides researchers with objective targets for future therapeutic interventions. While epilepsy management has traditionally focused on seizure control, these findings underscore the importance of protecting the sleeping brain and the respiratory drive as a primary goal in mortality prevention.
Funding and References
This research was supported by the National Institutes of Health (NIH) and the National Institute of Neurological Disorders and Stroke (NINDS).
Reference
Magana-Tellez O, Maganti R, Hupp NJ, Luo X, Rani S, Hampson JP, Ochoa-Urrea M, Tallavajhula SS, Sainju RK, Friedman D, Nei M, Gehlbach BK, Schuele S, Harper RM, Diehl B, Bateman LM, Devinsky O, Richerson GB, Lhatoo SD, Lacuey N. Sleep EEG and respiratory biomarkers of sudden unexpected death in epilepsy (SUDEP): a case-control study. Lancet Neurol. 2025 Oct;24(10):840-849. doi: 10.1016/S1474-4422(25)00273-X IF: 45.5 Q1 . PMID: 40975100 IF: 45.5 Q1 ; PMCID: PMC12707159 IF: 45.5 Q1 .