OBJECTIVE: Epilepsy is a highly heterogeneous neurological disorder with significant prognostic variability. Accurate long-term outcome prediction remains a clinical challenge. We investigated pharmacotherapeutic prognosis and key predictors, particularly baseline seizure timing, to guide individualized treatment. METHODS: This prospective cohort study consecutively enrolled newly diagnosed focal epilepsy patients from the epilepsy specialist clinic of the Department of Neurology in our hospital between September 2015 and September 2025. We applied the Cox proportional hazards regression model to analyze the association between baseline factors and prognosis to identify independent predictors. RESULTS: A total of 563 patients with ≥36-month follow-up (median = 48.0 months, interquartile range IQR = 36.0-78.0) were analyzed. Among them, 64.8% (365/563) achieved seizure freedom at least once (median = 27.0 months, 95% confidence interval CI = 22.5-31.5), and 35.9% (131/365) relapsed after seizure freedom; 25.4% (143/563) developed drug resistance, with 6.3% (9/143) later converting to drug-responsive status. At the end of follow-up, 50.8% (286/563) were classified as drug-responsive, 23.8% (134/563) as having drug-resistant epilepsy (DRE), and 25.4% (143/563) as having an undetermined prognosis. Among patients with drug-responsive epilepsy, 63.6% (182/286) achieved seizure freedom with monotherapy. Multivariate analysis identified baseline wake-related seizures (hazard ratio HR = 1.743, 95% CI = 1.101-2.760, p = .018), cluster seizures (HR = 2.630, 95% CI = 1.845-3.749, p < .001), status epilepticus (HR = 2.126, 95% CI = 1.394-3.242, p < .001), and sleep problems (HR = 1.740, 95% CI = 1.178-2.572, p = .005) as independent DRE risk factors. SIGNIFICANCE: This prospective large-sample study revealed the inherent complexity and dynamic nature of prognostic assessment in patients with newly diagnosed focal epilepsy. We identified independent prognostic predictors for this condition, first confirming baseline wake-related seizures as a predictor of poor prognosis to guide early risk stratification and personalized treatment.
Sun et al. (Tue,) studied this question.