Epilepsy-Related Mortality After Laser Interstitial Thermal Therapy in Patients With Drug-Resistant Epilepsy

Abstract

Objectives

The purpose of this report was to study the incidence of sudden unexpected death in epilepsy (SUDEP) after laser interstitial thermal therapy (LITT) for drug-resistant epilepsy (DRE).

Methods

A prospective observational study of consecutive patients treated with LITT between 2013 and 2021 was conducted. The primary outcome was the occurrence of SUDEP during postoperative follow-up. Surgical outcome was classified according to the Engel scale.

Results

There were 5 deaths, including 4 SUDEPs, among 135 patients with a median follow-up duration of 3.5 (range 0.1–9.0) years and a total of 501.3 person-years at risk. The estimated incidence of SUDEP was 8.0 (95% CI 2.2–20.4) per 1,000 person-years. Three SUDEPs occurred in patients with poor seizure outcomes, whereas 1 patient was seizure-free. Compared with pooled historical data, SUDEP occurred at a higher rate than in cohorts treated with resective surgery and at a rate similar to nonsurgical controls.

Discussion

SUDEP occurred early and late after mesial temporal LITT. The SUDEP rate was comparable with rates reported in epilepsy surgery candidates who did not receive intervention. These findings reinforce targeting seizure freedom to decrease SUDEP risk, including early consideration for further intervention.

Classification of Evidence

This study provides Class IV evidence that LITT is not effective in reducing SUDEP incidence in patients with DRE.

Introduction

Reducing the frequency of focal-to-bilateral tonic-clonic seizures (FBTCSs) is currently the most effective strategy to prevent sudden unexpected death in epilepsy (SUDEP).¹ Resective epilepsy surgery has been associated with reduction in SUDEP risk,^2-7 although certain factors, such as extratemporal epilepsy, may carry an increased risk of postoperative SUDEP.⁸ It is unclear whether similar associations exist for laser interstitial thermal therapy (LITT), particularly given more targeted ablative volumes and relatively modest seizure-free rates compared with resective surgery.⁹ We asked whether LITT reduces the risk of SUDEP and how recurrent seizures after LITT affect this risk, which may inform presurgical prognostication and individualized management strategies. In this study, we report post-LITT SUDEP outcomes in 135 patients with drug-resistant epilepsy (DRE).

Methods

Study Design

A prospective observational study of 135 consecutive patients with DRE treated with LITT at the University of Washington between 2013 and 2021 was conducted. Six patients participating in the SLATE trial (NCT02844465) were excluded as a condition of site participation and interim reporting. The study was approved by the University of Washington Institutional Review Board with waiver of informed consent. LITT was performed using the Visualase MRI-guided laser ablation system (Medtronic) by 3 neurosurgeons.¹⁰ Clinical data and outcomes were assessed through database query, chart review of clinical encounters, and direct communication with patients and community neurologists as applicable. All outcomes were determined from the time of LITT. For patients requiring multiple LITT procedures (n = 3) to achieve adequate ablation, data were analyzed from the last ablation. Engel class outcome was evaluated up to the time of subsequent resective surgery if applicable, otherwise at the latest neurology/neurosurgery clinical encounter. Mortality outcomes were determined by all clinical data available. Deaths were adjudicated by 2 epileptologists. SUDEP was classified according to the Nashef criteria.¹¹ Standardized mortality ratios (SMRs) for all-cause deaths were calculated using 2020 US age-adjusted expected death rates.

SUDEP Incidence Rate Pooled Analysis

Using a random-effects model, SUDEP rate was compared with incidence rates reported in independent cohorts undergoing resective surgery (including corpus callosotomy) vs cohorts of patients who were considered for epilepsy surgery but did not have intervention.^2-4 Pooled data were obtained from prior studies that (1) reported SUDEP mortality rates, (2) included both surgical and nonsurgical control groups, and (3) were not limited to a specific epilepsy etiology.

Statistical Analysis

Associations between categorical data were evaluated using the Fisher exact test. Data were analyzed using R 4.0.3.

Data Availability

Data are available on request.

Results

Patient Characteristics and Surgical Outcomes

Baseline demographics and clinical characteristics are summarized in Table 1. Among patients with at least 1-year follow-up, 58 of 127 patients (45.7%) had Engel class 1 outcome at 1 year and 44 of 127 (34.6%) remained Engel class 1 at the latest follow-up.

Table 1.

Demographics and Clinical Characteristics of the Overall Cohort (N = 135)

Mortality

There were 5 deaths among 135 patients with a median follow-up duration of 3.5 (range 0.1–9.0) years and a total of 501.3 person-years at risk. The all-cause SMR was 5.1 (95% CI 1.7–12.0). One patient died in a car crash while driving. SUDEP occurred in 4 patients (3.0%) (Table 2), corresponding to a rate of 8.0 (95% CI 2.2–20.4) per 1,000 person-years. SUDEP occurred in patients who were seizure-free (n = 1) and those with unchanged seizure burden (n = 3) after a median follow-up interval of 3.2 (range 0.1–7.0) years (Figure). All patients with SUDEP were believed to have unifocal mesial temporal lobe epilepsy on preoperative evaluation, and postoperative MRI revealed successful ablations in each case. Antiseizure medications were continued in all patients. All 3 patients with poor surgical outcomes and SUDEP were undergoing planning for repeat surgical intervention at the time of death. Patients with unchanged seizure burden (n = 41) had a higher all-cause death rate than those with improved seizure control (n = 94) after LITT (33.1 per 1,000 person-years [95% CI 9.0–84.9] vs 2.6 [95% CI 0.1–14.6], p = 0.01); all-cause SMR was also higher (25.8 [95% CI 6.9–66.1] vs 1.2 [95% CI 0.02–6.8]). Patients with ≥3 FBTCSs per year (n = 32) had a trend toward higher all-cause SMR than those with <3 FBTCSs (n = 103) (15.1 [95% CI 1.7–54.5] vs 3.6 [95% CI 0.7–10.4]).

Table 2.

Clinical Characteristics of Patients Who Had SUDEP

Figure. Timing of the Clinical End Points of SUDEP and Repeat Intervention With Resective Surgery During Long-term Follow-up After LITT.

LITT = laser interstitial thermal therapy; SUDEP = sudden unexpected death in epilepsy.

SUDEP After LITT vs Resective Surgery

The historical pooled SUDEP incidence rate was 2.0 (1.4–2.9) per 1,000 person-years after resective surgery, compared with 6.1 (4.3–8.7) per 1,000 person-years for epilepsy surgery candidates who did not receive intervention. Compared with resective surgery, the SUDEP rate after LITT was significantly higher (p = 0.02), but similar to the nonsurgical cohort (p = 0.55).

Classification of Evidence

This study provides Class IV evidence that LITT is not effective in reducing SUDEP incidence in patients with DRE.

Discussion

We evaluated postoperative mortality in a large epilepsy cohort treated with LITT. Most patients experienced improved seizure burden after LITT. There was minimal postoperative morbidity and no surgery-related mortality. SUDEP occurred at a higher rate than historical cohorts undergoing resective surgery, comparable with nonsurgically treated patients with DRE or those with seizure recurrence after surgery.^2-4 All-cause deaths were significantly higher in patients who had no improvement after LITT compared with those with improved seizure burden.

This report adds evidence that SUDEP can occur early or late after epilepsy surgery, including in persons with no reported seizure before death or possibly with the first seizure after a period of seizure freedom. Postoperative seizure freedom, especially for FBTCSs, can significantly reduce SUDEP risk.⁷ Similar to previous reports on resective epilepsy surgery,^5-7,12 most deaths occurred in patients with persistent seizures. These findings reinforce the goal of minimizing the absolute number of seizures to decrease SUDEP risk, including early consideration for repeat surgical intervention in patients with poor outcomes after the initial operation. Three SUDEP deaths occurred during presurgical evaluation for repeat surgery while no patients who underwent subsequent resective surgery died during the study.

Long-term seizure-free rates in this cohort were relatively low, although similar to other LITT series with prolonged follow-up.¹³ The modest seizure outcomes after mesial temporal LITT compared with temporal lobectomy,⁹ as well as with the historical resection cohorts used in this study,^3,4,7 may explain the higher SUDEP rate in this series. However, SUDEP has also been described without preceding seizures.¹⁴ Whether postoperative epileptic network changes differ between LITT and resective surgery and affect SUDEP risk independent from seizure outcome remains unclear. Given the decreased morbidity and shorter recovery time of LITT compared with resective surgery, it is likely that patients and physicians may adopt a strategy of LITT as initial curative or palliative intervention, followed by resection, if disabling seizures persist. This strategy should be used with caution, weighing the risk of delays in achieving seizure freedom against the risk of resective surgery on an individual basis.

The number of SUDEPs in this cohort was small. Although comparison with pooled SUDEP incidence rates was used to reduce bias, some variation between cohorts is inevitable. In comparison with multisite surgical series and population-based studies, long-term mortality outcomes were available in most patients raising the possibility of ascertainment bias, although in the absence of complete follow-up, these data may still underestimate SUDEP rates. Furthermore, the historical resection cohorts included extratemporal epilepsies and procedures like corpus callosotomy, which would bias toward worse seizure outcomes. Further investigation is required to determine whether SUDEP risk is elevated after LITT compared with resective surgery after controlling for postoperative seizure frequency and surgical outcomes.

In conclusion, early mortality after epilepsy surgery is driven predominantly by SUDEP. SUDEP rates may be less favorable with LITT compared with resective surgery, although the effect of seizure outcome and advances in LITT technique requires further study. Repeat interventions should be considered early for people with seizure recurrence after the initial procedure.

Appendix. Authors

Study Funding

No targeted funding reported.

Disclosure

B. Esmaeili and S. Tobochnik report no disclosures relevant to the manuscript. S. Hakimian and A.L. Ko have served as the local investigators for the SLATE trial. J.S. Hauptman has served as a consultant for Medtronic and BK medical. J.G. Ojemann has received research funding from NIH. J.W. Miller has received royalties from CRC press, Wiley Blackwell, and Up-To-Date. He has received research support from NIH, CURE/DOD, Medtronic 2016–2020, SK Biopharmaceuticals 2016–2020, and Xenon Pharmaceutics 2019–2020. He has equity interests in Therma Neurosciences. Go to Neurology.org/N for full disclosures.