Reviews/Epilepsy Management/Resective surgery for drug-resistant temporal lobe epilepsy
Surgical Dre · EP-SRG-01 · v2
Two randomised trials compared surgery against continued drug treatment in people whose temporal lobe epilepsy had not responded to medication, and both found far more patients became seizure-free after surgery [16][1]. A large non-randomised comparison of surgically versus medically managed patients pointed the same way [7], and many independent hospital and registry series report that most patients are seizure-free or close to it after resection [24][27][12]. The direction of the effect is consistent everywhere in this corpus. What varies is the size of the benefit: it depends heavily on the patient subtype and it fades somewhat over ten years of follow-up [2][33]. The panel treated the existence and causal direction of the effect as settled while noting that the randomised trials were small and the long-term magnitude less firmly pinned down.
The randomised evidence is directionally decisive: in Wiebe 58% of surgical patients versus 8% of medical patients were free of awareness-impairing seizures at one year (P<0.001) [16], and in ERSET 11 of 15 surgical versus 0 of 23 medical patients were seizure-free in year 2 (P<.001) [1]. A 671-patient comparative study reproduced the gap, with 73% versus 16.1% seizure-free at 5 years [7]. Uncontrolled cohorts and pooled multicentre analyses anchor the post-surgical magnitude, reporting 69% seizure-free [24], 74.2% at ILAE class 1 or 2 across 1167 cases [12], and 83.6% favourable outcome at one year across 1119 cases [27]. Across every design the direction is uniform; only the magnitude moves with population and time. The contested point: The division is over magnitude, not direction: a reduction relative to continued medication is demonstrated only by the three comparator studies [16][1][7], while most of the corpus reports post-surgical seizure-freedom proportions with no medical arm. The durable and subtype-specific size of the benefit is where the evidence weakens, with freedom declining from 47% at 5 years to 38% at 10 years [2] and temporal-plus epilepsy reaching only 14.8% Engel I at 10 years [33].
All 33 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Epilepsy Management, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
Both randomised trials against the correct comparator (medical therapy) are small and one, ERSET [1], was 'halted prematurely due to slow accrual' (n=38 vs planned 200); Wiebe [16] followed patients only 1 year. Early termination and short follow-up tend to inflate effect estimates.
Most of the corpus is uncontrolled single- or multi-arm cohorts reporting post-surgical seizure-freedom proportions with no medical comparator, so they evidence outcomes after surgery rather than a reduction relative to continued medication. The only non-RCT comparative study [7] is non-randomised (surgery vs those who 'opted for' medical management), exposing it to strong selection bias.
Where it remains uncertain. A reduction versus medication is shown only by the comparator studies [16][1][7]; most papers are single-arm cohorts, and the benefit erodes over 10 years [2] and is far lower in temporal-plus epilepsy [33].
Aleatory
Even among initially seizure-free patients, a fraction relapse as follow-up lengthens (47% at 5 years to 38% at 10 years [2]).
Epistemic
ERSET enrolled 38 of a planned 200 and was halted for slow accrual [1]; Wiebe randomised 80 patients with one-year follow-up [16], leaving the randomised base thin.
Inferential
Single-arm cohorts [24][27][12] report seizure-freedom after surgery with no medical comparator, so they evidence post-surgical outcome rather than reduction relative to continued medication.
The load-bearing papers from a corpus of 33, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Second RCT (ERSET) confirming surgery vs continued AED in newly intractable MTLE.
“Zero of 23 participants in the medical group and 11 of 15 in the surgical group were seizure free during year 2 of follow-up (odds ratio = ∞; 95% CI, 11.8 to ∞; P < .001).”
Panel auditInternal validity: medium · scope of inference: Newly intractable disabling mesial TLE, patients aged ≥12, 2-year outcome.Key limitations: very small · stopped early · crossovers between arms
Large comparative cohort reproducing the surgery-vs-medical gap out to 10 years in drug-resistant TLE-HS.
“Seizure freedom at 3 (81.8 vs 19.0%; P < .001), 5 (73% vs 16.1%; P < .001), and 10 years (78.3% vs 18.5; P < .001)”
Panel auditInternal validity: medium · scope of inference: Drug-resistant TLE with hippocampal sclerosis, long-term outcome.Key limitations: non-randomized · baseline differences · retrospective
Large (n=1167) 20-center cohort confirming high seizure-control rate and strong safety profile.
“After 1 year, 74.2% of patients achieved seizure outcomes classified as ILAE class 1 or 2.”
Panel auditInternal validity: medium · scope of inference: Drug-resistant mesial TLE undergoing microsurgical resection; safety and efficacy.Key limitations: no control arm · retrospective
Long-term outcomes; documents both substantial reduction and its erosion over time.
“Kaplan-Meier estimates showed that 47% (95% CI 40-58) remained seizure free (apart from simple partial seizures) at 5 years and 38% (95% CI 31-45) at 10 years after surgery.”
Panel auditInternal validity: medium · scope of inference: Patients undergoing resective epilepsy surgery (mostly temporal) at a single UK centre.Key limitations: no control arm · mixed resection types
4 of 33 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Randomised, comparator-controlled evidence plus uniformly concordant large cohorts carry a modest directional claim to the Established tier, held just below ceiling by the RCTs' small size and early termination.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Adequately powered RCT against medical therapy. The randomised evidence is two small trials, one terminated early [1][16], which the panel cited as the main constraint on certainty. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: A multicentre randomised trial of resection versus optimised continued medication reaching planned enrolment, with follow-up of 5 years or more.
Open question: Does resective surgery reduce seizures relative to continued medication beyond two years, under randomisation? Why it is still open: Both randomised comparisons against medication are short (1 year [16], 2 years [1]) and the longer comparison [7] is non-randomised.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19