Prognostic model for withdrawal of anti-seizure medication following epilepsy surgery in adults.
Abstract number :
148
Submission category :
9. Surgery / 9A. Adult
Year :
2020
Submission ID :
2422496
Source :
www.aesnet.org
Presentation date :
12/5/2020 9:07:12 AM
Published date :
Nov 21, 2020, 02:24 AM
Authors :
Carolina Ferreira-Atuesta, University College London Queen Square Institute of Neurology; Jane De Tisi - University College London; Chalfont Centre for Epilepsy; Fergus Rugg-Gunn - University College London Queen Square Institute of Neurology; Andrew W Mc
Rationale:
More than half of adult epilepsy patients who undergo resective surgery achieve long-term seizure freedom and might consider withdrawing anti-seizure medication (ASM). We developed a prognostic model to predict outcomes after reducing ASMs postoperatively.
Method:
We included adults who had resective epilepsy surgery and were free from seizures other than focal non-motor aware seizures (auras) before the beginning of ASM withdrawal. The cohort was randomly divided into a derivation and validation cohort (2:1 ratio). We used Cox proportional hazards regression to develop a model predicting recurrent seizures other than aura during and following postoperative ASM withdrawal.
Results:
The final cohort included 350 participants, of whom 88% remained free from seizures other than auras two years after the beginning of ASM withdrawal and 77% after five years. Median follow up time was 11 years (interquartile range 6-16). Independent predictors of seizure recurrence during and following the beginning of withdrawal of ASMs were auras after surgery and before withdrawal (adjusted hazards ratio [aHR] 5.5, 95% confidence interval [CI] 2.7-11·1), history of generalized seizures before surgery (aHR 1.6, 95% CI 0.9-2.8), shorter time from surgery to the beginning of withdrawal (aHR 0.9, 95% CI 0.8-0.9), and number of ASMs at time of surgery (aHR 1.2, 95% CI 0.9-1.6). Model discrimination, i.e. the ability to correctly identify patients with vs. without seizure recurrence after beginning ASM withdrawal, showed an overoptimism corrected concordance statistic of 0.67 in the derivation cohort and 0.68 in the validation cohort. Calibration plots indicated high agreement of predicted and observed outcomes.
Conclusion:
A simple algorithm, available as graphical nomogram and online tool, can predict seizure outcome during and following postoperative withdrawal of ASMs. This model might guide decisions on whether and when to consider ASM withdrawal after surgery and is a step towards more personalized medicine.
Funding:
:None
Surgery