Abstracts

Rationale:
Deep brain stimulation (DBS) to the thalamus is an emerging treatment for drug-refractory epilepsy. Whilst reported rates of surgical complications for DBS are low, data are largely derived from movement disorders and studies of anterior nucleus (AnT) stimulation for focal epilepsy. Centromedian nucleus (CM) DBS has been performed in patients with Lennox-Gastaut syndrome (LGS) (Velasco et al., Epilepsia (2006) 47:1203-12), however the complications and safety considerations in this cohort are not well-described.    We aimed to characterise the peri-operative complications of LGS patients undergoing CM-DBS as part of the ESTEL (Electrical Stimulation of the Thalamus in Epilepsy of Lennox-Gastaut phenotype) study, due for completion in August 2020. This is the largest cohort of patients to undergo a randomised, placebo-controlled clinical trial of CM-DBS.  
Method:
Twenty adult patients (mean age±1SD= 25±6.3 years; 13 females) with LGS underwent bilateral CM-DBS insertion (Medtronic Activa® PC with 3389 electrodes) between November 2017 to December 2019 at Austin Health, Melbourne, Australia. We performed neurologic evaluations pre-operatively and three-months post-operatively, as well as  prospectively documenting any peri-operative complications. Devices are not turned on until at least 3-months post-implantation, so complications described here represent surgical, not stimulation related side effects.
Results:
No patient had a serious surgical complication resulting in death or permanent neurologic deficit. Cerebral staphylococcus aureas infection necessitated DBS hardware removal in one patient, 58 days post-operatively. One patient required emergency evacuation of a sub-dural haematoma following a seizure/fall 41 days post-operatively, prior to the device being switched on. DBS hardware remained in-situ.   Median hospital length-of-stay post-operatively was 3 days (range=1-17). 60% (12/20) of patients had transient post-operative drowsiness, ranging 24 hours to 13 days. Three of these patients had more profound drowsiness, prolonging bed-stay and requiring temporary naso-gastric feeding to maintain nutrition. Acute imaging of these patients showed quite marked cerebral oedema along the course of DBS electrodes. Post-operative seizures were observed in 70% (14/20) of patients during the hospital admission, consistent with the high seizure burden experienced by these patients pre-implantation. No patient had status epilepticus.
Conclusion:
Transient post-operative drowsiness is common after CM-DBS for LGS, and although it can be quite significant, resolves with no specific treatment over days to weeks. Potential explanations for the post-operative drowsiness include a ‘stun effect’ on the CM nucleus, and/or bi-frontal cerebral oedema, along the electrode tract. Lead oedema is a recognised issue in Parkinson’s DBS, but the extent of drowsiness we observed seemed more prominent. This may reflect increased vasoreactivity in our younger population, or reduced cognitive reserve. Long-term follow-up to assess for hardware-related complications is required.
Funding:
:NHMRC project grant #1108881 Australian Government Research Training Program Scholarship