Authors :
Ahmed Yassin, MD – Jordan University of Science and Technology; Leen Al-Kraimeen, Medical Student – Jordan University of Science and Technology; Aref Qarqash, Medical Student – Jordan University of Science and Technology; Hassan AbuShukair, Medical Student – Jordan University of Science and Technology; Obada Ababneh, Medical Student – Jordan University of Science and Technology; Salma Al-Aomar, Medical Student – Jordan University of Science and Technology; Mohammad Abu-Rub, MD – The George Washington University; Khalid Alsherbini, MD – University of Tennessee Health Science Center
Rationale: Anterior nucleus of the thalamus (ANT) is the only deep brain stimulation (DBS) target that is approved by the FDA for treatment of drug-resistant epilepsy (DRE). Hippocampus (HC) and centromedian nucleus (CMN) have been reported as potential DBS targets for DRE. This study aimed to assess the effectiveness and predictors of response among DRE patients treated with DBS in general and among ANT, HC and CMN DBS-targets. To the best of our knowledge, this is the first meta-analysis utilizing individual patient data (IPD) on this topic.
Methods: A systematic literature search by two independent reviewers was executed on studies published between Jan 1, 2000 and June 29, 2020. Patients with DRE who underwent DBS treatment with at least three months of follow-up were included. Meta-analysis was conducted on DBS studies with available IPD. Response was defined as ≥ 50% reduction in seizures frequency. Responders group was compared with non-responders group in terms of demographics, epilepsy/seizure characteristics, MRI findings, and DBS targets and duration of use. Subsequently, predictors of response to different DBS targets were investigated.
Results: Thirty-nine studies with a total of 296 patients (ANT: 69%, HC: 11%, CMN: 21%) were included. The responders group constituted of 209 patients (70.6%).
The response was significantly higher in patients with generalized seizures compared to those with focal seizures (93.2% vs 63.9%; p< 0.001). Response was significantly higher with CMN (83.9%) and HC (77.4%) compared with ANT (65.5%) as DBS targets (p=0.014). Response was also significantly associated with longer duration of DBS use (p=0.008). The responder rate was higher among the patients with lesional MRIs (76.7%) than those with non-lesional MRIs (66.7%), but with no statistically significant difference (p=0.134). Age, gender, epilepsy etiology, onset zone of focal seizures, and previous use of VNS had no significant differences between the responders and non-responders. A binary logistic regression including the seizure type, MRI findings, DBS targets, and DBS duration showed, after controlling for confounders, that the duration of DBS use was the only significant predictor of response (adjusted OR 1.061; 95% CI 1.019-1.106; p=0.005). Regarding DBS targets, the response rate in patients with symptomatic etiology was significantly higher with HC or CMN targets than the ANT (p=0.003). In patients with non-lesional MRI, response rate was significantly higher with the CMN target compared to the other two targets (p=0.008).