Brain structural MRI changes before and after 6 months of preventive treatment with erenumab and onabotulinumtoxinA...

Structural neuroimaging findings in migraine remain heterogeneous, and the impact of preventive treatments such as erenumab and onabotulinumtoxinA (OnabotA) on brain plasticity is poorly understood. We aimed to comprehensively characterize brain morphology, including less explored regions such as the cerebellum and corpus callosum (CC), compare people with migraine with healthy controls (HC), identify baseline predictors of treatment response, and evaluate longitudinal structural trajectories following preventive therapy.

This prospective study included participants with migraine with ≥ 8 headache days/month eligible for erenumab or OnabotA, together with HC. T1-weighted MRI was acquired on 3T MRI scanners at baseline (M0) in all participants and repeated at 6 months (M6) in eligible participants with migraine. Regions of interest (ROIs) were segmented, and cortical thickness (CTh), cortical surface area (SA) and brain volume values were extracted using FastSurfer. Data were analysed using generalized linear mixed models (GLMM). Response was defined as a ≥ 50% reduction in monthly headache days at M6.

We included 189 participants with migraine and 63 HC; 87.9% were female (222/252), and age was similar between groups (adjusted-p = 0.627). Compared with HC, individuals with migraine showed larger volumes in the mid-posterior (adjusted-p = 0.001) and central CC segments (adjusted-p = 0.030). 2) At baseline, non-responders (NR) showed higher CTh in the left parahippocampal gyrus than responders (R) (adjusted-p = 0.008). In addition, a significant response-by-treatment interaction was observed in the bilateral caudate volume: post-hoc contrast showed higher caudate volumes in OnabotA responders than in OnabotA NR (adjusted-p = 0.041 for both hemispheres), whereas no such difference was obsvered with erenumab. Longitudinally, several time-by-treatment interactions indicated different structural trajectories between treatments. In most of these regions, post-hoc contrasts showed lower M6 values in the erenumab group, including SA in the left fusiform, left pericalcarine, left rostral anterior cingulate, and right lingual regions, and volume in the bilateral cerebellar cortex, right hippocampus, and left amygdala. The left choroid plexus showed an opposite pattern, with increased volume in the erenumab group. Comparable longitudinal changes were not observed in the OnabotA group.

Our results identify CC alterations and baseline parahippocampal and caudate morphometry as potential markers of migraine and treatment response. Erenumab and OnabotA showed different longitudinal structural trajectories, with more evident changes observed in the erenumab group in regions related to in pain, limbic, memory, and cerebellar processing.

EudraCT: 2019-002224-32, Date: 2019-11-12. ID: 2020-003650-77 available at REEC, Date: 2021-02-09. ID: 2020-002639-31 available at REEC, Date: 2020-10-23.