Structural Alterations Associated With Cardiovascular Autonomic Failure in Multiple System Atrophy

Abstract

Background: Early severe autonomic failure (AF) in multiple system atrophy (MSA) is a risk factor for poor survival. Postmortem studies suggested that AF is related to the degeneration of preganglionic autonomic neurons of the brainstem and spinal cord.

Objectives: Characterize cerebral alterations on brain imaging associated with cardiovascular AF.

Methods: Cardiovascular sympathetic failure was evaluated through orthostatic hypotension (OH) based on changes in systolic and diastolic blood pressure during tilt-test (ΔSBP and ΔDBP). Reduced heart rate (HR) variability reflecting cardio-vagal impairment was assessed with a composite score formed by the root-mean square differences of successive R-R intervals (RMSSD) and HR changes during deep breathing. Voxel-based morphometry (SPM12), volumetry, and cortical thickness measurements (FreeSurfer 7.0) of T1-weighted anatomical images were used to assess gray matter (GM) atrophy in sub-tentorial structures. Multivariate analysis included age, disease severity (UMSARS), and total intracranial volume as confounding factors.

Results: A total of 62 MSA patients followed at the French Reference Center were retrospectively included, aged 67.3 ± 8.6 years, 69.4% MSA-P, disease duration 4.2 ± 2.1 years. Medulla atrophy was correlated to OH (p < 0.006). Decrease in GM volume in the left anterior cerebellum (lobule V) was correlated to ΔDBP (pFWEc = 0.017). GM loss in the left interposed nucleus was correlated to ΔSBP (p < 0.003), whereas atrophy of the right dentate was associated with decreased HR variability (p < 0.003).

Conclusion: Medulla volume was strongly correlated with OH. Cerebellar degeneration was associated with the severity of cardiovascular AF.

Keywords: MRI; VBM; atrophy; autonomic failure; multiple system atrophy; volumetry.

FIGURE 1

Overview of the two steps of image analysis. Upper panel: In the non a priori analysis, correlations between local GM density (i.e., atrophy) and autonomic variables were performed using a voxel‐to‐voxel general linear model. Lower panel: In the region of interest analysis, correlations between autonomic variables and (i) mean values of gray matter concentration within six pairs of nuclei extracted from probability maps obtained from SPM12, and (ii) normalized volumes from the brainstem, the cerebellum, and both putamens were performed. iMRT, inferior medullary reticular formation hosting caudal region of the ventrolateral medulla; sMRT, superior medullary reticular formation hosting rostral region of the ventrolateral medulla; VSM, viscero‐sensori‐motor nuclei complex encompassing nucleus of the solitary tract, vagus, and hypoglossal nerves.

FIGURE 2

Infratentorial regional gray matter density changes associated with autonomic cardiovascular function (p ≤ 0.05, family wise error corrected at a peak level), evaluated by voxel‐based morphometry using SPM12 and DARTEL. Results are overlaid onto study‐specific gray matter template image for visualization, considering all patients (A, C) and only MSA‐P (B). Slice locations are indicated in Montreal Neurological Institute space. Scatters plots of mean gray matter density derived from each cluster and maximum variation in diastolic blood pressure (ΔDBP) during head up tilt‐test or heart rate (HR) composite score are shown in D.