Brain metabolites associate with white matter integrity and cognitive deficits in patients recovered from Covid-19: a magnetic resonance spectroscopy study.

Evidence on the neuropsychiatric consequences of COVID-19 are growing, and depressive symptoms and cognitive deficits now appear to be a relatively frequent outcome of the infection, being strictly interrelated; much less however is known about their pathophysiology, although some recent MRI studies started to associate post-COVID neuropsychiatric symptoms with alterations in brain structure and function. On the other hand, to date only few studies with limited sample size used brain spectroscopy to investigate COVID-related neuropathology. Therefore in the present study we employed single voxel brain spectroscopy to investigate brain metabolites levels and their relationship with neuropsychiatric sequelae and white matter integrity in patients recovering from COVID-19 infection. Our study was performed on 64 patients who were hospitalized at San Raffaele Hospital for a SARS-CoV-2 infection, recruited in the context of an ongoing prospective cohort study. The presence of depressive and cognitive symptomatology was assessed in the context of an unstructured clinical interview conducted by well-trained psychologists. Patients underwent 3-Tesla MRI scanning acquiring T1-weighted and Diffusion Weighted Images (DWI). 1H-MRS was performed on a voxel placed in the DLPFC and concentration of brain N-acetyl-aspartate (NAA) and Glutamate (Glu) were estimated using LCModel. DWI normalization, correction for motion and eddy current, and tensor fitting were carried out using FSL 6.0. The association between brain metabolites levels, cognitive and depressive symptomatology and time elapsed from the infection to the MRI scan were investigated using age, sex and BMI as nuisance variables. Furthermore the associations between brain metabolites levels and Fractional Anisotropy (FA), Axial (AD), Mean (MD), and Radial Diffusivity (RD) were investigated, again entering age, sex and BMI as covariates. Time from infection to MRI scan was found to positively affect brain NAA levels (β = 0.33, p=0.008), while it had no effect on Glu (β = 0.03, p=0.815). Subjects with cognitive deficits had significantly lower levels of NAA (p=0.005) and Glu (p=0.013). When testing the effect of brain metabolites on white matter integrity, both NAA and Glu levels were associated with higher levels of FA (NAA: p<0.001; Glu: p<0.001) and lower of RD (NAA: p=0.004; Glu: p=0.007) and MD (NAA: p=0.011; Glu: p=0.018). In our study we found NAA to progressively increase as a function of time elapsed from the infection to the MRI. Taking into account that NAA is considered to be a marker of neuronal viability, this finding could be suggestive of a brain recovery process after the infection. In line with this hypothesis, NAA and Glu levels were found to be lower in subjects with post-covid cognitive symptoms, and to be positively associated with white matter integrity. Our study therefore suggests that brain spectroscopy could provide novel insight into the pathophysiology of post-covid neuropsychiatric sequelae.