HIV-Associated Alterations in Normal-Appearing White Matter

Abstract

There are conflicting reports of adverse HIV-associated alterations in white matter integrity as measured by diffusion tensor imaging (DTI). We sought to address these conflicting reports by assessing, on a voxel-by-voxel basis, HIV-associated regional changes in radiologically defined normal-appearing white matter (NAWM) integrity using high-resolution DTI. 30 HIV-seropositive (SP) and 30 HIV-seronegative (SN) nondemented, community-dwelling participants underwent DTI to derive whole-brain measures of white matter integrity (fractional anisotropy [FA] and mean diffusivity [MD]). For each participant, the white matter T2 volume was thresholded to remove regions of abnormal signal, resulting in a NAWM mask, which was then applied to the FA and MD volumes to extract voxel-wise NAWM measures of white matter integrity. Voxel-wise group comparisons of FA and MD were conducted (P < 0.005, extent threshold 5 voxels) while controlling for age and substance-abuse history. There were no significant differences between the groups for demographic or cognitive performance variables. Summary whole-brain measures of FA and MD were equivalent between the SP and SN samples. Among the SP sample, history of substance abuse was associated with significantly increased whole-brain NAWM MD, and coinfection with hepatitis C virus (HCV) was associated with a trend for increased MD. Correlations between whole-brain NAWM FA and MD with cognitive performance measures were not significant. Regional analyses of DTI measures revealed variable differences in NAWM FA in the SP sample, with findings of both decreased and increased FA. Differences in NAWM MD were more consistent, with widespread increases noted in the SP sample compared to the SN sample. Eight of the 10 regions displaying significantly increased FA in the SP sample were also found to have significantly increased MD compared to the SN sample. Decreased white matter integrity is present even in radiologically defined NAWM in nondemented, community-dwelling patients with HIV. The decrease in NAWM integrity is best seen in increases in MD, a measure of generalized tissue breakdown. Indications of NAWM axonal integrity (FA) present a more complicated picture, with both decreased FA and increased FA in the SP sample. Our findings of variable HIV-associated FA changes in NAWM may account for previous conflicting reports of changes in DTI parameters in this population. The results of our study suggest that HIV infection contributes to variable changes in DTI values, reflecting both direct loss of axonal integrity and a loss of complexity to the underlying axonal matrix.