Association of urine dicarboxylic acid levels with intracranial volume–normalized brain regions and with white matter hypo‐intensities

Abstract Background Non‐invasive molecules related to AD imaging biomarkers can reveal disease mechanisms and may help in the mass screening of an aging population. Our study aims to determine if urine dicarboxylic acids (DCA) that changed in AD are coupled with brain volume changes. Method Study participants (> 65 years) were recruited in an ongoing HMRI Brain aging study, and MRI data were acquired and analyzed using NeuroQuant. The data were normalized to the intracranial volume (icv). DCAs were quantified using gas chromatography/mass spectrometry. We compared DCA and icv‐normalized brain region volumes in two ways: first, we used descriptive statistics to obtain the median (Q2) for short‐chain DCAs (C4+C5), medium‐chain DCAs (C7+C8+C9) and the ratio of short to medium chain DCAs ((C4+C5)/(C7+C8+C9)). Brain regions were then sorted to correspond to less than (<Q2) or greater than the DCA medians (>Q2). Differences were determined using two‐sample t‐tests and Wilcoxon rank‐sum tests with adjustment for multiple comparisons using FDR. Secondly, icv‐normalized brain regions and T1 hypo‐intensity were compared with DCA levels using a two‐tailed student’s t ‐test. Result There was no significant difference in the DCA levels of participants with low (icv<Q2) or high icv (icv>Q2) (Fig.1A). In contrast, C4+C5 levels were higher, while C7+C8+C9 were lower in participants with higher whole brain volumes (WBV) (Fig. 1B). In addition, hippocampal volumes were greater when participants had a higher urine C4+C5 and lower C7+C8+C9. We identified 19 other icv‐normalized brain regions that were bigger when urine C4+C5 levels were high. In addition, participants with higher white matter hypo‐intensities had lower urine C4+C5 and a higher C7+C8+C9 DCA levels (Table 1). Conclusion These data suggest that urine lipid composition is linked to icv‐normalized brain volumes. Higher mitochondrial‐associated lipids reflected larger brain volumes, while oxidized lipids indicated lower brain volumes. Thus, urine lipid levels can be used as surrogate biomarkers in clinical trials targeting brain volume changes. We will use a bigger sample size normalized for ApoE genotype, age, and sex to further examine lipids and brain volumes in participants with different cognitive states.