Fecal Metabolomics-Based Dietary Profiling Reveals Food Markers of HIV Brain Health Disorders

BACKGROUND: Despite widespread viral suppression, nearly half of people with HIV (PWH) experience brain health disorders, such as neurocognitive impairment (NCI) and depression. Dietary interventions can improve neuropsychiatric outcomes in other populations, but there is limited research in PWH. OBJECTIVE: Recent advances allow the detection of food molecules from fecal samples to provide semi-quantitative, high-resolution measures of dietary exposure. We applied a metabolomics-based method to identify food items associated with brain health outcomes in PWH. METHODS: Participants were 170 PWH (mean age=56; 84% male; 97% virally suppressed) enrolled in UC San Diego cohort studies (2019-2022). Fecal samples were analyzed using liquid-chromatography tandem mass-spectrometry. A food readout tool mapped fecal metabolomics features to ∼6,000 dietary molecular ions, predicting exposure with "food scores" scaled from 1 (lowest) to 4 (highest). Neurocognition was assessed with demographically adjusted T-scores across 7 ability domains. Depressive symptoms were measured by the Beck Depression Inventory-II and its subscales (i.e., affective, apathy, cognitive, somatic). Correlation analysis was followed by multivariate analysis adjusting for relevant covariates. Exploratory analysis in a subgroup of 37 participants with repeated neurocognitive measures examined longitudinal diet-neurocognition relationships. RESULTS: Eighty-six unique food items were detected based on fecal metabolomics profiles. Higher abundance of certain vegetables/herbs, fruits, eggs, and grains was associated with better neurocognitive outcomes, with longitudinal patterns consistent with possible diet-neurocognition relationships. As for depression, higher abundance of eggs, shrimp, and certain vegetables/fruits was associated with lower depression, particularly affective, apathy, and somatic symptoms. CONCLUSIONS: Metabolomics-based dietary profiling revealed differential food-brain health associations in PWH that align with prior self-report studies. This approach provides semi-quantitative estimates, suggesting that even modest differences in food exposure may meaningfully relate to brain health outcomes. These results provide preliminary evidence towards potential dietary targets for future interventions aimed at improving the brain health of at-risk PWH.