In a cross-sectional cohort of 218 participants—61 HIV-positive individuals with distal sensory polyneuropathy (DSP), 93 HIV-positive without DSP, and 64 HIV-negative controls—gut microbiome composition via 16S-rRNA sequencing differentiated DSP status more powerfully than serum cytokine profiles. Bayesian differential abundance analysis identified enrichment of Enterocloster and Bacteroides_H genera specifically in HIV-positive individuals with DSP. An XGBoost machine-learning model integrating demographic, clinical, inflammatory, and microbial features classified DSP with an AUC of 0.83, with a microbial log-ratio and age as the dominant predictors.
The finding that serum inflammatory markers failed to distinguish DSP within HIV-positive individuals—while gut microbial composition did—is genuinely striking. It suggests that peripheral neuropathy vulnerability in treated HIV may be rooted more in gut-neural axis disruption than in systemic cytokine storms, reframing where clinicians and researchers should look. Enterocloster species have been previously linked to intestinal barrier disruption and systemic translocation of bacterial products, offering a plausible mechanistic pathway to neuroinflammation. However, the cross-sectional design is a critical limitation: causality cannot be inferred, and reverse causation (neuropathy altering gut motility and microbiome) remains entirely possible. With only 61 DSP cases, the model's generalizability needs external validation. Still, an AUC of 0.83 using microbiome-derived features is clinically promising and elevates gut-targeted interventions—probiotics, dietary fiber, fecal transplant—as genuinely testable therapeutic hypotheses for HIV neuropathy. Confirmatory longitudinal work is urgently warranted.