Abstract
Opioid use disorder (OUD) is a chronic condition that can be characterized by an overwhelming desire to use opioids, heightened opioid tolerance, and withdrawal symptoms when opioid use is stopped. Given the relapsing nature of the condition, early intervention is crucial to enable timely treatment. Recent studies suggest that OUD involves dynamic interactions between the brain, immune system, and gut, resulting in significant alterations of the gut microbiome and changes in systemic biomarkers such as cytokines, chemokines, and other metabolic markers. However, the integration of combining gut microbiome data with blood-based biomarker data and their potential as diagnostic markers remain unclear. To address this gap, we compared alpha and beta diversity between OUD-positive and OUD-negative patients and found no significant differences in gut microbiome composition. We then conducted a differential abundance analysis using DESeq2 to identify gut microbial taxa that differed significantly between the two groups and combined the top 20 taxa with significant biomarkers associated with OUD to develop a predictive model using the Random Forest machine learning algorithm for identifying OUD status. We identified 203 taxa that were differentially abundant and 6 out of 21 statistically significant biomarkers that were associated with OUD. However, both individual or combined gut microbial and biomarker profiles showed no predictive power for OUD status. Our results indicate that gut microbiome and blood metabolic biomarker data, either individually or in combination, are insufficient for reliably predicting OUD status and therefore lack utility as a reliable clinical diagnostic tool.