Safety and Effectiveness of Medications for Alcohol Use Disorder among HIV+/-

HARP PROJECT 2 SUMMARY Although rarely addressed by HIV healthcare providers, alcohol use disorder (AUD) is a major cause of morbidity and mortality among people aging with HIV (PAH).

There are only three Food and Drug Administration-approved medications for AUD and these have had limited acceptability and uptake for several reasons.

HIV providers are unfamiliar with AUD medications and reticent to add an unfamiliar medication in the complex context of pre-existing polypharmacy and physiologic frailty.

Even in less complex contexts, the effectiveness of these medications is modest on average, likely due in part to variations in genetic liability.

Further, despite PAH?s excess liability for hepatic injury from alcohol-associated liver disease (AALD), we lack medications to mitigate this injury.

As more people are aging with HIV and they continue to drink alcohol, there is growing need to identify and evaluate candidate AUD and AALD medications that will be safe and effective in the context of polypharmacy and physiologic frailty.

Analyses of large scale, real-world data may help us identify and evaluate candidate medications for repurposing (i.e., medications commonly used for other purposes that may also decrease drinking or hepatic injury).

Building on our strong collaborative history and supported by the larger HIV and Alcohol Research center focused on Polypharmacy (HARP) Program, we will apply innovative approaches to enhance timely discovery and initial evaluation of candidate repurposed medications for AUD and AALD.

In Aim 1 we apply Explainable Artificial Intelligence methods and systems biology to data from the national Veterans Administration Healthcare (VA) electronic health record linked to genetic data and public data to identify candidate repurposed AUD and AALD medications in the context of gene networks, which may provide mechanistic details at the biological level.

In Aim 2 we use data from the well-characterized VA family of cohorts encompassing >60,000 PAH and millions of uninfected controls, to conduct propensity score matched analyses to examine the safety and effectiveness of four candidate medications (prazosin, pioglitazone, and verapamil for AUD and metformin for AALD) among PAH and uninfected individuals.

In Aim 3, informed by these analyses, our prior alcohol interventions in HIV treatment settings, and in collaboration with the Risk Communication and Administrative/Data Analytic Cores, we will conduct three pilot studies employing clinical pharmacist-delivered personalized risk messaging and identified candidate medications to generate time sensitive data on the feasibility, acceptability, and efficacy of these medications to address AUD or AALD.

Together, these studies will collectively expedite the pipeline from discovery to evaluation of novel medications to address AUD and AALD among PAH.