Stable Vulnerability vs. Momentary Triggers: Dissecting the Temporal Dynamics of Opioid Craving and Use

Abstract

Opioid use disorder (OUD) remains a major public health challenge, characterized by persistent relapse risk even among individuals receiving medications for opioid use disorder (MOUD). Although craving is widely studied as a predictor of opioid use, it remains unclear whether risk is driven primarily by momentary increases in craving or by stable individual differences in baseline craving. This study examined the temporal dynamics of craving and opioid use in a longitudinal clinical cohort of adults with OUD engaged in MOUD treatment. Across 190 participants and 937 study sessions, generalized linear mixed-effects models were used to examine concurrent and prospective associations between craving and opioid use. Craving was decomposed into between-person (“trait-like”) and within-person (“state-like”) components using subject-mean centering, and assessed across multiple clinically relevant timescales (current moment, past 24 hours, past week). Both trait and state craving were significantly associated with opioid use; however, trait-level craving emerged as the dominant predictor, showing substantially larger and more consistent effects across all timescales. Within-person increases in craving were associated with opioid use, but their effects were comparatively modest. In prospective analyses, current opioid use was the strongest predictor of subsequent use, while trait craving retained additional predictive value. Exploratory analyses further examined individual differences in biological sex and neuromelanin-sensitive MRI (NM-MRI) signal. Sex differences were observed in overall opioid use, with males exhibiting higher average use, but no evidence of sex moderation in craving-use relationships. NM-MRI signal in the substantia nigra/ventral tegmental area was associated with mean craving in a subset of participants, suggesting a potential link between dopaminergic function and trait-like craving vulnerability. Taken together, these findings indicate that opioid use risk is structured primarily by stable individual differences in craving rather than momentary fluctuations. This temporal dissociation has important implications for intervention, supporting approaches that target persistent vulnerability—such as sustained treatment engagement and harm reduction—rather than relying solely on strategies focused on acute craving episodes.