Pavlovian conditioning can be used to model maladaptive associations seen in anxiety/trauma and substance use disorders. One approach to attenuate conditioned responses is extinction learning (which underlies exposure therapy), wherein cues are repeatedly presented without the expected fearful or rewarding outcome. Extinction is not effective for all; therefore, identifying biomarkers that can phenotype non-responders is necessary to optimize treatment. The orexin system is involved in fear and reward extinction and responses to CO exposure. We previously found that CO reactivity predicts fear extinction memory and CO-induced orexin activity, and orexin activity is associated with extinction memory. In a separate study, we replicated the finding that CO reactivity predicts fear extinction memory and extended this finding to appetitive extinction memory. Here, we combined behavioral and orexin activity data from these three studies in male rats to examine whether we might identify new or common associations of fear and reward extinction with CO reactivity and orexin in a larger combined sample. We found that neither CO reactivity nor CO-induced orexin activity associate with extinction memory in the combined fear and appetitive sample. We found common CO reactivity predictors in the combined fear sample, including a new predictor associated with orexin activity. In an expanded analysis, we found that prior CO exposure may affect subsequent CO reactivity and decrease orexin activity. Our findings support the potential of CO reactivity to serve as a screening tool for identifying likely responders to exposure-based therapy, though specific predictors may differ by reinforcer valence.