Evaluating outcome measures and candidate biomarkers characterizing the human panic and anxiety response to the experimental CO2 inhalation model in healthy participants: a systematic literature review

This is the authors' abstract. We don't add key points for this paper.

Background The CO 2 inhalation challenge is a translational model of panic- and anxiety-like states used in early-phase drug development. We evaluated subjective reported outcome measures and objective candidate biomarkers. Methods We searched PubMed, Embase, Web of Science, the Cochrane Library, Emcare, CINAHL, and PsycINFO from inception through 15 July 2026 for CO 2 -challenge studies in healthy participants. Eligible outcome-measure rows were grouped into clusters. Sensitivity was the proportion showing the dominant statistically significant direction; measured but unreported results were classified as not significant. Risk of bias was assessed with revised Joanna Briggs Institute tools. Results 112 publications represented 117 trials, 126 conditions, and 821 eligible rows comprising 175 outcome measures in 49 clusters. In panic-model studies, heart-rate variability (86%) and the Panic scale (72%) met the prespecified criteria of at least 70% sensitivity and at least five rows. In generalized-anxiety-model studies, eight clusters met both criteria: expired or end-tidal CO 2 (89%), heart rate (86%), Anxiety scale (85%), Panic scale (80%), systolic blood pressure (80%), working memory (80%), postural or gait stability (75%), and Calmness scale (74%). Greater CO 2 exposure was generally associated with higher sensitivity. Risk-of-bias concerns most often involved allocation, blinding, and retention. Conclusions The most informative outcomes differed by challenge model. Near-real-time subjective measures and prespecified physiological or performance-based candidates provide complementary information. Standardized acquisition timing, clearer reporting, responder enrichment in panic-model studies, and direct concentration- and volume-comparison studies may improve the reliability and translational value of CO 2 challenges.