The 35% CO2 inhalation panic challenge in early clinical anxiolytic drug development: Addressing current methodological limitations

We evaluated methodological properties of the 35% CO 2 inhalation challenge as a panic model to optimize use in anxiolytic drug development. Twenty CO 2 -sensitive healthy volunteers – who participated 3-5 years ago and were sensitive to the fear-inducing effects of the CO2 challenge - completed a randomized open-label crossover comparing single versus double vital-capacity (VC) inhalations, followed by three additional weekly double-VC challenges. Subjective panic symptoms and fear were assessed using the Panic Symptom List-IV, standard visual analog scales, and a continuous real-time fear rating, alongside beat-to-beat cardiovascular monitoring and repeated plasma and salivary neuroendocrine measures. Binary panic-attack occurrence did not differ between single and double VC inhalation, but double VC produced significantly larger intensity-based subjective responses and broader physiological activation, including blood pressure and endocrine changes. Across repeated double-VC challenges, several subjective and endocrine outcomes showed significant attenuation consistent with habituation, whereas cardiovascular peak responses were relatively stable and cardiovascular parameters demonstrated the highest test–retest reliability. Individual sensitivity to CO 2 -induced panic remained largely stable over several years within this population. These findings support using double VC to maximize signal in pharmacology studies, incorporating habituation into design and analysis, and prioritizing robust cardiovascular endpoints; single VC may be sufficient for screening.