Twenty-six healthy volunteers received either 25 micrograms of cholecystokinin-tetrapeptide (CCK-4) or a mixture of 35% carbon dioxide in oxygen (CO2). DSM-III-R criteria including anxiety, apprehension and/or fear of at least moderate intensity were used to determine the occurrence of a panic attack. Results for the entire sample revealed that CCK-4 produced significantly more intense symptoms than CO2, but not a significantly greater number of symptoms. The incidence of DSM-III-R panic attacks was similar with both substances; 21% (3/14) for CO2 and 17% (2/12) for CCK-4. This study indicates that CCK-4 is at least as potent as CO2 in producing panic symptoms in healthy volunteers and is a useful challenge paradigm for comparative research of pharmacologic agents which possess distinct neurobiologic properties.