The objectives of this study were to a) replicate our prior finding of a decreased number (Bmax) of platelet alpha 2-adrenoreceptors in panic disorder, b) determine if binding is also decreased in asymptomatic first-degree relatives of panic patients (known to be at increased risk for developing panic), and c) evaluate the effect of treatment on the presumptive decrease in binding (i.e., is the decrease a state or a trait marker for panic?). Panic patients had clonidine and yohimbine platelet-binding assays, symptom ratings, and measurement of lying and standing plasma epinephrine, norepinephrine, systolic and diastolic blood pressure, and heart rate before treatment, after approximately 2 months of medication (fluoxetine, tricyclics, or alprazolam) and/or cognitive behavioral treatment, and after symptom remission while drug free; normal subjects had determinations of the same measures at approximately the same time intervals. Relatives of both groups had one determination only of all measures. Tritiated clonidine binding was decreased and lying heart rate was increased in patients before treatment. Magnitude of binding decrease was correlated with symptom severity and standing norepinephrine. No binding abnormality was seen in first-degree relatives of patients. Treatment increased clonidine binding in patients. Both patients and relatives of patients showed significantly increased standing plasma norepinephrine in comparison to controls. There is a state-related decrease in binding, associated with symptom severity and norepinephrine, in panic disorder. Abnormal reactivity of norepinephrine to standing might be a marker for increased likelihood of panic development in individuals at risk.