Infusion of sodium lactate has been shown by a number of investigators to induce panic in patients with panic disorder, but the pathophysiology underlying this phenomenon is unknown. One theory to explain lactate's anxiety-producing effects involves its ability to induce alkalosis because of metabolic conversion to bicarbonate. To test this hypothesis, we administered both sodium lactate and sodium bicarbonate infusions in counterbalanced order to patients with panic disorder. Thirteen of 22 subjects panicked in response to lactate and nine of 20 subjects panicked in response to bicarbonate. Although the rate of panic between the two infusion responses was not significantly different, several aspects of response to the two infusions indicated that lactate may be a more potent producer of anxiety than bicarbonate. An unexpected finding was that bicarbonate panickers had a reduction in arterial carbon dioxide pressure during the infusion, while bicarbonate nonpanickers had an increase in arterial carbon dioxide pressure during the infusion. Induction of hyperventilation and subsequent hypocapnia appears to be a common denominator between lactate- and bicarbonate-induced panic.