Basic lipophilic drugs such as propranolol and lidocaine are strongly bound by alpha(1)-acid glycoprotein, also called orosomucoid. Although the liver is known to rapidly clear plasma protein-bound propranolol or lidocaine, it is generally regarded that peripheral tissues, such as brain or heart, are only exposed to the small fraction of drug that is free or dialyzable in vitro. The "free drug" hypothesis is subjected to direct empiric testing in the present studies using human sera and an in vivo rat brain paradigm. Serum from 27 human subjects (normal individuals, newborns, or patients with either metastatic cancer or rheumatoid arthritis) were found to have up to a sevenfold variation in orosomucoid concentrations. The free propranolol or lidocaine as determined in vitro by equilibrium dialysis at 37 degrees C varied inversely with the orosomucoid concentration. Similarly the rate of transport of propranolol or lidocaine through the blood-brain barrier (BBB) was inversely related to the existing serum concentration of orosomucoid. However, the inhibition of rat brain extraction of drug by orosomucoid in vivo was only about one-fifth of that predicted by free drug measurements in vitro. This large discrepancy suggested orosomucoid-bound drug was readily available for transport into brain in vivo. Studies using purified human orosomucoid in the rat brain extraction assay also showed that orosomucoidbound propranolol or lidocaine is readily transported through the BBB. Conversely, albumin-bound propranolol or lidocaine was not transported through the BBB. The studies using albumin provide evidence that the in vivo rat brain paradigm used in the present investigations is capable of confirming, when possible, predictions made by the "free drug" hypothesis. These data suggest that the amount of circulating propranolol or lidocaine that is available for transport into a peripheral tissue such as brain is not restricted to the free (dialyzable) moiety but includes the much larger globulin-bound fraction. Therefore, existing pharmacokinetic models should be expanded to account for the transport of protein-bound drugs into peripheral tissues similar to what is known to occur in liver.