The aim of the present study was to evaluate whether metabolic responses to inhaled salbutamol may be used to measure the cardioselectivity of beta-adrenoceptor antagonists. We therefore studied the effects of oral doses of atenolol 50 mg, 100 mg, 200 mg (A50, A100, A200), propranolol 40 mg (P40), and placebo (Pl) on the hypokalaemic (K) and hyperglycaemic (Glu) responses to inhaled salbutamol in five healthy subjects. Increasing doses of atenolol were associated with a progressive attenuation of delta K compared with placebo: -0.72 mmol.l-1 (Pl) vs -0.20 mmol.l-1 (A200). However, delta K with A200 was significantly different from the response with P40: +0.12 mmol.l-1. There were partial reductions in the hyperglycaemic response with the beta-adrenoceptor antagonists, although this was only significant (compared with Pl) for P40: delta Glu 1.92 mmol.l-1 (Pl) vs 0.76 mmol.l-1 (P40). These results show that beta 2-adrenoceptor blockade by atenolol is a dose-dependent phenomenon, which may be measured by the attenuation of salbutamol-induced hypokalaemia. However, beta 2-adrenoceptor blockade by atenolol 200 mg was less than that by propranolol 40 mg. The glucose response to salbutamol was only partially blocked by propranolol and may therefore not be suitable to assess beta 2-adrenoceptor antagonism.