Aims To examine whether systemic beta (2)-adrenoceptor responses, such as tachycardia, tremor and hypokalaemia, can be used as a surrogate for the 20 min pharmacokinetic profile of inhaled salbutamol. Methods A retrospective analysis of eight separate published studies in healthy volunteers was performed, each with an identical protocol evaluating the early lung absorption profile of a nominal 1200 mug dose of salbutamol given by different inhaler devices. Peak postural finger tremor, plasma potassium and heart rate were assessed. Results We found the maximum (C-max) and average (C-av) plasma concentrations of salbutamol to be correlated (P < 0.0001) to change in plasma potassium (C-max r = 0.904; C-av r = 0.899) and tremor (C-max r = 0.875; C-av r = 0.857). No significant correlations existed between change in heart rate and C-max (r = 0.425) or C-av (r = 0.415). Conclusions Systemic <beta>(2)-adrenoceptor responses, in particular hypokalaemia and tremor, but not heart rate, appear to be good surrogates for evaluating the lung delivery of inhaled salbutamol. Consequently it is suggested that potassium or tremor responses may be used to evaluate the relative lung delivery of salbutamol from different inhaler devices.