Adenosine 5'-monophosphate (AMP) bronchoprovocation reproduces the lung function abnormalities that occur spontaneously during acute asthma and detects peripheral airway inflammation better than direct bronchoconstrictive agents. Pulmonary gas exchange disturbances may reflect changes in small airways related to airway inflammation rather than bronchoconstriction alone. The present authors investigated whether AMP induced a greater imbalance in the ventilation/perfusion ratio than methacholine (MCh), at an equivalent degree of bronchoconstriction, with and without salbutamol pre-medication. In total, 36 asthmatics were studied in three randomised, double-blind, crossover studies: 1) before and after AMP or MCh; 2) before and 30 min after salbutamol or placebo, followed by AMP; or 3) MCh challenge. Sputum was collected before and 4 h post-challenge. Compared with MCh, AMP provoked similar pulmonary gas exchange abnormalities at an equivalent degree of intense bronchoconstriction (forced expiratory volume in one second decrease of 28-44%). While salbutamol blocked AMP- or MCh-induced bronchoconstriction, arterial oxygen tension (P(a,O(2))) and alveolar-arterial oxygen tension difference (P(A-a,O(2))) disturbances induced by AMP and MCh were only partially blocked (P(a,O(2)) by 46 and 42%, respectively; P(A-a,O(2)) by 58 and 57%, respectively). Compared with MCh, AMP increased the percentage of neutrophils (mean+/-se increased from 28+/-4% to 38+/-4%), but this increase did not occur after salbutamol pre-treatment. Both adenosine 5'-monophosphate and methacholine induced similar peripheral airway dysfunction. The fully inhibited adenosine 5'-monophosphate-induced neutrophilia with residual hypoxaemia observed after salbutamol treatment is probably related to beta(2)-agonists acting on both bronchial and pulmonary circulation.