Carbon monoxide (CO) can be detected in exhaled air and is increased in asthmatic patients not treated with corticosteroids. However, it is uncertain whether exhaled CO is related to severity of asthma. To study whether exhaled CO is related to severity of asthma in clinical courses, exhaled CO concentrations were measured on a CO monitor by vital capacity manoeuvre in 20 mild asthmatics treated with inhaled beta2-agonists alone, 20 moderate asthmatics treated with inhaled corticosteroids, and 15 stable asthmatics treated with high dose inhaled corticosteroids and oral corticosteroids once a month over 1 years. Exhaled CO concentrations were also measured in 16 unstable severe asthmatics who visited the hospital every 7 or 14 days for treatment with high dose inhaled corticosteroids and oral corticosteroids. The mean values of exhaled CO in severe asthma over 1 year were 6.7 +/- 9.5 p.p.m. (n = 31, mean +/- SD) and significantly higher than those of non-smoking control subjects (1.2 +/- 0.9 p.p.m., n = 20, P < 0.01). Exhaled CO concentrations in unstable severe asthmatics were significantly higher than those in stable severe asthmatics. However, exhaled CO concentrations in mild and moderate asthmatics did not differ significantly from those in non-smoking control subjects (P > 0.20). There was a significant relationship between the exhaled CO concentrations and forced expiratory volume in one second in all asthmatic patients. These findings suggest that exhaled CO concentrations may relate to the severity of asthma and measurements of exhaled CO concentrations may be a useful means of monitoring airway inflammation in asthma.