Asthma mortality increased sharply in New Zealand in 1977, prompting a national investigation into circumstances of asthma deaths. Subsequent observations of improved asthma control in subjects withdrawn from regular beta2-agonist treatment raised the question of whether asthma severity and, therefore, mortality could relate to frequent beta-agonist use. A randomized controlled trial of regular inhaled fenoterol versus as-needed bronchodilator use showed worsened asthma control during regular treatment despite concomitant use of inhaled corticosteroids. Assessment of these findings led to delay in the publishing of the American Asthma Guidelines, which were modified to suggest caution in using beta2-agonist treatments. Simultaneously, case control studies in New Zealand suggested that prescription of fenoterol was a substantial risk factor for asthma mortality. The causal association was hotly debated, but increasing evidence pointed to an adverse effect of fenoterol on asthma severity and, hence, mortality. This was supported by dramatic decreases in both morbidity and mortality when fenoterol was effectively withdrawn from use in New Zealand. The link between worsening asthma morbidity and mortality, and the use of potent short-acting beta2-agonists fulfills the Bradford Hill criteria for attributing causality. Application of evidence from randomized, controlled trials of short-acting beta-agonist use has led to a major shift in therapy in asthma to the recommendation of as-needed use only of short-acting beta-agonists and decreased patient reliance on regular bronchodilator therapy.