The association between inhaled fenoterol and death from asthma has been investigated further by studying 112 asthma deaths (cases) during 1981-7 in patients aged 5-45 years who had been admitted to a major hospital for asthma during the 12 months before death. Two age matched control groups were chosen. Control group A comprised 427 patients who had been admitted to hospital for asthma during the calendar year that the corresponding death occurred and who had also had a previous admission for asthma in the previous 12 months. Control group B comprised 448 patients admitted to hospital for asthma during the calendar year in which the admission of the corresponding case occurred. The inhaled fenoterol odds ratio was 2.11 (95% confidence interval (CI) 1.37-3.23, p less than 0.01) when group A was used as the control (the approach used in previous studies), and 2.66 (95% CI 1.74-4.06, p less than 0.01) with group B as the control (the approach recommended by critics of previous studies). Markers of chronic asthma severity were associated with asthma death when control group B was used, but not when control group A was used (which indicates that these markers were indirectly matched for when control group A was used). Information was also collected on various markers of acute asthma severity and prescription of psychotropic drugs, but it was found that these were not important confounders. These findings address the major criticisms of previous case-control studies of this issue, and add support to the hypothesis that inhaled fenoterol increases the risk of death in patients with severe asthma.