Background Asthma mortality rates have increased in Canada and worldwide. Within Canada, the highest rates were seen in the prairie provinces. Objective The objective was to determine risk factors for fatal asthma by comparing those who died of an acute exacerbation with those who attended an emergency department for treatment of asthma. Methods The case-control study included all deaths from asthma among those aged 5 to 50 years in Alberta, Saskatchewan and Manitoba from November, 1992 through October, 1995 (cases). The 35 fatalities were matched to 209 controls by age, gender, time of the index event and residence. Results Cases were more likely than controls to have had severe asthma, an unscheduled physician visit in the past year, a past hospitalization for asthma, and to have been intubated. Both groups reported frequent, regular asthma symptoms. Beta-agonist bronchodilator use was more common among cases, as was use in excess of prescribed amounts. Use of inhaled steroids did not differ between groups. Prior to the index event controls were more likely to report a cold or flu (OR = 0.27; 95% CI: 0.10 to 0.72) and that medications were "not working" (OR = 0.30; 95% CI: 0.12 to 0.71). Cases were more often sad and depressed (OR = 2.88; 95% CI: 1.03 to 8.05). Time between onset/recognition of symptoms and the event was significantly shorter for cases than controls. Conclusions Both groups tolerated high levels of regular symptoms, suggesting poor management. Opportunities for intervention existed for both groups near the time of the event. The short time between recognition of symptoms and death suggests patients at increased risk should monitor their condition closely and take action in response to predetermined criteria.