Slow-onset asthma deaths have more eosinophils and health care utilization than rapid-onset deaths

Patients with fatal asthma have been hypothesized as representing two distinct subgroups according to the onset of symptoms prior to death.

To determine if these two groups may be reliably identified and determine if they differ clinically and pathologically.

Patients with autopsies, 2-34 years of age, dying from asthma were classified slow-onset (Type 1) or rapid-onset (Type 2). The consistency of classification by two independent observers was determined. Clinical and pathologic data was compared between groups.

Among 37 subjects there were 21 Type 1 and 16 Type 2 mortalities. Inter-observer agreement was good; kappa 0.63 (95% CI 0.38, 0.87). Subjects did not differ by age, race, sex, presence of obesity, insurance status, or use of corticosteroids. Type 1 mortalities were hospitalized more (2.5+1.3 versus 0.1+0.1; p=0.048) and made more emergency room visits (10.5+4.9 versus 0.8+0.2; p=0.023) in the year prior to death than Type 2 mortalities. At autopsy, bronchial eosinophils (p<0.025) and bronchiolar basement membrane thickening (p<0.05) were more predominant in Type 1 mortalities compared with Type 2. Airway wall neutrophils were uncommon in both groups.

Good inter-observer agreement exists in classifying patients dying from asthma according to whether the onset of symptoms was slow or rapid in onset. Slow-onset patients had a predominance of eosinophils and basement membrane thickening and higher health care utilization. These differences support the validity of classifying asthma mortalities into these distinct subgroups.