Seventy Years of Asthma in Italy: Age, Period and Cohort Effects on Incidence and Remission of Self-Reported Asthma from 1940 to 2010.

BACKGROUND

It is well known that asthma prevalence has been increasing all over the world in the last decades. However, few data are available on temporal trends of incidence and remission of asthma.

OBJECTIVE

To evaluate the rates of asthma incidence and remission in Italy from 1940 to 2010.

METHODS

The subjects were randomly sampled from the general Italian population between 1991 and 2010 in the three population-based multicentre studies: ECRHS, ISAYA, and GEIRD. Individual information on the history of asthma (age at onset, age at the last attack, use of drugs for asthma control, co-presence of hay-fever) was collected on 35,495 subjects aged 20-84 and born between 1925-1989. Temporal changes in rates of asthma incidence and remission in relation to age, birth cohort and calendar period (APC) were modelled using Poisson regression and APC models.

RESULTS

The average yearly rate of asthma incidence was 2.6/1000 (3,297 new cases among 1,263,885 person-years). The incidence rates have been linearly increasing, with a percentage increase of +3.9% (95%CI: 3.1-4.5), from 1940 up to the year 1995, when the rates begun to level off. The stabilization of asthma incidence was mainly due to a decrease in the rates of atopic asthma after 1995, while non-atopic asthma has continued to increase. The overall rate of remission was 43.2/1000person-years, and it did not vary significantly across generations, but was associated with atopy, age at asthma onset and duration of the disease.

CONCLUSIONS

After 50 years of a continuous upward trend, the rates of asthma incidence underwent a substantial stabilization in the late 90s. Despite remarkable improvements in the treatment of asthma, the rate of remission did not change significantly in the last seventy years. Some caveats are required in interpreting our results, given that our estimates are based on self-reported events that could be affected by the recall bias.