We found that the incidence of allergic asthma was highest in childhood and gradually decreased in older age whereas the incidence of non-allergic asthma peaked in late adulthood. Most subjects with childhood-onset asthma were allergic while most subjects with asthma onset after 40 years of age were non-allergic.
Childhood-onset asthma is commonly associated with allergy.[8,9] Our results supported this as close to 70% of responders with asthma diagnosed before 20 years of age had also allergic rhinitis in adulthood. It has been commonly recognized that asthma has its origin in childhood and the atopic early-onset asthma is the most important and widely recognized phenotype.[25,26] However, a recent U.S.-based study showed adult-onset asthma as being the dominant phenotype among women in middle age. In Finland during 2012–2013, 70% of new asthma diagnoses were made in adults indicating that adult-onset asthma is a clinically relevant phenotype. Nevertheless, studies on adult-onset asthma are still scarce. In our study, the combined incidences of allergic and non-allergic asthma were highest after middle age and this was mainly driven by new cases of non-allergic asthma.
In the current study, the proportion of allergic asthma among new cases decreased steadily with advancing age at asthma diagnosis. There are some previous results that are in line with our finding, but in those studies age at asthma diagnosis was classified into two or three classes instead of looking at the prevalence of allergy on a wide range of asthma onset age classes.[7,29] According to a European multi-center population-based study, atopy explained a minority (12–21%) of adult-onset asthma. A Swedish and a Dutch study both reported prevalence rates of around 45% for atopy in adult-onset asthma.[15,16] Warm et al. divided Swedish asthmatics into three age groups according to the age of asthma onset (≤ 6 years, 7–19 years and ≥ 20 years) and reported a decrease in the frequency of allergic sensitization in adulthood with increasing age of asthma onset (86, 56 and 26%, respectively). Our study results were similar as proportion of allergic rhinitis among new cases of asthma declined below 50% after the age of 35 years. To our knowledge this is the first study that shows the proportions of allergic and non-allergic asthma in consequent 10-year age groups at ages 0–69 years at asthma diagnosis.
Several studies using cluster analysis have also reported early-onset atopic asthma as a distinct phenotype.[3,6,30] The heterogeneity of phenotypes appears to increase with advancing age at asthma onset, resulting in the recognition of new phenotypes and risk factors especially in adult-onset asthma.[5,16,31] In our study, the responders with non-allergic asthma were more obese than allergic asthmatics. A phenotype of older, obese and less atopic women with frequent exacerbations and symptoms but at most only a moderate reduction in lung function has previously been reported.[3,5,6] Late-onset asthma is suggested to be more often non-allergic, severe and having a lower lung function than early-onset asthma.[7,32] Accordingly, a cluster of severe late-onset less atopic asthma with eosinophilic inflammation and male-dominance has been identified. A Finnish study with a follow-up of 12 years reported two adult-onset asthma phenotypes consisting predominantly of males. The first cluster had non-atopic males with moderate smoking history who developed persistent airflow limitation at follow-up but with the lowest number of uncontrolled asthma. The other cluster contained older men with a heavy-smoking history, poor lung function and mostly uncontrolled asthma. Late-onset asthma phenotypes with milder clinical picture has also been reported such as a mild, untreated and less atopic adult-onset asthma and a nonsmoking female-predominant cluster with good lung function and well-controlled/partially controlled asthma. Generally thought, phenotypes that present more severe or symptomatic disease are especially identified in late-onset asthma.[5,6,16,33]
The reason for the decline in the incidence of allergic asthma with increasing age at asthma diagnosis may be related to at least two factors. Firstly, atopic allergy often begins in childhood and early adulthood while non-allergic asthma may be related to cumulative exposure to irritating factors such as occupational exposures and smoking and thereby becomes evident only later in life span after sufficient exposure times. Indeed, the higher proportion of ex-smokers among the non-allergic asthmatics in our study would support this hypothesis. Another reason may be the cohort effect that may affect incidence when estimated from cross-sectional data. Those subjects who have lived long enough to be able to have got late-onset asthma, were born earlier and represent different cohorts with lower overall prevalence of allergic sensitization than in younger generations. Moreover, according to population-based studies, the prevalence of allergic sensitization in general decreases with increasing age due to low incidence and higher remission. Prevalence rates over 50% for allergic sensitization are reported among young adults compared to 26–39% in adults over 50 years of age.[18,20,34] A database search survey from the U.S. (2005–2006) showed a lower rate of allergic sensitization in asthmatics ≥55 years old when compared to 20–40 years old (65 and 75%, respectively). In the same survey, it was shown that if the analysis of asthma onset age and allergy is restricted to only subjects who were at least 55 years old, the difference in the frequencies of atopic sensitization between asthmatics with onset before or after 40 years of age was low (72% vs 63%, respectively), but the numbers of subjects were quite small (12 and 31, respectively) to make firm conclusions. A recent Swedish study reported a continued increase in the prevalence of allergic asthma in the last 20 years (from 5.0% in 1996 to 7.3% in 2016), which may reflect the overall increase in the prevalence of atopic sensitization in new generations. Accordingly, in our sample the prevalence of allergic rhinitis was significantly lower in the oldest age cohort. However, in all age cohorts of our sample, asthma diagnosed before the age of 40 years was more often allergic than non-allergic, and the difference between age cohorts was not statistically significant in this regard. In the future, when the currently young generations with higher prevalence of atopy grow old, also late-onset asthma may have higher proportions of allergic cases.
In the present study, a responder was defined to have allergic asthma if he had both physician-diagnosed asthma and physician-diagnosed allergic rhinitis, but we did not have any objective tests on allergic sensitizations. According to a Swedish study, 83% of persons aged 21–40 years with an atopic tendency (production of IgE antibodies against any allergen) had allergic rhinitis, concluding that allergic rhinitis is a good marker for allergic sensitization and clinical allergy. In another study, both allergic sensitization and allergic rhinitis were significantly associated with incidence of asthma in adulthood in bivariate analyses, but interestingly only allergic rhinitis was associated with asthma incidence in multivariate analysis. This indicates that a clinical allergy is a better predictor of asthma than atopic sensitization, and allergic rhinoconjunctivitis doubles the risk for incident asthma among adults. Furthermore, allergic conjunctivitis is often associated with allergic rhinitis and the term rhinoconjunctivitis is used. This was also observed in the present study as the sensitivity analysis for allergic and non-allergic asthma in association with the age at asthma diagnosis did not noticeably change when allergic conjunctivitis was included into the definition of allergy (see Additional files 1 and 2). We did not take allergic skin diseases into account when defining allergy as the definition of allergic dermatitis/eczema is not as uniform as respiratory tract allergic diseases and poses a higher risk for misdiagnosis of allergy.
The present study was cross-sectional in nature and we did not specifically inquire whether the responder had allergic rhinitis or other allergy symptoms at the time of asthma diagnosis. For this reason, responders with asthma and allergic rhinitis diagnosed at some point of their life might not have had allergic rhinitis at the time of their asthma diagnosis. However, it has been reported that both allergic and non-allergic rhinitis usually precede asthma onset in children and adults.[38,39]
Study limitations also include that no objective measurements of lung function to confirm asthma diagnosis were obtained. In Finland, a person with persistent asthma is entitled to special reimbursements for expenses of asthma medication if he/she has an objectively confirmed variable expiratory airflow limitation or bronchial hyperresponsiveness as determined in international asthma diagnostic guidelines. Therefore, most of the asthma diagnoses in Finland are based on lung function measurements and the reliability of the reported asthma diagnosis used in our study is probably high. There is a risk for recall bias as we requested a self-reported age at asthma diagnosis, which in Australia was estimated as most often inaccurate while in Sweden it has been estimated as most often accurate.
BMC Pulm Med. 2020;20(9) © 2020 BioMed Central, Ltd.