In this population-based case-control study of adult-onset asthma, severe asthma was associated with male sex, smoking, NERD, comorbidity, and the number of siblings. Noteworthy, associations for smoking, NERD and comorbidity were found in a sensitivity analysis based on another asthma severity definition, indicating that these associations were robust to the asthma severity definition. The effects of these risk factors seem to be additive; each additional risk factor gradually increases the risk of severe asthma.
In our study population the prevalence of severe asthma was 7.4% of patients with adult-onset asthma. This is in line with previous observations, in which the prevalence of severe asthma has been estimated to vary between 5–10% of patients with asthma (8). A Finnish single-center (central hospital) cohort of unselected patients with adult-onset asthma estimated that 5.9% fulfilled ERS/ATS severe asthma definition (9). This lower proportion could be explained by slightly different definition of severe asthma than in our study.
Our results reinforced previous findings regarding the impact of NERD, smoking and comorbidities in severe adult-onset asthma. A systematic review identified 27 publications, in which the prevalence of NERD among asthmatics was about 7%, and it was the highest among severe asthmatics (17). Korea Severe Asthma Registry analysis (n = 489) showed that individuals with severe asthma (including early- and adult-onset) have comorbidities such as allergic rhinitis (59%), atopy (39%), Aspirin hypersensitivity (14%) (39). Severe asthma and/or NERD has been shown to be associated with Type2 inflammation (11). Our findings could support that NERD, often characterize by a Type2 inflammation in literature (40), is one of the important risk factors of severe adult-onset asthma. Consistently with the literature, smoking and comorbidities were important independent factors of severe adult-onset asthma in our study. The deleterious effect of smoking in subjects with asthma has well been demonstrated in the literature, with decreased lung function (37), increased asthma severity (41), and risk of mortality (8). A cohort of Finnish middle-age asthmatics (including early- and late-onset) (n = 529) showed that 8% of the asthmatics, with more severe asthma and comorbidities, had poorer Work Ability Score during 10-year-follow-up (1).
Our study identified that independently of age and other factors, the presence of ≥ 2 siblings was a risk factor of severe adult-onset asthma. Professional training was associated with a lower risk of severe asthma, although the association was borderline significant in the multiple logistic regression. In our study population (who was born between 1904 and 1966), it could be speculated that the presence of ≥ 2 siblings could reflect poorer early living conditions predisposing to lower SES also in adulthood (42), which thus may have an impact on asthma self-care behavior (43). A study of elderly French women (n = 2258) showed that a low educational level (11%) was associated with an increased risk of uncontrolled asthma (including early- and late-onset cases) (15). In agreement with the hygiene hypothesis, number of siblings has been suspected to protect against the development of childhood asthma (44) and other atopic diseases (45). On the other hand, number of siblings might be a risk factor for asthma and lower lung function because it might lead to increase contact with pathogens that cause lower respiratory infections (46) and may lead to exacerbated asthma especially in genetically predisposed individuals (47).
Regarding gender, the main analysis showed a higher risk of severe adult-onset asthma in men as compared to women, but the sensitivity analysis using the OCS-based definition did not. This could reflect that male asthmatics reporting more difficult symptoms, could have OCS-resistant asthma such as smoking-related inflammation (48).
Our study did not detect an association between self-reported history of severe childhood infection(s) and severe adult-onset asthma, which might be due to the fact that childhood infections increase risk of exacerbated childhood onset asthma (47), more than severe adult-onset asthma. Another possible explanation might be related to measurement error/cohort effect as during the decades when our study population was born, the treatment of childhood infections has changed due to increased availability of public children´s counseling, doctors and antibiotics. There is no or little previous evidence of childhood infections on severe adult-onset asthma. In Tasmanian Longitudinal Health Study (n = 7312) a history of pneumonia before age of 7 year was ascertained from parents and measles, rubella, mumps, chickenpox, diphtheria, and pertussis from school medical records (36). Greater infectious diseases load was negatively associated with persisting asthma at all ages (36).
Presence of allergic disease(s) (AR and/or AC and/or AD) was not associated with the risk of severe adult-onset asthma in our study. In terms of mortality among asthmatics and matched controls, our previous study has shown that presence of AR or AC did not explain excess mortality among asthmatic adults (7), which is in line with our current findings. Overall, it is likely that asthma with allergic multimorbidity represent phenotypes that considerably differ from asthma alone in terms of mechanisms, severity and prognosis. Further studies in younger populations are needed as we demonstrated earlier that the association between allergic multimorbidity and asthma differs with age, with a stronger association observed in the youngest (24).
Although there is growing evidence that early-life factors play a role in the development of asthma (i.e. parental smoking, infection, nutrition, rural environment) (49), whether these early-life factors are associated with severe adult-onset asthma remain to be addressed. In our study, growing in the countryside, or parental asthma/allergy/smoking were not associated with severe adult-onset asthma in the main analysis, but the sensitivity analysis resting on the OCS-based definition of severe asthma showed a significant association with growing in countryside/farm. This could reflect OCS-sensitive inflammation (such as allergic fungal asthma). Other studies have shown that farm environments represent a source of fungi and increase asthma risk, and that sensitization to fungi might be related to severe asthma (20), and the severity of allergic fungal asthma can be decreased by OCS (50). Our previous analysis discovered association between sensitization to Aspergillus fumigatus and asthma in adult population (51).
Previous studies have shown that obesity increase the odds of a more persistent and severe asthma phenotype (52), and that obesity-associated severe asthma may represent a distinct clinical phenotype (53). Yet, we did not detect an association between BMI and severe adult-onset asthma in the main analysis, but a trend for a positive association was observed when using the medication-based definition of severe asthma.
This study presents several strengths, among which the outcome definition. In this population-based study, a specific definition of asthma was used, based on lung-function test confirmed doctor-diagnosed asthma. In addition, asthma is a spectrum of inflammatory lower airway conditions with variable obstruction, and heterogeneous background. Many epidemiological studies aiming at identifying risk factors for asthma did not consider this heterogeneity in the disease expression, which may affect interpretation and comparison of results between studies. In this study, we look for risk factors of severe adult-onset asthma, a specific phenotype associated with poor prognosis, with a definition of severity combining several domains of the disease. A further strength relies in the analytical approach consisting in estimating the joint effect of multiple risk factors and in addressing the robustness of the results to the definition of severe asthma. We chose two severe asthma definitions, based on data available in order to test the robustness of our findings to the definition of severe asthma. Although only 41% of cases had severe asthma by both Sev-Q and Sev-OCS definitions, these two severe asthma definitions showed consistency for three risk factors (smoking, comorbidities and NERD), supporting the role of these factors in severe asthma.
The weaknesses of our study include the limited statistical power of multivariable analyses. Due to cross-sectional design we were not able to evaluate the causal direction of associations, however because asthma was recently diagnosed (past year) given to the study design, covariates should have occurred before development of adult-onset asthma. We acknowledge that we lacked the lung function test data as an additional objective measurement of severe asthma and that a small portion of asthmatics might had childhood-onset asthma that relapsed in adulthood. In addition, a memory bias in the report of risk factors might have occurred. Finally, the associations reported could be partly biased by residual for confounding, either due to missing potential confounders in the regression model (such as occupational exposure), or due to limited accuracy in the assessment of some independent variables (i.e. the smoking variable does not consider the amount and the duration of smoking). Thus, before extrapolating our results, replication studies are needed.