If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Medical School, Faculty of Health and Medical Sciences, The University of Western Australia, Perth, AustraliaDepartment of Pulmonary Physiology and Sleep Medicine, Sir Charles Gairdner Hospital, Perth, Australia
Childhood asthma increases respiratory morbidity in adulthood.
•
There is no impact of childhood asthma on all-cause mortality in adulthood.
•
Long term impact of adequately treated childhood asthma should be investigated.
Abstract
Background
Long-term childhood asthma studies that investigate adult outcomes other than lung function are lacking. This study examines the associations of childhood asthma and the occurrence of respiratory events and all-cause mortality in adulthood.
Methods
A cohort of 4430 school children (aged to 17 years) who attended the Busselton Health Study between 1967 and 1983 were analysed. Self-reported history of asthma was determined using questionnaires. Participants were followed until 2014 for respiratory disease-related events (hospital admissions or death) and all-cause mortality using the Western Australia Data Linkage System. Cox regression models were used to investigate the impact of childhood asthma on respiratory events and all-cause mortality in adulthood. A subgroup of 2153 participants who re-attended a survey in young adulthood was also analysed.
Results
A total of 462 (10%) of the cohort had childhood asthma. During follow-up 791 participants experienced a respiratory event and 140 participants died. Childhood asthma was associated with an increased risk of respiratory events in adulthood (unadjusted HR 1.84, 95% CI 1.52 to 2.23; P < 0.0001). The result remained significant after adjusting for adult-onset asthma, FEV1, body mass index, smoking, dusty job, hay fever, and respiratory symptoms (adjusted HR 1.68, 95% CI 1.07 to 2.64; P = 0.0247). Childhood asthma was not associated with all-cause mortality in adulthood (unadjusted HR 1.08, 95% CI 0.63 to 1.84; P = 0.7821).
Conclusion
Childhood asthma is associated with increased risk of respiratory disease-related hospital admissions and death but not all-cause mortality in adulthood.
]. Childhood asthma is characterised by chronic airway inflammation and structural remodelling, which results in irreversible lung function reduction in adulthood [
], which can provide prognostic information for patients, parents and clinicians.
An Australian cohort study of 346 participants found that children with severe asthma at age 7 years had 32 times higher risk of developing chronic obstructive pulmonary disease (COPD) by age 50 compared with non-asthmatics [
]. Another Australian cohort study of 5170 participants found that children with asthma at age 7 years had double the risk of developing “chronic bronchitis” (usual cough and phlegm) in middle age compared with non-asthmatics [
]. A Japanese case-control study of 700 participants concluded that childhood asthma was associated with 3 times higher risk of COPD between ages 50 to 75 [
Childhood asthma studies that investigate adult outcomes beyond lung function, such as overall respiratory disease morbidity and all-cause mortality, are lacking. This study examined the associations of childhood asthma with the occurrence of respiratory disease-related events (hospital admissions or death) and all-cause mortality in adulthood for children in the Shire of Busselton in Western Australia.
2. Methods
2.1 Study design and population
We analysed asthma and confounder data on a cohort of school-aged children who participated in the Busselton Health Study (BHS). The BHS consists of multiple cross-sectional surveys of adults and school children in the Shire of Busselton in Western Australia (WA). The BHS aims to collect data on common diseases and establish biochemical reference ranges [
]. The BHS invited, via schools, all school children to participate in cross-sectional surveys in 1967, 1970, 1973, and 1983. Signed parental consent was required and response rates were around 90% [
]. A follow-up survey of all survivors from previous child surveys was conducted in 1994/95 and the response rate was 49%. Using the WA Data Linkage System, hospital admission records (from January 1980) and death records (from January 1969) for surveyed individuals were obtained. The WA Hospital Morbidity Data System includes all WA public and private hospital admission records while the WA Death Register includes all death records among WA residents [
]. To confirm participants were still in WA, hospital records and electoral registration were checked and where there was evidence of departure from WA follow-up was censored at that time. We excluded participants who were older than 17 years in their childhood survey and those who died before age 18 years or before January 1, 1980. The study was approved by the Human Research Ethics Committees of the University of Western Australia (Reference RA/4/1/6682) and WA Department of Health (Project 2011/60).
2.2 Asthma variables
Childhood asthma was defined by an affirmative response by the children or parents to “Has this child ever been diagnosed as having asthma by a doctor?” in any of the childhood surveys.
2.3 Other covariates
The childhood questionnaires were not identical: to overcome this limitation, we also analysed a subgroup of 2153 children who returned as adults (aged above 18 years) in the 1994/95 follow-up survey where some potential confounders were measured. These potential confounders were derived from previous studies [
]. Adult-onset asthma was defined as an affirmative response to “have you ever been told by a doctor that you have asthma” in the 1994/95 survey but did not have childhood asthma. Spirometry tests were performed in the 1994/95 survey according to American Thoracic Society standards and the forced expiratory volume in 1 second (FEV1) was measured [
]. Participants’ height and weight, wearing only light clothing without shoes, were measured by staff using stadiometer and scales. Body mass index was derived from weight divided by height squared. Smoking status was categorised as never smoked, ex-smoker, < 15 cigarettes per day, or ≥ 15 cigarettes per day. Alcohol consumption (one week drinking diary) was categorised as never drink, ex-drinker, ≤ 140 g per week, or > 140 g per week. Self-reported hours per week of physical activity at work, home, and leisure-time (and total) were assessed. History of working in any job that had dust or fumes exposure was recorded. Allergy to common inhaled allergens (grasses, moulds, animal proteins and house dust mites) (Hollister-Stier, WA, USA) was defined using forearm skin prick test as any wheal of 3 mm greater than negative control (10 mg/mL histamine). History of hay fever was recorded if participants had ever been told by a doctor that they had hay fever or allergic rhinitis. A respiratory symptoms variable, derived from the ECRHS II study [
], was calculated as the number of respiratory symptoms or history present at the 1994/95 survey (maximum of 7): dyspnoea, cough or sputum production, history of doctor-diagnosed bronchitis, pneumonia, pleurisy, COPD, and being on respiratory medications.
2.4 Outcome variables
Hospital admissions (from 1980) and death records (from 1969) to 2014 were used to define respiratory event and all-cause mortality outcome variables. Hospital admissions diagnoses and causes of deaths were coded using the International Classification of Diseases (ICD). Respiratory event was defined as either hospital admission with primary diagnosis of, or death from, a respiratory disease (ICD-9 460–519, ICD-10 J00-99). Time to (first) respiratory event was defined as time from the later of either age 18 or January 1, 1980 (or in the subgroup, time from 1994/95 survey attendance) to first respiratory event.
Time to all-cause mortality was defined as time from age 18 or January 1, 1980, whichever is later, to death. Given the limited number of deaths during follow-up, only the full cohort was analysed for this outcome.
2.5 Analysis
Analyses were conducted in SAS 9.4 (SAS Institute, NC, USA). Participants’ characteristics were presented as means (±SD) or percentages as appropriate. Chi-square and ANOVA tests were used to compare characteristics across groups. A P-value of < 0.05 was regarded as significant.
Incidence rates of respiratory events and deaths per 1000 person-year of follow-up from adulthood (as defined above) were calculated and compared using Poisson regression. Unadjusted Kaplan-Meier survival curves with age as the time scale were calculated to show overall differences between childhood asthmatics and non-asthmatics. By applying the PHREG procedure in SAS, these calculations and the Cox regression models (see below) recognised left truncation at the higher of age 18 or age in 1980 (full cohort for respiratory event), at age in the 1994/95 survey (subgroup for respiratory event), and at age 18 (all-cause mortality).
For the full cohort and subgroup, Cox proportional-hazard models were used to estimate hazard ratios (HRs) and 95% confidence interval (CI) for childhood asthma in relation to time to first respiratory event and all-cause mortality. HRs after adjusting for birth year, age at child survey and gender (model 1) were also calculated. For the subgroup, covariates measured in 1994/95 were assessed for significant relationships with childhood asthma and with time-to-respiratory event, after adjusting for birth year, age at child survey and gender. Covariates were deemed potential confounders if they had P-value of < 0.10 in both the Cox regression analysis with respiratory events and the chi-square or ANOVA tests with childhood asthma. Identified potential confounders were incorporated in the final multivariate model, in addition to body mass index and smoking given their biological significance to respiratory morbidity (model 2). Proportional-hazard assumptions were confirmed using the Schoenfeld residual plots.
3. Results
There were 4528 school children who attended at least one of the childhood surveys. After excluding 25 children who only attended surveys after age 18 years or died before age 18 years, 66 children with incomplete asthma data, and seven participants who died before January 1, 1980, there remained 4430 children of whom 2153 attended the 1994/95 survey (Supplementary Fig. S1). Hospital admissions and deaths records were linked for all 4430 participants.
The full cohort had 462 (10%) asthmatic and was 50% male with a mean age of 11.8 (±2.9) years at childhood assessment (Table 1). The 1994/95 subgroup included 240 (11%) asthmatic and was 46% male with a mean age of 12.1 (±2.8) years at childhood assessment (Table 2). The subgroup mean age at the 1994/95 assessment was 32.4 years and 255 (13%) participants reported adult-onset asthma (Table 2). Adults with childhood asthma had a higher prevalence of working in a dusty job (P = 0.0550), lower FEV1 (P < 0.0001), having allergy (P < 0.0001), having a history of hay fever (P < 0.0001) and had a higher number of respiratory symptoms (P < 0.0001, Table 2).
Table 1Characteristics of full cohort by childhood asthma status and overall (n = 4430).
Characteristics
Childhood asthma
Yes (n = 462, 10%)
No (n = 3968, 90%)
P-value
Overall
Male n (%)
281 (61%)
1931 (49%)
<0.0001
2212 (50%)
Characteristics at child survey
Birth year
1964 (7.6)
1963 (7.4)
0.0632
1963 (7.4)
Age (years)
11.8 (2.9)
11.8 (2.9)
0.5704
11.8 (2.9)
Follow-up
Age at start of follow up for event outcome (years)
During the 119,245 person-years of follow up, 791 participants experienced a respiratory event (Supplementary Table S1). Kaplan-Meier plots for respiratory disease-related events (hospitalisation or death) are displayed in Fig. 1 for the full cohort and Fig. 2 for the subgroup. There was a significant association between childhood asthma and respiratory events for the full cohort (P < 0.0001) and the subgroup (P = 0.0003).
Fig. 1Kaplan-Meier plot of respiratory disease-related events (hospitalisation or death) according to childhood asthma status for the full cohort, accounting for left truncation.
Fig. 2Kaplan-Meier plot of respiratory disease-related events (hospitalisation or death) by childhood asthma status for the 1994/95 subgroup, accounting for left truncation.
In the full cohort, childhood asthma was associated with respiratory events in adulthood (unadjusted HR 1.84, 95% CI: 1.52–2.23, P < 0.0001). After adjusting for birth year, age at the child survey, and gender, the HR of childhood asthma for respiratory events was 1.91 (95% CI: 1.58–2.31, P < 0.0001).
In the 1994/95 subgroup, there was a significant unadjusted association between childhood asthma and respiratory events (HR 1.81, 95% CI: 1.31–2.51, P = 0.0003, Table 3). After adjusting for birth year, age at child survey, and gender, the association remained significant (HR 1.89, 95% CI: 1.36–2.62, P = 0.0001, Table 3).
Table 3Risk of respiratory events in adulthood by childhood asthma and covariates in the 1994/95 subgroup.
Hazard ratios (95% CI) for respiratory events
Unadjusted (n = 2153)
Model 1‡ (n = 2153)
Model 2§ (n = 1358)
Childhood asthma
1.81 (1.31–2.51)***
1.89 (1.36–2.62)***
1.68 (1.07–2.64)*
Birth year
0.96 (0.93–0.99)*
0.99 (0.95–1.03)
Age at child survey
1.02 (0.96–1.07)
1.05 (0.98–1.12)
Male gender
1.32 (1.03–1.70)*
1.28 (0.90–1.81)
Adult-onset asthma
1.66 (1.07–2.56)*
FEV1
0.99 (0.98–1.00)
Body mass index
1.02 (0.98–1.06)
Smoking
Never
1 (ref)
Ex
1.36 (0.94–1.96)
< 15 cigarettes/day
1.15 (0.69–1.93)
≥ 15 cigarettes/day
1.16 (0.70–1.90)
Dusty job
1.41 (1.00–1.98)*
Hay fever
1.20 (0.87–1.66)
Number of respiratory symptoms
1.13 (0.96–1.33)
P-value of childhood asthma
0.0003
0.0001
0.0247
*p < 0.05; **p < 0.01; ***p < 0.001.
‡Model 1 adjusted for birth year, age at child survey, and gender.
§Model 2 further adjusted for adult-onset asthma, FEV1, body mass index, smoking, dusty job, hay fever and number of respiratory symptoms in addition to Model 1.
After adjusting for birth year, age at child survey and gender, respiratory events were associated with the following covariates: adult-onset asthma (P < 0.0001), FEV1 (P = 0.0004), body mass index (P = 0.0206), alcohol consumption (P = 0.0893), physical activity at work (P = 0.0365), total physical activity (P = 0.0348), history of working in a dusty job (P = 0.0381), history of hay fever (P = 0.0003), and number of respiratory symptoms (P < 0.0001, Supplementary Table S2). Adult-onset asthma, FEV1, history of working in a dusty job, history of hay fever and number of respiratory symptoms were also associated with childhood asthma (Table 2) and were hence regarded as potential confounders. Further adjustment for these five potential confounders, body mass index and smoking (Table 3) reveal a significant relationship between childhood asthma and respiratory events (P = 0.0247). The adjusted HR for respiratory events in subjects with childhood asthma, compared with those without, was 1.68 (95% CI: 1.07–2.64).
3.2 All-cause mortality
During the 142,538 person-years of follow up, 140 participants died. Supplementary Table S3 summarises the all-cause mortality incidence rates. Kaplan-Meier plot for all-cause mortality (Fig. 3) demonstrated there was no significant association between childhood asthma and all-cause mortality (P = 0.7821).
Fig. 3Kaplan-Meier plot of all-cause mortality according to childhood asthma status for the full cohort, accounting for left truncation.
Childhood asthma was not associated with risk of all-cause mortality in adulthood (unadjusted HR 1.08, 95% CI: 0.63–1.84, P = 0.7821). After adjusting for birth year, age at child survey, and gender, the relationship remained insignificant (HR 1.02, 95% CI: 0.60–1.75, P = 0.9316).
4. Discussion
This cohort study of 4430 individuals showed asthmatic children had an 84% increased risk of respiratory events (hospitalisation or death) in adulthood. We found no impact of childhood asthma on all-cause mortality in adulthood. This is the first report describing the associations between childhood asthma and overall respiratory morbidity, and all-cause mortality in adulthood.
Other longitudinal studies have consistently shown that childhood asthma increases the risk of fixed airflow obstruction (“COPD”) in adulthood by two to thirty-two times [
]. Studies investigating the childhood asthma impact on adult respiratory diseases beyond COPD are lacking. Nevertheless, asthmatic children are well-known to have a greater hospitalisation frequency compared with non-asthmatics [
]. This is reflected in our cohort, where the childhood asthma group had a lower FEV1 in early adulthood compared with non-asthmatics. Asthmatics are prone to exaggerated airway inflammatory response to infection which may trigger an asthma attack, making them susceptible to hospitalisation [
Could the relationship between childhood asthma and respiratory events be explained by childhood asthma persisting into adulthood? Despite the various definitions of remission, it is estimated that 11%–64% of children with asthma achieve remission by adulthood [
Gina 2019: a fundamental change in asthma management: treatment of asthma with short-acting bronchodilators alone is no longer recommended for adults and adolescents.
] have argued for the use of combination corticosteroid-β-agonist inhaler over bronchodilator-only therapy. Although likely, it is unknown if improved asthma control during childhood reduces asthma-related events in adulthood. Our results demonstrate a lasting impact of childhood asthma on respiratory morbidity and reinforce the need to shift the focus beyond symptomatic management to prevention or remission of asthma.
Consistent with previous asthma studies, our childhood asthma group had more boys than girls [
]. We also found that adult-onset asthma has an independent association with respiratory events. A Canadian study found that health care claims from respiratory diseases other than asthma were double in asthmatic adults compared with non-asthmatics, due to the higher rate of respiratory comorbidities [
]. Our results demonstrated that occupational exposure to dust or fumes had an independent impact on respiratory events, in keeping with evidence that suggests occupational exposure to dusts and fumes increases the incidence of COPD [
Occupational exposure to dusts, gases, and fumes and incidence of chronic obstructive pulmonary disease in the Swiss Cohort Study on Air Pollution and Lung and Heart Diseases in Adults.
Am. J. Respir. Crit. Care Med.2012; 185: 1292-1300
]. Our results reflected that smokers had higher risk of respiratory events compared with non-smokers (Supplementary Table S2), consistent with evidence that suggest smoking increases the risks of lung cancer and COPD [
Due to the relatively small number of deaths, this study sample had 70% power to detect a HR of 2.0 for all-cause mortality. The statistically insignificant finding for all-cause mortality means that a smaller effect size remains possible. Regardless, parents can be reassured that childhood asthma does not have clinical significance impact on death in adulthood. Indeed, a study found that the death incidence from childhood asthma was only 0.19 per million person-years [
]. The highest mortality rate from asthma occurs in elderlies beyond age 55 years, who have 5 times the mortality risk compared with younger asthmatics [
]. Our study was unable to capture that effect despite having a relatively long follow-up.
A strength of this study was its long-term follow-up extending 30 to 45 years from school-age. Endpoints were defined objectively, making measurement bias unlikely. Attrition or selection bias were minimised as the WA Data Linkage System identified all respiratory-related hospital admissions or deaths and all-cause mortality for all participants. This study was conducted on a community population, enhancing the generalisability of the results.
This study has several limitations. Self- or parental-reported asthma status is subject to recall bias. Misclassification of asthma status however would be random and would have biased the estimates towards the null. Doctor-diagnosed asthma was relied on to define childhood asthma status. A recent opinion article and the GINA guidelines [
Gina 2019: a fundamental change in asthma management: treatment of asthma with short-acting bronchodilators alone is no longer recommended for adults and adolescents.
] have argued for objective asthma diagnosis using spirometry and departure from the no-test culture, which has contributed to over-diagnosis of asthma [
]. Furthermore, we did not have socioeconomic status information to assess its confounding impact. Lower socioeconomic status has been linked to higher prevalence and mortality of COPD, as well as increased asthma hospitalisation [
]. However, standardised methods to incorporate socioeconomic status in determining respiratory disease prognosis are lacking.
This study did not have data for estimation of severity and duration of childhood asthma. Ideally a dose-response relationship of childhood asthma severity should be examined, however the classification of asthma severity is still a debatable subject [
], and it remains confounded by treatment. Lastly, we did not have accurate data available for treatment. Many of the children were first seen before the introduction of inhaled corticosteroids and long-before the introduction of long acting bronchodilators into routine asthma. In this retrospective longitudinal analysis, compliance data were not recorded and degree of asthma control (“well-managed” and “poorly-managed”) could not be assessed. It should be noted, however, that in general the severity and therefore treatment requirements of childhood asthma change little over many years [
]. The number of respiratory events in this cohort precluded analysis of the childhood asthma impact on specific respiratory diagnosis, which can be a focus on future research.
5. Conclusions
Children with asthma are at increased risk of respiratory disease-related events (hospitalisation or death) but not all-cause mortality in adulthood. Adult-onset asthma and occupational exposure to dusts are indicators of respiratory disease morbidity. Future research should examine the dose-response relationship and adequate treatment of childhood asthma on specific respiratory morbidities beyond fixed airflow obstruction (COPD).
Funding
The 1994/95 follow-up survey was supported by Healthway, Western Australia.
CRediT authorship contribution statement
Christopher A.C.M. Ng: Conceptualization, Methodology, Software, Formal analysis, Data curation, Writing - original draft, Writing - review & editing, Visualization, Project administration. Matthew W. Knuiman: Conceptualization, Methodology, Software, Investigation, Validation, Formal analysis, Writing - original draft, Writing - review & editing, Supervision. Kevin Murray: Conceptualization, Methodology, Writing - review & editing, Supervision. Mark L. Divitini: Software, Validation, Investigation, Resources, Data curation, Visualization. Arthur W. (Bill) Musk: Investigation, Writing - review & editing, Supervision. Alan L. James: Investigation, Writing - review & editing, Supervision, Funding acquisition.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We thank the Busselton Population Medical Research Institute for permission to access Busselton Health Study data and the WA Data Linkage Unit for facilitating access to the linked data on hospital admissions and deaths. We thank the Busselton population for their long-standing support and participation in the Busselton Health Study.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
Gina 2019: a fundamental change in asthma management: treatment of asthma with short-acting bronchodilators alone is no longer recommended for adults and adolescents.
Occupational exposure to dusts, gases, and fumes and incidence of chronic obstructive pulmonary disease in the Swiss Cohort Study on Air Pollution and Lung and Heart Diseases in Adults.
Am. J. Respir. Crit. Care Med.2012; 185: 1292-1300
30235-3&id=gr1.jpg){kind=link}
30235-3&id=gr2.jpg){kind=link}
30235-3&id=gr3.jpg){kind=link}