Childhood asthma is frequently perceived as a disease with uniform clinical pathways. This perception might be an oversimplification. The aim of the present study was to investigate the incidence and natural course of wheeze over the first 13 yrs of life and analyse the risk factors predicting wheeze at 11–13 yrs of age.
The Multicentre Allergy Study, a German birth cohort, recruited 1,314 children in 1990. Physical examinations, interviews on atopic diseases, immunoglobulin (Ig)E and lung function tests were performed up to 13 yrs of age.
Complete data on the course of wheeze were available for 441 children. It was found that incidence of wheezing declined with age. The first wheezing episode was reported by 29, 9 and 9% of participants at ≤3 (early wheezers), 3–6 (late wheezers), and >6 yrs (very late wheezers) of age, respectively. Wheezing at the age of 13 yrs was associated with parental atopy, and with IgE sensitisation to common allergens, elevated total IgE and exposure to high levels of indoor allergens in early life. All these associations were remarkably stronger among early wheezers than among early nonwheezers.
In conclusion, the relevance of an early expression of atopy as a predictor of wheezing at age 13 yrs declines with increasing age of wheezing onset.
Asthma and other wheezing disorders are among the most common childhood diseases 1. Understanding of the genetic, environmental and developmental factors of childhood asthma is essential for the development of tools for predicting and, hopefully, preventing recurrence of wheezing episodes and subsequent deficits in airway function growth 2, 3. Algorithms to predict asthma in wheezing children have been developed on the basis of observational longitudinal studies 4–7, with the most popular algorithm targeted to predict asthma in children who start wheezing before 3 yrs of age 4. However, many children start wheezing only after the age of 3 yrs; if atopy and other risk factors for asthma in these children are different from those predicting asthma among early wheezers, then additional algorithms should be formulated. The airways of children with early wheezing may be more susceptible to the negative consequences of an early sensitisation and exposure to allergens. Therefore, the present authors hypothesised that atopy and other risk factors may have a different impact on the natural history of wheezing disorders in children according to their propensity to wheeze. In order to test this hypothesis, data from the German Multicentre Allergy Study (MAS), a prospective observational birth cohort study followed up until 13 yrs of age 8, were analysed. On the basis of the present findings, a description of the natural course of wheezing up to 13 yrs of age in the MAS cohort is provided and a comparative analysis of the risk factors predicting wheeze at age 11–13 yrs in children both wheezing and not wheezing in the first 3 yrs of life is performed.
The German MAS recruited 1,314 infants born in 1990 in six German obstetric departments in five German cities (Dusseldorf, Mainz, Freiburg, Gaisach, Berlin). A detailed description of the stratified sampling scheme and study subjects is given elsewhere 8. Briefly, 499 newborns with presumed risk factors for atopy (cord blood immunoglobulin (Ig)E increased to ≥0.9 kU·L−1, at least two atopic family members, or both) and 815 newborns with none of these risk factors were included in the cohort. All children were followed up at 1, 3, 6, 12, 18 and 24 months, and then annually, within 4 weeks of the child's birthday, up to the age of 13 yrs. The present study was approved by the hospital’s ethics committee (University Children's Hospital, Munich, Germany) and written informed consent, at the time of enrolment, was obtained from the child’s parents.
Parental questionnaires and interviews
At each follow-up visit, parents either gave structured interviews to a study physician or completed postal questionnaires (at age 8, 9, 11 and 12 yrs). The main topics of these interviews were asthmatic and atopic symptoms and disease; questions on diet, development and psychological aspects were also included. From age 5 yrs onwards, questions relating to wheeze corresponded to the International Study of Asthma and Allergies in Childhood core questions 9. Foremost interest was in atopic symptoms and diseases. Furthermore, symptoms and diagnoses of other illnesses were assessed at every follow-up, as were maternal breast-feeding and feeding practices up to age 2 yrs. Parental smoking habits were assessed at age 1 month and pet keeping was assessed at age 3 months. Furthermore, the use of medication was assessed at every follow-up. Children regularly visited a physician independent of the study team.
Wheeze was defined as any parent-reported wheezing in the 12 months prior to any of the follow-ups. The definitions “early wheezers” and “late wheezers” were used to describe the children starting wheezing before 3 yrs of age and at 3–6 yrs of age, respectively 10. Children who started wheezing only after 6 yrs of age were defined as “very late wheezers”. Early wheezers were further classified on the basis of the recurrence of wheezing between 4 and 13 yrs of age; they were defined as: “early persistent wheezers” if they continued to wheeze every year up to the age of 13 yrs; “early intermittent wheezers” if they wheezed in some years but not in others between 3 and 13 yrs of age; and “early remittent wheezers” if they did not wheeze from 3 yrs up to the age of 13 yrs. For the purposes of the univariate and multivariate analyses, children not wheezing before the age of 3 yrs were defined as “early nonwheezers”. Parental atopy was defined as any reporting of asthma, hay fever or atopic dermatitis ever. Early atopic dermatitis was defined as present if the parents reported at least one of the following up to the age of 2 yrs: a reported diagnosis by the family physician; a parental reporting of symptoms of atopic dermatitis; or visible atopic dermatitis at the time of follow-up. In order to exclude very mild cases of atopic dermatitis, only children with parental reports of self-scratching were included in the definition of atopic dermatitis 11. Lower respiratory tract infections in the first 3 yrs of life were defined on the basis of parental reports of symptoms and diagnoses 12.
Cord blood, total and specific IgE
Serum samples were obtained from the children at birth and at 1, 2, 3, 5, 6, 7 and 10 yrs of age. Cord blood IgE, total IgE and specific IgE antibodies to food allergens (cow's milk, egg white, soy bean and wheat) and inhalant allergens (house dust mite (Dermatophagoides pteronyssinus), cat dander, mixed grass and birch pollen) were determined by using ImmunoCAP (Phadia, Freiburg, Germany). Sensitisation to a specific allergen was defined as a concentration ≥0.70 kU·L−1 of the respective specific IgE (ImmunoCAP class 2). This cut-off was chosen to constitute a stronger atopic effect than would have been by merely choosing evidence of any detectable specific IgE, i.e. ImmunoCAP class 1. Increased cord blood IgE was defined as a concentration of ≥0.9 kU·L−1, and increased total IgE at age 3 yrs was defined as a concentration ≥30.0 kU·L−1. This cut-off was based on the upper standard deviation limit of an equivalent aged normal infant population 13.
Assessment of indoor allergen exposure
At the ages of 6 and 18 months and 3, 4 and 5 yrs, Der p 1, Der f 1, and Fel d 1 allergens were extracted from dust samples collected by parents from the carpet and analysed with a sandwich ELISA (ALK, Copenhagen, Denmark) 14, 15. High level of exposure to a specific allergen at a specific age was defined as a measured value above the third quartile of the respective distribution in the total population.
Chi-squared tests were used to compare prevalence between groups and Mantel–Haenszel tests were used for analysing trends over categories. Generalised estimation equation models were used in order to adjust for repeated measures in the analysis of the effect of early childhood factors on wheezing at age 11, 12 and 13 yrs. Data are presented as odds ratios (ORs) and 95% confidence intervals (CI). All factors significant in univariate analysis were included in a stepwise procedure to elicit a final model. In order to take the stratified sampling scheme and a possible participation bias into account, all multivariate models were adjusted for parental atopy. Furthermore, sensitivity, specificity, and positive and negative predictive values of various early life factors were calculated for the outcome of wheezing at the age of 11, 12 or 13 yrs. For analyses on the incidence and natural course of wheeze over time, all children with complete data on wheeze up to the respective time point of analysis were included in the analyses. For all other analyses, all children with data on early wheeze were included.
Study population and response rates
Of the 1,314 enrolled children, 441 (33.6%) had information on wheeze for every year of the follow-up until age 13 yrs (fig. 1⇓). For 391 (88.7%) of these children, data on specific IgE in the first 3 yrs of life were available. In order to assess potential participation bias, the study population of 441 children was compared with children with incomplete data on the course of wheeze based on data collected at birth (table 1⇓). No significant differences were found with respect to parental history of atopy, sex, cord-blood IgE, older siblings or parental education. However, children in the study population were more likely to have nonsmoking parents (p<0.001). Furthermore, the study population of 441 children was compared with children with complete data up to the age of 3 yrs and incomplete data thereafter, based on data collected in the first years of life. No significant differences were found with respect to early atopic dermatitis, early wheeze and early atopic sensitisation. However, children in the study population were more likely to be breastfed for at ≥4 weeks (p = 0.020). A detailed description of the population sample is reported elsewhere 14.
Incidence of first wheezing episode from birth to 13 yrs of age
In total, 80 (18%) children out of the study population of 441 wheezed in the first year of life. The incidence of new wheeze sharply declined in the second (33 (9%) out of 361) and third year of life (13 (4%) out of 328). Overall, 126 (29%) out of 441 of the children in the MAS cohort started wheezing in their first 3 yrs of life (early wheezers). The incidence of new wheezing disorders was stable thereafter up to 7 yrs of age at ∼4% (range 3.5–5.0%). Then the incidence further declined to <1% at 10–12 yrs and abruptly rose again at 13 yrs to about 4% (fig. 2⇓). Overall, 40 (9%) out of the population sample of 441 started wheezing between 3 and 6 yrs of age (late wheezers) and another 42 (9%) out of 441 started wheezing after 6 yrs of age (very late wheezers).
The natural course of wheezing
Six different longitudinal patterns of wheezing in the population sample were distinguished (fig. 1⇑). Among early wheezers, remission of symptoms after early childhood was very high: 79 (63%) out of 126 did not wheeze after the age of 3 yrs, (early remittent wheezers); 43 (34%) out of 126 continued to wheeze after the age of 3 yrs but not every year (early intermittent wheezers); and only 4 (3%) out of 126 of the early wheezers continued wheezing every single year from birth up to age 13 yrs. Among children who never wheezed before the age of 3 yrs: 233 (74%) out of 315 also did not wheeze up to the age of 13 yrs (never wheezers); 40 (13%) out of 315 started wheezing between 3 and 6 yrs of age (late wheezers); and 42 (13%) out of 315 started wheezing after the age of 6 yrs (very late wheezers). Overall, the prevalence of children wheezing at age 11, 12 or 13 yrs was 22% (28 out of 126) among early wheezers and only 10% (31 out of 315) among early nonwheezers. Among these 59 children wheezing at age 11–13 yrs, 28 (47%) of them were early wheezers, 6 (10%) were late wheezers and 25 (42%) were very late wheezers.
The frequency of an atopic family background, early sensitisation and current sensitisation was analysed in relation to the six different longitudinal patterns of wheezing (table 2⇓). The prevalence of atopy at 3 yrs of age was very low among never wheezers (12%), low among early remittent wheezers and very late wheezers (both 19%), intermediate among late and early intermittent wheezers (32 and 36%, respectively) and high among early persistent wheezers (75%). The prevalence of atopy at 10 yrs of age was low among never and early remittent wheezers (27 and 30%, respectively), intermediate among late and very late wheezers (45 and 53%, respectively) and highest among early intermittent and early persistent wheezers (69 and 75%, respectively). Interestingly, among very late onset wheezers, the prevalence of early sensitisation was relatively low (18.7%), whereas sensitisation at age 10 yrs was moderately high (53.1%; table 2⇓).
Different patterns of risk factors for wheezing at the age of 13 yrs
The association of early atopic sensitisation (and related risk factors) with wheezing at 11–13 yrs of age in children wheezing or not wheezing before 3 yrs of age was examined (table 3⇓). In the MAS cohort, cat and/or mite exposure and sensitisation were strongly linked 14 and their concurrence contributed to a reduced lung function at school age 16; therefore, in the present study, their interaction in the multivariate analysis was accounted for. In univariate analysis, the spectrum of early-life risk factors associated with wheezing at age 13 yrs was remarkably broader in early wheezers than in early nonwheezers; indeed, among early wheezers, wheezing at puberty was associated with parental asthma, hay fever and atopic dermatitis, allergic sensitisation against food or airborne allergens, elevated total IgE and early atopic dermatitis. In contrast, among early nonwheezers, wheezing at puberty was significantly associated with only parental hay fever and asthma, sensitisation to indoor allergens, and early atopic dermatitis. In addition, the strength of the associations observed and their statistical significance was much stronger among early wheezers than among early nonwheezers. This trend was maintained in the multivariate analyses; perennial sensitisation to indoor (mites, cat and dog) allergens persisted as a very strong predictor of wheezing at 11–13 yrs among early wheezers but no longer showed a significant association among early nonwheezers, and the association of wheezing at 11–13 yrs with parental atopy was three times stronger among early wheezers (adjusted OR 8.32, 95% CI 2.74–25.20) than among early nonwheezers (adjusted OR 2.48, 95% CI 1.01–6.09). However, whereas atopic dermatitis early in life remained a significant predictor for wheezing at 11–13 yrs among early nonwheezers, it was no longer significant in the group of children with early wheeze. In contrast, in the group of early wheezers, male sex emerged as a significant predictor for wheezing at 11–13 yrs. All other early life factors analysed (i.e. cord blood IgE, siblings, breast feeding, parental smoking, parental education, early pets, early allergen exposure and early infections) did not show a significant association with wheezing at 11–13 yrs in either groups (data not shown).
In a model including the whole population, wheezing at 11–13 yrs of age was best described by the following factors in early life: early wheeze (adjusted OR 1.87, 95% CI 0.98–3.54; p = 0.056) parental atopy (adjusted OR 3.22, 95% CI 1.54–6.72; p = 0.002), early indoor (mites, cat and dog) sensitisation (adjusted OR 4.04, 95% CI 1.87–8.73; p≤0.001) and early AD (adjusted OR 2.99, 95% CI 1.52–5.88; p = 0.002).
Prediction of wheezing at 11–13 yrs of age
Table 4⇓ shows the sensitivity, specificity, and positive and negative predictive values of early-life factors for wheezing at 11–13 yrs of age. Best overall values for all prediction parameters were observed for early sensitisation to indoor (mites, cat and dog) allergens (positive predicted value of 57%). When early wheeze was a factor, prediction was even improved to some extent, with a positive predicted value of 75%. Thus, a child with sensitisation to indoor (mites, cat and dog) allergens as well as at least one episode of wheezing before the age of 3 yrs had a probability as high as 75% of still having wheeze at 13 yrs. The positive predicted value increased to 83% when a child with early wheezing and early sensitisation against indoor allergens was also exposed to high concentrations of indoor allergens in the first 3 yrs of life.
In the present analysis of the German MAS cohort, the natural course of wheezing was found to be highly heterogeneous. In an attempt to analyse and classify this heterogeneity, six different longitudinal patterns of wheezing between birth and the age of 13 yrs were described. The present findings add further complexity to the previous classification, which was based on three longitudinal patterns (early, late and persistent wheezers) and covered only the period between birth and 6 yrs of age. It was found that the relative role of early atopic sensitisation and of related risk factors for the long-term prognosis of wheezing change according to the age of wheezing onset. Taken together, the present results support the concept that childhood asthma is more a syndrome, rather than a single, uniform disease 17.
The incidence of wheezing was quite high in the first 3 yrs of life; however, the long-term prognosis of early wheezing was, in most cases, excellent. The vast majority of early wheezers either wheezed no more after the age of 3 yrs (early remittent wheezers) or only in some years (early intermittent wheezers) up to the age of 13 yrs. Only a very small fraction of early wheezers kept on wheezing every single year up to the age of 13 yrs (early persistent wheezers; fig. 1⇑).
The major question of the present study was how to identify, at 3 yrs of age, among early wheezers, those at risk for long-term persistent wheezing. In keeping with outcomes from previous birth cohort studies 4–7, the present study found that a strong predisposition to atopy (a positive family history or early IgE sensitisation) is the strongest factor predicting whether children wheezing early in life will keep on wheezing until puberty (tables 3⇑ and 4⇑). The present authors have previously shown that lung function at 7 yrs of age is compromised in children with early wheezing combined with early atopic sensitisation 16.
The long-term natural course of wheezing was highly heterogeneous among children who did not wheeze in the first 3 yrs of life. Most of them remained free from wheeze up to the age of 13 yrs, while a minority started wheezing between the age of 3 and 6 yrs (late wheezers) or after the age of 6 yrs (very late wheezers). The question here was how to identify, at 3 yrs of age, among early nonwheezers, those who will start wheezing later, up to the age of 13 yrs. The spectrum of early life atopy-related risk factors associated with wheezing at puberty was remarkably smaller both qualitatively and quantitatively in early nonwheezers compared with early wheezers (table 3⇑). In a consistent proportion of children, both atopy and wheeze started after 3 or 6 yrs of age. In particular, among very late onset wheezers, the prevalence of sensitisation at 10 yrs of age was three times higher than in the first 3 yrs of life, suggesting that this wheezing pattern was associated with later development of sensitisation. This trend is likely to be maintained at older ages because new atopic sensitisation acquired at school age is also a significant independent predictor of unremitting asthma after puberty 18, 19.
By testing just a small number of parameters (sensitisation to indoor allergens, such as mites, cat and dog, and high exposure to the same allergens) at 3 yrs of age, a subset of wheezing children with an extremely high probability of wheezing up to age 13 yrs could be identified (table 4⇑). In the Tucson cohort study, presence of wheeze at or before 3 yrs of age was also used to predict outcomes at 13 yrs 4. The present study used most of the “loose” criteria of that study (i.e. wheeze at any stage in the first 3 yrs, plus other factors 4) but found comparable sensitivities, specificities and negative predictive values to the “stringent” criteria, and even more favourable positive predictive value (table 4⇑). The additional inclusion in the present study of allergen exposure may have contributed to this outcome. The age-dependent variation of risk factors for wheezing at age 11–13 yrs suggests that multiple, distinct algorithms are required to predict persistence of wheezing starting at different ages. This is also in accord with other studies, suggesting that phenotype-specific strategies to predict and prevent asthma in children should be aimed at 20, 21.
There are limitations to the present study, for example, the population study is enriched with infants at high risk for atopy, so the results may not be relevant to the general population. Losses in follow-up may have further limited a general value of the present conclusions. Unfortunately, objective evaluation of viral infections on swabs was not available in the MAS. Other studies have found direct evidence of respiratory syncytical virus and rhinovirus infection in infancy to be associated with persistence of asthma in teenagers 22, 23.
Recent trials have provided evidence that prolonged anti-inflammatory treatment has an impact on infant wheeze but not on its natural history 24, 25. Generalised, aggressive intervention for secondary and tertiary prevention of wheezing cannot, therefore, be suggested on the basis of the present findings. However, it was observed that a small subset of children with early wheezing, early sensitisation and high exposure to indoor allergens showed a very high positive predictive (83.3%) for wheezing at the age of 11–13 yrs. More studies are necessary in order to ascertain whether this particular small, subset of early wheezers might benefit from preventive strategies yet to be identified.
In conclusion, the present data support the concept that the natural course of wheezing disorders in childhood is quite heterogeneous and that prediction of long-term outcomes may take advantage of not only early diagnosis but also algorithms based on the age of wheezing onset.
The present study was supported by the German Ministry of Education and Research (BMBF), grant number 01EE9406.
Statement of interest
The present authors would like to thank the study participants, their parents and the collaborators of the MAS group: V. Wahn, M. Groeger and A. Schuster (Düsseldorf, Germany); F. Zepp, I. Bieber and W. Kamin (Mainz, Germany); J. Forster and U. Tacke (Freiburg, Germany); C-P. Bauer (Gaissach, Germany); R. Bergmann (Berlin, Germany). The present authors are indebted to A. Reich (Institute for Social Medicine, Epidemiology and Health Economics, Charité University Medical Centre, Berlin, Germany) for data management and to P. Wagner (Dept of Paediatric Pulmonology and Immunology, Charité University Medical Centre) as the coordinating study nurse in MAS and to C. Clawson for his language expertise.
- Received June 1, 2007.
- Accepted April 27, 2008.
- © ERS Journals Ltd