- Top of page
Aims: Although atopic sensitization is common in childhood, its relationship to clinical allergic disease remains incompletely understood. We therefore sought to explore this relationship by defining sensitization based atopic phenotypes.
Methods: Children were recruited at birth (n = 1456) and reviewed at 1, 2, 4 and 10 years. Skin prick testing (SPT) to common allergens was done at 4 (n = 980) and 10 years (n = 1036) with lung function (n = 981), bronchial challenge (n = 784) and serum IgE (n = 953) testing at 10. Atopic phenotypes were defined, by sensitization pattern, for children with SPT at both 4 and 10 years (n = 823).
Results: Of phenotyped children, 68.0% were never atopic, 4.3% early childhood atopic (only atopic at age 4), 16.5% chronic childhood atopics (at 4 and 10 years) and 11.2% delayed childhood atopics (only at 10). Never atopics showed small but identifiable prevalence of allergic diseases such as asthma, eczema and rhinitis. Amongst allergen-sensitized subjects, aeroallergen predominated over food sensitization throughout childhood. Chronic childhood atopics showed highest prevalence of lifetime plus persistent wheeze, eczema and rhinitis, increased prevalence of aeroallergen sensitization, some evidence of persistent food sensitization, significantly greater cord IgE than never atopics (P = 0.006), plus higher total IgE (P < 0.001) and bronchial hyper-responsiveness (P < 0.001) at 10 years than other phenotypes.
Conclusion: A proportion of childhood eczema, rhinitis and asthma is nonatopic. The commonest childhood pattern of atopy is chronic sensitization, associated with early, persisting and clinically significant allergic disease. The currently accepted childhood ‘Allergic March’ may oversimplify the natural history of childhood atopy and allergic disease.
Allergic diseases, such as eczema, food allergy, rhinitis and asthma are common in childhood (1). Atopic sensitization, the genetically mediated predisposition to produce specific IgE antibody against antigens, is thought fundamental to these disorders. In turn, atopic sensitization and allergic disease are believed to follow a characteristic evolution or ‘Allergic March’ in early life (2). This notion suggests that susceptible individuals initially develop eczema and food allergy due to early food allergen sensitization, which is then replaced by aeroallergen sensitization, with a parallel shift in allergic disease burden to that of asthma and rhinitis (3–6).
Longitudinal population studies have potential to enhance our understanding of childhood atopic sensitization and its relationship to allergic conditions such as asthma, eczema and rhinitis. Specifically, such work could substantiate a natural history-based phenotypic classification of atopy, identifying clinically significant atopic phenotypes and clarifying the nature of any ‘Allergic March’. Such classification has previously identified (7–9), and characterized (10), troublesome forms of early onset persistent wheezing. Presence of similar relevant atopic phenotypes is plausible, but needs evidence-based confirmation.
Complex associations of genetic predisposition and environmental exposure in early childhood are thought crucial for developing eczema (11), rhinitis (12) and asthma (13). Identification of influences associated with different atopic phenotypes could provide further insight into the aetiology of childhood atopy.
In this paper, we define and characterize atopic phenotypes, using skin test sensitization, for the Isle of Wight Whole Population Birth Cohort Study.
- Top of page
A whole population birth cohort was established on the Isle of Wight, UK, in 1989 to prospectively study the natural history of allergic diseases. Ethical approval was obtained from the Local Research Ethics Committee. Of 1536 children born between 1 January 1989 and 28 February 1990, informed consent was obtained to enroll 1456 subjects. The majority (99.1%) of these were Anglo-Saxon Caucasian. Enrolment took place at birth when asthmatic/allergic family history (parental or sibling), household pet ownership, smoking habit during pregnancy, social class (Registrar General's Classification) and birth weight were recorded. Cord IgE was also measured (n = 1358).
Children were followed at 1 (n = 1167; 80.2% of 1456), 2 (n = 1174; 80.6%), 4 (n = 1218; 83.7%) and 10 years (n = 1373; 94.3%). Results of cohort follow-up (1, 2, 4 and 10 years) have been reported previously (14–17). At each follow-up, detailed questionnaires were completed with the parents for every child regarding allergic symptoms and diagnoses. ‘Current wheeze’ occurred on at least one occasion in the prior 12 months. Exposure to relevant environmental factors (domestic pets and tobacco smoke) was noted. Method of feeding was obtained at 1 and 2 years. A history of recurrent chest infections (more than one in the past year) was assessed at 1 and 2 years of age. Investigators diagnoses of eczema (chronic or chronically relapsing, itchy dermatitis lasting more than 6 weeks with characteristic morphology and distribution), recurrent nasal symptoms/rhinitis (recurrent nasal discharge or blockage with attacks of sneezing and itchy eyes) and food allergy (vomiting, diarrhoea, colic or rash within 4 hours of ingesting a particular food on at least two occasions) were made each time. At 1, 2 and 4 years, asthma was diagnosed as at least three separate episodes of wheezing, each at least 3 days in duration during the past year. At 10 years, asthma was diagnosed as combined current wheeze and ever being diagnosed asthmatic.
Skin prick testing (SPT) was performed to common inhaled and food allergens at 4 (Biodiagnostics, Germany) and 10 years (ALK, Denmark). At 4 years, this comprised house dust mite (Dermatophagoides pteronyssinus), grass pollen mix, cat and dog epithelia, Alternaria alternata, Cladosporium herbarum, milk, hens’ egg, soya, cod, wheat and peanut. At 10 years, it included house dust mite (D. pteronyssinus), grass pollen mix, tree pollen mix, cat and dog epithelia, A. alternata, C. herbarum, milk, hens’ egg, soya, cod and peanut. Histamine and physiological saline were used as positive and negative controls, respectively. Any reaction with mean wheal diameter at least 3 mm greater than negative control was regarded positive and defined atopy. At 10, serum was also analysed for Total IgE (Phadiatop – Pharmacia Diagnostics, Uppsala, Sweden).
At 10 years, spirometry was performed in 981 subjects and methacholine bronchial challenge in 484 children with wheezing histories, plus 300 children who never wheezed. For bronchial challenge, a Koko dosimeter (Pds Instrumentation, Louisville, CO) was used with compressed air source at 8 l/min and nebulizer output 0.8 l/min. Initial inhalation of 0.9% Saline was followed 1 min later by spirometry to obtain a baseline value. Incremental concentrations from 0.0625 to 16 mg/ml of methacholine were then serially administered. The concentration causing a stable 20% fall in FEV1 from postsaline value was interpolated and expressed as PC20 FEV1.
- Top of page
Data were double entered onto SPSS (V10.0; SPSS Inc., Chicago, IL). Atopic phenotypes were assigned in children with SPT at both 4 and 10 years (n = 823). Thus never-atopics had negative SPT on both occasions, early childhood atopics had positive SPT only at 4 years, chronic childhood atopics had positive SPT both times and delayed childhood atopics only at 10 years. Children were also categorized at 10 years by presence of ‘current wheeze’ at each follow-up; nonwheezers (children who never wheezed in the first decade of life), early transient wheezers (wheezing onset in the first 4 years of life which ceased and was not present within 12 months of assessment at 10 years), persistent wheezers (children with wheezing onset during the first 4 years of life who still wheezed at 10) and late-onset wheezers (children with wheezing onset from 5 years onwards who still wheezed at 10 years). Persistent eczema and rhinitis were defined by the respective physician diagnosed state at both 4 and 10 years. Cumulative diagnoses of asthma, eczema, rhinitis and food allergy ‘ever’ were created by amalgamating prospective records.
For analysis of continuous variables, one-way anova followed by post hoc Bonferroni testing to counter multiple comparison effects was used. For categorical variables chi-square testing (with Fisher's exact test where indicated by low expected cell counts) was conducted. A continuous measure of bronchial hyper-responsiveness (BHR) was estimated as least square dose–response slope using least-squares regression of percentage change in FEV1 upon methacholine concentration for each child. Inverse transformation of this dose–response slope (inverse slope) was used to satisfy normality and homoscedasticity with low values of 1/(10-slope) extrapolating to high BHR.
Risk factors for each atopic state were identified by univariate analysis comparing factors between each phenotype and the never-atopic state. To obtain the independent effect of risk factors with significance at univariate testing (P < 0.2), separate logistic regression models were created for early childhood atopy, chronic childhood atopy and delayed childhood atopy, using stepwise backward (likelihood ratio) logistic regression.
- Top of page
Of the original 1456 study participants, 83.7% (1218) and 94.3% (1373) were re-assessed by questionnaire at 4 and 10 years, respectively. SPT was conducted in 980 subjects at 4 years and 1036 at 10 years. Atopy, as defined above, was found in 19.7% (193/980) at 4 years and 26.9% (279/1036) at 10 years. A total of 823 children underwent SPT at both 4 and 10 years and were eligible for phenotypic assessment. Children included in this analysis were broadly representative, in demographic and disease characteristics, of subjects seen at the largest (10 year) follow-up visit (data not presented).
Of children with defined phenotypes, 68% were never atopic, 17%chronic childhood atopic, 11%delayed childhood atopic and 4%early childhood atopic.
The prevalence of both food and aeroallergen sensitization did not differ significantly between early and chronic childhood atopics at 4 years or between chronic and delayed childhood atopics at 10 years (Table 1). Most children with food allergen sensitivity at 4 years were chronic childhood atopics (84.8%– 28/33). In chronic childhood atopics with food sensitivity at 4 years (n = 28), 35.7% (10) still had food sensitivity at age 10. Peanut (7) was the commonest persistent sensitivity. Five peanut sensitive chronic childhood atopics lost sensitivity by 10 years, whilst six new cases of peanut sensitivity developed in this phenotype at 10 years. Of chronic childhood atopics with food sensitivity at 4, 96.4% were simultaneously aeroallergen positive. By age 10, all were aeroallergen positive.
Table 1. The proportion of food and aeroallergen sensitization for phenotypes at 4 and 10 years of age
|Early childhood atopic (n = 35)||Chronic childhood atopic (n = 137)|
|Four years SPT|
| Food allergen (n)||14.3% (5)||20.4% (28)||0.409|
| Aeroallergen (n)||94.3% (33)||99.3% (136)||0.106|
| ||Delayed childhood atopic (n = 91)||Chronic childhood atopic (n = 137)|| |
|Ten years SPT|
| Food allergen (n)||6.6% (6)||13.1% (18)||0.115|
| Aeroallergen (n)||98.9% (90)||100.0% (137)||0.399|
Amongst individual allergens, house dust mite sensitivity predominated throughout childhood. At 4 years (Fig. 1), several aeroallergen sensitivities [grass (P = 0.001; OR = 4.82; 95% CI = 1.76–13.16), cat (P = 0.021; 3.13; 1.14–8.61) and house dust mite (P = 0.03; 2.27; 1.07–4.83)] were significantly more prevalent in chronic than early childhood atopics. There were no significant differences in prevalence of food sensitization between chronic and early childhood atopics. Mean size of any skin test wheal elicited by individual allergens was also assessed (data not presented). House dust mite (4.17 mm vs 3.15 mm, P < 0.001) and grass (4.15 mm vs 2.83 mm, P < 0.001) were significantly larger in chronic than early childhood atopics at 4. At 10 years (Fig. 2), cat (P < 0.001; 3.45; 1.79–6.67), dog (P = 0.001; 3.85; 1.61–9.09), grass (P = 0.019; 1.89; 1.11–3.23) and peanut (P = 0.029; 4.76; 1.03–20) sensitization were significantly more prevalent in chronic than delayed childhood atopics. However, in terms of mean wheal size, only house dust mite differed significantly between the phenotypes at 10 years (4.42 mm vs 3.91 mm, P = 0.009).
The IgE and lung function for the atopic phenotypes are given in Table 2. Chronic childhood atopics had significantly higher Cord IgE at birth than never atopics plus higher total IgE at 10 years than all phenotypes. Early and delayed childhood atopics also had higher total IgE at 10 years than never atopics. Chronic childhood atopics showed significantly greater airflow obstruction at spirometry than never atopics. The BHR was significantly higher in chronic childhood atopics compared to both never and delayed childhood atopics. Lifetime prevalence of diagnosed asthma, eczema and rhinitis was also greatest in chronic childhood atopics (Table 3) though food allergy proved slightly more frequent in early childhood atopics. Diagnosed allergic states were still observed in never atopics, albeit at lower prevalence. Chronic childhood atopics had significantly higher prevalence of persistent wheeze than other phenotypes, plus significantly higher prevalence of both persistent eczema and rhinitis than never and delayed childhood atopics. Never atopics showed significantly higher prevalence of never wheeze than chronic or delayed childhood atopics.
Table 2. Immunological and lung function characteristics of atopic phenotypes
|Never atopic||Early childhood atopic||Chronic childhood atopic||Delayed childhood atopic|
|Cord IgE at birth (KU/l)||0.26||0.30||0.62‡||0.41|
|95% CI (n)||0.19–0.33 (484)||0.18–0.43 (29)||0.26–0.98 (120)||0.22–0.60 (78)|
|Log10 total IgE at 10 year||1.66||2.17†||2.66*¶**||2.20†|
|95% CI (n)||1.61–1.72 (521)||1.91–2.43 (31)||2.56–2.75 (124)||2.07–2.34 (86)|
|95% CI (n)||2.00–2.05 (554)||1.94–2.13 (35)||1.93–2.04 (136)||1.98–2.09 (91)|
|95% CI (n)||2.26–2.32 (554)||2.18–2.42 (35)||2.22–2.34 (136)||2.23–2.35 (91)|
|95% CI (n)||0.88–0.89 (554)||0.87–0.91 (35)||0.86–0.88 (136)||0.88–0.90 (91)|
|95%CI (n)||4.03–4.16 (554)||3.79–4.34 (35)||4.06–4.33 (136)||4.07–4.35 (91)|
|Inverse slope (BHR)||0.058||0.049||0.026*¶||0.050|
|95% CI (n)||0.055–0.062 (400)||0.027–0.054 (22)||0.021–0.031 (123)||0.039–0.061 (71)|
Table 3. Clinical expression of allergic disease for atopic phenotypes
|Never atopic||Early childhood atopic||Chronic childhood atopic||Delayed childhood atopic|
|Never wheeze (n)||63.4% (335/528)||62.5% (20/32)||34.8%*¶ (46/132)||48.2%‡ (41/85)|
|Early transient wheeze (n)||23.1% (122/528)||25.0% (8/32)||12.1% (16/132)||23.5% (20/85)|
|Early persistent wheeze (n)||8.3% (44/528)||6.3% (2/32)||37.1%*††** (49/132)||9.4% (8/85)|
|Late onset wheeze (n)||5.1% (27/528)||6.3% (2/32)||15.9%* (21/132)||18.8%* (16/85)|
|Asthma ever (n)||13.4% (75/559)||5.7% (2/35)||56.9%*§†† (78/137)||23.1% (21/91)|
|Eczema ever (n)||24.7% (129/522)||36.7% (11/30)||48.8%*¶¶ (63/129)||29.9% (26/87)|
|Persistent eczema (n)||3.9% (22/559)||11.4% (4/35)||21.2%*‡‡ (29/137)||6.6% (6/91)|
|Food allergy ever (n)||13.9% (72/518)||36.7%† (11/30)||32.3%* (41/127)||20.5% (17/83)|
|Rhinitis ever (n)||31.4% (164/523)||30.3% (10/33)||67.7%*§ (90/133)||45.3%** (39/86)|
|Persistent rhinitis (n)||0.5% (3/559)||5.7% (2/35)||13.1%*§§ (18/137)||2.2% (2/91)|
Multivariate logistic regression was used to identify early life risk factors for each atopic phenotype (Table 4). For early childhood atopics, asthma at 4 year was of independent significance, whilst parental smoking held a negative association. For chronic childhood atopics, male gender plus early life allergic disease were independently significant, with recurrent nasal symptoms in infancy giving reduced risk. For delayed childhood atopics, male gender, sibling food allergy, high Social Class at birth and low birth weight showed independent significance.
Table 4. Early life risk factors for atopic phenotypes
|Risk factor||OR||95% CI||P-value|
|Early childhood atopic|
| Asthma at 4 years||2.81||1.19–6.65||0.018|
| Parental smoking at 4 years||0.32||0.13–0.80||0.014|
|Chronic childhood atopic|
| Male gender||2.00||1.09–3.68||0.026|
| Eczema at 1 year||4.05||1.66–9.87||0.002|
| Nasal symptoms at 2 years||0.34||0.13–0.89||0.028|
| Asthma at 4 years||7.81||3.77–16.18||<0.001|
| Rhinitis at 4 years||3.31||1.07–10.27||0.038|
|Delayed childhood atopic|
| Male gender||2.15||1.06–4.37||0.033|
| Sibling food allergy||2.67||1.09–6.56||0.032|
| Social class I–III at birth||2.12||1.05–4.30||0.037|
| Low birth weight (<2.5 kg)||4.26||1.22–14.93||0.023|
- Top of page
This paper employed a phenotypic classification of childhood atopy to define the extent, nature and clinical relevance of atopic sensitization during the first decade of life. Several noteworthy findings emerge. First, only 33% of our population showed atopic sensitization during their first decade. Second, amongst childhood atopics, chronic sensitization proved the commonest and most clinically important phenotype. Third, our results failed to confirm the classical notion of an ‘Allergic March’, instead suggesting more complex patterns of atopic sensitization and allergic disease. Finally, although two thirds of our population never showed atopic sensitization, they still had identifiable prevalence of allergic conditions, highlighting the need to further investigate the role of atopic sensitization in allergic disease.
Before interpreting our findings some methodological issues should be considered. Our definition of atopy was made by SPT alone as serum sampling was not conducted at 4 years. A combination of SPT and serum specific IgE to define atopy might have yielded higher prevalence of atopy. It is unlikely however that this would have changed the overall nature of our results. Our SPT at each stage followed a validated technique limiting possible false positive or negative results.
House dust mite was the most prevalent sensitization at 4 years for early childhood atopics, at 10 years for delayed childhood atopics and throughout childhood for chronic childhood atopics. Furthermore, skin test wheal size was greatest for dust mite in these phenotypes. Collectively this suggests that house dust mite was the major local aeroallergen, consistent with findings reported (18) in a similar Southern English environment. However, in different environments, alternative allergens could prove relevant to the phenotypes identified in this paper.
Associated with male gender, chronic childhood atopics experienced considerable allergic disease. The finding that Cord IgE was significantly elevated in chronic childhood atopics suggested that the immunological pathway for such subjects is defined at an early stage. Independent associations of this phenotype with early clinical allergy like infantile eczema, plus diagnosed asthma and rhinitis at age 4 years demonstrate that the clinical manifestations of the chronic childhood atopic emerge quickly. Associations with persistent wheeze, eczema and rhinitis highlight the continuing clinical impact of this phenotype. The apparently protective effect of early nasal symptoms against chronic childhood atopy may appear contradictory but echoes prior findings from the German Multicenter Asthma Study (19) that children with such symptoms suffer less asthma and atopy. Considerable interest has surrounded a ‘hygiene hypothesis’ (20) for atopy development and these findings may reflect that process. Relationships of chronic childhood atopy with diagnosed asthma, persistent wheeze, airflow obstruction and enhanced BHR indicate presence of a significant childhood persistent atopic wheezing/asthma phenotype. Using different defining criteria, we extend prior characterization of early persistent atopy with respect to airways disease reported by Illi et al. (21), Peat et al. (22) and Sherrill et al. (23).
The ‘Allergic March’ (2) from food sensitization-related early life eczema and food allergy to later aeroallergen mediated airways disease has gained acceptance. Consistent with this notion we confirmed that most food allergen sensitization at 4 years is transient. However, not all food sensitization proved transient; one third of food-sensitive chronic childhood atopics at 4 years were still food sensitive at age 10. Peanut emerged as chief persistent food allergen. We noted a larger proportion (40%) of transient peanut sensitization in our population than previously (24) (20%), though we also found new peanut sensitivity in several chronic childhood atopics at 10 years. Chronic childhood atopics therefore appear at ongoing risk of this important clinical problem. Also out of keeping with a classical ‘Allergic March’, our skin testing at age 4 years identified predominance of aeroallergen over food sensitization. As suggested by prior work (3, 25) it is possible that transient food sensitization presaged aeroallergen sensitization in infancy, not detected at 4-year skin test. Unfortunately our cohort lacks skin prick data in infancy. An alternative interpretation is that food and aeroallergen sensitization, plus associated clinical diseases, occur concurrently rather than sequentially in early life. In this context, chronic childhood atopics had high prevalence of early persistent wheeze, a state previously shown (9) to frequently commence in infancy. It is plausible that aeroallergen sensitivity might manifest in infancy amongst chronic childhood atopics. This could reflect that early food sensitization is associated with accelerated aeroallergen sensitization as suggested by Rhodes et al. (26). Recently, Illi et al. (27) noted that childhood wheeze often began before, or concurrently with, atopic dermatitis in early life rather than as a sequel. Instead of ‘cause and effect’, might food sensitization, food allergy and eczema be similar ‘end organ’ expressions to asthma and rhinitis that arise from shared genetic and environmental triggers? Does the ‘Allergic March’ oversimplify the true natural history of atopic sensitization and allergic disease in childhood and merit modification (28)?
Whilst the chronic childhood atopic emerged as clinically most important, most children remained never atopic. It is worth noting that many never atopics still had eczema, rhinitis, wheeze and asthma. Overall most wheezing in our population was ‘never atopic’ with diagnosed asthma proving more prevalent amongst this group than early childhood atopics. Does this support argument against (29) the role of atopy in childhood asthma and wheeze? We recently (30) demonstrated substantial nonatopic wheeze at 10 years, defining atopy by SPT at that age and wheeze as within the past 12 months at that age. Such nonatopic wheezers were less often diagnosed asthmatic, or treated, but still had considerable morbidity. In this paper, we performed longitudinal analyses of atopy and wheeze, showing that, whilst constituting a substantial proportion of early persistent wheezers, most never atopic wheezers fell into the early transient wheezing group. Lower prevalence of persistent eczema and rhinitis were also found in never atopics suggesting that persistent ‘allergic’ disease still occurs in never atopics. Whilst not disputing the role of atopic sensitization in allergic disease our findings indicate significant presence of non-IgE mediated asthma, eczema and rhinitis.
In addition to the chronic childhood atopic we defined two further sensitized phenotypes. The early childhood atopics showed transient sensitization mainly to dust mite. They showed smaller positive skin test responses than their chronic childhood atopic counterparts. Whilst their lifetime prevalence of eczema, food allergy and rhinitis was substantial, their prevalence of persistent eczema, rhinitis and wheeze was low. This phenotype was independently associated with diagnosed asthma at 4 years but apparently protected against by parental smoking. The latter finding could reflect reverse causation, with parental smoking cessation after development of airways disease in the offspring. We propose that early childhood atopics bear a weaker, more ‘losable’, form of sensitization with transient clinical expression. Continued surveillance of this phenotype may prove interesting given their persistently raised IgE at 10 years, despite loss of sensitization. The delayed childhood atopics identified in this paper may also warrant future interest. Like chronic childhood atopics they showed male predominance and strong aeroallergen sensitivity. Independent associations with low birth weight and high Social Class at birth indicate early influences on this state. Yet their lung function was preserved, BHR not elevated, and asthma prevalence relatively low. Their increased prevalence of late onset wheeze and rhinitis suggest future potential morbidity, perhaps as atopy mediated inflammation takes effect.
In conclusion, much remains to be understood about the modern epidemic of asthma and atopy, even though disease trends (31) may be abating. Our paper highlights a new phenotypic classification of childhood atopy that could shed more light on the association of atopy and allergic disease. Allergic disease is not necessarily all atopic, nor does it always follow a classical ‘Allergic March’. In terms of clinical expression, the chronic childhood atopic appears the paramount atopic phenotype in the first decade and deserves recognition.
- Top of page
The authors gratefully acknowledge the cooperation of the children and parents who have participated in this study. The 10-year follow-up of this study was funded with the assistance of the National Asthma Campaign, UK (Grant no. 364). We also thank Monica Fenn, Linda Waterhouse, Linda Terry, Gail Poulton, Heidi Savory, Tessa Booth, Andrew Gallini, Cathy Wilby, Rosemary Lisseter, and Roger Twiselton for their considerable assistance with many aspects of the 10-year follow-up of this study. Finally we would like to highlight the role of the late Dr David Hide in starting this study.