• allergy;
  • birth cohort;
  • childhood;
  • Multi-centre Allergy Study;
  • specific IgE;
  • total IgE


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Background:  The development and the quantitative relationship between allergen-specific IgE (S-IgE) responses and total IgE (T-IgE), during childhood and adolescence have not been described and understood in detail. The objective of this study was to describe and compare the longitudinal trends of serum levels of S-IgE and T-IgE during childhood.

Methods:  We analysed data from participants in the MAS birth cohort study at 2, 5, 7 and 10 years of age (n = 273) and at 1, 3, 5, 6, 7, 10 and 13 years (n = 84). Total-IgE and the overall level of specific-IgE against nine locally relevant airborne and food allergens were determined by FEIA (ImmunoCAP). Allergic rhino-conjunctivitis and asthma were ascertained by questionnaires.

Results:  Longitudinal patterns of T-IgE levels from age 1 to 13 years were highly heterogeneous (declining, flat or increasing with different profiles). From 5 years of age, logarithmic (log10) transformed values of T-IgE and of S-IgE levels tend to follow a parallel trend, so that their relation remained constant throughout school age. A flat trend of T-IgE vs a constantly increasing trend of T-IgE was associated with a low or, respectively, high rate of wheezing at 13 years of age.

Conclusions:  Beginning at the age of 5 years, total serum IgE levels in children from an industrialized country evolved in parallel with overall S-IgE levels. Therefore, variations in T-IgE levels at school age closely reflect variations in overall S-IgE levels. Further studies are required to strengthen the biological and clinical implication of this novel finding.


confidence interval


enzyme-linked immunosorbent assay




Multi-centre Allergy Study


IgE with undefined specificity


coefficient of determination


allergen-specific IgE


total IgE


Delta E: difference between log10 total IgE and log10 overall allergen-specific IgE

Soon after the discovery of immunoglobulin E (1), heightened levels of it were found to be related to allergic diseases (2). Assays to determine IgE concentration in the serum were soon prepared (3, 4) and widely used as a diagnostic tool for allergic diseases. ‘Normal’ values of total IgE (T-IgE) were proposed (5) and thresholds established to label children of different ages (6) or adults (7) as ‘allergic’ or ‘nonallergic’. This diagnostic use of T-IgE as a ‘yes/no’ test was soon abandoned, because the frequency distribution slopes of T-IgE levels among normal or allergic subjects overlap too much in adults (8, 9) as in children (9, 10). Total IgE determination was then disregarded as less useful than others’ tests and it is currently excluded from guidelines on the diagnosis of asthma (11, 12).

The low diagnostic efficiency of T-IgE determination has been easily explained (8). In the absence of parasitic infections, total serum IgE levels generally result from the combination of two pools of IgE: (i) those of unknown specificity, observed in both, atopic and nonatopic individuals [IgE with undefined specificity (N-IgE)]; (ii) those recognizing allergens, observed only in atopic and allergic individuals [allergen-specific IgE (S-IgE)] (13). Some subjects who show no evidence of allergic diseases can have high levels of T-IgE due mainly to high levels of N-IgE, and some subjects with allergic diseases have very low levels of T-IgE due mainly to low levels of N-IgE. Therefore, a single determination of T-IgE levels cannot efficiently predict the presence or absence of IgE sensitization (8).

Multiple, serial determinations of T-IgE levels may have a better diagnostic value. Actually, atopic sensitization is not a ‘yes/no’ and static phenomenon (14) but a very dynamic process, characterized by a sequence of new sensitizations and remissions, especially in childhood (15). Therefore, changes (increase, decline) of T-IgE may reflect clinically relevant variations in S-IgE levels. However, the same changes may correspond to clinically irrelevant variations in N-IgE levels. Unfortunately, little information is available so far on the ‘natural course’ of T-IgE levels in childhood and its relation to the evolution of S-IgE levels.

The Multi-centre Allergy Study (MAS) provides a unique opportunity to investigate this area. Indeed, in this birth cohort study, peripheral blood drawing was scheduled at several time-points up to puberty (16, 17). A preliminary analysis of the relevance of T-IgE levels in this cohort focused on investigating the hierarchy of each value of T-IgE levels in relation to the whole group of subjects (18). With this new data analysis, we aimed to: (i) describe the time course of T-IgE and S-IgE in childhood; and (ii) ascertain whether, during childhood, serum levels of T-IgE and S-IgE evolve in a closely interrelated or in a totally independent fashion.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

The multi-centre allergy study

The MAS, a prospective observational birth cohort, recruited 1314 of 7609 infants born in 1990, in six delivery wards in five German cities. A detailed description of the stratified sampling scheme and study subjects is given elsewhere (16). Briefly, 499 newborns with risk factors for atopy [increased cord blood IgE (≥ 0.9 kU/l), 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 ages 1, 3, 6, 12, 18 and 24 months and from then on yearly within 4 weeks of the child’s birthday up to the age of 13 years. The study was approved by the local ethics committee. Each child’s parents gave written informed consent at the time of enrolment. At 13 years of age, parents gave structured interviews to a study physician that included standardized questions on asthmatic and atopic symptoms according to the International Study of Asthma and Allergy in Childhood (19). Blood samples were collected at 1, 2, 3, 5, 6, 7, 10 and 13 years of age.

Study population

For the purposes of this study, we examined two subsets of the MAS cohort. The first subset is composed of all the 273 children who participated at the follow-up visits at 2, 5, 7 and 10 years and whose data set, in relation to these four follow-up points, was complete. The vast majority of these children (255/273; 93.4%) also participated in the interview and examination at 13 years. The second subset is composed of all the 84 children who participated at the follow-up visits at 1, 3, 5, 6, 7, 10 and 13 years and whose data set, in relation to these seven follow-up points, was complete.

IgE assays

Total IgE was examined with a fluorescent enzyme immunoassay (ImmunoCAP FEIA, Phadia, Freiburg, Germany) and expressed in kU/l. The detection range of T-IgE was 2–2000 kU/l. Sera, with a concentration above the detection limit, were diluted 1 : 10 to obtain a precise determination, while sera with undetectable T-IgE were assigned an arbitrary value of 1 kU/l. A full set of IgE testing against five airborne allergens (house dust mite (Dermatophagoides pteronyssinus), cat and dog dander, mixed grass and birch pollen) and four food allergens (cow’s milk, hen’s egg, wheat and soy) was performed with ImmunoCAP FEIA (Phadia). Results were expressed in kU/l (detection range 0.35–100 kU/l). Sera, with a concentration of specific IgE antibodies, above the detection limit, were diluted 1 : 10 to obtain a precise determination, and those with a concentration below the detection limit were assigned an arbitrary value of 0.01 kU/l. The overall level of specific IgE was calculated by summation of the individual values of IgE-Ab against the nine allergen tested.


Longitudinal patterns of total serum IgE levels were defined according to the following criteria: a flat trend was arbitrarily defined by a total variation (increase or decrease) of 0.5 log10 kU/l or less of T-IgE, between any of the time-points considered. An increase or a decrease was arbitrarily designated as a variation of >0.5 log10 kU/l. Early, late or very late increase or decrease was defined as a variation occurring in the first 3 years of life, between 4 and 7 years of age, and after 7 years of age, respectively. A continuous increase was defined as a combination of both early and late increases. The six longitudinal patterns of T-IgE levels were examined in relation to the clinical status of children at 13 years of age in the subset of 84 participants. To test whether an increase in T-IgE level is temporally linked at individual level to the onset of respiratory allergies or of IgE sensitization, children with allergic rhinitis and/or asthma at 13 years of age were stratified by age at disease onset and variations of their T-IgE values examined. Values of overall specific IgE antibodies (S-IgE) were calculated as the sum of IgE-Ab against mites, cat, dog, birch and grass pollens, cow’s milk, hen’s egg, wheat and soy. Log-transformed values were used wherever indicated. The difference between log-transformed T-IgE levels (log10 T-IgE) and log-transformed overall specific IgE antibodies levels (log10 S-IgE) was defined as ‘ΔE’, and this value was used to explore relations in the evolution of log10 T-IgE and log10 S-IgE at population and individual levels throughout childhood.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

Characteristics of the study population

A comparison of the 273 children with complete sets of sera at 2, 5, 7 and 10 years of age with the rest of the MAS cohort revealed that participating children were not significantly different from nonparticipants with regard to sex ratio (boys 53%vs 52%), older siblings (43%vs 41%) and level of parental school education (parents with ≥12 years of schooling: 55%vs 51%), but had significantly more atopic parents (59%vs 50%, P = 0.01). A comparison of the 84 children with complete sets of sera at 1, 3, 5, 6, 7, 10 and 13 years of age with the rest of the MAS cohort revealed that participating children were not significantly different from nonparticipants with regard to sex ratio (boys 52%vs 52%), older siblings (39%vs 41%) or atopic parents (55%vs 53%) but had significantly better educated parents (parents with ≥12 years of schooling: 59%vs 52%, P < 0.05).

Longitudinal patterns of T-IgE levels

Longitudinal patterns of T-IgE levels were examined in the subset of 84 children with a complete dataset at 1, 3, 5, 6, 7, 10 and 13 years of age. A flat trend over the full follow-up period was observed in 19 participants (22.6%); an early, a late, a very late and a continuous increase was observed in 13 (15.5%), 12 (14.3%), 11 (13.1%) and 13 (15.5%) participants, respectively; a declining trend was observed in three (3.6%) children. In some children, dramatic variations in T-IgE levels were observed between two consecutive time-points. Longitudinal trends from 13 (15.5%) participants did not fit any of the six patterns described. In these children transient peaks, declines followed by an increase or irregular variations were observed.

Relationship between total and overall specific IgE levels

The relation of T-IgE and S-IgE levels over time was examined in individual participants. In most participants, the two slopes were poorly related in the first years of life, but their trends tended to become parallel from age of 5 years onwards. Geometric mean values of T-IgE and S-IgE levels were then calculated at each scheduled time-point within the six subsets of children with a homogeneous longitudinal pattern of T-IgE. A parallel trend of geometric mean values of T-IgE and S-IgE started clearly from the age of 5 years in all the six groups (Fig. 1). This trend was confirmed by monitoring over time the differences between T-IgE levels and S-IgE levels (ΔE value, see Methods for definition) at any time-point in the larger subset of 273 participants with a full dataset at 2, 5, 7 and 10 years of age. ΔE values were rather unstable between 2 and 5 years of age; by contrast, they were rather stable after 5 years of age (Fig. 2). This demonstrates that log10 T-IgE and log10 S-IgE run in parallel at school age not only at group level (Fig. 1), but also at individual level (Fig. 2).


Figure 1.  Longitudinal patterns of serum total IgE and serum-specific IgE-Ab to common airborne and food allergens at group level. Values are the average of each group (A–F) of participants with a similar pattern. The number of participants contributing to each group is 19, 13, 12, 11, 13 and 3 for panels A, B, C, D, E and F, respectively. Levels of overall specific IgE are the sum of the levels of IgE-Ab directed against nine common food and airborne allergens.

Download figure to PowerPoint


Figure 2.  Parallel trends of log-transformed values of T-IgE and S-IgE. Each dot represents the difference between log10 total IgE and log10 overall specific IgE levels measured at 5, 7 and 10 years of age in relation to the variation observed since the previous time-point (2, 5 and 7 years of age, respectively). Values on the identity line correspond to a perfect parallel trend of total and specific IgE changes in the respective time period. R2 = coefficient of determination.

Download figure to PowerPoint

Relevance of variation of T-IgE levels

There was a trend for children with wheezing at 13 years of age to have higher average values of T-IgE at any age than children with allergic rhinitis at 13 years. The course of T-IgE average values of nonatopic children with neither wheezing nor allergic rhinitis at 13 years of age was almost flat (Fig. 3A). The lowest frequencies of wheezing at 13 years of age were observed in children with a flat evolution of T-IgE level; by contrast, the highest frequencies of wheezing at 13 years of age were observed in children with a continuous increase of T-IgE throughout childhood (Table 1). Interestingly, the maximum increase of T-IgE levels throughout childhood coincided in all these groups with the onset of their disease (Table 2). In a second analysis, children with atopic sensitization at 10 years of age were stratified by the age at onset of their sensitization. Interestingly, T-IgE average values of atopic children clearly deviate from the trend-line of the nonatopic children at the age at onset of their atopic sensitization (Fig. 3B). Taken together, all these data suggest that in childhood not only actual levels, but especially variations also (i.e. increases) in T-IgE levels are temporally related to the onset of respiratory allergies.


Figure 3.  Longitudinal trends of total serum IgE by clinical and atopic status at age 13. Trends of total IgE level (geometric means) in: (A) children with no atopic sensitization/allergy up to the age of 13 years, children with current wheezing (WZ) or current allergic rhinitis (AR) at 13 years of age (at 2 years: AR vs none, P = 0.044; WZ vs none, P = 0.004; AR vs WZ, P = 0.110; at 5 years: AR vs none, P < 0.001; WZ vs none, P < 0.001; AR vs WZ, P = 0.031; at 7 years: AR vs none, P < 0.001; WZ vs none, P < 0.001; AR vs WZ, P = 0.078; at 10 years: AR vs none, P < 0.001; WZ vs none, P < 0.001; AR vs WZ, P = 0.124); (B) atopic children, stratified by age at sensitization onset and children who have never developed atopic sensitization.

Download figure to PowerPoint

Table 1.   Allergic diseases and atopy at 13 years of age, by longitudinal pattern of T-lgE serum levels
Pattern of T-lgE evolution*nAR n (%)WZ n (%)Atopy n (%)T-lgE* (kU/l)
  1. AR, allergic rhinitis with or without wheezing; WZ, wheezing with or without allergic rhinitis; atopy, children with S-lgE (> 0.7 kU/I) to at least one of the nine allergen examined at the age of l3 years.

  2. *Geometric mean.

Flat196 (31.5)2 (10.5)3 (15.9)25.70
Early increase136 (46.1)3 (23.1)7 (53.8)145.30
Late increase127 (58.3)2 (16.7)9 (75.0)180.60
Very late increase114 (35.4)1 (9.1)7 (63.6)185.80
Continuous increase138 (61.5)6 (46.1)8 (61.5)211.20
Declining31 (33.3)1 (33.3)0 (0)21.70
Irregular135 (38.4)0 (0)5 (38.5)34.70
All8437 (44.0)15 (17.9)39 (46.4)83.00
Table 2.   Variations in log10 transformed total IgE levels, by age and onset of persistent allergic rhinitis or asthma
Age at onset of allergic rhinitis and/or asthma (years)Actual value of T-lgE at 2 yearsIncrease in the level of T-lgE from one follow-up point to the next*
2–5 years5–7 years7–10 years
  1. Peak values at each age (2, 2–5, 5–7, 7–10 years) are in bold.

  2. *Each value is the result of the formula: log10 T-lgE at age (x) − log10 T-IgE at age (z), where z and x are two consecutive follow-up points.

  3. †log10(kU/l).



  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

In this study, we describe the evolution of serum T-IgE and S-IgE at various time-points from birth up to 13 years of age and their relation to symptomatic respiratory allergies at 13 years of age. We distinguished six different longitudinal patterns of T-IgE and found that they are paralleled by similar trends of S-IgE responses. Increases in T-IgE levels were temporally related to the onset of respiratory allergies and of specific sensitization. Taken together, our data suggest that variations, rather than actual levels, of serum T-IgE deserve further attention as a potentially useful parameter for the prediction and prevention of paediatric allergy.

Our data show that, opposite to all the other immunoglobulin isotypes, the evolution of T-IgE levels is extremely heterogeneous in childhood. The simple fact that we could discriminate six longitudinal IgE patterns in only 84 children makes questionable whether by adopting more stringent criteria, additional patterns would have been identified. However, it is clear that at least the group of children with no atopic sensitization and no allergic diseases displayed a quite regular pattern of its average T-IgE levels, characterized by a very slow growth with age. Such a ‘normal’ trend was already described a quarter of a century ago (10). Our data suggest that, in atopic children, the slope of T-IgE levels ‘deviates’ upwards from the ‘normal’ trend when allergen-specific responses and allergic disease initiate. This is a consequence of the fact that, opposite to other immunoglobulins classes, IgE lack of homeostatic mechanisms keeping their serum levels at a steady level over time independently from the number and intensity of IgE antibody responses. Therefore, in westernized children, growing without parasitic infections, T-IgE serum levels change with changes in S-IgE levels.

Is there any fixed rule linking T-IgE levels to S-IgE levels? Our data suggest that the relationship between T-IgE and S-IgE levels is regulated in the long-term, so that the ratio between the total and overall S-IgE is kept relatively constant. This ratio is expressed by the difference in log10 T-IgE value and log10 S-IgE values in our analyses (ΔE). This is quite evident at group level, but is also rather clear at individual level.

From a biological point of view, they may be interpreted in light of the model proposed a long time ago by David Marsh (13). According to this model, total serum IgE levels result from the combination of two IgE pools: one produced through antigen-specific, cognate interactions, the other produced through nonspecific, noncognate (bystander) interactions; the two pools would be regulated by different genes. In keeping with this model, we may speculate that, in nonatopic children, IgE levels are because of IgE produced through noncognate interaction. Our data may suggest that during atopic responses this pool of nonallergen-specific IgE is expanded, but in a fixed proportion with respect to S-IgE. In this light, a higher ΔE value would indicate a higher genetic propensity to produce polyclonal IgE through noncognate processes.

Our study population was not selected in a clinical environment; therefore little can be said on clinical implications of our results. However, the observed parallelism between T-IgE and S-IgE longitudinal trends may inspire clinical studies based on T-IgE monitoring. In principle, serial determination of T-IgE levels may be more useful than determinations at a single time-point. For example, a strong increase of T-IgE may lead to suspect the onset of atopic sensitization, even in the case of negative skin tests to a standard panel of common allergens. Lack of variation of T-IgE, in nonatopic subjects, may contribute to exclude the onset of new sensitizations. These hypotheses need to be tested in larger cohorts and in a clinical setting.

This study has some limitations to be acknowledged. A proportion of the T-IgE labelled as ‘nonallergen specific’ (ΔE value) may be directed against allergens not included in the panel we analysed. This panel, however, included the most important airborne and food allergens in Germany, and we assumed that the relative contribution of IgE-Ab against additional allergens to T-IgE levels was small or negligible in most of the children. Further, the MAS birth cohort population sample is not representative of the German population of the same age, and that only a rather small subset of participants could be included in the present data analysis. However, to our knowledge, this is the largest sera-bank targeted to the study of allergic diseases and covering the whole childhood from birth to puberty with so dense time-points.

In conclusion, total serum IgE levels of children in an industrialized country evolve in parallel with the overall S-IgE levels since the age of 5 years. Therefore, consistent variations in T-IgE levels at school age reflect consistent variations in overall S-IgE levels. Further studies are needed to investigate the biological and clinical implication of this novel finding.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References

The authors thank the study participants, their parents and the collaborators of the MAS group: Volker Wahn and Antje Schuster, Düsseldorf; Fred Zepp, Imke Bieber and Wolfgang Kamin, Mainz; Johannes Forster and Uta Tacke, Freiburg; Carl-Peter Bauer, Gaissach; Renate Bergmann, Berlin. The authors are indebted to Andreas Reich for data management and to Petra Wagner as the coordinating study nurse in MAS and to Charles Clawson for his language expertise. This study was supported by the German Ministry of Education and Research (BMBF), grant number 01EE9406.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgments
  7. References
  • 1
    Johansson SGO, Bennich H. Immunological studies of an atypical (myeloma) immunoglobulin. Immunology 1967;13:381.
  • 2
    Johansson SGO. Raised levels of a new Immunoglobulin class (IgND) in asthma. Lancet 1967;ii:951953.
  • 3
    Wide L, Bennich H, Johansson SG. Diagnosis of allergy by an in-vitro test for allergen antibodies. Lancet 1967;2:11051107.
  • 4
    Kjellman NM, Johansson SG, Roth A. Serum IgE levels in healthy children quantified by a sandwich technique (PRIST). Clin Allergy 1976;6:5159.
  • 5
    Kjellman NM. Predictive value of high IgE levels in healthy children. Acta Pediatr Scand 1976;65:465.
  • 6
    Lindberg RE, Arroyave C. Levels of IgE in serum from normal children and allergic children as measured by an enzyme immunoassay. J Allergy Clin Immunol 1986;78:614618.
  • 7
    Zetterström O, Johansson SG. IgE concentrations measured by PRIST in serum of healthy adults and in patients with respiratory allergy. A diagnostic approach. Allergy 1981;36:537547.
  • 8
    Klink M, Cline MG, Halonen M, Burrows B. Problems in defining normal limits for serum IgE. J Allergy Clin Immunol 1990;85:440444.
  • 9
    Dati F, Ringel EP. Reference values for serum IgE in healthy non-atopic children and adults. Clin Chem 1982;28:1556.
  • 10
    Saarinen UM, Juntunen K, Kajosaari M, Björkstén F. Serum immunoglobulin E in atopic and non-atopic children aged 6 months to 5 years. A follow-up study. Acta Paediatr Scand 1982;71:489494.
  • 11
    Homburger HA, Katzman JA. Methods in laboratory immunology. In: MiddletonE et al., editors. Allergy. Principles and practice. 4th edn. St Louis, MO: Mosby, 1993:554572.
  • 12
    National Heart Lung and Blood Institute. Global initiative for asthma. Global strategy for asthma management and prevention. NHLBI/VVHO workshop report. Bethesda: National Institutes of Health, 1995. National Heart Lung and Blood Institute publication number 95-3659.
  • 13
    Marsh DG, Neely JD, Breazeale DR, Ghosh B, Freidhoff LR, Ehrlich-Kautzky E et al. Linkage analysis of IL4 and other chromosome 5q31.1 markers and total serum immunoglobulin E concentrations. Science 1994;264:11521156.
  • 14
    Matricardi PM, Nisini R, Biselli R, D’Amelio R. Evaluation of the overall degree of sensitization to airborne allergens by a single serologic test: implications for epidemiologic studies of allergy. J Allergy Clin Immunol 1994;93:6879.
  • 15
    Stern DA, Lohman IC, Wright AL, Taussig LM, Martinez FD, Halonen M. Dynamic changes in sensitization to specific aeroallergens in children raised in a desert environment. Clin Exp Allergy 2004;34:15631669.
  • 16
    Bergmann RL, Bergmann KE, Lau-Schadensdorf S, Luck W, Dannemann A, Bauer CP et al. Atopic diseases in infancy. The German multicenter atopy study (MAS-90). Pediatr Allergy Immunol 1994;5(6 Suppl):1925.
  • 17
    Illi S, Von Mutius E, Lau S, Niggemann B, Gruber C, Wahn U et al. Perennial allergen sensitization early in life and chronic asthma in children: a birth cohort study. Lancet 2006;368:763770.
  • 18
    Nickel R, Illi S, Lau S, Sommerfeld C, Bergmann R, Kamin W et al. Variability of total serum immunoglobulin E levels from birth to the age of 10 years. A prospective evaluation in a large birth cohort (German Multicenter Allergy Study). Clin Exp Allergy 2005;35:619623.
  • 19
    Asher MI, Keil U, Anderson HR, Beasley R, Crane J, Martinez F et al. International Study of Asthma and Allergies in Childhood (ISAAC): rationale and methods. Eur Respir J 1995;8:483491.