SEARCH

SEARCH BY CITATION

Keywords:

  • asthma;
  • epidemiology;
  • IgE;
  • Staphylococcus aureus ;
  • superantigens

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References

Background

Specific IgE to Staphylococcus aureus enterotoxins (SE-IgE) has been associated with asthma. In the general population, we aimed to determine the prevalence of and risk factors for serum SE-IgE and to examine the association with asthma.

Methods

A postal questionnaire was sent to a random sample of adults in 19 centers across Europe. A random sample of respondents was invited for clinical examination upon which they answered a questionnaire, underwent skin prick tests (SPTs) for common aeroallergens, and provided blood for measurement of total IgE and SE-IgE. Risks were analyzed within centers using weighted logistic regression, and overall estimates calculated using fixed-effects meta-analysis.

Results

2908 subjects were included in this analysis. Prevalence of positive SE-IgE was 29.3%; no significant geographic variation was observed. In contrast to positive skin prick tests, SE-IgE was more common in smokers (<15 pack-year: OR 1.11, P = 0.079, ≥15 pack-year: OR 1.70, P < 0.001), and prevalence did not decrease in older age-groups or in those with many siblings. Total IgE concentrations were higher in those with positive SE-IgE than in those with positive SPT. SE-IgE was associated with asthma (OR 2.10, 95% confidence interval [1.60–2.76], P = 0.001) in a concentration-dependent manner. This effect was independent of SPT result and homogeneous across all centers.

Conclusions

We report for the first time that SE-IgE is common in the general population throughout Europe and that its risk factors differ from those of IgE against aeroallergens. This is the first study to show that SE-IgE is significantly and independently associated with asthma in the general population.

Abbreviations
CI

confidence interval

CRS

chronic rhinosinusitis

IQR

interquartile range

SE-IgE

specific IgE for S. aureus enterotoxin mix (SEA, SEC, TSST1)

SPT

skin prick test

SPT-ANY

skin prick test to any of the tested allergens

SPT-HDM

skin prick test for house dust mite

In the general population, asthma or bronchial hyper-responsiveness is strongly associated with positive skin prick tests to common environmental perennial allergens (such as house dust mite and cat dander) [1] and with elevated levels of serum IgE to these allergens [2, 3]. High total IgE concentrations have also been associated with asthma, even in the absence of detectable specific IgE [4]. More recently, specific IgE antibodies directed against bacterial products such as Staphylococcal enterotoxins have been described locally in the upper airways [5]. Enterotoxins are proteins secreted by most strains of Staphylococcus aureus and have classically been associated with food poisoning and toxic shock syndrome. Specific IgE to S. aureus enterotoxins (SE-IgE) has been found in the serum of patients with atopic dermatitis [6] and in nasal polyp tissue [5] and in the latter has been associated with more severe eosinophilic inflammation and the presence of comorbid asthma [7, 8]. Small clinical studies have shown that SE-IgE is present in the peripheral blood from adults [9] and teenagers [10] with asthma and in those with severe asthma [11, 12].

The prevalence of Saureus enterotoxin IgE and its association with asthma and sensitization to common allergens in the general population is unknown. The Global Allergy and Asthma Research Network (GA²LEN) conducted an international multicenter population-based study to determine the prevalence of and risk factors for asthma, allergy, and sinusitis in the European adult population and included measures of SE-IgE. In this report, we describe the prevalence of and risk factors for this S. aureus enterotoxin IgE and examine its association with asthma and total IgE concentration. We also compare the epidemiology of SE-IgE with that of sensitization to other allergens.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References

Study design

A postal questionnaire about symptoms of asthma, rhinitis, and chronic rhinosinusitis (CRS) was sent to a random sample of participants aged 15–75 years, identified from a population-based sampling frame in 19 centers (as described previously [13, 14] – see Fig. 1). Except for Montpellier, all centers conducted a clinical examination of a sample of responders (n = 3505), which was selected randomly from four groups – those with asthma, with CRS, with asthma and CRS, and with neither asthma nor CRS, according to the initial questionnaire.

image

Figure 1. Study flowchart. Results from the first phase of the study (GA²LEN Survey, upper part) have been published before [13, 14]. *This number only includes centers that participated in both the postal and clinical study. **One center excluded for overall low rate of response and low rate of blood sampling (n = 50), one center excluded for low rate of sampling of controls (n = 152).

Download figure to PowerPoint

The same protocol was used in all centers, with centralized fieldworker training. The postal survey was conducted between September 2007 and December 2008, and the clinical examinations were held between August 2008 and August 2010. In each center, the study was approved by local ethical review boards.

Variables

At the clinical interview, participants underwent interviewer-administered questionnaires, skin prick tests (SPTs) (grass mix, timothy grass, birch, Blattella, olive, Artemisia, Parietaria, Dermatophagoides pteronyssinus, Dermatophagoides farinae, Alternaria, cat, dog) using allergens as described previously [15], and venepuncture for measurement of serum total IgE and specific IgE to a mixture [9] of S. aureus enterotoxin A, S. aureus enterotoxin C, and toxic shock syndrome toxin 1 (further referred to as SE-IgE). Assays were conducted in a single laboratory (Academic Medical Center, Amsterdam) using ImmunoCAP® (Phadia, Uppsala, Sweden).

Definitions

Asthma was considered present if participants reported they ‘ever’ had asthma and they had experienced at least one of the following in the preceding 12 months: wheezing, waking up with chest tightness, waking up with shortness of breath, or waking up with cough [13]. SPT was considered positive if the mean wheal diameter at 15 min was greater than 0 mm [16]. SPT to house dust mite was considered positive if there was a wheal for either D. pteronyssinus or D. farinae.

As per manufacturers' recommendations, a serum SE-IgE concentration above 0.10 kUA/L was considered positive [9, 11]. Total IgE concentrations greater than 100 kUA/L were categorized as high.

Statistical methods

As responders to the postal survey were selected for clinical interview using a case–control design, inverse sampling-probability weights [17] were applied to standardize the joint distribution at baseline (of asthma, CRS and study center status) to that of the postal survey. This means that the prevalence and effect estimates presented in this report reflect those we would see, as if all postal survey responders in the centers sampled would have been seen in the clinical examination. As CRS was not considered an outcome variable in this study, sampling status of CRS was used in the weighting process, but no further analyses of CRS were performed.

Unadjusted prevalences, and unadjusted and confounder-adjusted odds ratios for SE-IgE, house dust mite-positive SPT, any positive SPT, and asthma were estimated using logistic regression with sampling-probability weights and Huber variances. For ordinal predictor variables (age, pack-year, sibship size, specific IgE concentration tertiles), linear trend in effect was tested by applying logistic regression on the category rank value of the predictor. Unadjusted geometric means and adjusted geometric mean ratios for SE-IgE and total IgE were estimated using linear regression on the logged values. Tests for interaction of two predictor variables were performed by including the product of the predictors in the model.

Estimates were mutually adjusted for confounders, which included gender, age-group, smoking pack-years, sibship size, and parental history of allergy. Only subjects with serum sampled were included in the analyses. Other missing data were deleted pairwise.

All analyses were carried out within centers. Because of small sample sizes (n < 5) in some case groups in the two UK and three Polish centers, data were pooled to country level for these centers. For estimating overall effects, within-center estimates were combined using fixed-effects meta-analysis, and heterogeneity was assessed with the chi-square and I2 statistic [18]. Analyses were conducted using STATA v11.1 (StataCorp, College Station, TX, USA). Confidence intervals are 95% wide and indicated by square brackets.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References

The overall response to the postal survey in the 18 centers that also took part in the clinical survey was 56936/113938. Of these, 12461 were invited for further tests, and 3505 (28.1%) responded (Fig. 1). Three centers were excluded from analysis (one collected no serum samples, one had too few serum samples, one had too few subjects sampled in the control group). No serum measure was available from 268 subjects, leaving a final sample of 2908 subjects in 15 centers. The median age was 48.9 years (IQR 36.0–60.3 year), and 56.7% of subjects were female.

Prevalence and risk factors of SE-IgE and sensitization to aeroallergen

Table 1 shows the weighted population prevalence of specific IgE to S. aureus enterotoxins (SE-IgE), positive skin prick tests to house dust mite (SPT-HDM), and positive SPT to any of the tested allergens (SPT-ANY) in each center. The overall prevalence of SE-IgE was 29.3%, ranging from 21.5% (Umea) to 36.1% (Coimbra). Confidence intervals around estimates were wide but there was little evidence of geographic variation – only one center (Uppsala) had a confidence interval that was lower than the overall estimate. In contrast, there was considerable geographic variation in the prevalence of SPT-HDM (14.9%, ranging from 8.6% in Stockholm to 28.2% in Ghent) and SPT-ANY (44.4%, ranging from 37.9% in Poland to 54.0% in Duisburg).

Table 1. Overall weighted prevalences of outcome parameters, as assessed in the follow-up phase. In each area, prevalence estimates were weighted for sampling by case status (control, asthma, CRS, or asthma + CRS) in survey phase
Area N SE-IgEHDM-positive SPTAny positive SPTAsthma
Prevalence (%)95% CIPrevalence (%)95% CIPrevalence (%)95% CIPrevalence (%)95% CI
  1. N: total number of participants who provided blood samples. SE-IgE: detectable serum immunoglobulin E (>0.10 kUA/l) against mixture of SEA, SEC, and TSST-1. HDM: house dust mite. SPT: skin prick test. CI: confidence interval.

  2. a

    Estimates pooled on country level because of low within-center sample sizes in Poland (Katowice, Lodz, Krakow) and UK (Southampton, London).

Belgium (Ghent)14828.3[20.5–37.6]28.18[20.2–37.9]45.7[35.9–55.9]9.1[5.7–14.1]
Denmark (Odense)36328.1[21.8–35.5]9.55[6.0–14.8]34.2[27.3–41.9]8.3[6.1–11.2]
Finland (Helsinki)16234.5[26.0–44.2]10.47[5.8–18.1]47.4[37.9–57.2]8.6[5.9–12.3]
Germany (Brandenburg)17727.7[19.5–37.8]10.54[5.8–18.6]51.1[40.8–61.3]4.6[3.3–6.5]
Germany (Duisburg)19127.2[20.9–34.4]19.82[14.4–26.7]54.0[46.2–61.7]9.3[6.7–12.9]
The Netherlands (Amsterdam)21530.6[22.9–39.6]22.83[16.1–31.3]49.3[40.0–58.6]8.5[5.3–13.3]
Poland a24235.2[28.0–43.2]16.54[11.2–23.7]37.9[30.1–46.4]6.4[4.4–9.3]
Portugal (Coimbra)25836.1[29.2–43.7]26.6[20.5–33.8]48.9[41.0–56.9]14.4[10.9–18.7]
Sweden (Stockholm)33735.7[28.8–43.2]8.61[5.3–13.8]41.4[34.1–49.0]9.9[7.3–13.2]
Sweden (Umea)27621.5[14.3–30.9]8.97[4.6–16.8]45.2[35.1–55.6]15.4[10.6–21.9]
Sweden (Uppsala)37822.0[16.1–29.1]14.38[9.5–21.1]50.3[42.2–58.3]10.9[8.3–14.2]
United Kingdom a16129.9[20.8–40.9]22.58[14.6–33.2]39.1[28.6–50.8]17.9[12.0–26.0]
Overall 2908 29.3 [26.8–31.8] 14.92 [13.1–17.0] 44.4 [41.5–47.2] 10.6 [9.4–11.9]

Risk factors for SE-IgE, SPT-ANY, and SPT-HDM were identified and compared using a common logistic regression model (Fig. 2 and Table 2). SPT-ANY was more common in those with a familial history of allergy, and less common in older age-groups (ptrend = 0.001), in smokers with >15 pack-year, and in those from larger families (>1 sibling). SPT to HDM was less common in older subjects (ptrend < 0.001), more common in those with a familial history of allergy, and was not associated with smoking (ptrend in pack-year = 0.570) or with family size (OR per extra sibling = 1.01, ptrend = 0.872). Sensitization to SE was more common in those with a familial history of allergy, but unlike SPT-ANY or SPT-HDM, it was more common in smokers, particularly in current smokers (OR current smokers = 2.02 [1.42–2.88], OR ex-smokers = 1.41 [1.06–1.89]). There was a dose-dependent relationship with the number of pack-years (ptrend < 0.001). Furthermore, there was some evidence that SE-IgE was more common in those with more siblings (OR 1 sibling = 1.52, OR > 1 sibling = 1.47; no overall significance was reached, due to between-center variation in the association: I2 = 46–62%). The prevalence of SE-IgE was similar in all age-groups and less common in women. The associations were unaltered when adjusted for SPT-ANY or SPT-HDM, and analyses in which SE-IgE was considered as continuous outcome showed similar associations. There was no evidence (P > 0.05) that SE-IgE was associated with parental smoking during pregnancy or childhood, birth order, history of severe childhood respiratory infections, day care attendance, bedroom sharing with siblings, and rural vs urban living during childhood (data not shown).

Table 2. Multiple logistic regression analysis for the presence of (A) Staphylococcus aureus enterotoxin-specific IgE, (B) positive skin prick tests (SPTs), and (C) house dust mite-positive SPT
 Unadjusted estimatesAdjusted estimatesHeterogeneity
Prevalence (%)OROR95% CI P I²P (I²)
  1. All estimates were weighted for case sampling. Adjusted odds ratios were mutually adjusted for all predictor variables, estimated within each center, and meta-analyzed.

  2. a

    One center (Umea, n = 374) excluded due to empty groups.

(A) S. aureus enterotoxin IgE
Male sex34.7(1)(1)[0.45–0.77]<0.0010%0.599
Female sex24.70.650.59
Age 15–3427.8(1)(1)    
Age 35–5428.20.940.79[0.56–1.13]0.1930%0.657
Age 55–7431.00.920.71[0.49–1.04]0.07634%0.122
Nonsmoker25.2(1)(1)    
1–15 pack-year27.61.111.33[0.97–1.84]0.0790%0.455
>15 pack-year38.31.701.87[1.34–2.60]<0.0017%0.381
No sibling22.2(1)(1)    
1 sibling30.31.751.52[0.97–2.38]0.07046%0.041
>1 siblings29.51.631.47[0.95–2.26]0.08162%0.003
Negative parental history of allergy26.9(1)(1)    
Positive parental history of allergy30.51.271.38[1.06–1.81]0.0180%0.445
(B) Positive skin prick tests
Male sex44.7(1)(1)    
Female sex44.10.980.88[0.68–1.14]0.33819%0.258
Age 15–3455.6(1)(1)    
Age 35–5547.90.740.71[0.51–1.00]0.05058%0.007
Age 55–7435.00.380.42[0.29–0.60]<0.00152%0.019
Nonsmoker46.4(1)(1)    
1–15 pack-year49.21.131.07[0.79–1.45]0.66427%0.183
>15 pack-year36.40.600.70[0.51–0.96]0.02738%0.085
No siblings48.6(1)(1)    
1 sibling47.01.080.72[0.48–1.08]0.11633%0.125
>1 siblings41.60.750.59[0.40–0.88]0.01054%0.013
Negative parental history of allergy41.6(1)(1)    
Positive parental history of allergy50.11.371.39[1.08–1.8]0.0110%0.666
(C) House dust mite-positive SPT
Male sex15.3(1)(1)    
Female sex14.60.950.91[0.65–1.26]0.5530%0.953
Age 15–3421.0(1)(1)    
Age 35–5514.80.550.46[0.31–0.68]<0.00147%0.035
Age 55–7411.90.330.27[0.17–0.42]<0.00159%0.005
Nonsmoker15.3(1)(1)    
1–15 pack-year16.91.061.06[0.73–1.56]0.74919%0.254
>15 pack-year12.30.670.81[0.53–1.23]0.3140%0.502
No siblingsa18.9(1)(1)    
1 siblinga16.91.000.76[0.47–1.24]0.26955%0.013
>1 siblingsa13.80.750.68[0.42–1.12]0.13161%0.004
Negative parental history of allergy13.3(1)(1)    
Positive parental history of allergy18.51.631.61[1.15–2.23]0.0050%0.911
image

Figure 2. Risk factors for Staphylococcus aureus enterotoxin IgE, for positive skin prick tests to any allergen, and to house dust mite. Odds ratios and 95% confidence intervals are obtained from logistic regression model as shown in Table 2. Estimates were mutually adjusted for all predictor variables. Estimates were weighted for case sampling and were done within center after which they were meta-analyzed.

Download figure to PowerPoint

Association of SE-IgE with sensitization to aeroallergens and total IgE

SE-IgE-positive subjects were more likely than SE-IgE-negative subjects to have SPT-HDM or SPT-ANY (age-/sex-adjusted OR: 1.97 [1.47–2.65] and 2.95 [2.23–3.90], respectively). In those sensitized to SE, 22.4% [18.6–26.7%] were also sensitized to house dust mite. However, 42.2% [37.3–47.4] of those positive to SE-IgE had negative SPT to all tested aeroallergens, suggesting that under this testing regime, 16.9% [14.9–19.1%] of the overall population could be considered monosensitized to SE-IgE.

Increased total IgE levels were associated with the presence of SE-IgE and SPT-HDM. This was particularly marked for SE-IgE (adjusted ratio of geometric mean total IgE in those with SE-IgE compared with those without SE-IgE 4.26 [3.77–4.81]) in contrast to SPT-HDM (ratio 2.01 [1.74–2.32]).

SE-IgE, atopy, and asthma

Asthma was present in 10.6% of the general population. Without mutual adjustment, there was a significant association of asthma with SE-IgE, with total IgE, and with positive skin prick tests (SPT-ANY and SPT-HDM) (Table 3A). The prevalence of SE-IgE was higher in asthmatics than in nonasthmatics (40.7% vs 28.0%, OR 2.10 [1.60–2.76], P = 0.001).

Table 3. (A) Prevalence of asthma and unadjusted and adjusted odds for asthma, for different sensitization states, obtained from logistic regression. (B) Association of SE-IgE with asthma, additionally adjusted for different markers of atopy
AUnadjusted estimatesAdjusted estimatesHeterogeneity
PrevalenceOROR95% CI P I 2P(I 2)
SE-IgE negative8.9(1)(1)[1.6–2.76]<0.0014.2%0.403
SE-IgE positive14.71.762.10
HDM SPT negative8.9(1)(1)[1.91–3.62]<0.00148.0%0.032
HDM SPT positive20.72.432.63
All SPT negative5.5(1)(1)[3.12–5.43]<0.0010.5%0.438
Any SPT positive17.03.424.12
Total IgE <100 kUA/l8.3(1)(1)[2.57–4.49]<0.00148.3%0.031
Total IgE >100 kUA/l19.32.893.40
B. OR for asthma in SE-IgE positive subjectsAdjusted estimatesHeterogeneity
OR95% CI P I 2P(I 2)
  1. A: All estimates were adjusted for age, sex, smoking, family history of atopy, sibship size, and history of severe childhood respiratory infections and were weighted for case–control status. OR: odds ratio.

  2. B: All estimates were weighted and adjusted as in Table 3A. The natural logarithm of total IgE concentration was used.

Additionally adjusted for any positive SPT1.75[1.31–2.32]<0.0010.0%0.449
Additionally adjusted for HDM-positive SPT2.01[1.51–2.67]<0.0019.2%0.355
Additionally adjusted for total IgE (log)0.93[0.66–1.3]0.6620.0%0.528

To assess dose dependency in the association of specific IgE with asthma, IgE values were grouped in tertiles (Fig. 3). Increasing SE-IgE level was associated with increasing risk of asthma (OR 1.20, 1.74, 2.57 for respectively first, second, and third tertile above 0.10 kUA/L; ptrend = 0.010). Intervals of the tertiles were [0.10–0.18[, [0.18–0.42[, and ≥0.42 kUA/L. Total IgE showed a concentration-dependent association with asthma (OR per naturally logged unit increase 1.69 [1.52–1.87]).

image

Figure 3. Odds ratios for asthma presence, for each of the tertiles of SE-IgE. The same logistic regression models were used as illustrated in Fig. 2. Estimates were mutually adjusted. Intervals of the tertiles were [0.10–0.18[, [0.18–0.42[, and ≥0.42 kUA/l.

Download figure to PowerPoint

To assess possible confounding or interaction, the association of SE-IgE with asthma was adjusted for SPT sensitization and for total IgE (Table 3B). The association of SE-IgE remained unchanged after adjustment for SPT-ANY or HDM-positive SPT. However, the association of SE-IgE with asthma was attenuated by adjusting for total IgE concentration (OR 0.93 [0.67–1.30], P = 0.662). Adjustment for SE-IgE had little effect on the association of total IgE with asthma (OR for high total IgE = 1.69 [1.48–1.91], P < 0.001). Similarly, the association of total IgE with asthma was little altered after controlling for SPT-ANY (OR 1.47 [1.32–1.64], P < 0.001) or SPT-HDM (OR 1.57 [1.41–1.76], P < 0.001).

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References

To our knowledge, this is the first large-scale population-based epidemiological study to demonstrate that sensitization to S. aureus enterotoxins is common in European adults, occurs independently of sensitization to other common aeroallergens, and is associated with high total IgE concentrations and asthma. In contrast to sensitization to other aeroallergens, SE-IgE is more common in smokers and is as common, and probably more common, in those with siblings compared to those without siblings.

Participants in this multicenter international population-based epidemiological study were initially identified by random sampling from relevant local population-based sampling frames and can be considered representative of the general population [14]. Although they were recruited into the clinical part of the study on the basis of symptoms reported in a postal survey, we have used appropriate statistical techniques to derive estimates that reflect the epidemiological pattern of sensitization and disease in the population. Response to the postal survey varied between centers, but we have previously shown that response was not related to disease status [14].

There is accumulating evidence that S. aureus enterotoxins might play a role in pathophysiology of chronic airway disease [19, 20]. Enterotoxins act as superantigens, which, unlike allergens, provoke an intense polyclonal immune response by nonspecifically binding the major histocompatibility complex (MHC) class II molecules with the T-cell receptor. This interaction is independent of specific antigen recognition, resulting in a polyclonal T- and B-cell activation [21].

We have shown that almost one out of three adults has serum IgE antibodies to a mixture of enterotoxins secreted by Staphylococcus aureus, and this proportion does not vary across European countries to the same extent as do other common aeroallergen sensitizations such as house dust mite. Sensitization to house dust mite and other common aeroallergens showed considerable geographic variation, consistent with previous epidemiological observations [15, 22, 23]. The relatively high prevalence of SE-IgE might indicate that exposure to S. aureus is common, even though some studies on colonization of the nasal vestibulum with S. aureus in healthy subjects reported that only 12–30% carry the bacterium (24). Little is known about the relationship of sensitization with enterotoxins and colonization by S. aureus in the airways, although it is commonly held that colonization is more frequent than SE-IgE and that colonization alone is insufficient to generate an IgE response.

Some groups of the population (men, smokers, and those with a history of allergic disease in the family) are particularly at risk of SE sensitization. Of particular note is the observation that SE-IgE was more common in heavy smokers, did not show a sharp decline in elderly, and was not decreased in those with many siblings. A lower risk of sensitization to aeroallergens in smokers (as seen in our study) has been reported previously [25], and although this could have been related to a healthy smoker bias, the same study showed that house dust mite-specific IgE was more common in smokers, an observation we could not confirm with skin prick test. Cigarette smoke has a disruptive effect on the epithelial barrier of both nasal and bronchial respiratory mucosa, and sensitization, particularly to SE, may be enhanced by loss of this epithelial barrier. Disease-related loss of epithelial barrier integrity is thought to explain the association of SE sensitization with atopic dermatitis [26].

The protective effect of larger family size on allergic sensitization and allergic airway disease is well documented [27, 28] and has been attributed to increased microbial exposure from siblings. We found that family size had no significant, but a possible increasing effect on sensitization to SE. We speculate this may be because large families may be a risk factor for an increased risk of colonization early in life with SE, outweighing any other immunologic benefit.

Subjects sensitized to SE were more likely to be sensitized to other aeroallergens, but even so, about one in six adults was considered to be only sensitized to SE-IgE after skin prick testing to ten of the most common aeroallergens in Europe. Total IgE levels were high in those who were SE sensitized, much higher than seen in those sensitized to aeroallergens, and this probably reflects superantigen-induced polyclonal IgE formation due to polyclonal B-cell proliferation and antigen-unspecific activation of V-beta subsets of T cells [21].

In line with previous small clinical studies, we have shown that SE-IgE is a risk factor for asthma in the general population [9, 10, 29, 30]. Unlike these earlier studies, the asthmatics in this study represent the full spectrum of disease, which may explain why the strength of our associations is smaller than previously reported. The presence of asthma was ascertained by questionnaire, based on the self-reporting of a set of symptoms, as used previously [13], but similar associations were found when self-reported physician-diagnosed asthma was considered (data not shown). Our symptom-based definition has been validated as described previously [31, 32]. Sensitization to SE-IgE has also been associated with CRS with nasal polyps and atopic dermatitis. Although no symptom-based definition of nasal polyposis is available, adjustment of the association of SE-IgE with asthma, for either CRS (OR 2.06 [1.57–2.71]), allergic rhinitis (OR 1.76, [1.31–2.36]), or atopic dermatitis (OR 2.09 [1.58–2.77]) did not significantly alter the estimates, indicating an independent relationship.

Our study is the first to show that SE-IgE is associated with asthma in a concentration-dependent manner, independent of the increased tendency of those with SE-IgE to also be sensitized to aeroallergens. The relationship with asthma was attenuated by adjustment for total IgE. As SE-IgE and total IgE are considered to be very closely related on the same causal pathway [12], this effect of adjustment cannot be interpreted as evidence of confounding [33]. In contrast, it suggests that the pathophysiologic effect of SE-IgE in asthma is predominantly mediated through high total IgE. Supporting this, treatment with omalizumab, a monoclonal antibody against IgE, has been shown to be effective in nasal polyposis, a disease characterized by S. aureus enterotoxins [34].

Conclusion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References

This is the first population-based study to describe the epidemiology of specific IgE to Staphylococcus aureus enterotoxins. Of relevance to public health, we here show that IgE sensitization to SE is common in Europe, may occur in the absence of sensitization to other allergens, and follows an epidemiological pattern different to that seen for aero allergen such as house dust mite sensitization, probably reflecting a different pathophysiologic basis. Of great clinical relevance is that IgE to SE is associated with the presence of asthma, independent of sensitization to other allergens. This effect may be mediated through its association with strongly increased total IgE concentrations via the polyclonal superantigen action of enterotoxins.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References

The authors wish to thank all GA²LEN Survey Follow-Up contributors: Jaap H. Akkerdaas3, Abir Al-Kalemji15, Kjell Alving (Department of Women's and Children's Health, Uppsala University), Jesper Baelum15, Heidrun Behrendt (Zentrum Allergie und Umwelt, Technische Universität und Helmholtz Zentrum München, München), Helén Bertilsson4, Anders Bjerg (Krefting Research Centre, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg), Angelina Bockelbrink11, Barbro Dahlén (Department of Medicine Huddinge: Lung and Allergy Research, Karolinska Institutet, Stockholm), Sven-Erik Dahlén12, Alexandra Ek (Institute of Environmental Medicine: Experimental Asthma and Allergy Research, Karolinska Institutet, Stockholm), Linda Ekerljung (Krefting Research Centre, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg), Wytske Fokkens17, Karl Franklin (Department of Public Health and Clinical Medicine: Medicine, and Department of Surgery, Umeå University, Umeå), Philippe Gevaert1, Joanna Gluck16, Louisa Gnatuic2, María Gunnbjörnsdottir7, Tari Haahtela (Skin and Allergy Hospital, University of Helsinki, Helsinki, Finland), Gunilla Hägg7, Gunilla Hedlin (Centre for Allergy Research and Department of Women's and Children's Health, Karolinska Institutet, Stockholm), Teresa Jäger10, Bengt Järvholm4, Thomas Keil11, Anette D. Kjeldsen (Department of Otorhinolaryngology, Odense University Hospital, Odense, Denmark), Olga Krysko1, Bibi Lange (Department of Otorhinolaryngology, Odense University Hospital, Odense, Denmark), Kjell Larsson (Institute of Environmental Medicine: Lung and Allergy Research, Karolinska Institutet, Stockholm), Maria Leibl10, Jan Lötvall (Krefting Research Centre, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg), Carlos Loureiro13, Bo Lundbäck (Krefting Research Centre, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg), Joanna Makowska5, Andrei Malinovschi (Department of Medical Sciences: Clinical Physiology, Uppsala University), Mareike McIntyre10, Lars Modig4, Katarina Nisser7, Dan Norbäck (Department of Medical Sciences: Occupational and Environmental Medicine, Uppsala University), Inga Sif Olafsdottir7, Claudina Perez-Novo1, Florian Pfab10, James Potts2, Johannes Ring10, Barbara Rogala16, Hitasha Rupani5, Barbara Rymarczyk16, Monica Salagean5, Ulrike Spetz Nyström7, Pär Stjärne (Department of Otorhinolaryngology, Karolinska Institutet, Stockholm), Kjell Torén (Environmental and Occupational Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg), Nicholas Van Bruaene1, Paul Van Cauwenberge1, Koen Van Crombruggen1, Thibaut Van Zele1, Serge A. Versteeg3, Chatrin Wahlgren4, Thorsten Wehrend11, Jan E. Zejda16, Nan Zhang1.

Author contributions

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References

PB, DJ, and CB designed the study and the instruments that were used. PT, RN, and DJ analyzed the data. PT, CB, DJ, and PB interpreted the results. PT and DJ wrote the manuscript to which they equally contributed. All authors took part in data collection and reviewed the manuscript.

Funding

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References

This study was funded by the Sixth European Union Framework Program for Research, contract no. FOOD-CT-2004-506378 and by Fonds Wetenschappelijk Onderzoek Vlaanderen (Flemish Government, Belgium), contract no. G.0854.09.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Author contributions
  9. Funding
  10. Conflict of interest
  11. References