To cite this article: Blum S, Gunzinger A, Müller UR, Helbling A. Influence of total and specific IgE, serum tryptase, and age on severity of allergic reactions to Hymenoptera stings. Allergy 2011; 66: 221–228.
Background: The aim of this study was to analyze the influence of total serum IgE and other potential risk factors on severity of systemic allergic Hymenoptera sting reactions.
Methods: In a retrospective analysis of one thousand and two patients referred for insect allergy over 5 years, 865 reported systemic allergic sting reactions, most often by honey bees and wasps. In 758, total IgE, venom-specific IgE, and baseline tryptase levels were available and analyzed together with atopy state, age, and sex in relation to severity of sting reactions according to H. L. Mueller.
Results: In a binary logistic regression model considering, besides IgE, also other risk factors for severity, an influence of total and specific IgE on severity of systemic allergic sting reactions could not be shown, while high severity of systemic allergic sting reactions was significantly more often reported in patients with a baseline tryptase of ≥11.4 μg/l (P < 0.0001) and higher age (P = 0.026). In a bivariate analysis, however, in patients with grade IV reactions total IgE (P = 0.003) and honey bee venom-specific IgE (P = 0.001) were significantly lower than in lower severity grades. Bee venom-specific mean IgE rank was significantly higher in bee than in Vespula venom allergic patients (P = 0.0001).
Conclusions: Connection of high severity sting reactions with lower IgE is mainly because of older age, which is associated with lower total IgE, and moreover with cardiovascular disease and elevated baseline serum tryptase, which are both risk factors for severe reactions.
baseline serum tryptase
honey bee venom
large local reaction
systemic allergic reaction
specific serum IgE
total serum IgE
Based on the results of several studies, various risk factors may facilitate the development of Hymenoptera venom allergy and increase the severity of systemic allergic reactions (SARs) to Hymenoptera stings (1, 2). A short interval between two stings is often followed by SAR to the second sting (3). However, being stung very frequently appears to induce tolerance. Beekeepers stung less than 10 times a year have a higher risk of SAR than those stung more than 200 times a year (4–6). Severe SAR is significantly associated with an elevated baseline serum tryptase (bT) (7–9), which indicates an elevated whole body mast cell load. In systemic mastocytosis, Hymenoptera sting reactions are most often severe (10–12). Severity of SAR is also associated with age. In children, 60% of systemic reactions are mild (13), whereas in adults respiratory or cardiovascular symptoms occur in about 70% (14). An atopic constitution may influence symptoms of SAR (15, 16). More severe SAR may be favored in older people by pre-existing cardiovascular and pulmonary disease and their medication (17–19). Studies on patients who died from Hymenoptera sting reactions revealed that most of those patients were between 50 and 70 years old, and according to autopsies over 90% suffered from coronary heart disease and/or chronic obstructive lung disease (20). Also, medical treatment of cardiovascular disease, especially with beta-blockers or angiotensin-converting enzyme (ACE) inhibitors, may result in more severe SAR (17, 18).
In a recent study from Austria (21), it was observed that patients with Hymenoptera venom allergy and high levels (>250 kUa/l) of total IgE (tIgE), predominantly develop milder SAR, while patients with low tIgE (below 50 kUa/l) more often develop severe SAR. A tendency for lower venom-specific IgE (sIgE) in patients with severe SAR was also described (21). While it is well known that the presence of venom sIgE is significantly associated with a history of SAR to Hymenoptera stings, the level of venom sIgE did not seem to be related to severity (1, 2). The influence of tIgE on severity has not been analyzed so far by other studies. Atopy is associated with an increased serum level of tIgE, but is not more frequent in Hymenoptera sting allergic patients than in the general population (15, 16). However, atopic patients were reported to more often develop respiratory allergic sting reactions (22).
The first aim of our study was to examine the hypothesis (21) that tIgE and sIgE levels influence the severity of SAR to Hymenoptera stings. Compared to the study of Sturm (21), we included a larger population of patients with SAR to Hymenoptera stings and venom allergy documented by intracutaneous skin tests to venoms and presence of venom sIgE serum antibodies. Additional goals of the study were to look at other potential risk factors for severe reactions, namely age, sex, atopy, and elevated bT.
Material and methods
In this retrospective study, the charts of all patients with insect sting allergy referred to the Zieglerspital over 5 years from January 1, 2003 to December 31, 2007 have been evaluated. Of the 1002 referred patients, 865 showed SARs to Hymenoptera stings. One hundred and thirty-seven patients were excluded with only large local reactions (LLRs), reactions to non-Hymenoptera insects, or symptoms suggesting a non-IgE-mediated pathogenesis. Of the remaining 865 patients, 758 had also complete diagnostic data (total IgE, specific IgE, degree of severity, baseline tryptase, age, sex), and 731 of these were also tested for atopy by prick tests with frequent local inhalant allergens. We did not exclude patients with cardiovascular disease, on beta-blockers, ACE-inhibitors, or any other medical treatment and patients with elevated bT (≥11.4 μg/l).
Because age was also reported to be a risk factor (1, 13), we examined the influence of age on the severity of SAR. We divided patients into the following five categories: ≤16, >16 to ≤30, >30 to ≤40, >40 to ≤55, and >55 years.
The degree of severity of SAR was classified in four grades according to H. L. Mueller with modifications (1, 23).
Intracutaneous skin tests with honey bee venom (BV) and Vespula venom (VV) (Pharmalgen; ALK-Abello, Horsholm, Denmark) were performed at least 3 weeks after a SAR to avoid false-negative results during the refractory period. Their results were analyzed by intradermal skin test endpoint titration (24): 0.02 ml of venom solution were injected intradermally in increasing concentrations – from 0.00001 to 1 μg/ml – in the volar surface of the forearm. A wheal ≥5 mm in diameter with surrounding erythema was defined as a positive reaction.
Serum venom-specific IgE, total IgE, and baseline tryptase
Serum levels of tIgE, sIgE to BV and VV, and bT were measured using UNICAP (Phadia, Uppsala, Sweden) according to the manufacturer’s instructions. Patients were grouped according to tIgE levels as in the Austrian paper (21): Group 1 < 50 kUa/l, Group 2 50–250 kUa/l, Group 3 > 250 kUa/l. sIgE were categorized according to CAP classes 0, 1, 2, 3, 4, and ≥ 5. For the statistical analysis, individual values of tIgE and sIgE were, however, also included. Serum tryptase levels ≥ 11.4 μg/l (95th percentile of the general population) were considered to be elevated.
The atopy score was based on skin-prick tests (ALK-Abello) with 14 common regional inhaled allergens. If the patient had a typical wheal-and-flare reaction (wheal >3 mm in diameter with surrounding erythema) to at least one of the allergens, an atopic disposition was assumed.
Modeling severity using logistic regression Two binary logistic regression models were used to predict the occurrence of a grade IV reaction using total and specific IgE, one model using the IgE data in categories and the other model using uncategorized IgE values. Both models further included age, sex, and bT as potential other risk factors for the occurrence of a grade IV reaction. All analyses were performed in sas 9.1.3 (SAS Institute Inc., Cary, NC, USA) and graphics using R 2.7.1 (R Foundation for Statistical Computing, Vienna, Austria).
Bivariate analysis Wilcoxon rank sum test was used to compare individual values of tIgE and sIgE of patients with severity grade IV vs grade I–III reactions. Wilcoxon rank sum test was also used for the analysis of (categorized) age vs severity, for bT vs severity, and for atopy vs tIgE.
Specific IgE to BV and specific IgE to VV were compared by Mann–Whitney U-test. The Spearman correlation coefficient between (uncategorized) age and individual bT values was computed to investigate a possible association. The Spearman correlation coefficient was also used for the analysis of correlation between individual values of tIgE and sIgE and between individual tIgE values and age. Chi-square statistics were used for elevated bT vs atopy, for severity vs atopy, and for age classes vs atopy.
Clinical and laboratory data of the 758 patients with complete laboratory data are given in Table 1.
|Patients with SAR and complete laboratory data|
|Number of patients||758|
|Mean age (range)||37.9 (4–82 years)|
|Severity of SAR||Grade I: 95|
|Grade II: 144|
|Grade III: 201|
|Grade IV: 318|
|History of SAR to sting of|
|Insect not identified||266|
|i.c. skin tests positive at 1 μg/l to|
|Both insects||463 (61%)|
|sIgE ≥ 0.35 kUa/l to|
|Both insects||421 (56%)|
|bT < 11.4 μg/l||714|
|bT ≥ 11.4 μg/l||44|
Total and specific IgE vs degree of severity
In the 758 patients with complete data for the laboratory tests (tIgE, sIgE and bT), a cumulative logistic regression model was used to predict the severity of an allergic reaction using tIgE, sIgE (BV and VV), bT, sex and age.
In this group, the SAR was caused from history by Vespula stings in 256 (33.6%), by honeybee stings in 212 patients (28.1%), and 290 patients (38.3%) couldn’t identify the stinging insect or identified other Hymenoptera insects like hornets or bumblebees. Of the 758 patients, 547 had positive skin tests to BV and 651 to VV; BVsIgE was detected in 551 and VVsIgE in 579 patients (Table 1). Of those 758 patients, 519 suffered from a severe reaction (grade III or IV) and 239 showed a mild reaction (grade I or II) after a Hymenoptera sting. Double positivity to BV and VV in skin tests (61%) and in sIgE (56%) was frequently observed (Table 1).
We could not find a significant effect of tIgE on the occurrence of a grade IV reaction in the logistic regression models either by using categorized (P = 0.531) or by using uncategorized (P = 0.592) tIgE values (Table 2, Fig. 1A). There was only a slight tendency for lower tIgE in patients with grade IV reactions and higher tIgE in grade III reactions (Fig. 1A). When using the bivariate analysis by Wilcoxon rank sum test; however, significantly lower tIgE was found in grade IV than in grade I–III reactions (P = 0.003) (Table 3).
|Variable||Degree of freedom||χ2||P|
|sIgE BV categories||5||4.475||0.483|
|sIgE VV categories||5||4.217||0.519|
|sIgE BV ind values||1||8.739||0.003|
|sIgE VV ind values||1||0.123||0.726|
|Grade IV vs grade I–III||Age||Baseline tryptase||Total IgE||Bee venom-specific IgE||Vespula venom-specific IgE|
|Wilcoxon rank sum P-value||0.0019||0.004||0.0028||0.0001||0.268|
Using categorized sIgE values, also no significant effect of sIgE to venoms on the occurrence of a grade IV reaction was found (P = 0.483). However, uncategorized BVsIgE values had a significant effect on grade IV reactions (P = 0.0031), whereas there was no significant effect for uncategorized VVsIgE. (Table 2, Fig. 1B,C). These findings were confirmed in Wilcoxon rank sum test (BVsIgE P = 0.0001; VVsIgE P = 0.27) (Table 3). Mean values of BVsIgE were significantly higher than those of VVsIgE (Mann–Whitney U-test; P < 0.0001).
Age vs severity
Statistical analysis revealed a significant increase in severity with higher age both in the logistic regression model (χ2 = 10.988, P = 0.0267) and in the Wilcoxon rank sum test (P = 0.0019). Children reported more often mild SAR of grade I and II in 52/140 (40.8%) vs 36/146 (24.7%) of adults over 55 years. Adults over 55, however, reported reactions of grade IV in 73/146 (50.0%), children in 41/140 (29.3%) (Table 4, Fig. 2).
|Category of age (years)||Category of severity||Total|
n (% row)
n (% row)
n (% row)
n (% row)
|n (% column)|
|Total (% row)||95 (12.53)||144 (19)||201 (26.52)||318 (41.95)||758 (100)|
Tryptase as a risk factor
Of the 758 patients, 44 (5.8%) had an elevated bT (>11.4 μg/l) and 714 patients (94.2%) were in the normal range. The multivariate analysis showed a significant association between bT and severity (P < 0.0001). This significant association was confirmed by Wilcoxon rank sum test (P = 0.0019). No association between elevated bT and atopy was found (χ2 = 0.0378; P = 0.846). However, there is a significant positive correlation according to Spearman rank between bT and age (r = 0.1599; P < 0.0001).
According to skin test criteria, 253 (34.6%) of the 731 patients tested were atopic. Atopy has an influence on severity on the whole (χ2 = 8.037; P = 0.045). Atopic patients more often report grade III reactions (41.7%) than nonatopic patients (32.1%) (χ2 = 5.0193; P = 0.0251). Atopy was strongly associated with an increased total IgE (Wilcoxon rank sum test; P = 0.001), but was not more frequent in patients below 30 (34.3%) than over 30 years (34.7%).
The prevalence of SAR to Hymenoptera stings in Europe varies from 0.3% to 7.5% (2) in different countries. In Switzerland, it was found to be at 3.5% (25). In our study, 519 of 758 patients suffered from a severe SAR grade III and IV, but only 239 from a mild grade I and II SAR. The high proportion of severe reactions is surprising and probably because of the fact that patients with severe SAR more often search medical aid and more often are referred to allergy specialists.
Among several other risk factors for severe SAR, like older age, concurrent cardiovascular disease (17), mastocytosis, and elevated baseline serum tryptase (7–12), a low serum level of tIgE was recently claimed to be associated with a history of severe SAR (21). The authors also observed a tendency – although not significant – for a similar association between venom sIgE and severe SAR. In our study, we could, however, only partly confirm these findings. We observed a tendency to lower tIgE and to lower BVsIgE in BV allergic patients with a history of SAR grade IV (Fig. 1A–C) in the logistic regression program when using categorized tIgE and sIgE levels, which did not reach statistical significance. When using individual values for total and sIgE, however, BV sIgE was significantly reduced in grade IV reactions, while tIgE and VVsIgE were not. According to a bivariate analysis, however, both total and BV-specific IgE were significantly lower in patients with grade IV than with grade I-III reactions. The differences between the results of the logistic regression program and the bivariate analysis can be explained by the fact that the latter does not consider other risk factors like age and bT: older age is associated with lower tIgE, and moreover with more frequent cardiovascular disease and higher bT, which are both associated with severe allergic sting reactions (7–12).
Several reasons for some discrepancies between the results in the two studies must be discussed:
- 1 The classification of severity of SAR used was different: in the Austrian study (21) grading after Ring and Messmer (26) was used, in our study grading after H. L. Mueller (1, 23) (Table 5). In the Ring classification, respiratory and cardiovascular symptoms are included in grades II–IV; in our classification, respiratory symptoms are predominant in grade III, cardiovascular symptoms in grade IV. The observation of highest tIgE mean values in grade III in our system is probably because of the fact that respiratory sting reactions are more prevalent in atopic patients (5, 22), who have a significantly higher mean total IgE than nonatopics. The significantly increased prevalence of atopy found in patients with grade III reactions when compared to the other reaction grades confirms this. In both grading systems, the grade may not always be easy to assign to the individual patient and different doctors may decide differently.
- 2 In the Austrian study (21), 150 patients referred during only one season were included, when compared to 758 referred over 5 years in our study. The much larger number is in favor of our study, owing to the longer observation period; however, more different doctors will have classified the severity grades in our study, and this may have been a disadvantage.
- 3 The inclusion criteria were different: in the Austrian study (21) but not in ours, patients with cardiovascular disease, on beta-blockers or ACE-inhibitors, and those with elevated bT were excluded. Most patients with cardiovascular disease are elderly and have been shown to develop severe SAR more frequently than younger individuals (17). Concurrent treatment with beta-blockers and ACE-inhibitors is well known to increase severity of SAR (10), and an elevated bT is definitely associated with more severe SAR (7–9). Moreover, we have shown in this study that bT levels increase while tIgE decrease in correlation with age. The increase in bT with age was first described by Kucharewicz (9) and confirmed in this study. The exclusion of 70 of the original 220 patients for cardiovascular disease, elevated bT, and age matching could have influenced the results of the Austrian study considerably.
- 4 The responsible Hymenoptera species differed in Austria and in Switzerland: while 73% of the patients in Austria were primarily sensitized to VV, almost equal numbers were primarily sensitized to BV and VV in Switzerland. Differences in venom allergy to these two species with regard to re-sting reactions (27, 28), safety, and efficacy of venom immunotherapy (29) are well documented and in this study we also found significantly higher sIgE levels in BV than in VV allergic patients. This is most likely because of sIgE to cross-reactive carbohydrate determinants, which are probably of little clinical relevance and are more often found in BV allergic patients. Major BV allergens like phospholipase A2, hyaluronidase, and acid phosphatase are all glycosylated, while the two major VV allergens, antigen 5 and phospholipase A1, are not, and glycosylated hyaluronidase of VV is probably not a major allergen (30–32). The lacking association between severity and sIgE in patients with vespid venom when compared to patients with BV allergy may be related to these differences, but will have to be analyzed in further studies.
- 5 The statistical analysis of the data was different: while in the Austrian study (21) the two sample Kolmogorow Smirnow analysis was used to process the data, we preferred a logistic regression model considering the multiple factors influencing tIgE levels in a better way. With increasing age tIgE decrease; in patients with an atopic constitution they are significantly increased; in BV allergic patients, sIgE is significantly higher than in VV allergic patients. Moreover, we analyzed individual levels of tIgE and sIgE besides the categorized IgE values used in the Austrian studies (21).
|Classification of SARs to Hymenopter stings after H. L. Mueller (23), with modifications (1)|
|Grade I||Generalized urticaria, itching, malaise, anxiety|
|Grade II||Any of the above, plus 2 or more of the following:|
Angioedema (grade II also if alone), constriction in chest, nausea, vomiting, diarrhea, abdominal pain, dizziness
|Grade III||Any of the above, plus 2 or more of the following:|
Dyspnea, wheezing, stridor (any of these alone are grade III), dysphagia, dysarthria, hoarseness, weakness, confusion, feeling of impending disaster
|Grade IV||Any of the above, plus 2 or more of the following:|
Fall in blood pressure, collapse, loss of consciousness, incontinence (urine, stool), cyanosis
|Classification of severity of anaphylactic reactions after Ring und Messmer (26)|
|Respiratory collapse||Heart/circulation collapse|
In conclusion, based on our study in a large number of Hymenoptera venom allergic patients, we could confirm a connection between lower total IgE serum levels and severity grade IV SARs to Hymenoptera stings, which is mainly because of older age associated with cardiovascular disease, lower total IgE, and higher baseline serum tryptase.
We thank Chris Kopp from the Institute of Mathematical Statistics and Actuarial Science of the University of Bern, Switzerland for his most valuable cooperation in statistics.
Allergy foundation, Spital Ziegler, Spitalnetz Bern, Switzerland.