Patients

Nineteen patients (15 males, 4 females, age range 16–75 years, mean age 45.9 years) with systemic reactions to Hymenoptera stings of at least Mueller grade II [3], positive skin tests results to Polistes venom, and who had not previously been treated by VIT (to avoid treatment-induced changes in specific IgE), were included in the in vitro study, which aimed to assess the level of cross-reactivity to P. dominulus and mAP. A result from skin tests or CAP system showing monosensitization to P. dominulus was an exclusion criterion.

In vivo tests

Skin tests were performed by P. dominulus venom from Anallergo (San Piero a Sieve, Florence, Italy) and mAP venom from Stallergenes (Antony, France), by an initial prick test at 100 mcg/ml, followed, if negative, by intradermal testing at 0.1 mcg/ml and 1 mcg/ml.

In vitro tests

Sera from all patients were analyzed using CAP inhibition to determine sensitization, following the method previously described by Caruso et al. [7], Savi et al [9)] and Straumann et al [11]. Briefly, two 100 μL aliquots of serum were incubated separately for 12 h at 4°C with 200 mL of P. dominulus or mAP venom at increasing dilutions (0 μg/mL; 25 μg/mL; 50 μg/mL; 100 μg/mL; 200 μg/mL). Subsequently, specific IgE values (sIgE) against each of the venoms were determined in the prepared samples [9]. The extent of homologous (blockage of venom-specific IgE by the same venom) and heterologous (blockage of the venom-specific IgE by the other venom) inhibition was computed using the following formula: % inhibition = 100 - [IgE inhibited sample (kU/L) x 100/IgE anti-venom (kU/L) at zero concentration of venom]. Serum samples that produced an homologous inhibition of less than 70% were excluded from evaluation (11). Heterologous inhibition >70–75% was considered to be strongly suggestive of cross-reactivity among P. dominulus and mAP venoms.(9) The protocol used for CAP inhibition was deposited in protocols.io with the identifier dx.doi-org/10.17504/protocols.io.ht7b6rn.

The venom preparations tested were the same as those used for the skin tests. Owing to the limited number of patients, the level of the inhibition obtained with P. dominulus and mAP were statistically compared using the non-parametric Wilcoxon test.

Venom immunotherapy

For the clinical part of the study, 191 patients (138 males, 53 females, age range 10–82 years, mean age 48.5 years) with positive tests results to Polistes venom prior to VIT and subsequently restung by an identified insect were retrospectively analyzed to determine the rate of protection from Polistes venom. Monosensitization to P. dominulus was an the exclusion. Of the 191 patients, 89 were treated with mAP and 102 were treated with P. dominulus venom. All patients were instructed to recognize Polistes spp, and distinguish it from Vespula spp., based on anatomical characteristics. In particular, the wasp waist, e.g. the spindle-shaped morphology of the thorax-abdomen joint, is typical of Polistes spp, while in Vespula and Vespa spp this joint is truncated. Only stings from Polistes spp. were considered in the analysis. The difference in the rate of protection from stings were analyzed using the chi square test.

All participants or parents/tutors for patients of pediatric age gave their written informed consent.

Comparing the observations from studies on P. dominulus in Italy

In this study we found that in sera from patients with allergic reactions to Polistes stings the CAP-inhibition obtained with P. dominulus was higher than that obtained with mAP. Namely, it was 79.8% for sIgE P. dominulus with the homologous venom vs. 64.2% with the mAP (p <0.01), and 73.6% for the sIgE mAP with the homologous venom vs 80.1% with P. dominulus (n.s.). These differences are lower than those reported by Severino et al, who found a value using RAST inhibition of 80.9% on average with P. dominulus, whereas the inhibition by with mAP was 46.1% on average [10]. To estimate the reliability of these findings we must consider two important issues. First, the number of patients was low (19 in our study and 10 in the study by Severino et al), thus, the results could be stochastic. Indeed, it is quite surprising that observations based on only 10 patients have led to the suggestion to use exclusively P. dominulus for VIT in patients allergic to Polistes venom.

International knowledge on Polistes venom.

The consensus document from the European Academy of Allergy and Clinical Immunology (EAACI) on Hymenoptera venom allergy stated that the cross-reactivity among European species of Polistes (P. dominulus, P. gallicus) was stronger than that between European and American species, but there was no recommendation to use only P. dominulus venom [4]. Second, the importance of the data from RAST- or CAP-inhibition has to must be measured against clinical data. Before P. dominulus venom became available for VIT in Europe, all patients were treated using mAP with no report of complete treatment failure, i.e. fatal reactions to stings; therefore, the suggestion to use only P. dominulus venom is flawed. In addition, P. dominulus is present in the United States [14] and also in this country no fatal reaction has been reported in patients treated with VIT by mAP venom, which is the only Polistes venom available. The rate of incomplete protection, i.e. to have a systemic reaction to a sting following VIT, depends upon the kind of venom. Muller et al reported that honeybee VIT provided complete protection in 77% of patients compared with 91% in patients treated with Vespula spp. VIT [15]. More recently, better results were observed in a multicenter survey in Europe, with a rate of complete protection of 89% with honeybee VIT and 96% with vespid VIT; however, for vespids, there was no distinction between Vespula spp. and Polistes spp. [16]. In the present study, we evaluated the rate of complete protection from stings in patients allergic to Polistes venom, distinguishing P. dominulus VIT from mAP VIT. The results showed no statistically significant difference, based on a rate of incomplete protection of 3.9% and 4.5%, respectively, with no life-threatening reaction. This is comparable to the 4% rate of incomplete protection for vespids reported by Rueff et al, though they did not separately analyze Vespula and Polistes stings [16]. Again surprisingly, in the study by Severino et al. on 104 patients monosensitized to P. dominulus and treated by VIT when Polistes was recognized as the stinging insect, 11 (10.6%) had systemic reactions, two of which were severe (one grade III and one grade IV) [17].

These contrasting data should stimulate discussion regarding the allergens from Polistes venom. The first study dates back to 1982,; when studying sera from 62 American patients allergic to Polistes venom, four species' venoms were more closely related to each other than various Vespula spp were related to each other. The authors concluded that the commercial Polistes venom mix appeared to contain all of the relevant allergens [18]. In the following years the major allergens were identified as hyaluronidase, phospholipase A and B and antigen 5, all showing variable levels of cross-reactivity [19–20]. Concerning Polistes spp, these allergens are classified as follows: phospholipase A1 is Pol a 1 (from P. annularis) and Pol d 1 (from P. dominulus), hyaluronidase is Pol a 2, and antigen 5 is Pol a 5 and Pol d 5. Hyaluronidase is the main responsible of the cross-reactivity among venom from different species, including not only other vespids but also Apis mellifera, while antigen 5 is considered to be species-specific [21]. However, in a study that determined the complete amino acid sequence of antigen 5, the alignment of antigen 5 from the European species P. gallicus and the American species P. annularis showed an 85% identity. The sequence identity increased to 98% within the same subgenus, suggesting that the presence of specific epitopes on antigen 5 should be related to the variations on the superficial loops [13]. Remarkably, although antigen 5 was identified more than 30 years ago and investigated in a number of studies, its protein nature and biological role had not yet been ascertained. Of relevant interest, in the venom from the wasp Polybia scutellaris, which belongs to the Polistinae subfamily, but does not cause allergic reactions, the amino acid analysis of antigen 5 showed a close similarity with antigen 5 of vespids from different species. However, two forms of antigen 5 were isolated, with the amino acids residues at positions 5 and 11 in P. scutellaris antigen 5 differing from those of the previously known sequences for antigen 5, suggesting that one or the other might be responsible for the lack of allergenicity of the P. scutellaris venom [22]. Vinzon et al. reappraised the issue, demonstrating that in mice Poly s 5 from P. scutellaris induced IgG antibodies which cross-reacted with Pol a 5, while Poly s 5 induced only minimal amounts of IgE and was a poor inducer of basophil-mediator release. This finding led the authors to propose Poly s 5 as a candidate for VIT [23], but it should also stimulate investigations on the grade of allergenicity of antigen 5 from more common Polistinae including P. dominulus. Indeed, a different allergenicity of antigen 5 from Polistes spp. could be a possible factor explaining the different clinical outcome observed with Polistes VIT. This is similar to what is known for food allergy; for example, a divergent allergenicity of apple strains mainly depended on different expression levels of the major allergen Mal d 1, with a proline residue in position 111 causing a drastic reduction of allergenicity [24].

Conclusions

Currently, there is no scientific report supporting the use of only P. dominulus for VIT in patients allergic to Polistes venom, unless the patient is monosensitized to the latter. In patients with positive test results to both P. dominulus and mAP, performing VIT with any of the two venoms has the same odds of clinical success.