LLRs to Hymenoptera stings may be IgE-mediated or not. Green et al. in 1980 reported that, in 22 patients who had LLRs from Hymenoptera stings, approximately half of them had venom-specific IgE antibodies, and that no correlations could be found between the presence of venom-specific IgE and age, sex, sting location, atopic history, or prior stings . This suggested to the authors that after an LLR from an insect sting patients must be individually assessed for the presence of venom-specific IgE and considered for specific immunotherapy. Subsequent studies focused the interest on IgE-mediated LLRs and particularly on the risk to develop SRs after an LLR. Four studies on this issue are available thus far. The first two studies have similar results, with a rate of SRs of 4 % in children  and 5 % in adults . Instead, contrasting data were reported in the two more recent prospective studies, with a rate of SRs of 7 % in children  but no SR at all in a group of patients including both children and adults . The major issues concerning these studies are the known limitations of retrospective studies and the low number of patients in the prospective studies. Concerning the first issue, the biases which can negatively impact the reliability of this type of study include the selection bias, the difficulty in assessing the temporal relationship and control of outcomes, and the need of large sample size for rare outcomes . In the case of allergic reactions to insect stings, the psychological characteristics of patients may influence their decision to undergo a medical evaluation, being possible that patients claim such evaluation after an LLR or feel it unnecessary after an SR. These bias weaken the findings of retrospective studies, while the prospective evaluation allows more reliable observations. Still, a small number of patients makes it unlikely for statistical reasons to achieve a sound consistency also for prospective studies. In particular, the contrasting outcome of the two prospective studies in patients with LLRs—7 vs. 0 %—were observed in two groups of 44 children and 23 children and adults, respectively. In the prospective part of our study we included 81 adult patients with at least two LLRs, but only 53 were evaluated for 3 years, by annual control visits, 31 of them being re-stung with no SR, that confirms the finding by Fernandez et al. . Concerning the retrospective part of the study, in the 396 patients with an SR only 4.2 % had had a previous single LLR as debut of allergy, that is the identical rate observed in the study by Mauriello et al.  in 118 patients with SR.
The most interesting data of our study is the correlation between the number of LLRs and the risk to develop an SR. In fact, we observed that in patients with a single LLR as the first manifestation of venom allergy there is a risk, although low, of SR to a subsequent hymenoptera sting, while there is no risk of SR in presence of at least two previous consecutive LLRs.
These findings may be useful in practical management of patients sensitized to hymenoptera venom, especially concerning the prescription of epinephrine auto-injectors. Actually, patients with at least two LLRs to stings are unlikely to need an auto-injector because there is no apparent risk of SR, while in patients evaluated after a single LLR the risk of SR cannot be ruled out and the availability of epinephrine for auto-injection is worthwhile. However, only a prospective study with a prolonged follow-up on a large number of patients with LLRs to stings could definitely settle the issue.