Diagnosis and allergen immunotherapy treatment of polysensitised patients with respiratory allergy in Spain: an Allergists’ Consensus

Background Polysensitisation is common in patients with respiratory allergy in Spain. Selection of the best allergen immunotherapy (AIT) is difficult in polysensitised patients. The present study was designed to help allergists better identify relevant allergens in these patients and to improve the selection of AIT in Spain. Methods Sixty-two Spanish allergists answered a survey containing 88 items divided into four groups: 1) general approach to polysensitised subjects; 2) sensitisation profile involving mite, animal dander and moulds; 3) grass and olive pollen co-sensitisation, and 4) other pollen polysensitisation profile (weed and tree pollen). The Delphi method was used. Results A consensus was achieved for 83% of items (92%, 81%, 83% and 73% of the four groups analysed, respectively). Only polysensitised patients with clinical relevance should be considered polyallergic. A detailed medical history (clinical symptoms and medication) together with a profound knowledge of allergens present in the patient’s environment are essential for diagnosis. Skin prick tests (SPTs) are not adequate to decide the clinical relevance of each allergen. Serum specific IgE against allergen sources adds value to SPT but molecular diagnosis, when possible, is strongly recommended, especially in pollen-allergic patients. Specific allergen challenge tests are difficult to perform and not recommended for daily practice. Regarding AIT composition, up to three allergens can be used in the same vaccine, but only related allergens may be mixed. In some cases more than one vaccine may be needed. Conclusion Some criteria have been established to improve diagnosis and AIT prescription in polysensitised patients.


Background
Respiratory allergy represents an overall public health problem because of its prevalence, morbidity, impact on quality of life and cost [1][2][3][4]. Pollen [5,6] and house-dust mites (HDM) [7] are the most common allergens related to respiratory diseases, followed by animal dander and moulds [8]. Polysensitisation, defined as the presence of more than one specific IgE sensitisation against nonrelated allergen sources, is frequent in Spain [6]. However, the exact clinical relevance of each sensitisation is sometimes difficult to establish and, consequently, allergists may have difficulty in deciding which is the most appropriate composition for allergen immunotherapy (AIT) for each patient. For this purpose, the recognition of the sensitisation profile (identification of primary sensitisation markers with respect to detection of specific IgE against cross-reactive allergen molecules) has been suggested as a tool to define better the relevant or irrelevant allergens in each patient [5].
The efficacy of AIT in monosensitised patients has been proven in both children and adults and with both routes of administration, subcutaneous (SCIT) and sublingual (SLIT) [9][10][11]. There is no agreement on the therapeutic approach to polysensitised subjects. Thus, while monoallergen AIT is preferred in European countries, the use of more than eight allergens in the same vaccine is common in the US [12]. In this regard, the European Medicine Agency (EMA) recommends that allergists restrict the mixture of non-related allergens to a minimum and advises not to mix seasonal and perennial allergens, or allergens with proteolytic activity without justification [13]. Taking these concepts together, define precisely the diagnosis seems important before prescribing AIT [14]. The present study was designed to explore the medical opinion of a panel of 62 Spanish allergists in order to achieve consensus regarding diagnosis and treatment with AIT in polysensitised patients, aiming to improve allergic treatment.

Methods
To investigate the opinion of a sample of 62 allergists from Spain ( Figure 1), a modified Delphi method [15] was used. First, a steering committee of five allergists (authors of this manuscript) reviewed the medical literature on the topic and discussed the main items to be included in a structured questionnaire. Fifty-seven more allergists with a solid clinical experience in AIT in adults and/or children and a teaching clinical practice of AIT in fellows in training and in continuing medical education Spanish programmes were selected by the steering committee. This selection process began with a specific proposal from each member of the steering committee, who named 30 allergists throughout the country known for their interest in AIT to work on the project. All names repeated at least twice were included as "panelists by agreement". The steering committee together with this team of allergists defined the final questionnaire.

Questionnaire and Delphi methodology
The final version of the questionnaire included 88 items distributed in four topics: 1) general approach to polysensitised subjects (n = 24) ( Table 1); 2) sensitisation profile regarding mite, animal dander and moulds (n = 26) ( Table 2); 3) grass and olive pollen co-sensitisation (n = 18) ( Table 3), and 4) other pollen polysensitisation profile (n = 20) (Table 4) representing the main geographical patterns of sensitisation in Spain. For all topics including allergens, items were grouped according to epidemiology, clinical relevance and therapeutic strategy (Tables 2A, 3A  and 4A). An independent team, consisting of 3 nonmedical people worked as facilitators, requesting the opinion of each allergist, using the online survey (88 items), on an individual and anonymous basis. Answers were collected from May 2013 to June 2013. After analysing the results of the first round, one of the facilitators provided an anonymous summary of the forecasts as well as the reasons allergists provided for their judgements. Thus, allergists were encouraged to revise their earlier answers in light of the replies of other members of the panel and a second round was performed dealing with those questions lacking. A nine-point, single, ordinal, Likert-type scale was used to assess the opinion on each item. Following Delphi categorization, responses were classified in three regions: (1-3) = "disagree"; (4-6) = "neither agree nor disagree"; (7-9) = "agreement". The survey also offered the possibility of adding individual explanatory observations to every answer.  Once the second round was finished, the results were analysed. The median position of scores and the "level of agreement or disagreement" [16] achieved was measured according to the following criteria: an item is considered to have consensus when no more than a third of the scores are found outside of the region of three points (1-3),(4-6), (7)(8)(9) where the median is located. In this case, the value of the median score determines the group consensus reached: "agreement" majority with medians ≥7; "disagreement" majority with medians ≤3; "no consensus" items with medians in the region 4-6 and when the scores of a third or more of the participants are in the region 1-3, and another third or more in the region 7-9. Also considered for reassessment were the items where a high dispersion of opinions (interquartile-range ≥ 4points) (Figure 2) was observed.

Results
Detailed results of each item (mean, median, percentage of distribution of respondents located outside the region of the median, interquartile range and consensus result) are depicted in Tables 1B, 2B, 3B and 4B. Overall, consensus was reached in 73/88 items (83%), 55 of them (63%) were in terms of agreement and the remaining 18 items (20%) in terms of disagreement with the assertion presented.

General approach to polysensitized subjects
Consensus was reached on 22/24 items, 16 in terms of agreement and six items in terms of disagreement. Two items were left without consensus. Specifically, when considering the size of the wheal obtained in SPT, some doctors supported its usefulness to define the clinical relevance of an allergen while others did not. Consequently, no consensus was achieved for this specific statement (item number 2, Table 1B). Even though participants gave more value to serum specific IgE determinations than to SPT for identifying clinically relevant allergens, they seemed to rely more on molecular diagnosis but the lack of an exact cut-off level of serum specific IgE against molecular compounds that could assist this purpose is still not established. As regards to AIT, the prescription should be based on scientific evidence and only extracts with proven efficacy should be used. Vaccines containing up to two to three allergen sources could be used if more than one allergen appears to be relevant but no consensus was found when deciding if a mixture has or not a nonspecific positive therapeutic effect.

Sensitisation profile involving mite, animal dander and moulds
Consensus was reached on 21/26 items, 18 with overall agreement and three in terms of disagreement. Consensus was not reached on the remaining five items (Table 2B). Thus, even though allergists think that serum specific IgE against groups Der 1 and Der 2 could be helpful if these allergens are quantified in the allergen extract, no consensus was achieved regarding the value of these molecular compounds to better identify the sensitisation profile of patients allergic to mites (item number 31). Minor mites such as Lepidoglyphus and Blomia represent a diagnostic problem because of the lack of clinical trials using AIT for them and consensus was not reached on this point. Nevertheless, in circumstances where sensitisation to Dermatophagoides and minor mites coexists and clinical relevance for both types of mites is suspected, a vaccine containing both species of mites could be suitable.
Participants also gave their opinion on mould and animal dander allergy in this section. As far as mould allergy is concerned, Quantification of Alt a 1 in the vaccine is highly recommended, even after not reaching consensus with respect to the need for quantifying serum specific IgE against this main allergen of Alternaria in sensitised patients. In case of double sensitisation to mites and Alternaria, no consensus was reached regarding the value of the evaluation of symptoms after exposure to both. Finally, in relation to animal dander sensitisation, AIT with cat and dog epithelia has demonstrated its efficacy and both may be used but no consensus was achieved in case of animals different from cat and dog.

Grass and olive pollen co-sensitisation
Consensus was achieved in 15/18 items, eight of them in terms of agreement and seven in terms of disagreement with the assertion mentioned (Table 3B). Results obtained in this section of the questionnaire marked the limitations of SPT and serum specific IgE against grass and olive pollens as complete allergenic sources for clearly establishing the allergenic profile of these patients. However, performing specific IgE against specific allergens such as Phl p 1, Phl p 5 and Ole e 1, considered as markers of primary Table 2 Items included in the questionnaire about "Allergy to mites, animal dander and moulds" (2A) and results (2B) (Continued) 49-In the same situation as previously described, the use of a vaccine containing one allergen could be recommended followed by the consideration for a second AIT sensitisation, helps to discriminate between a genuine sensitisation pattern and cross-reactivity without clinical relevance. Selection of the vaccine should be made depending on such results, and only when double genuine sensitisation assessed by molecular diagnosis was confirmed should vaccination with both allergens be performed. The use of personalised formulas with different percentages of olive and grass extracts in the same vaccine, based on size of the wheal or the value of specific IgE, is still a matter of debate and no consensus was reached.
In fact, some panellists are still in favour of varying the proportion of allergens for individual patients.  (Table 4B). The lack of a clearly established seasonal pattern for some pollen hinders the diagnosis. And, in this case, molecular diagnosis is of little help because of the low availability of molecular compounds of pollens apart from grass and olive. Interpretation of polcalcin sensitisation is not always possible and could act as a confounding factor. For that reason, a deep knowledge of all allergenic sources from each zone (Figure 3), including pollen calendars and specific taxons is essential. No consensus was achieved regarding the utility of the sIgE to panallergens (profilin or polcalcin) in complex pollen areas because of difficulties in recognising the relevant and primary sensitiser. The same occurred when profilin or polcalcin were assessed as relevant grass sensitisation markers that would allow excluding other allergens in the AIT. In the case of mixing several allergens, a consensus was reached regarding the use of those pollens with demonstrated clinical relevance, notable exposure, no more than 3 allergens in the same vaccine, and always with quality extracts. Defining the composition of the mixtures is an ongoing problem without consensus. An accepted alternative is the possibility to administer ≥2 vaccines in some cases (although in practice it is assumed that the administration of 2 vaccines is not always acceptable due to the economic cost).

General approach to polysensitised subjects
Polysensitisation is a serious problem in countries like Spain, where different allergenic sources may influence clinical symptoms in allergic patients. Allergists from Spain are used to dealing with this problem but, on some occasions, lack of evidence could make them hesitate about the best treatment option in each case. Other Mediterranean countries such as Italy or Greece share the same problem [17,18]. This lack of scientific evidence led us to promote the present Consensus Report aiming at helping allergists with diagnosis and treatment in complex areas from an allergological point of view.
As authors, we are aware of the limitations of expert opinion statements and for this reason the well-known and highly contrasted Delphi Method was used to gather and analyse the proposed items reflecting common problems in clinical practice. As a whole, 83% of the  items agreed, but consensus was not similarly reached if we consider each group of items separately. It seems that decisions are easier when considering general situations (91.6% agreement) than when deciding specific solutions for specific allergens. It seems that experts have the skills to decide how to act but they fail to agree when diagnostic tools are limited or are not available. According to this, the lowest level of agreement, 75%, was reached in the "other pollen sensitisation profile", reflecting difficulties in recognising relevant allergens in polysensitised patients in those areas, such as Mediterranean Spanish coast, with a high number of allergens and high allergen pressure [19]. According to what has been published, experts agree with respect to the limited value of SPT in firmly recognising relevant allergens [20]. SPT is commonly used in daily clinical practice because of its simplicity and reliability to quickly identify IgE sensitisation to inhalant allergens. The relationship between SPT wheal sizes and the clinical relevance of positive results is debatable [21,22] and, consequently, it is understandable that consensus was not reached on this point. In fact, many factors can influence SPT results. A seasonal variation in wheal sizes has been described for several allergens [23]. The site where SPTs are performed may influence the results. The ulnar side of the arm is more reactive than the radial side, and the antecubital fossa is more reactive than the wrist [24]. Finally, the SPT technique may influence the size of the wheal reaction [22]. Experts consider serum specific IgE more reliable in recognising clinically relevant allergens, as it has been considered by others [25]. Recent advances in molecular diagnosis are supposed to help allergists in the recognition of genuine sensitisations, and experts agree on the value of this diagnostic tool, but the lack of a cut-off level is a limitation of the test. The fact that sensitisations are not always accompanied by clinical symptoms is well-recognised [26]. For that reason, experts agree that polysensitisation is not the same as polyallergy, meaning that polyallergic patients are only those with clinically relevant symptoms related to allergen exposure. To identify relevant allergens, symptom scores are useful in a similar way than when performing clinical trials. The recognition of a cause-effect relationship should be done according to the allergen exposure. Aerobiological information is essential for this purpose. Even though allergen challenge tests can be useful in clinical trials and even in proving the clinical relevance of purified allergens (profilin, polcalcin and so on) [27], they are time-consuming, and therefore the implementation of these tests in daily clinical practice before prescribing AIT is not necessary. The use of allergenic extracts with proven efficacy is strongly supported and clinical studies should follow the guidelines of regulatory agencies [13]. Taking into account differences among products from different allergen manufacturers, each one must perform their own clinical trials for every single product, since results from one study with one allergen cannot be extrapolated to a different allergen, or even the same allergen produced by another manufacturer [28][29][30]. Finally, when considering mixing allergens and following EMA suggestions, non-related allergens with possible enzymatic activity should not be mixed; even for related allergens, no more than three should be used in the vaccine.

Sensitisation profile involving mite, animal dander and moulds
According to what has been accepted in the general approach section, experts agree with the idea of knowing the species of mites predominant in one area. It is wellknown that mites are not uniformly distributed all over the world and, even for the same Dermatophagoides genus, distributions of pteronyssinus and farinae species vary from place to place [7,31]. However, the high degree of cross-reactivity proven between them [7,32] seems not to be a problem when selecting the composition of AIT. That is not the case for minor mites. The lack of crossreactivity among them and Dermatophagoides [33][34][35] represents a problem difficult to resolve. Thus, while experts could recommend the prescription of a vaccine containing minor mites and Dermatophagoides in equal proportions, they recognise the need for more studies showing clinical efficacy of AIT with minor mites alone before strongly supporting their use as a therapeutic option. Consequently, no agreement was achieved on this specific item.
Molecular diagnosis, as far as mite and mould allergy is concerned, seems to be of little or no value, and therefore experts show no agreement regarding the need for measuring sIgE before prescribing AIT with them. In fact, no consensus was achieved when specifically asked about the need of measuring sIgE against Der p 1, Der p 2 or Alt a 1. However, quantification of major molecular compounds of both mites and moulds in the vaccine is evaluated positively by participants. One possible explanation for these contradictory results may be related to the recent advances in the recognition of the pattern of sensitisation against Der p 1 and Der p 2 and the lack of correlation between them [7]. More studies are needed to balance the importance of both molecules in each patient, but in the meantime quantification of major allergen compounds in the vaccine could be useful to explain adverse effects or even efficacy.
Biological standardisation of moulds is difficult [36] but participants consider Alternaria and Aspergillus spp the only relevant allergens from a respiratory point of view. Moreover, only Alternaria could be considered as a potential allergen for AIT. For that reason mixture of moulds should never be an option accepted by experts. The main problem involving Alternaria is co-sensitisation with mites. There is no consensus regarding the utility of symptom and medication scores or the recognition of sources of exposure when dealing with patients sensitised to mites and Alternaria. If needed, more than one vaccine could be used and this recommendation applies not only for mites and Alternaria but also for animal dander allergy. Vaccines with cat or dog allergens may be used if these allergens are clinically relevant and avoidance is not an option. AIT with horse extract is not supported by all experts, and no consensus was reached despite proven efficacy. The main limitation could be that there was only one published study when the questionnaire was done, and it is not a double-blind placebo-controlled study [37]. Recently, Nanda and Wasam have published their experience with a horse extract at different doses in eight patients, with positive results [38].

Grass and olive pollen cosensitisation
As far as grass and olive pollen is concerned, participants agree that measurement of specific IgE to Phl p1, Phl p 5 and Ole e 1, molecular compounds considered as phenotypic markers of genuine grass and olive pollen sensitisation, respectively are useful in deciding the composition of AIT. In addition, some allergens considered minor because of their lower prevalence in terms of sensitisation, such as Ole e 7 and Ole e 9, could be of importance in regions with high allergenic pressure to olive pollen. Moreover, the presence of high levels of specific IgE against them could explain some serious adverse reactions during AIT with olive pollen extracts [38]. AIT is claimed to be a first line therapy option in patients sensitised to grass and olive, if both are clinically relevant. The low cross reactivity between pollens of different species, including oleaceae and grass [39,40], but particularly the presence of genuine sensitizers [41,42], justifies the consensus on the suitability of individual dosages of each extract. Finally, no consensus was achieved when experts were asked about the use of personalised immunotherapy. The term "personalised" was used to express formulations with different proportions of each allergen as opposed to the idea of fixed combinations in equal proportions (50%-50%), but the statement recorded was not clearly defined and could have been misunderstood.

Other pollen (weeds and trees) polysensitisation profiles
Pollens from trees and weeds such as Betula spp, Platanus spp, Cupressus spp, Artemisia spp, Chenopodium spp, Parietaria spp, Plantago lanceolata and Salsola kali may induce IgE sensitisation in atopic subjects in exposed individuals in Spain and other Mediterranean countries [43]. Exposure to these pollens varies from place to place. This is why allergists from different regions may have different opinions about how to deal with polysensitised subjects, depending on the different taxons present in a particular area. In fact, the level of consensus was lower in this section compared with the others. It is noteworthy that experts do not have a recommendation regarding the usefulness of molecular diagnosis because of difficulties in distinguishing primary sensitisation and cross-reactivities. Even symptom and medication scores are of little value. Polcalcin has been considered a confusion factor, but the actual meaning of polcalcin sensitisation is far from known. There are no published articles investigating the prognostic value of such sensitisation or the possible interference on efficacy or safety of immunotherapy. As regards profilin, some authors have demonstrated a relationship between profilin sensitisation and symptoms [19] while others consider profilin sensitisation a confounding factor that could interfere with immunotherapy efficacy, because of its role as a confounding factor for the diagnosis [6]. More studies are needed to clarify these questions. In the meanwhile, patients allergic to more than one non-related pollen could be treated with more than one high-quality vaccine or a mixture of pollens with clinically relevant exposure.

Conclusion
In conclusion, choosing the best allergen for AIT in polysensitized patients is a difficult task. Consensus like the one presented here could be of help to allergists before making decisions on AIT composition. From the results of this survey we can conclude that: a. SPTs are not enough to accurately diagnose allergy in polysensitised patients. b. An approach to molecular diagnosis seems to be useful in pollen-allergic patients if all relevant molecules were available. c. AIT prescription should be based on scientific evidence, and only be indicated if relevant allergens are identified. d. Data from clinical records and knowledge of allergen exposure in each area are essential to define AIT composition. e. No more than three allergenic sources should be mixed in the same vaccine. f. Each vaccine should have its own safety and efficacy studies.