National clinical practice guidelines for allergen immunotherapy: An international assessment applying AGREE‐II

Since 1988, numerous allergen immunotherapy guidelines (AIT‐GLs) have been developed by national and international organizations to guide physicians in AIT. Even so, AIT is still severely underused.


| INTRODUCTION
Allergen immunotherapy (AIT), as a causal treatment for IgEmediated allergic diseases, such as rhinoconjunctivitis, allergic asthma, venom allergy, and in some regions atopic dermatitis, has a history of more than a century. In the course of its developing process, details on AIT practice have evolved differently in various parts of the globe, with a striking difference in the practice of AIT, especially subcutaneous AIT (SCIT), between the American and European continent, 1,2 both with their supporting evidence. With the growing number of physicians administering AIT, subcutaneous and lately also sublingually (SLIT), several national and regional allergy societies have tried to give some guidance on the practice of AIT since the 1980s. 3 Depending on the local regulatory situation and extract availability, in some regions the American school and in others the European school of the practice of AIT have been found to be more suitable. However, in general, AIT is still severely underused. In 2009, the first World Allergy Organization (WAO) consensus was published on SLIT, 4   Applicability, and (6) Editorial Independence. At the end, the AGREE II tool asks the reviewer to rate the overall quality of the guideline and to state whether the guideline would be recommendable for local use. The first guideline in the allergy world adhering to the AGREE II principals and following the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) system for evidence evaluation was the Allergic Rhinitis and its impact on Asthma (ARIA) 2010 update, which includes a complete evaluation at that time-point of the evidence on AIT for allergic rhinitis and for asthma. 7 In this article, we aim to evaluate guidelines specifically focused on allergen immunotherapy (AIT-GLs) with AGREE-II, to distinguish between AIT-GLs with a higher and a lower AGREE II quality score.
This leads us to suggest which AIT-GLs might be best to adhere to, or might be most suitable for local adaptation in countries where no high-quality, evidence-based AIT guidelines exist. Finally, our study also shows in which domains existing AIT guidelines still flaw and could be improved in the future.
Guidelines on allergic rhinitis or asthma, which mention AIT, but apart from it analyze many other treatment options, were not included in this analysis, as their scope goes beyond the sole evaluation of AIT. As such, the ARIA 7,8 and the Global Initiative on Asthma (GINA) 9 guidelines are not included in the here presented evaluation. As the team consisted of allergists from various parts of the world, language was not a restriction. When needed, GLs were translated into English. The collection of AIT-GLs was cross-checked and completed against those identified in an active search by team members contacting local and regional AIT experts to also identify those papers published in gray literature. In online-discussion rounds, the investigators selected the GL papers from among the found articles.

| ME TH ODS
Each selected publication was scored on 23 items, divided into the 6 domains of AGREE II, related to GL content, presentation, involved authors, declaration of conflict of interest, among others. In a first round, each AIT-GL was AGREE-II-evaluated by 2 independent reviewers, scoring each item on an ordinal Likert-type scale from 1 to 7, and sustaining the given score with a detailed explanation, for example, where the item could be found in the GL or what was missing; discrepancies were resolved in a second round in which mutual evaluations and their reasoning were sent back to both reviewers. After that, unresolved issues were decided in a third round by team discussion or methodologists' consulting. After

| RESULTS
We found 31 publications of AIT-GLs, see Table 1. Fifteen of them were guidelines published from 2010 onward.
The papers covered the range from local consensus reports to position papers issued by regional and global allergy organizations, such as the European Academy of Allergy and Clinical Immunology (EAACI), the joint council of the American Academy of Allergy, Asthma and Immunology (AAAAI) and the American College of Allergy, Asthma and Immunology (ACAAI), and the World Allergy Organization (WAO). Some smaller consensus documents, written in the local language, were translated into English before evaluation.
The total scores of the pre-2010 GLs ranged from 1.6 to 5.2 and of the 2010-onward GLs from 2.1 to 6 (ordinal scoring from 1-7), see Some specific comments of the reviewers on the guidelines can be found in Table S1 (see the online supporting information tab for this article).  Concerning the low scoring domains 2 and 5, these are closely linked to dissemination and implementation of a GL. Domain 5 (applicability) seems the most interesting domain, as here the pre-2010 AIT-GLs scored better than the newer ones. A possible explanation for this finding might be that older GLs were more focused on informing clinical assistants on how to treat allergic diseases, while newer GLs might be focusing more on improving knowledge, reducing the focus on dissemination.
It is generally accepted that AIT is still underused, 41  Involve them early-on in the process. A core group of only a few that prepares the initial lay-out and drafts might be a solution, to avoid slowing the process when the GDG becomes too big. But an initial, mid-, and final face-to-face meeting with the whole group is recommendable. (eg, GRADE approach, SIGN) For transculturized GLs: see ADAPTE approach, all evidence comes from the selected "mother" GL(s).

Domain 4. Clarity of presentation
Bullet-point presentation at the beginning/end of the document on the recommendations and suggestions, so they can easily be found by readers without going through the whole document.

Domain 5. Applicability
Discuss with the whole GDG barriers and obstacles for GL implementation, and how to solve them. Facilitate GL implementation with user-friendly (downloadable) application sheets, in which recommended steps are plotted. (eg, US AIT practice parameters' sheets). Domain 6. Editorial independence Apply a method (and describe in the GL) to reduce bias from GDG members with conflicts of interest and from the financing body.

LARENAS-LINNEMANN ET AL.
| 669 the applicability of the GL. Paying special attention to improving these domains in future AIT-GLs would be highly desirable, see Table 2. In the previously mentioned AGREE II evaluation of allergic rhinitis GLs, the same domains 2 and 5 were found to rank low. 40 After analyzing all results, it amazed the investigators to see how little difference there is between older and the newer AIT-GLs in the global mean guideline AGREE-II-quality-scoring. Taking a closer look, we realized that this could partly be due to the fact that within each group of newer vs older AIT-GLs, some GLs are of higher and others are of much lower quality (see Figure 1) resembling more consensus documents and thus reducing the mean scoring of the whole guidelines' group. Therefore, in a post hoc analysis, we selected from the 2010-onward GLs only those with a mean score above 40%. Recalculating statistics based on this group of 9 AIT-GLs all domains showed statistically significant improvement vs the older GLs.
Under low-resource conditions (both economically and intellectually, because of time restraint of local experts), it can be very difficult to develop a GL from zero, based on all published literature. In such situations, the ADAPTE tool can be very useful, allowing high-quality existing GLs to be adapted to local reality. 42 In ADAPTE, the literature search is thus for GLs published specifically in the field of interest, which subsequently are analyzed with AGREE II to select the ones with best-quality, most suitable for adaptation. As such, a locally adjusted high-quality GL can be developed without too much investment in the collection and evaluation of evidence. This might leave some resources for inviting more stakeholders and thus enhancing applicability.
The last domain, 6, on editorial independence, though much improved in the newer GLs, still only scores 4/7 points, as often conflict of interests and financing bodies are declared, without describing a method to avoid that these interfere with the GL content.
In this discussion, a word must be said about the tool we used, AGREE II, as well. Although it analyzes guidelines through a structured item list, with clear explanations in a manual on how to apply them, 43 finally it is based on the reviewer's criterion of scoring, a subjective measure. We came across this issue, as most GLs were revised in groups of 2 and on some items discrepancies arose, that in 3 cases continued even after the second round of discussion and re-revision of both reviewers. In these cases, we asked orientation of the methodologists. At that point, it became clear that some items could be sensitive to personal interpretation and thus not be uniformly scored by all reviewers. Here, the group decided to interpret as unanimously as possible and analyze all GLs with the same criterion.
Resuming, Table 2 shows per domain suggestions of the reviewers for developers of future AIT-GLs that might enhance the overall quality and thus implementation. These are only suggestions, as till now no AIT guideline has been subject to validation in the real world, nor can we assure that the best guidelines as described by AGREE II are more able to be used more often and at their best in the real life. But till now, it is the best tool we have.
During the past months, several systematic reviews have been published by a Task Force of the European Academy of Allergy and Clinical Immunology (EAACI) on AIT for allergic rhinitis, 44 allergic asthma, 45 and for allergy prevention, 46 to name some. Authors anticipate these are the preparatory documents for a set of guidelines from the EAACI on the different aspects of AIT. This kind of preparation seems adequate for the development of a high-quality guideline.
The final EAACI guideline documents shall have to be awaited to evaluate them with AGREE II and to confirm or reject this suspicion.

| CONCLUSION S
In AIT-GLs, there is still a lot of room for improvement, especially in 2 domains, crucial in dissemination (stakeholder involvement and applicability). For some, the "Scientific rigor" domain flawed. In situations with limited resources, local adjustment ("transculturizing"), a high-quality AIT-GL might be appropriate and preferable over the development of a low-quality one. The AGREE II instrument could help to pick quality candidate AIT-GLs for such a procedure.

ACKNOWLEDG MENTS
We would like to thank M. Brouwers

PATIENTS' DATA PROTECTI ON
The authors declare that no patient data appear in this article.