Box 3: Recommendations on eligible food allergens and types of clinical outcomes that can be achieved by OIT	Ethical imperative, data or other considerations in support of the recommendation Level of evidence (applicable when recommendations are based on outcome data from clinical studies)
There is no convincing evidence of a clinically significant difference between food allergens in terms of safety and efficacy outcomes in OIT for the treatment of IgE-mediated food allergy. Therefore, all recommendations in these CPGs are generally applicable to all food allergens, unless there is specific evidence to demonstrate otherwise	This recommendation is based on the principle of equity of eligibility It is supported by large amount of consistent clinical evidence, considering the absence of demonstrated lack of efficacy or of a consistent safety issue for any specific food despite a large number of clinical studies for a variety of foods [22–24, 29–38]. Level of evidence: HIGH It is also supported by the lack of biological plausibility that the mechanism of OIT would differ from one allergen to another.
OIT is recommended as a treatment to achieve desensitization. A majority of patients will achieve a level desensitization to a daily dose of the allergen that will be sufficient to provide protection against trace exposure, while a sizable proportion of patients will be able to tolerate a full serving	This recommendation is based on the principle of beneficence. Whether or not the outcomes (i.e. desensitization, continued consumption or sustained unresponsiveness) are worth pursuing remains patients’ prerogative, in line with the principle of patient autonomy It is supported by a large amount of consistent clinical evidence, which includes three meta-analyses [22–24] (together covering 31 published OIT controlled clinical trials), three additional RCTs [31, 34, 39] and two non-randomized controlled clinical trials (CCTs) [36, 38], five large case series in clinical practice (N > 150) [29, 30, 32, 33, 40], and two smaller case series [35, 37]. This body of evidence includes 10 RCTs rated as being at low risk of bias [34, 41–49] and is coherent with data from consultations. Level of evidence: HIGH
OIT can be recommended for long term management since a sizable proportion of patients will continue to regularly consume a sufficient amount of the food to maintain desensitization after reaching maintenance, without reverting to complete avoidance	This recommendation is based on the principle of beneficence. Whether or not the outcomes (i.e. desensitization, continued consumption or sustained unresponsiveness) are worth pursuing remains patients’ prerogative, in line with the principle of patient autonomy It is supported by a moderate amount of consistent clinical evidence, from three long-term, 2-arm follow-up studies of RCTs (high risk of bias due to open-label; mean follow-up 2.5 to 5 years; completeness of follow-up: 77 to 90%) [50–52], three large case series in clinical practice (N ≥ 145) (moderate risk of bias using the IHE tool due to single-center, retrospective design; median follow-up: 1.5 to 6.5 years; completeness of follow-up: 83 to 99%) [33, 53, 54], and four smaller single-arm, follow-up studies (N = 43 to 100) at low to high risk of bias (completeness of follow-up: 82 to 100%) [35–37, 55]. This is coherent with data from consultations. Level of evidence: MODERATE
OIT may be recommended to achieve sustained unresponsiveness, but data is limited and variable	This recommendation is based on the principle of beneficence. Whether or not the outcomes (i.e. desensitization, continued consumption or sustained unresponsiveness) are worth pursuing remains patients’ prerogative, in line with the principle of patient autonomy It is supported by a small amount of evidence with some incoherence in findings, including one meta-analysis (covering 4 RCTs) [22], three additional RCTs [34, 46, 56], one CCT [57], one prospective trial with retrospectively- matched controls [58] and one case series in clinical practice [40]. This is coherent with data from consultation concerning pre-school children. Level of evidence: LOW