World Allergy Organization (WAO) Diagnosis and Rationale for Action against Cow’s Milk Allergy (DRACMA) Guidelines update – IV – A quality appraisal with the AGREE II instrument

Background Since the publication of The World Allergy Organization (WAO) Diagnosis and Rationale for Action against Cow's Milk Allergy (DRACMA) Guidelines in 2010, a number of other guidelines, expert opinions, and position papers relating to the management of cow's milk allergy (CMA) have been published. We aimed to systematically review the quality of the guidelines on CMA diagnosis and management in children and/or adults published between 2010 and 2020. Methods The MEDLINE, EMBASE, ISI Web of Science, World Health Organization Global Index Medicus, and Turning Research into Practice databases as well as website guideline repositories were searched from January 2010 until May 2020. Any clinical practice recommendations and/or guidelines focusing on the diagnosis and management of CMA in children and/or adults developed or endorsed by professional scientific societies or organizations were included. The guidelines were evaluated using the Appraisal of Guidelines for Research and Evaluation (AGREE II) tool, a 23-item tool organized within 6 domains and 2 global rating items. Results We included 12 guidelines; 8 were developed by national and 4 by international organizations. The quality scores for each domain varied: of all domains, the clarity of presentation domain had the highest median score (92%; Q1-Q3 81–100%), whereas rigor of development had the lowest median score (30%; Q1-Q3 15–67%). The median scores (Q1-Q3) for individual domains were as follows: scope and purpose 82% (70–99%), stakeholder involvement 63% (21–79%), rigor of development 30% (15–67%), clarity of presentation 92% (81–100%), applicability 68% (57–75%), and editorial independence 75% (69–100%). The median overall score was 70% (58–89%). Only 1 guideline (from the National Institute for Health and Care Excellence [NICE]) achieved top ratings (100%) in five domains and the overall score. Three guidelines (from the NICE, the British Society for Allergy & Clinical Immunology [BSACI] and WAO) achieved the highest ratings (100%) in at least 3 domains and the overall score. Conclusion The majority of identified guidelines were of good or very good quality. However, the weakest point was the rigor of development domain, mostly due to unclear description of strengths and limitations of the body of evidence and the procedure for updating the guidelines.


INTRODUCTION
Since the publication of the 2010 DRACMA guidelines, a number of other guidelines, expert opinions, and position papers for the management of CMA have been published. However, their quality has not been formally appraised. In 2016 a systematic review assessed the quality of guidelines on cow's milk allergy (CMA) published from 2010 through November 2015 using the Appraisal of Guidelines for Research and Evaluation (AGREE II) tool. 1 Fifteen guidelines were included. Only the guidelines developed by recognized professional/ scientific organizations such as the British Society for Allergy and Clinical Immunology (BSACI) and the European Academy of Allergy and Clinical Immunology (EAACI) were of the highest quality. In addition, the 2010 World Allergy Organization (WAO) Diagnosis and Rationale for Action against Cow's Milk Allergy (DRACMA) Guidelines, 2 the only Grading of Recommendations Assessment, Development and Evaluation (GRADE) guidelines for CMA, were considered to be of high quality.
In 2018, the DRACMA panel committee reassembled in order to update the DRACMA guidelines. The aim of this study was to systematically review the quality of the guidelines on CMA diagnosis and management in children and/or adults published from 2010 onwards, and to summarize specific recommendations.

METHODS
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement 3 was followed during each stage of this review. The protocol was pre-defined and submitted to PROSPERO; however, it was not accepted for registration, as it was assessed as being outside of the scope of included protocols due to the lack of at least 1 outcome of direct patient or clinical relevance. The AGREE II User's Manual 4 was followed during the quality assessment of the included guidelines.

Search for guidelines
The MEDLINE (through PubMed), EMBASE, ISI Web of Science (Thomson Web of Knowledge), World Health Organization Global Index Medicus (GIM) (https://www.globalindexmedicus.net/), and Turning Research into Practice (TRIP) (https://www. tripdatabase.com/) databases were searched from January 2010 up until May 2020, and then the search was updated in April 2021. The rationale for choosing 2010 as the start date was that this is the issue date of the DRACMA guidelines. However, we recognized that an update of any guidelines/recommendations is generally required from 2 to 5 years after the issue date, 5 and therefore, some of the earlier guidelines could be outdated. MEDLINE and EMBASE were searched following a pre-specified search-strategy (see Supplemental Appendix 1). The websites of guideline repositories were also searched including: National Institute for Clinical Excellence (NICE, https://www.nice.org.uk/), The Guideline International Network (GIN, https:// guidelines.ebmportal.com/), Scottish Intercollegiate Guidelines Network (SIGN) (https://www.sign.ac. uk), and Agency for Healthcare Research and Quality (AHRQ, https://www.ahrq.gov/).
References of all included guidelines and guideline publisher's websites were also searched for any supporting documents (ie, technical reports, methodological manuals).
The search was carried out independently by four reviewers (AS, AH, LD, and MR). No filters or restrictions other than English language were imposed.

Inclusion criteria & exclusion criteria
Any clinical practice recommendations and/or guidelines focusing on the diagnosis and management of CMA in children and/or adults developed or endorsed by recognized scientific societies or organizations were included. In case of an updated version of a guideline, only the most recent document was considered for inclusion.
Guidelines were included, regardless of CMA mechanism (ie, IgE-mediated, non-IgE-mediated, mixed); however, if feasible, they were assessed separately. Guidelines focusing on food allergy or a single disease (eg, food protein-induced enterocolitis syndrome [FPIES]) were not considered for inclusion in this review, unless there was a section focusing explicitly on CMA or cow's milk proteins.
Consensus-based and expert opinion clinical practice guidelines, if not endorsed by recognized scientific or professional organizations, were excluded based on their limited generalizability as well as our limited capability to evaluate the level of expertise, that these publications represent, and the audience addressed. Guidelines focused on a single specific management option (eg, immunotherapy) or prevention were excluded. Guidelines which were ongoing or unpublished were also excluded.

Data selection
As recommended, 4 reviewers (AS, AH, MR, and LD) screened the titles and abstracts of articles identified in the search to identify potentially eligible guidelines. The full texts of all potentially relevant articles were retrieved and critically assessed against the pre-defined inclusion criteria independently by each of the reviewers. Any discrepancies were first discussed by the 4 reviewers (AS, AH, MR, and HS).
Initially, members of the DRACMA panel not involved in the earlier process (AF, ANW, RS, JS, YV, CV, LD) provided their comments on the included and questionable documents and, if feasible, any unidentified papers, via an online survey using Google Forms. The list of excluded papers was also reviewed. Guidelines were included if at least 90% agreement was reached; in case of agreement 50%, a paper document was excluded. All of the comments were discussed. Then, all questionable documents (between 50% and 90% agreement) were put to a second vote by the members of DRACMA panel to determine eligibility for inclusion. Any discrepancies, as well as all other disagreements between the reviewers, were resolved through discussion until a consensus was reached.

Data extraction
Three reviewers (AS, MR, and LD) independently extracted data from all included guidelines. The reviewers extracted the following information: title, year of publication, organization (country), level of guideline development (ie, local, regional, national, or international), financial support, and conflicts of interest (number of people who obtained financial support and/or had conflicts of interest/number of all authors). Data extraction was performed using data-extraction forms developed by the reviewers. Any discrepancies were discussed until a consensus was reached.
Specific recommendations were summarized in a comparative table, focusing on possible gaps and common messages. A "List of specific recommendations to be assessed" had been prespecified in the protocol. If feasible, recommendations were extracted separately for IgEmediated, non-IgE-mediated, and mixed CMA, as well for each age group (ie, children, adults).

Assessment of guidelines using AGREE II
All appraisals were made using My AGREE PLUS interactive guideline appraisal platform (www. agreetrust.org) by 3 reviewers (AS, AH, and MR). Two authors had previous experience with the AGREE II instrument, 6 and one reviewer (AS) underwent the online AGREE II tutorial before the review (available at: http://www.agreetrust.org/).
The AGREE II is a 23-item tool organized within 6 domains: (1) scope and purpose; (2) stakeholder involvement; (3) rigor of development; (4) clarity of presentation; (5) applicability, and (6) editorial independence. The AGREE II instrument also contains 2 global rating items: (1) overall guideline assessment (that requires the appraiser to make an overall judgement of the practice guideline while considering how they rated the 23 key items) and (2) a question on whether the appraiser would recommend a guideline for use in practice (assessed on a 3-point scale [ie, yes, yes with modification, and no]). All of the AGREE II items and the overall guideline assessment item are assessed using a 7-point Likert agreement scale ranging from 1 (strongly disagree) to 7 (strongly agree). The reviewers discussed all scores that differed by 2 or more points among themselves, until a consensus was reached.
For each item and domain, the score was summed and calculated as a percentage of the maximum possible score for that item/domain using the formula provided by the AGREE II consortium: 4 [(score obtainedminimum possible score)/(maximum possible scoreminimum possible score)] x 100. The possible standardized scores range from 0% (the minimum) to 100% (the maximum).
The AGREE II does not provide a minimum or maximum range for domain score quality to differentiate high-and low-quality guidelines and recommends that it should be done by the reviewer. In agreement with a previous quality appraisal with the AGREE II of the same clinical question carried out by members of the current review group, 1 a standardized domain score of above 60% for each domain has been chosen as the threshold.

Statistical analysis and data synthesis
Normality of quality scores was assessed using the Shapiro-Wilk test and based on visual assessment of histograms. Due to the lack of a normal distribution of scores, data are presented as the median followed by the quartiles (upper [Q3] and lower [Q1]) and IQR (interquartile range). Agreement between raters (inter-rater reliability) was analyzed using Fleiss' Kappa and intraclass correlation coefficient (ICC) estimates. The ICC calculation was based on a single rating, absolute agreement, two-way random effects model including a 95% confidence interval (CI). Analysis was conducted in R software, version 3.5.1 (http:// cran.r-project.org). by an independent statistician. Although Kendall's W coefficient was pre-specified in the protocol to assess agreement between raters, after consultation with the statistician, it was changed to Fleiss' Kappa that is suitable for analysis of the agreement using ordinal or nominal parameters (either dichotomous or not). 7

RESULTS
For the guideline selection process, see Fig. 1. Excluded guidelines with reasons for exclusion are summarized in Supplementary Table 1.

Characteristics of included guidelines
We included 12 guidelines (for characteristics, see Eight guidelines were focused only on children. [8][9][10][11][12][13][14][15] Two guidelines were not only on the management of CMA in children, but also in adults. 16,17 One set of guidelines, although developed with regard to all ages, was focused especially on young ones; 2 the second was directed mostly at children aged 5 years and younger, 18 however, older children and adults were also discussed. Three guidelines were focused on the diagnosis and management of infants with any CMA. 8,9,14 Among 2 Spanish guidelines, one 15 included recommendations for management of infants only with non-IgE-mediated CMA, and one 16 for infants only with IgE-mediated CMA. Five guidelines provided recommendations with regard to IgE-mediated and non-IgE-mediated CMA separately. 10 recommendation only for IgE-mediated CMA (and non-IgE-mediated CMA recommendations were in a review). One set of guidelines reported recommendations on the diagnosis and management of infants only with FPIES. 13 Half of the included guidelines 9,10,12,13,16,18 were published in the last 5 years.
Quality of included guidelines (the AGREE II quality scores) The median score for the scope and purpose domain was 82% (Q1-Q3: 70-99%) across all guidelines. Three guidelines (British Society for Allergy and Clinical Immunology [BSACI], Spanish on non-IgE-mediated CMA and WAO) 2,15,17 achieved the highest median score (100%), and one set of guidelines (NICE) 18 achieved a median score equal to 98%. Two guidelines with the lowest ratings achieved median scores for this domain below 60%. 9,14 Low scores were mainly due to a lack of proper reporting, including a non-specified overall objective and a poor description of a target population.
AGREE II domain and overall scores were calculated by summing up the individual scores for all items of each domain/all ratings of the overall quality and calculating as a percentage of the maximum possible score for that domain (where 0% was the minimum, and 100% was the maximum), using the formula provided by the AGREE II consortium 4 : [(score obtained -minimum possible score)/(maximum possible score Recommendation for a collection of detailed history of symptoms to establish suspicion of CMA.
WAO 2010 2 Not as official recommendation.

ESPGHAN 2012 11
Recommendation for a collection of detailed history of symptoms and physical examination.

BSACI 2014 17
Recommendation for a collection of detailed history of symptoms (including severity evaluation).

SEICAP 2015 16
Recommendation for a collection of detailed history of symptoms and physical examination.
AAAAI and I-FPIES 2017 13 Recommendation for a collection of clinical history of typical signs and symptoms for both acute and chronic FPIES, and to consider a broad differential for a patient with acute vomiting in a diagnosis of FPIES.
SEGHPN, AEPAP, SEPEAP, and SEICAP 2019 15 Recommendation for a collection of detailed history of symptoms, physical examination, growth assessment, and feeding history.
GPIFN and MAP 2019 10 Recommendation for a specifically allergy-focused clinical history and physical examination.

(continued)
Clinical history and physical examination to establish suspicion of CMA NICE 2019 18 Recommendation for a specifically allergy-focused clinical history and physical examination, including: nutritional status and growth (weight, length/height, and calculation of BMI), any signs of a clinical reaction, or comorbid conditions such as atopic eczema, asthma, and/or allergic rhinitis, or suggesting an alternative diagnosis.

ISPGHAN 2020 12
Recommendation for a collection of detailed history of symptoms and physical examination.
Other guidelines 9,14 Not reported. Elimination-reintroduction EWPGAG 2010 8 Recommendation for use of CMP elimination diet and, in case of resolution of symptoms, confirmation with OFC. If IgE-mediated CMA, supervised challenge in minority of cases. Not recommended in: -exclusively breastfed infants with bloody stools (proctocolitis), -with suspected reaction to CMA and mild symptoms, -with mild AD and negative history for CM reactions. Children with any severe symptoms should be referred to a specialized center.

WAO 2010 2
Suspected IgE-mediated CMA: In settings in which an OFC is not a requirement, in patients with an average pretest probability of IgE-mediated CMA, suggestion for use of OFC with CM as the only test without measuring milk sIgE levels as a triage or add-on test.
Finnish guidelines 2012 14 Recommendation for use of elimination diet with no milk or egg and, in case of resolution of symptoms, referral to a specialist who will supervise an OFC.

Breastfeeding
Treatment of non-IgE CMA (mild to moderate): strict adherence to CM-free diet for the mother/ infant until the child is 9-12 month and for at least 6 months with support of dietitian.
NICE 2019 18 -Exclusively breastfed infants: recommendation for continuation of breastfeeding with maternal elimination diet without CMP. -Mixed-fed infant: revert to exclusive breastfeeding. -Infants asymptomatic on exclusive breastfeeding: recommendation against maternal elimination diet. -Infants with severe non-IgE-mediated allergy and/ or AD: consider seeking specialist advice to also exclude soy protein and egg. -Supplementation: calcium and vitamin D according to local protocols. Treatment of non-IgE CMA (mild to moderate), strict adherence to CM-free diet for the mother/ infant until the child is 9-12 month and for at least 6 months.

ISPGHAN 2020 12
Recommendation for continuation of breastfeeding with maternal CMP elimination diet. -A child fed with CM formula with mild-moderate symptoms: if the oral food challenge is positive, the child elimination diet and re-challenged after 6 months (a shorter period for GORD) and in any case, after 9-12 months of age. -A child fed with CM formula with severe symptoms: the OFC for tolerance acquisition performed not before 6-12 months after the last reaction. Child elimination of CM until 12 months of age, but in those with enterocolitis syndrome, until 2-3 years of age. -A breasted child with moderate-severe symptoms: food challenge after 6-12 months of avoidance. If lack of symptoms after the reintroduction of CM in mother's diet, the introduction of excluded foods one by one in the diet.
WAO 2010 2 Not reported as recommendation. Re-evaluation of all dietary interventions and avoidance strategies with patients and their families on a yearly basis, ideally through an OFC carried out under medical. Convincing symptoms after accidental ingestion equivalent to positive OFC and reschedule of the follow-up procedure accordingly.
Finnish guidelines 2012 14 Not discussed in CMA section but with regard to food allergies in general.
(continued) tolerance -A follow up of a child with food allergy by the basic health service. In case of a serious allergy for an important food (milk, grain), a follow up at the specialist-level health service. -In milk allergy, a trial with small amount milk made at home at the age of 18 months. -If CMA first appeared in the form of a serious allergy symptom, then milk provocation at specialist-level health care. Return of eliminated foods into the diet tried at 6-month intervals during the first 3 years and then at 12-month intervals. -Child 5-year visit (if not earlier): the examination of diet to ascertain whether based on an elimination-provocation trial and assess a need for consultation with a specialist.
BSACI 2014 17 -Reassessment of individuals at 6-12 monthly intervals from 12 months of age to assess for suitability of reintroduction. -The challenge food in CMA: either baked or fresh milk. Baked milk for initial use (less allergenic, reactions less likely severe). -Home reintroduction using a ladder approach in children who have had only mild symptoms (only cutaneous symptoms) on noteworthy exposure (eg, a mouthful of fresh milk) and no reaction to milk in the past 6 months and in IgE-mediated disease, a significant reduction in sIgE/SPT weal diameter.
AAAAI and I-FPIES 2017 13 Evaluation of patients with FPIES at regular intervals according to the patient's age and food allergen to determine whether she or he is still allergic. Recognition the age of development of tolerance varies by type of food trigger and country of origin. Development of tolerance in patients with CM-FPIES at an earlier age than tolerance in cereal grain-or other solid food induced FPIES.

CNSFP 2018 9
A challenge under medical supervision to test the tolerance of baked milk in children from 1 year of age. The appropriateness and timing of its introduction assessed individually. Not reported as recommendation. Infants with proven CMA: a CM-free diet until 9-12 months of age and for at least 6 months before attempting to reintroduce it.
SEGHPN, AEPAP, SEPEAP, and SEICAP 2019 15 Periods of treatment with a CMP-free diet: from 3 to 6 months in mild forms, to up to 12 months in the most severe cases. Unfavorable response to reintroduction of CMP: periodical re-evaluation of tolerance every 6 to 12 months. Mild cases: testing for tolerance at home under medical supervision. (continued) Cow's milk re-challenge to test for acquired tolerance Child with a personal history of atopy, immediate reactions (onset within 2 h from ingestion), FPIES and all severe forms of allergy: a sIgE test and/or a SPT before reintroducing CMP. Based on the results of specific IgE or SPT: tolerance tested in a hospital setting.
GPIFN and MAP 2019 10 Confirmed CMA: CM-free diet until 9-12 months of age and for at least 6 monthswith a support of a dietitian. Then a planned reintroduction or supervised challenge using a ladder approach to determine tolerance acquisition.
No current AD and no history at any time of immediate onset symptoms: no need to test IgE or SPT: reintroduction at home, using a milk ladder. Current AD: check serum sIgE or SPT. If negative: and still no history at any age of immediate onset symptoms -reintroduction at home using a milk ladder. If positive, refer to local pediatric allergy service. History of immediate onset symptoms at any time: sIgE or SPT. If negative, referral to local allergy service for re-challenge. If positive or test not available, refer to local pediatric allergy service.

NICE 2019 18
Re-testing: arranged every 12-18 months depending on local pathways and protocols. Strict adherence to a CM-free diet for the mother/ infant until the child is 9-12 months old and for at least 6 months. If symptoms do not improve over this time: (1) and CMA no longer suspected, the mother/infant resume normal feeding -referral to a pediatrician if symptom persist; (2) and CMA still suspected, referral to an allergology specialist and seeking specialist advice to avoid soy protein and egg. Child with non-IgE-mediated allergy: following a CMfree diet, a planned home reintroduction of cow's milk into the mother's or infant's diet. Tolerance to CMP e assessed using a 'milk ladder' and monitoring the symptoms (baked milk products reintroduced first (heating reduces allergenicity)). Signs of current atopic eczema or any history at any time of immediate-onset symptoms: home reintroduction contradicted and referral to an allergy specialist for allergy testing. Established tolerance: greater exposure of less processed milk gradually encouraged, ending in the reintroduction of fresh CM. Oral antihistamines available at home, in case of symptoms on reintroduction.
Symptoms return on reintroduction of CM: a CM-free diet continued, and re-evaluation after a 6 to 12 months. Confirmed IgE-mediated CMA: follow-up arranged by the specialist allergy service (may include serial allergy testing and subsequent OFC).

(continued)
Cow's milk re-challenge to test for acquired tolerance ISPGHAN 2020 12 OFC required before reintroduction of the allergen after therapeutic elimination period to confirm development of tolerance. Infant with IgE-mediated CMA: the elimination diet continued for at least one year and re-evaluation every 6 months subsequently.
Other guidelines 11,16 Not reported. Introduction of complementary feeding in infants with CMA EWPGAG 2010 8 -Home-made meals a dietary option after 4 months of age. -Breastfed infants: weaned as recommended for healthy children, but with avoidance of CM until 9-12 months of age and for at least 6 months from the beginning of the diet.
Finnish guidelines 2012 14 Not discussed in CMA section but with regard to food allergies in general.
-Introduction of additional foods in all children on a child-by-child basis beginning at the age of 4-6 months while breastfeeding is continued. Recommendation for introduction of wheat and oats before 6 months. Oral tolerance induction as a novel treatment option to the small but clinically significant proportion of affected individuals whose CMA persists.  Other guidelines 8,9,[11][12][13][14] Not reported. Nutritional deficiencies in CMA EWPGAG 2010 8 Diets must be nutritionally balanced. A supplementation with Ca must be evaluated.
WAO 2010  not achieve a median score of 60% for this domain. The main reason for such low scores was a lack of assessment of the views and preferences of the target population (patient, public, etc.).

Rigor of development (domain 3)
For this domain, the median score was 30% (Q1-Q3: 15-67%). The highest median score (100%) was achieved only by 1 set of guidelines (NICE). 18 The median score for 8 guidelines 8-12,14-16 did not exceed 44%. The main reasons for low scores for this domain were unclear description of strengths and limitations of the body of evidence and a lack of reporting of the procedures for updating the guidelines.

WAO 2010 2
Not reported as a recommendation. Some studies suggested a positive effect of probiotic interventions on atopic dermatitis, but meta-analyses did not confirm it. More RCTs need to be conducted to elucidate whether probiotics are useful for the treatment of AD.
Finnish guidelines 2012 14 Not discussed in CMA section but with regard to food allergies in general. Lactobacillus rhamnosus GG inhibits and ameliorates atopic eczema to some extent. There is no consistent evidence for the usefulness of probiotic bacteria in airways allergies. Not reported as recommendation. The use of PUFAs to treat CMA could be attempted in well-defined cases, but there is a need for more and comprehensive (pre-clinical) data for widespread recommendation.

WAO 2010 2
Not reported as recommendation. Studies are in the preclinical stage to treat food allergy with a traditional Chinese herbal remedy.

Clarity of presentation (domain 4)
The median score for this domain was 92% (Q1-Q3: 81-100%). Five guidelines (AAAAI and I-FPIES, BSACI, NICE, Spanish on non-IgEmediated CMA and WAO) 2,13,15,17,18 achieved the highest median score (100%). Only 1 set of guidelines 14 did not exceed a median score of 60%, in which the main reason for the low score was the lack of easily identifiable key recommendations.

Diagnosis
Recommendations for clinical history and physical examination to establish suspicion of CMA were presented in 9 guidelines. 8,10-13,15-18 Use of oral food challenge (OFC) and/or home reintroduction of baked-milk for diagnosis of CMA was recommended mostly in cases with suspicion of non-IgE-mediated CMA in four guidelines; 10,11,17,18 in 3 guidelines, it was advised regardless of IgE-mechanism, 8,11,12,14 and, in 4 guidelines, 2,13,15,16 in only defined specific cases. Skin prick or specific serologic IgE tests were recommended only if IgE-mediated CMA was suspected according to 9 guidelines; 2,10-13,15-18 however, in 1 set of guidelines, it was recommended regardless of type of CMA reaction. 8

Maternal elimination of cow's milk during breastfeeding
A continuation of breastfeeding with a maternal cow's milk elimination diet was recommended in 8 guidelines. 8,[10][11][12][14][15][16]18 Six of the included guidelines 10,11,13,15,16,18 recommended against a maternal elimination diet if the infant was asymptomatic on breastfeeding alone; in an additional one, 8 it was recommended against elimination diet in case of mild symptoms. Supplementation of the maternal elimination diet with calcium was recommended in 7 guidelines, 2,10-12,15,16,18 including 3, 10,15,18 that also recommended supplementation of vitamin D.

Use of cow's milk formula
Extensively hydrolyzed formulas (EHFs) were recommended as a first-line treatment in formula-fed children with CMA in 5 guidelines; 2,11,13,16,18 however, in 3 guidelines, 10,12,15 EHFs were only recommended for infants with mild-to moderate CMA. Amino acid formula was recommended in the management of children with severe CMA symptoms in 6 guidelines, 8,10,12,15,17,18 and/or in those with a high risk of anaphylaxis according to 4 guidelines, 2,11,16,18 and/or in case of no response to or refusing EHF according to 5 guidelines. 11,12,[15][16][17] Use of plant-based formula Rice-based formula was recommended as the treatment of choice in selected infants according to 3 guidelines, 8,11,16 and, in 1 additional set of guidelines, 15 hydrolyzed rice formula was recommended as an alternative if the infant refuses or does not respond to EHF. Use of soy formula was not recommended in infants below 6 months of age in 10 guidelines. [8][9][10][11][12][14][15][16][17][18] Use of plant based beverages and mammalian milks Use of other mammalian milks was not recommended in children with CMA according to 7 guidelines; 8,11,12,[15][16][17][18] however, in 1 of these, 16 an exception was made for equine milk with modified fat content, which could be used as an alternative. Five guidelines 11,15-18 recommended against use of soy plant-based beverage in infants with CMA. According to 3 guidelines, 10,17,18 use of rice plantbased beverage is not advised in children under 4.5 years of age. Two guidelines, 11,15 recommend against any plant-based beverages.

Acquisition of tolerance
Eight guidelines [8][9][10]12,13,15,17,18 recommended periodic re-assessments of acquisition of tolerance with oral food challenges in children with CMA; however, the recommended period varied across the documents. According to 4 guidelines, 8,13,15,16 complementary feeding should be introduced similarly as in healthy children. Five guidelines recommended supervision of the elimination diet by a dietitian (ie, assessment of one or more specific nutrients intake). 10,11,[16][17][18] Pre-, pro-and synbiotic and nutrient supplementations There were no recommendations with regard to probiotics, prebiotics, synbiotics, polyunsaturated fatty acids, or other non-pharmacological methods (ie, Chinese herbal medicine) for management of CMA.

DISCUSSION
This systematic review assessed CMA guideline quality using the AGREE II instrument and summarized specific recommendations for the diagnosis and management of CMA. While the quality of the CMA guidelines published in the past 10 years varied, the median score for almost all domains exceeded 60%, except the rigor of development domain, that had the median score 30%; Q1-Q3: 15-67%. The clarity of presentation domain had the highest median score (92%; Q1-Q3: 81-100%). Three guidelines (BSACI, NICE, WAO) achieved the highest ratings (100%) in at least 3 domains and for the overall score.

Agreement with other systematic reviews
Compared to the previous similar systematic review by Ruszczy nski et al, 1 which assessed CMA guidelines published from 2010 to November 2015, we included fewer full-text articles (12 compared to 15) despite the longer years of publication inclusion period. This is explained by our decision to only include the guidelines endorsed by the recognized scientific societies or organizations. Similar to Ruszczy nski et al, 1 we found the clarity of presentation to be the domain with the highest median score. We also found an improvement over time in the score for the applicability domain (68%, Q1-Q3: 57-75%) compared to the previous systematic review 1 (32%, range: 6-100%).
In the recent, non-systematic review of CMA guidelines by Munblit et al, 19 commercial involvement was reported as an important issue; 81% of authors in nine guidelines had financial conflict of interest with formula manufacturers and three CMA guidelines were directly supported by formula manufacturers. However, these 3 guidelines 20-22 with financial support from pharmaceutical companies were not endorsed by any scientific organizations. In our review, the editorial independence (domain 6) was of good quality in the majority of included guidelines. Sixty-seven percent of authors in six guidelines [9][10][11]13,15,17 declared conflicts of interest, in two 2,14 individual conflict of interest was not reported, in the other four, 8,12,16,18 there was nothing to declare.
The AGREE-II instrument is widely used to evaluate the methodological quality and transparency of guidelines that are used in clinical practice. 6,23 It was also designed to inform development and reporting requirements for practice guidelines (ie, prioritization of the update of high quality guidelines and improvement of quality of guidelines during update, if necessary). 5 Rigorous development and strategies for reporting are key predictors of successful implementation of the recommendations. 24 Although the AGREE-II instrument focuses on methodological issues around guideline development and reporting, these issues are insufficient to ensure that recommendations are appropriate and valid, as methodological rigor and validity are not necessarily correlated. 25 Therefore, when a clinician is choosing guidelines, some other factors may need to be considered according to the individual clinical situation, including guideline applicability (that may differ with regard to geographical region and/or available resources), scope of guidelines, and year of publication (preferably updated no later than 2 to 5 years from issue date). 5

Strengths and limitations
The search was limited to guidelines published in English only (risk of language bias). No blinding to the authors/organizations was implemented. However, the review team was well aware of available guidelines; thus, blinding, while feasible, might have been artificial. We used the AGREE II instrument to appraise all guidelines. However, the AGREE methodology has its limitations. For example, it does not provide a threshold for discrimination between good-and poor-quality guidelines, thus, the judgment is left to the appraisers. Of note, previous reviewers/appraisers, including members of the current panel, 1 have provided input as to an appropriate quality threshold (ie, standardized domain score of above 60%).
Some of the authors who contributed to this systematic review were also authors of some of the included guidelines. However, the appraisal of methodological quality using the AGREE II instrument was performed by independent reviewers.

CONCLUSIONS
The majority of the included CMA guidelines published from 2010 to 2020 were of good or very good quality. However, the weakest domain was the rigor of development, mostly due to the poorly described strengths and limitations of the body of evidence and the procedure for updating the guidelines.