Introduction

Cow’s milk protein allergy (CMPA) is an adverse and reproducible reaction to cow’s milk proteins that occurs mainly in infants and is a consequence of immediate (immunoglobulin E [IgE]-mediated), late (non-IgE-mediated), or mixed immune reactions.^1–3 The clinical manifestations of non-IgE-mediated gastrointestinal CMPA (non-IgE-GI-CMPA) are unspecific, and its differential diagnosis includes functional gastrointestinal disorders, reflux gastroesophageal disease, and infections, among others,^1–5 as well as symptoms related to gastrointestinal development in infancy.⁶ CMPA can cause poor growth, inadequate intake of energy and nutrients, micronutrient deficiencies, a predisposition to the development of feeding difficulties, and affect quality of life.^3,7–17

The prevalence of CMPA is below 1%³; however, some infants have clinical manifestations similar to that of CMPA which are secondary to other disorders.^1–7 Hence, overdiagnosis is likely to occur more frequently than underdiagnosis.^3,4

Several guidelines agree that the diagnoses of non-IgE-GI-CMPA must be performed in three stages:

1. Clinical evaluation that may eventually include additional laboratory tests

2. Assessment of the response to 2–4 weeks of cow’s milk proteins elimination diet

3. If there is a clinical recovery under the diagnostic elimination diet, an oral food challenge (OFC) test must be performed to confirm the diagnosis.^1–4

A positive OFC test is defined by the recurrence of symptoms after oral exposure to cow’s milk proteins that could last for 2–4 weeks.^2,4 The diagnostic OFC (D-OFC) test is strongly recommended, as approximately 50% of infants who begin a cow’s milk elimination diet do not have a positive D-OFC test.^7,18 In infants with suspected non-IgE-GI-CMPA, D-OFC test is of importance since the clinical manifestations are nonspecific and often secondary to causes other than food allergy. Furthermore, it is not possible to count on the support of the allergy tests available for patients with IgE-mediated CMPA, and the clinical manifestations are not immediate, making it difficult to establish the association between milk consumption and the occurrence of symptoms.^1–7

Infants with non-IgE-GI-CMPA usually develop spontaneous oral tolerance (OT) to cow’s milk proteins after 6–12 months of an elimination diet, which should be assessed using an OT-OFC test.^1–4

Surveys performed in Brazil and other countries have shown gaps in knowledge^19–22 and adherence^23–25 of physicians to CMPA guidelines; however, to the best of our knowledge, no published study has focused specifically on diagnosing and treating infants with non-IgE-GI-CMPA in Brazil. Thus, this study aims at evaluating how physicians diagnose and treat non-IgE-GI-CMPA in infants. These results can be used to plan medical education activities, public healthcare programs, and economic studies.

Materials and Methods

Study design and subjects

This cross-sectional study included physicians recruited by email invitations sent to 26,814 Brazilian Society of Pediatrics (BSP) members. The Ethics Committee of the Federal University of São Paulo (CAAE 40956720.4.0000.5505) approved the project, and only those who signed the Informed Consent Form had access to the questionnaire. A sample size of 379 pediatricians was calculated according to Serdar et al. (2021) from a total of 26,814 members of the BSP.²⁶ We assumed that 50% of respondents would not indicate D-OFC performance, a 95% confidence level, and a 5% margin of error. With a 15% safety margin, the sample expectation was 435. Eligibility criteria included graduating from a medical program and handling babies with CMPA in Brazil, whereas the exclusion criterion was managing less than one infant with CMPA per month.

Questionnaire

To evaluate the applicability of the questionnaire, a pilot study was conducted with 28 physicians from five regions in Brazil. The questionnaire contained information on professional specialization and experience, state, whether physicians worked in public and/or private healthcare services, diagnostic and therapeutic management practices for infants with clinical manifestations of non-IgE-GI-CMPA, and the supply of hypoallergenic infant formulas by the public healthcare system. The Informed Consent Form and the questionnaire were hosted on Google Forms (Google, Menlo Park, CA, US).

Questions regarding the diagnosis and treatment of non-IgE-GI-CMPA were asked based on two clinical cases. The first (Case 1) was of an infant with mild-to-moderate clinical manifestations: “The infant received exclusive breastfeeding until 3 months of age, when an infant formula based on cow’s milk proteins was introduced. At 3.5 months of age, the onset of gastrointestinal symptoms compatible with non-IgE-GI-CMPA (without malnutrition) was observed. At four months, the infant was evaluated by a physician and presented non-IgE-GI-CMPA symptoms, being fed with cow’s milk infant formula and without the possibility of relactation.” The questionnaire included the following possible gastrointestinal symptoms: regurgitation, vomiting, diarrhea, constipation, blood in the stool, food refusal, excessive colic/crying, and irritability.

The second clinical case (Case 2) was an infant with severe clinical manifestations—the clinical history was the same, but the symptoms were described as “severe” and associated with weight loss. The infant did not present with anaphylaxis or food protein-induced enterocolitis syndrome.

The same questions were asked for both clinical cases, including requests for tests, prescription of drugs, dietary management, referrals to specialists, and OFC performance.

In the instructions issued to respondents, they were encouraged to provide answers that reflected clinical practice with respect to their routine care activities and not knowledge obtained from guidelines or medical education activities. To this end, it was ensured that the answers could not be linked to the identity of the respondent, and 11 highlighted alerts appeared reiterating that their answers should be based on their clinical practice.

Statistical analysis

Statistical analyses were performed using Jamovi version 2.3 (https://www.jamovi.org). Bivariate and multiple analyses were performed to identify factors associated with using the D-OFC test and a significance level of 5% was used for all the analyses.

Results

The questionnaire was completed by 478 participants between February and October 2022. However, 31 physicians were excluded because they had not managed any infants with CMPA per month. Table 1 presents the interviewees’ characteristics and workplaces and Table 2 presents information on infants’ access to hypoallergenic formulas.

Diagnosing infants with non-IgE-GI-CMPA

Table 3 presents the answers related to the management of non-IgE-GI-CMPA in cases 1 and 2. Detailed information on the statistical analyses is provided in Supplementary File 1, and Supplementary File 2 shows the requested laboratory tests and the prescribed drugs and probiotics.

If administering the elimination diet resulted in a lack of clinical recovery, 72.3% (311/430) of the physicians in Case 1 and 35.9% (133/370) in Case 2 reported that they would change the initially prescribed hypoallergenic formula. The most frequently reported formula change was from the extensively hydrolyzed formula (EHF) to the free amino acid formula (AAF) (91.6% [285/311] in Case 1 and 69.9% [93/133] in Case 2).

Another approach for addressing the lack of response to the elimination diet was the referral to another physician with an additional board certification: 21.9% (94/430) in Case 1 and 50.8% in Case 2 (188/370). For both the cases, most referrals were to pediatricians with certification in gastroenterology (73/94 [77.7%] and 140/188 [74.5%]) and allergology (19/94 [20.2%] and 39/188 [20.7%]).

The third most common approach by physicians for addressing the lack of response to the elimination diet was to rule out the diagnosis of CMPA in infants fed with AAF; however, only 21.0% (13/62) and 15.4% (36/224) of the physicians who indicated AAF in both the cases ruled out the diagnosis of CMPA. The number of participants who did not choose to prescribe a D-OFC test was relevant (Table 3).

Treating infants with non-IgE-GI-CMPA

During the follow-up of infants with CMPA, the practice of switching the hypoallergenic formula with a cheaper one or one with another type of protein was reported by 47.9% (206/430) physicians in Case 1 and 31.9% (118/370) in Case 2 (p < 0.001).

In Case 1, the most frequently reported changes were:

1. AAF to EHF: 65.1% (134/206)

2. EHF to partially hydrolyze protein-based infant formula (PHF): 37.9% (78/206)

3. EHF to soy protein-based infant formula (soy formula [SF]): 13.1% (27/206).

For Case 2, the same changes were observed:

1. AAF to EHF: 86.4% (102/118)

2. EHF to PHF: 15.3% (18/118)

3. EHF to SF: 7.6% (9/118).

More than 90% of the physicians recommended the OT-OFC (Table 3), and the most frequent milk protein sources used in the OT-OFC were cow’s milk infant formula in Case 1 and baked foods in Case 2.

Factors associated with the performance of the D-OFC test

Table 4 presents the bivariate and multivariate analyses performed to evaluate the variables associated with using the D-OFC test (complete data in Supplementary File 3). The multiple regression analyses included variables with p values < 0.20 as independent variables. For both cases, performing the D-OFC presented a positive association with having a certification in pediatric gastroenterology and < 20 years of professional practice, and was negatively associated with using baked foods as a protein source in the OT-OFC. Only in Case 1 did certification in nutrition present a negative association with the practice of performing the D-OFC.

Discussion

We evaluated the practices of 447 physicians with respect to the diagnostic and therapeutic management of non-IgE-GI-CMPA in infants. In general, the initial management, prescription of an elimination diet, and frequency of indication of the OT-OFC test followed the guidelines. Access to hypoallergenic formulas subsidized by the public healthcare system was reported by > 90% of the physicians. In contrast, the D-OFC was performed by approximately half of the physicians, > 90% indicating that a good response to diet does not require this test. The D-OFC was most frequently performed by physicians specializing in pediatric gastroenterology and with < 20 years of professional experience. The use of baked foods as a protein source in the OT-OFC test was reported by more than one-third of physicians and was associated with not performing the D-OFC.

Evidence shows that 20 years ago, the most commonly used formulas in Brazil were soy-based, in addition to inadequate options such as plant-based drinks, goat’s milk, lactose-free milk, and PHF or even the nonprescription of breast milk substitutes.^7,8,19,20 In the past 20 years, several guidelines have been published for CMPA,^1–4,27,28 and several scientific events have been held with the focus on this disease. Simultaneously, the public healthcare system established municipal and state protocol for the supply of hypoallergenic formulas. Based on data collected from 2014, a preference for EHF and AAF has been observed.^{11,13,14,22,24,29}

This study complements the published data, demonstrating temporal adherence to the guidelines regarding the choice of formulas according to the severity of the clinical condition. Increasing access to hypoallergenic formulas is essential for the implementation of these guidelines. Various studies that have evaluated the nutritional status and food intake of patients with CMPA have shown a positive change, which could have resulted from adopting appropriate elimination diets. In addition to the lower than recommended daily intake of energy, calcium, and phosphorus in half of the children and vitamins A and D in 38%,⁸ studies conducted 20 years ago identified underweight in 15–23% and stunting in 12–24% of infants.^8,9 However, recent studies have shown an improvement in nutrition with a decrease in the rates of underweight (0.0–6.7%) and stunting (6.7–16.7%).^11,12,14,29 Additionally, more than 90% of formula-fed infants were fed EHF or AAF^11,13,14,29 associated to a significant reduction in the frequency of intake below the daily recommendations for energy and macro- and micronutrients.^11,13

Regarding the formulas, adherence to the guidelines was observed in initial management and in the management of infants with a lack of response to the diet, where exchange of EHF for AAF or referral of the infant for management by a specialist predominated. Nevertheless, many interviews revealed a frequent misconception regarding the exchange of EHF for PHF after the confirmation of CMPA diagnosis. PHF is not recommended for the therapeutic management of CMPA^1–4 and tolerance to this type of product may indicate that the infant has already developed tolerance to cow’s milk proteins.

Regarding the duration of the diagnostic elimination diet, the most recent guidelines recommend 2–4 weeks for patients with non-IgE-GI-CMPA.^2–4 This period may be too short for patients with severe conditions to achieve complete clinical and nutritional recovery. In this study, the mean duration of the diagnostic diet was 6 weeks in Case 1 and 8 weeks in Case 2, which are consistent with those recommended in the Brazilian guideline.¹ Another highlight is that 76% of the physicians reported that access to formulas subsidized by the public system was not immediate and that the family needed to pay for the formulas at the beginning of the elimination diet. This can lead to the use of cheaper and inadequate milk substitutes such as plant-based milks, thereby extending the period required for clinical and nutritional recovery. Recommending D-OFC test 2 weeks after clinical and nutritional recovery would be a personalized and assertive strategy, respecting each patient’s severity and recovery speed.³⁰

In this study, 12.3% and 37.7% of the participants reported requesting tests and 8.3% and 12.1% reported prescribing drugs and/or probiotics, respectively, for cases 1 and 2 (Supplementary File 2). Most of these tests were probably requested for differential diagnosis or investigation of symptoms related to other conditions. The substantial diversity of tests requested suggests that different prior experiences of the respondents may influence the lab tests requested for differential diagnosis. Notably, the frequencies of lab tests requested and prescribed drugs were lower than those reported in an older study,²⁰ again demonstrating progress in the Brazilian medical practice.

The D-OFC test has the potential to rule out suspicion of CMPA in approximately half of the children,^7,18 provide several nutritional and quality of life benefits, and avoid unnecessary costs associated with formulas, special foods, consultations, and exams.^{3,4,15,23,31–34} This diagnostic option was reported by 31% of the physicians in a Brazilian study conducted more than 10 years ago²⁰ and by only half of the participants in this study. Another study conducted in 2017 reported a notable result, as only 18% of pediatricians adhered to the use of the OFC test with 20% aware of this recommendation.²⁴ An analysis of the national medical prescription databases of the United Kingdom, Norway, and Australia confirmed that the prescription of hypoallergenic formulas has been much higher than expected, given the prevalence of CMPA.³⁵ In this study, the types of formulas prescribed by the physicians complied with the guidelines, highlighting the lack of D-OFC test as a critical negative factor for unnecessary prescription of hypoallergenic formula. The provision of more public services for OFC performance may also contribute to avoiding overdiagnosis.

In this study, the parent’s refusal to perform the OFC and the difficulties in its implementation and evaluation were not assessed, such as possible false-negative and false-positive OFC tests, secondary, respectively, to a delay in performing the D-OFC or viral infections and other symptoms during the OFC test, suggesting the recurrence of reactions to cow’s milk proteins. Brazilian guidelines¹ advocate the use of the D-OFC test; however, additional modifications are necessary to ensure its comprehensive implementation. The OFC test was officially incorporated into the country’s public and private health systems in 2022, as stated in the Ordinance SCTIE/MS No. 32 dated 03/31/2022 and ANS Normative Resolution No. 536/22. Nevertheless, prior to the submission of this article for publication, we could not ascertain from the SIA/SUS (Ambulatory Information Systems of the Unified Health System) website the total number of OFC tests conducted from the time of the publication of the ordinance as well as the amount of reimbursement.

However, over 90% of physicians have reported indicating OT-OFC test predominantly after 6 months of an elimination diet or when the infant reaches the age of 12 months, as established by the guidelines.^1–4

This study revealed that for the most part, the physicians opted to conduct the OFC test at home for Case 1 and in a health service for Case 2, which is consistent with the latest guidelines.^2–4 The low availability of public and private services structured for the performance of the OFC may have contributed to these practices. However, it is essential to point out that the OFC test is not completely risk-free. Besides, fear and anxiety may cause parents to mistake a nonspecific manifestation to be secondary to a reaction to the milk. Thus, performing the test at home does not exempt the physician from interpreting the result³ and whenever possible, the OFC should be performed under medical supervision.¹

Regarding the food used in the OFC test, milk formulas predominated in the D-OFC test, whereas baked foods had a relevant frequency in the OT-OFC test. The use of baked foods in the OT-OFC test is not a substitute for a D-OFC test and is not recommended for confirming the diagnosis of CMPA.^3,4 Thus, although the use of baked foods in the OT-OFC test is an option in the most recent guidelines,^3,4 the association of this practice with the nonperformance of the D-OFC suggests a misunderstanding of the role of the use of baked foods.

Another point observed in this study was that the D-OFC test was used most frequently by physicians who had less professional experience (< 20 years). In Brazil, the OFC test has been recommended for the last 20 years; hence, more likely to be part of the practices of professionals with academic training during this period. Specialization in gastroenterology was also associated with the indication for the D-OFC test, possibly because of the nonspecific characteristics of non-IgE-mediated gastrointestinal manifestations compared to IgE-mediated reactions.

One limitation of this study was that the recruitment and data collection were done online; however, this had the advantage of having a low cost and ample coverage by including professionals from all regions of Brazil. Although the absolute number of respondents working in the Southeast was higher, the distribution across all regions was similar to that of the workplaces of members of the BSP and the medical demography of Brazil.³⁶ Notably, the aim of this study was to evaluate real-life medical practices; hence the strategies adopted to encourage responses that pertained to these care practices. These data should be confirmed in future studies with data collection from records of care provided by public health services. Additionally, to our knowledge, this is the first study to assess access to formulas, which is critical for the adequate treatment of CMPA.

Conclusion

Finally, unlike in the past, most participants followed the guidelines regarding the type of hypoallergenic formulas that should be used in the elimination diet according to the severity of clinical manifestations. This progress in pediatric practice in Brazil is likely the result of improvements in guidelines adherence, medical education campaigns, and access to hypoallergenic formulas subsidized by the public healthcare system. Administering the OT-OFC test was a common practice; however, the D-OFC test was done by only half of the physicians, indicating the need to promote this practice to avoid the negative impacts of unnecessary restrictive diets and provide better outcomes for individuals and society. Studies need to be carried out on the practices of physicians from other countries in performing the OFC for diagnosing CMPA in infants.