Food allergy (FA) as a common disease in children with increasing prevalence rates over the past two to three decades. In some countries, FA prevalence has reached as high as 10% (1). A systematic review estimated that direct medical costs for patients with FA in the United States alone amount to $4.3 billion annually (2). Cow’s milk allergy is one of the most common causes of FAs in children. The reported prevalence of CMA varies due to differences in diagnostic methods and epidemiological survey designs. CMA prevalence range of 1.8%–7.5% in the first year of life was reported (3). Some adult patients with CMA had allergies since childhood, but some studies have also reported adult-onset CMA (4).

Management of cow milk protein allergy involves strict avoidance of food allergens and the use of pharmacological treatments to address reactions from accidental exposure. In infancy, CMA is often managed by substituting cow milk with extensively hydrolyzed protein or amino acid-based formulas. However, these alternatives have poor taste and are costly, imposing a significant financial burden on families (4). Alongside the persistent risk of systemic allergic reactions, avoiding milk consumption may also have nutritional implications. Sinai et al. compared the adult height of CMA patients with a control group without dietary restrictions and found that CMA patients were, on average, 3.8 cm shorter and had a higher rate of vitamin D deficiency (5). Recently, there has been an increasing research focus on immunotherapy for CMA, aiming to achieve desensitization as early as possible (6). Various immunotherapy approaches for milk protein desensitization, including oral immunotherapy, sublingual immunotherapy, and subcutaneous immunotherapies, and OIT is the most common approach. OIT is a treatment method that induces immune tolerance by gradually increasing the intake of allergens. OIT likely induces immune tolerance via repeated allergen exposure by rebalancing the Th2-driven allergic response toward a Th1/Treg-dominated profile (7). However, findings regarding their effectiveness are inconsistent. Some studies have found that immunotherapy can facilitate desensitization to cow milk protein (8), while others have reported no such effects (9). Therefore, we aimed to conduct a systematic review and meta-analysis of oral immunotherapy for CMA in children to provide clinical recommendations for managing CMA.

This review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (10). The protocol and search strategies were registered with PROSPERO (York. ac. uk) (CRD42024541769).

Our meta-analysis included 19 studies to comprehensively evaluate the overall efficacy and safety of immunotherapy in patients with IgE-mediated cow milk protein allergy. Despite the high heterogeneity and generally low evidence levels among these studies, immunotherapy was more effective for achieving desensitization in patients with CMA compared to avoidance therapy, and it significantly altered the cumulative tolerance dose. The immunotherapy group did not show a higher risk of adverse effects than the control group, and severe adverse events were rare.

Like the Riggioni (38) study, oral immunotherapy was effective in the treatment of cow’s milk allergy compared to avoidance of milk protein, with the results exhibiting considerable heterogeneity. Unlike the Riggioni study, which included eight RCT studies and two case-control studies and did not perform subgroup analysis, our study exclusively included RCTs, yet it still exhibited a high degree of heterogeneity, suggesting that the experimental design is not the source of heterogeneity. The differences may be related to the high heterogeneity in study populations, intervention protocols, outcome measures, follow-up periods, and other methodological variations. Despite the heterogeneity, sensitivity analysis suggested the reliability of the results. Standardization of trial design, outcome measurement, and reporting practices is crucial to address these issues and enhance the reliability and applicability of research findings.

CMA pathogenesis can be mediated by IgE, non-IgE, or both. Most cow milk protein allergies are mediated by IgE. Cow milk contains caseins and whey proteins, which account for approximately 80% and 20% of the total milk protein, respectively. Most patients with cow milk protein allergy are allergic to multiple proteins in cow milk, with casein, β-lactoglobulin, and α-lactalbumin being the major allergens (39). Immunological evaluation in this meta-analysis showed that immunotherapy reduced cow milk protein-specific IgE and casein-specific IgE levels. Levels of cow milk-related IgG4 increased, likely due to enhanced IgG4 binding to milk epitopes, which in turn reduces IgE binding to these epitopes and induces tolerance (40).

This meta-analysis also examined how intervention type and dosage affect desensitization efficacy. Among the included studies, four types of milk interventions were used: baked milk, heated milk, raw cow milk, and glycated milk, with most studies using raw cow milk. Heating or baking milk at high temperatures can degrade certain structural epitopes, rendering them unrecognizable by allergen-specific IgE. Although one study (18) suggested that frequent consumption of baked products accelerates tolerance to cow milk protein, our study did not find this relationship, which is consistent with the meta-analysis results of Anagnostou et al. (41). Tolerance to heated or baked milk may be temporary and diminish within a few years. Dantzer et al. found that a protocol involving gradually increasing doses of baked milk was effective in promoting desensitization. This suggests that the initial dose may influence the efficacy of desensitization. Two studies evaluated the effects of low-dose and high-dose target protocols, and a pooled results indicated that the maximum dose did not appear to be associated with the efficacy of tolerance. Furthermore, baked or heated milk products are not standardized, and the extent and consistency of the allergen quantity or epitope structure in each preparation method are unknown, necessitating further research.

Although children with cow’s milk allergy have a higher prevalence of EoE than the general population, Ferreira Martins et al. (33) reported cases of children with CMA who had no pre-existing EoE developing EoE following oral immunotherapy. This suggests that EoE may represent a potential complication of OIT treatment. EoE typically emerges approximately 2.8 years after initiating the maintenance phase of OIT in CM-allergic patients (28). Therefore, studies with short follow-up periods may underestimate the true incidence of EoE. In our meta-analysis, the observation periods in the included RCTs ranged from 23 to 96 weeks, with only one study following patients for 96 weeks and the remainder not exceeding 52 weeks. This relatively short follow-up duration may have led to underestimation. After expanding our search to include non-RCT studies, we found that 4% of CM-allergic patients developed biopsy-confirmed EoE following OIT - a rate consistent with previous reports (42).

This study has several strengths: 1) inclusion of the latest research on immunotherapy and biological therapy for cow milk protein allergy, encompassing many participants, and 2) a comprehensive analysis with stratified evaluations of immunotherapy for CMA by region, age of inclusion, intervention methods, and control groups. However, the high heterogeneity across studies, including differences in study population, intervention protocols, outcome measures, and other methodological variations, presents a limitation. Nevertheless, the large sample size and the stability of sensitivity analysis results add robustness to our findings for clinical decision-making.

Limitations of this study included: 1) the included studies had a small sample size; 2) the exclusion of studies without a control group, potentially introducing bias in the evaluation of indicators before and after immunotherapy or biological therapy; 3) high heterogeneity among included studies and the failure to identify the sources of heterogeneity through subgroup analyses and other methods. The heterogeneity may be attributed to variations in target doses and maintenance duration across studies. The included studies in this analysis employed varying protocols, with target doses ranging from 3 mL to 250 mL. The high-dose group demonstrated a higher incidence of severe adverse reactions during the maintenance phase compared to the low-dose group. Longer durations of OIT were associated with higher rates of desensitization or non-responsiveness (43). Consequently, the treatment duration significantly impacts efficacy evaluation. The double-blind, placebo-controlled food challenge (DBPCFC) is the gold standard for diagnosing food allergies and is an objective tool for assessing desensitization therapy, allowing for the evaluation of changes from before to after immunotherapy and the assessment of thresholds. Notably, many articles did not use DBPCFC as an outcome measure, limiting the interpretation of the results. Inconsistent follow-up periods and potential publication bias further complicate the interpretation and generalization of the results. Optimizing the included population, standardizing trial design, and outcome measurements may help improve heterogeneity; 4) Sustained unresponsiveness is defined as the absence of allergic reactions upon food reintroduction after a prescribed period of allergen avoidance following OIT discontinuation, serving as a key metric for assessing long-term therapeutic success. However, none of the included RCTs reported sustained unresponsiveness rates, underscoring the need for future trials to incorporate this clinically meaningful endpoint;5) the lack of a systematic evaluation of long-term effects or side effects. The outcome measures included in this study had a relatively short duration, mostly within one year, and lacked long-term outcome evaluations, and 6) Among the included studies, only one assessed cytokine changes during OIT without in-depth analysis of the underlying immunomodulatory mechanisms. There are reports (44) that OIT may have long-term side effects, including eosinophilic esophagitis; however, most are case reports, and further attention is needed to monitor these side effects.

Future studies should establish standardized definitions for initial dosing (differentiating low versus standard starting doses), maintenance dosing (distinguishing low versus high maintenance doses), and efficacy evaluation criteria to reduce inter-study heterogeneity. A severity-stratified design should be implemented, with conservative low starting doses adopted for children with severe allergies. Furthermore, extended follow-up periods are recommended to properly evaluate long-term therapeutic efficacy and safety profiles.

This meta-analysis suggests that oral immunotherapy can effectively desensitize milk proteins in children with cow protein milk allergy, though data on long-term follow-up remains incomplete. Currently, there are limited data on the long-term use of immunotherapy agents as well as the impact of treatment interruption on milk allergy status, immunological changes, and improvements in quality of life. Although the reported adverse events are mainly mild to moderate, comprehensive studies on long-term safety are still lacking.