Serum Eosinophil-Derived Neurotoxin: A New Promising Biomarker for Cow’s Milk Allergy Diagnosis

Even though respiratory symptoms (wheezing, chronic cough), GI symptoms (regurgitation, vomiting, diarrhea, constipation, and bloody diarrhea), and general symptoms (poor growth and infantile colic) typically occur in 15-20% of infants, they can also be related to CMA diagnosis.²

CMA may be IgE-mediated, non-IgE-mediated, or mixed in which both IgE-mediated and cell-mediated immunological mechanisms are involved. Despite their frequency, non-IgE and mixed CMA are less understood.¹⁸ In our study, CM specific IgE was positive in only 33.3% of CMA group with no reported immediate clinical reactions. In contrast to IgE-mediated CMA, most cases of non-IgE and mixed CMA are diagnosed clinically, which is not always easy and often leads to a misdiagnosis or a diagnostic delay. This is owing to their potentially chronic presentation, delayed onset of symptoms, less obvious allergen link, and possibility for clinical overlap with GERD, infections, or FGIDs which are much more common in the first year of life with a reported prevalence of 25% to 77%.^7,19 In addition, apart from an OFC with CM, there are currently no optimal diagnostic tests and biomarkers for non-IgE-mediated CMA.²⁰ However, parents frequently refuse or delay OFC due to fear of a severe reaction.²¹

Failure to thrive (FTT) is an alarm symptom that can be caused by a variety of factors or underlying disease and requires a comprehensive diagnostic work-up and referral for full evaluation.¹⁰ In the current study, weight and weight for length were significantly lower in CMA group than the other 2 groups with prevalence of faltering growth in around one third of CMA infants. Despite CMA may have an impact on growth and lead to faltering growth, data are still limited.⁸

Considering that FGIDs infants and their families have a worse quality of life (QoL) and seek medical attention more frequently than asymptomatic controls, FGIDs have per definition no identifiable underlying organic cause.²² So, infants with poor weight gain or abnormal physical examination shouldn’t be diagnosed with FGIDs.¹⁹ This can explain the absence of significant difference between FGIDs and control groups in all anthropometric measurements in this study.

Patients with CMA often experience a wide range of symptoms that spread over many organ systems. FGIDs, on the other hand, only affect the GI tract and have general symptoms such as excessive crying and irritability. Thus, the presence of GI and/or general symptoms along with skin and/or respiratory signs increases the likelihood of CMA.¹⁰ Nevertheless, there is ongoing controversy regarding the relationship between CMA and FGIDs in the absence of further symptoms or signs of atopy.^19,23 Therefore, it is crucial to search for a reliable test or biomarker for CMA diagnosis to prevent both overdiagnosis and underdiagnosis.

In our study, CoMiSS demonstrated higher statistical significance in the CMA group than FGIDs and control groups (with a median of 14). Given that the majority of our cases presented with multiple symptoms and that 46.66% of them had more than one system affection, this high score in the CMA group makes sense. Also, the median CoMiSS in healthy controls was comparable with reported results of CoMiSS cut-off score in healthy infants aged ≤ 6 months, which varies between 3 and 4,^24,25 and CoMiSS in infants aged 6–12 months with an overall median of 3.²⁶ The total CoMiSS was significantly higher in the FGIDs group (median 9) than the control group (median 4). However, no prior research has reported a cut-off score for CoMiSS in FGIDs.

Despite high median CoMiSS in CMA group in our study, ROC curve analysis of CoMiSS identified score > 9 as the best cut-off for differentiating between infants with CMA and without CMA (FGIDs and control groups) with 97.78% sensitivity and 78.89% specificity and AUC: 0.964 which was consistent with the most recent CoMiSS update with a cut-off score of ≥10 suggestive of CMA.¹⁰ According to eleven studies, a score of at least 12 indicates a favorable response to CMFD and the estimated sensitivity and specificity for CMA diagnosis were 20% to 77% and 54% to 92%, respectively.² However, there is a disparity in the cut-off values of CoMiSS in different studies, ranging from ≥5.5 to ≥12, which can be attributed to differences in study design, some studies used symptoms as an inclusion criterion and calculated CoMiSS as supplementary information,^2,27,28 while other studies used a CoMiSS above a specific cut-off as an inclusion criterion.^29,30 This range of values indicates that the type of symptoms reported and the study design may have an impact on how CoMiSS operates. So, CoMiSS cannot be considered as a stand-alone CMA diagnostic tool.

Considering the lack of markers unique to CMA, CBC values may be regarded as a useful tool for inflammatory diagnosis and monitoring in numerous chronic disorders as well as in allergic reactions, where MPV and NLR have been investigated as inflammatory markers with contradictory findings.^11,31,32,33

In the current study, MPV and TLC, though with low sensitivity and specificity, revealed a statistically significant difference between the CMA and control groups but not between the CMA and FGIDs groups. Similarly, Çam³⁴ found that the MPV values of the allergic proctocolitis (AP) group were statistically higher than the control group. In contrast, Dogan and Sevinc³⁵ found no significant difference in MPV levels of infants with and without CMA. These findings suggest that alterations of MPV levels may be easily affected by the type and severity of inflammation.

Few studies have assessed NLR in children with most of these studies focusing on NLR in asthma.^36,37 Like studies assessing MPV levels, varied results for NLR in asthma have been documented.^38,39 Furthermore, there was no significant difference in NLR between infants with CMA and controls according to Çam.³⁴ In our results, despite NLR in the CMA group being statistically significant compared to the other groups, ROC curve showed low sensitivity and specificity. Therefore, its role in CMA diagnosis is debatable.

About 60% of all leukocytes in the bloodstream are neutrophils, which play a significant role in type 1 and type 3 immunological responses but have a controversial role in type 2 immunity.^40,41 Severe allergies, autoimmune and autoinflammatory diseases are influenced by neutrophil dysregulation.⁴² Furthermore, in FPIES, positive responses to the OFC are usually followed by a rise in the ANC of more than 3500 cells/mm3.⁴³ Our findings were in the same line as ANC showed a higher statistically significant difference in CMA group than other groups, and ROC curve revealed 80% sensitivity and 78.89% specificity at a cut-off point >3900 cells/mm3. Likewise, Kimura et al.⁴⁴ reported a significant increase in ANC in OFC-positive than in OFC-negative subjects. On the other hand, any bacterial infection, inflammation, burns, cancer, and certain medications, like cortisol, also cause an increase in neutrophils.⁴⁵ Furthermore, blood samples that are either EDTA-based or insufficiently anticoagulated with heparin or citrate might result in platelet clumping in an automated cell counter leading to falsely elevated neutrophil count (factitious neutrophilia).⁴⁶ As a result, ANC monitoring for infants with CMA is suggested. However, its usefulness as a CMA biomarker is questionable.

Non-IgE FA is characterized by increased intestinal permeability and inflammation, which lead to granulocytes and eosinophils migrating to the intestinal lumen. Fecal markers are therefore a helpful diagnostic tool for inflammatory disorders in gastroenterology and a non-invasive means of assessing intestinal inflammatory responses. Additionally, the identification of fecal biomarkers is becoming increasingly interesting due to the absence of reliable diagnostic techniques.¹⁷

Numerous studies have investigated the use of different fecal biomarkers for CMA diagnosis, including fecal IgA, EDN, eosinophilic cationic protein (ECP), and fecal calprotectin (FC).¹⁷ Unfortunately, studies regarding fecal EDN in non-IgE-mediated CMA children are controversial and intestinal permeability appears to interact with the composition of the microbiota, which has been implicated in the development of food allergies.⁴⁷ fEDN level was higher in CMA patients compared to controls.^1,17 Although this finding seems promising, these differences were not statistically significant.

The most reliable marker of allergic inflammation is eosinophils, which have big cytoplasmic granules that contain proteins such major basic protein, eosinophil protein X, EDN, and ECP. Numerous inflammatory and allergic conditions, including atopic dermatitis, asthma, and other conditions can raise the number of eosinophils in the blood.^48,49 Additionally, elevated EDN and ECP levels have been reported in asthma patients, and these levels are associated with exacerbating asthma symptoms. Therefore, EDN and ECP may aid in the diagnosis and monitoring of asthma.⁵⁰

In the same line, Çam³⁴ found significantly higher sECP levels in infants with CMA compared to controls (sensitivity 60.7% and specificity 97.5%). Furthermore, comparable results were observed by some studies.^51,52,53 Nonetheless, several studies have found that sEDN is a more accurate indicator of disease severity than sECP.^14,50 In a prior study, sEDN levels were assessed in 4 allergic diseases (bronchial asthma, atopic dermatitis, allergic rhinitis, and FA) in children aged 6 to 12 years, and the ROC curve of sEDN revealed 81.2% sensitivity, 69.8% specificity, and AUC: 0.790.¹⁴

But as far as we know, there are no studies in the literature on the role of sEDN in the diagnosis of infants with CMA or the relationship between sEDN, CoMiSS and hematological parameters. In our study, sEDN at a cut-off point > 14 ng/mL demonstrated 86.67% sensitivity, 60.00% specificity and AUC: 0.754 in differentiating infants with CMA than those without CMA. Additionally, Spearman’s coefficient revealed a significant positive correlation between sEDN with both CoMiSS and ANC. However, multivariate linear regression analysis found that CoMiSS was the only parameter that statistically and independently influenced the sEDN level in the CMA group. CoMiSS as an acknowledged awareness tool, demonstrated good screening performance in our study; however, its high sensitivity was attributed to the fact that most cases had multiple symptoms, increasing its total score. Furthermore, the heterogeneous presentation of CMA and the subjectivity of parental perception of severity and duration of crying (which may be over-reported) may limit the diagnostic value of CoMiSS in CMA diagnosis.

Therefore, given the increasing prevalence of CMA, we recommend more research on the role of sEDN in CMA diagnosis including participants with a broad range of symptoms with evaluation of its level prior to and following CMFD, and to evaluate sEDN role in disorders other than FGIDs which may overlap with CMA diagnosis. Additionally, to fully understand their significance in CMA diagnosis, it is advisable to correlate sEDN with CoMiSS and hematological parameters, particularly ANC.

The main strengths of our study were that we are the first study to investigate sEDN in CMA infants. Also, we compared sEDN in CMA infants with infants diagnosed with FGIDs which are much more common in the first year of life and have similar manifestations to CMA, in addition to healthy controls. Moreover, all samples were withdrawn in absence of any acute illness, dietary interventions, or drug intake to avoid their possible impact on results. Finally, we correlate sEDN with both basic hematological parameters and CoMiSS which is the available clinical awareness tool for CMA.