Cow’s Milk Protein Allergy in Term and Preterm Infants: Clinical Manifestations, Immunologic Pathophysiology, and Management Strategies

The prevalence of cow’s milk allergy in preterm infants is poorly established.

After completing this article, readers should be able to:

1. Discuss the wide range of clinical presentations of cow’s milk allergy from nonspecific gastrointestinal manifestations, atopic dermatitis, and urticaria to more severe symptoms including anaphylaxis and septic appearance.

2. Describe the management of suspected cow’s milk allergy in an infant.

3. Explain how cow’s milk allergy may clinically resemble necrotizing enterocolitis in preterm infants.

The most common food allergy among young children and infants is cow’s milk allergy (CMA). (1) Food allergy is characterized by an adverse immune reaction occurring in response to the interaction of the immune system with a food protein. There are immunoglobulin E (IgE)–mediated, non–IgE-mediated, and mixed mechanisms of food allergy. In IgE-mediated food allergy, specific IgE antibodies directed at epitopes of 1 or many allergens in cow’s milk trigger an allergic cascade upon exposure to the antigen. Symptoms of an IgE-mediated allergic reaction occur within minutes to 2 hours after ingestion and include urticaria, angioedema, wheezing, rhinoconjunctivitis, vomiting, or diarrhea. Symptoms may be mild or may progress to anaphylaxis, which can be life-threatening. Non–IgE-mediated allergy includes food protein–induced allergic proctocolitis (FPIAP), food protein–induced enterocolitis syndrome (FPIES), food protein–induced enteropathy (FPE), and Heiner syndrome (pulmonary hemosiderosis). These diagnoses comprise about half of all CMA. The mixed IgE- and non–IgE-mediated immune processes are seen in atopic dermatitis and eosinophilic esophagitis (EoE), both of which can be manifestations of food allergy. The terms “cow’s milk protein sensitivity” and “cow’s milk protein intolerance” are no longer suggested terminology. Instead, the terms that identify the specific type of non–IgE-mediated disorder should be used to help differentiate nonimmune mechanisms (such as lactose intolerance) from classic IgE-mediated immune mechanisms.

Cow’s milk (CM) is typically the first foreign protein in an infant’s diet, and CMA is the most common food allergy in young infants. CM contains many different proteins including caseins and whey. Caseins, including ⍺-casein, β-casein, and κ-casein, make up the majority of the protein in CM. Whey proteins include bovine immunoglobulin, lactoferrin, serum albumin, α-lactalbumin, and β-lactoglobulin (BLG). Although BLG, α-lactalbumin, and casein are responsible for most milk allergens, some patients may be sensitized to the other CM proteins. Whey and casein differ in that whey will denature when heated extensively. (2) Caseins are located in micellular complexes in milk and interact with the immune system via Peyer patches in the gut. On the other hand, whey is soluble and is quickly transported from the lumen to the other side of the intestinal epithelium. (3) It has been shown that in mice sensitized to both whey and casein, whey proteins stimulate a more severe allergic reaction as a result of rapid transport across the epithelium. (3) Caseins are more likely to produce a robust antibody response, specifically IgE. (3)

CM protein is secreted in human milk in varying concentrations among mothers. One study analyzed human milk from 53 lactating women and found that 15% of the women did not secrete detectable BLG in their milk. (4) Jarvinen et al found that 46% of breast milk samples that caused allergic symptoms in infants did not have detectable BLG after a maternal CM challenge. (5) Human milk samples reach maximum BLG content anywhere from 4 to 24 hours after ingestion. (6)(7) The variable kinetics seen in BLG secretion have been demonstrated with other food proteins as well, and there is high inter- and intraindividual variability of BLG secretion in human milk as there is for peanut and egg proteins. (8) A major peanut allergen known as Ara h2 is secreted in breast milk; however, its kinetics vary, with some mothers secreting the antigen rather quickly (in 1–6 hours) whereas a subset of lactating mothers has delayed secretion of peanut antigen (8–12 hours). (8) Multiple factors potentially influence the variation in the timing of food antigen secretion in breast milk, such as maternal intestinal permeability, extent of breast emptying at each feeding, maternal gastric motility, and the quantity of food antigen in the maternal diet. Lack of adherence to a strict dairy-free diet can contribute to the continued presence of CM proteins in breast milk.

CM proteins in breast milk can elicit various clinical symptoms in already sensitized infants. Mothers of infants with colic were found to have higher concentrations of BLG. (9) Maternal dietary proteins in breast milk can also provoke cutaneous symptoms in allergic infants. (5) Mothers of 17 exclusively breastfed infants with CMA eliminated CM for at least 1 week to achieve symptom resolution and were then given increasing amounts of CM followed by breastfeeding. Among infants with CMA, 94% displayed signs of an allergic reaction during the challenge. Symptoms occurred between 2 and 80 hours after maternal ingestion, with a median time of 21 hours. All infants who reacted displayed cutaneous symptoms (mostly a flare-up of eczema or maculopapular rash), and 5 infants had gastrointestinal symptoms (vomiting, loose stools, or an increase in regurgitation) as well. Infants in this study were exclusively breastfed, suggesting that the initial sensitization occurred in utero or through breastfeeding. Not all mothers of symptomatic infants had detectable BLG in their milk, which suggests that other CM proteins that were not measured may have played a role in their symptoms. (5) In conclusion, this study showed that CM proteins secreted in breast milk can cause symptoms in some already sensitized infants.

Some data suggest that early exposure to CM-based formula in the first few days after birth in otherwise exclusively breastfed infants may increase the risk of developing CMA (Table 1). CM given in the first days or weeks of age in the form of CM formula may initiate sensitization, triggering an allergic reaction upon subsequent exposure in the future, either through direct CM ingestion or through CM proteins in human milk. (5)(10) In a Danish population-based study of 1,749 infants, 39 infants had confirmed CMA. Of the 39, only 9 were solely breastfeeding at the time of diagnosis and all 9 had received CM formula briefly in the first 3 days after birth. Of the 1,749 infants, 210 did not receive CM-based formula in the first 3 days after birth, and none went on to develop CMA (P<.05 compared with those who developed CMA). (10) Saarinen et al conducted a prospective study of 6,209 full-term, healthy infants who required supplementary milk while in the hospital. (11) Infants were randomized to receive CM-based formula, pasteurized human milk, or whey hydrolysate formula. The control group was composed of exclusively breastfed infants. The main outcome was CMA at 18 to 36 months of age diagnosed by challenge-proven adverse reaction (urticaria, vomiting, diarrhea, wheezing, allergic rhinitis, atopic dermatitis, or exanthema) to CM after successful 2- to 4-week CM elimination. Both the pasteurized human milk and whey hydrolysate groups had a lower incidence of CMA compared with the CM-based formula group (odds ratio [OR] 0.70; 95% confidence interval [CI] 0.44–1.12 and OR 0.61; 95% CI 0.38–1.00, respectively). (11)

In a randomized trial, (12) newborns at high risk for atopy were randomized either to breastfeed and avoid supplementation with CM formula for 3 days after birth (with or without amino acid–based hypoallergenic elemental formula) or to breastfeed and receive CM-based supplemental formula. Infants in the CM formula group were given 5 mL of CM formula within 24 hours of delivery and daily, which was increased to 40 mL/day at 1 month of age. In the exclusively breastfed group, mothers were asked to exclusively breastfeed until at least 5 months of age. If mothers felt that breast milk was not enough, they were allowed to supplement with amino acid–based hypoallergenic formula. If the mother gave the infant more than 150 mL/day of amino acid–based formula for 3 consecutive days, they were allowed to switch to CM formula. At 24 months of age, significantly fewer infants in the breastfed/amino acid–based elemental formula group had sensitization to CM (measured by serum IgE >0.35 U/mL) compared with the breastfed/CM formula group (16.8% and 32.2%, respectively, risk ratio [RR] 0.52; 95% CI 0.34–0.81). A secondary outcome of the study was clinical food allergy, as determined by oral food challenges (with observed reaction) or by a strongly suggestive history of reaction in combination with serum-specific IgE. Clinical food allergy to CM was present in 0.7% of infants in the breastfed/amino acid–based elemental formula group, which was significantly less than the 6.6% of infants in the breastfed/CM group (RR 0.10; 95% CI 0.01–0.77). Urashima et al found that avoiding supplementation with CM formula for at least the first 3 days after birth decreases the risk of sensitization to CM and the development of CMA. (12) Conversely, in a large population-based prospective study of more than 13,000 infants, Katz et al investigated the timing of CM protein introduction and the relationship with IgE-mediated CMA. There was a higher risk of challenge-proven IgE-mediated CMA in infants who were introduced to CM protein at 15 to 194 days compared to before 14 days of age (OR 13.3, P<.001); however, this risk declines with introduction beyond 194 days. (13) Limitations of this study include reliance on self-reporting, lack of data on patients with atopic dermatitis, and not taking parental atopy into account. (13) It is also not known whether infants continued to receive CM formula regularly after introduction, which could be a main difference with an early bottle given at the newborn nursery without continued exposure thereafter. Koletzko et al applied an analysis similar to that of Katz et al to a large cohort of high atopy-risk infants to investigate the hypothesis that introduction of CM protein in the first 2 weeks after birth decreased the risk of IgE-mediated CMA. Newborns were randomized to receive either hydrolyzed CM protein–based study formula or CM-based formula at birth. CM protein sensitization was determined by specific IgE. In this cohort, the age of CM protein introduction did not correlate with decreased risk of sensitization. (14) In conclusion, early supplementation of CM formula during the first few days after birth may predispose infants to CMA. Data are not conclusive, and continued exposure after the first few days after birth may affect the rate of sensitization. However, routine early introduction of CM formula to exclusively breastfed infants is not recommended if the infant is healthy and maternal breast milk is available.

IgE-mediated CMA occurs when specific IgE antibodies recognize the food protein and trigger an immune response. CM allergen-specific IgE antibodies are bound to mast cells and basophils. When there is exposure to the protein, the mast cells degranulate, releasing mediators such as histamine and cytokines that promote mucus secretion, smooth muscle contraction, and peripheral vasodilation. This leads to the characteristic symptoms of an IgE-mediated allergic reaction, including urticaria, wheezing, angioedema, nasal congestion, ocular pruritis and tearing, vomiting, diarrhea, abdominal cramping, and potential hypotension, dizziness, and shock. (15) This reaction can be fatal. In infants, the most common presenting symptom of an IgE-mediated allergic reaction to food is urticaria, (16)(17) but anaphylaxis is relatively uncommon, occurring in 2.8% of positive reactions. (17) In a retrospective study of 375 anaphylactic food reactions in infants (<12 months) and preschool children (1–6 years), CM was the most common food allergen. (16)

It has been estimated that up to 60% of CMA is IgE-mediated; however, this estimate varies by population and age. (18) The prevalence of CMA is highest in younger children, with IgE-mediated CMA affecting 0.5% to 2.2% of children up to 2 years old. (13)(19) The evaluation of IgE-mediated CMA includes a thorough history, skin prick testing, and/or serum-specific IgE. Positive skin prick tests and IgE indicate sensitization, and larger values indicate a higher probability of clinical allergy. However, values do not correlate with the severity of reaction upon exposure to the food protein. In isolation, these tests are not sufficient for diagnosis alone. In an ideal world, an oral food challenge would be the gold standard for the diagnosis of food allergy. However, given that food challenges are time consuming, with a risk of anaphylaxis, they are not routinely used to diagnose food allergy. Evidence of sensitization and a convincing history consistent with clinical reaction are often sufficient to diagnose IgE-mediated CMA. In instances where the clinical history of a reaction is less clear and/or there is little to no evidence of sensitization, an oral food challenge in an allergist’s office or other supervised medical setting equipped to handle severe allergic reactions can be considered to determine if clinical allergy exists. Food challenges are also indicated when it is suspected that the patient may have outgrown the CMA, which happens in most patients over time.

Treatment of IgE-mediated CMA is avoidance and prompt treatment of reactions upon accidental exposure. Infants with IgE-mediated CMA should avoid CM formulas and instead be fed a hypoallergenic infant formula—an extensively hydrolyzed formula, soy-based formula, or even an amino acid formula in more severe cases. Goat milk should also be avoided because of the high risk of cross-reactivity. (20) If an infant with IgE-mediated CMA tolerates maternal ingestion of CM without symptoms, the infant can continue to breastfeed without maternal elimination of CM. Although infants have been reported to develop IgE-mediated severe allergic reactions from human milk, this is rare. To our knowledge, there has been 1 case report in which an infant developed anaphylaxis in the setting of maternal ingestion of CM. An infant developed anaphylaxis on different occasions, including after ingesting maternal breast milk and after consuming a casein hydrolysate-containing formula. The infant’s symptoms resolved with maternal elimination of CM. (21)

As treatment for mild reactions, antihistamines may be used. For more severe reactions such as anaphylaxis, intramuscular epinephrine should be administered. Patients may require intravenous fluid for volume resuscitation or supplemental oxygen. Adjuvant therapies include H1 and H2 antihistamines, glucocorticoids, and bronchodilators, and in epinephrine refractory anaphylaxis, glucagon can be used. The US Food and Drug Administration has not approved any pharmaceutical therapies for CMA, though trials investigating oral and epicutaneous immunotherapy are ongoing. The natural history of IgE-mediated CMA is favorable. Studies show that 29% to 78% of children with IgE-mediated CMA will develop tolerance to CM by 6 years of age. (22)

CM is a common trigger for various non–IgE-mediated food allergies (Table 2). Non–IgE-mediated food allergies include FPIAP, FPIES, FPE, and Heiner syndrome (pulmonary hemosiderosis).

The terms “cow’s milk intolerance” or “cow’s milk protein allergy” in infants often refer to FPIAP. CM is the most common trigger for FPIAP, which typically manifests with blood-streaked stools, mucous, and/or foam in stools. (23)(24) Significant anemia and failure to thrive are not typically present, though a mild anemia may occur with chronic bleeding. Generally, term infants with FPIAP can be well-appearing. The differential diagnosis for rectal bleeding is dependent on the extent of rectal bleeding. Severe rectal bleeding may be indicative of intussusception, volvulus, sepsis, necrotizing enterocolitis (NEC), coagulopathy, or possibly FPIES. Mild to moderate rectal bleeding may be seen in dermatitis, anal fissure, or infections.

Symptoms of FPIAP typically present in the first few months of age; breastfed infants are usually older at initial presentation compared with formula-fed infants. It is possible for symptoms to begin as early as the first week after birth, which suggests that in utero sensitization may occur. (25) At this time, the mechanism of in utero sensitization is not completely understood. Most term infants present with blood-streaked stools at 2 to 8 weeks of age. (26)

The prevalence of FPIAP is not well known and has been estimated to range from 0.16% to 64% of healthy infants with rectal bleeding. (27)(28) Sigmoidoscopy is not typically indicated, but if performed, it usually shows findings. Lesions affect the distal bowel including the colon and are most prominent in the rectosigmoid segment. (29) A range of findings may be present, including a friable mucosa, increased nodularity, and in more severe instances, mucosal ulceration and decreased vascularity. (29) The most prominent histologic finding is eosinophilic infiltrates throughout the mucosa, especially in the lamina propria. Often neutrophil- and eosinophil-containing abscesses are seen in the crypts as well as eosinophilic infiltration in the muscularis mucosa. (30) Peripheral eosinophilia is not associated with the eosinophilic infiltrate in the rectosigmoid, (31) but may be present, especially in preterm infants with FPIAP. (32)(33)

The exact mechanism of FPIAP is not well understood. The immune response causes inflammatory cells such as eosinophils to infiltrate the mucosa, leading to damaged epithelial cells, which causes rectal bleeding and sometimes diarrhea.

Anecdotally, fecal occult blood tests are often performed in infants with colic or suspicion of FPIAP. Because of the high rate of false positives in healthy infants, this practice is not recommended and may lead to unnecessary dietary restrictions in both the infant and breastfeeding mother. (34) No confirmatory laboratory test has been established for non–IgE-mediated food allergies, including FPIAP. The gold standard of diagnosis is resolution of symptoms with elimination of the offending food protein for 2 to 4 weeks and recurrence of symptoms with reintroduction of the offending food. Symptoms typically improve over the course of 48 to 96 hours. (33)(35) Reintroduction is a key step in confirming this diagnosis to avoid overdiagnosis leading to unnecessary maternal dietary eliminations and/or costly specialized formula. The Figure illustrates a suggested pathway for management of suspected CMA, including FPIAP.

The main treatment for FPIAP is avoidance of the culprit food protein. In breastfed infants, mothers should eliminate CM from their diet (including products containing baked or processed milk and all CM products) and be counseled on label reading, as well as calcium and vitamin D supplementation. Table 3 highlights sources of CM in the diet that should be avoided when following a CM elimination diet. CM-based breast milk fortifiers should be avoided. In infants who continue to be symptomatic despite strict maternal CM elimination diet, removal of soy and egg from the mother’s diet should be considered (rarely other foods). If there is no improvement or the infant is already formula fed, transition to a hypoallergenic formula is the next step. In infants who are fed CM-based formula and/or CM-containing foods, discontinuation of CM in the infant’s diet may be enough for symptom resolution, and removing CM from the maternal diet is not necessary if symptoms disappear. The prognosis of FPIAP is favorable, and most infants are able to tolerate CM introduction by 12 months of age.

Data on the role of probiotics in treating FPIAP are variable. Baldassarre et al found that adding lactobacillus GG in extensively hydrolyzed formula resulted in indirect evidence of enhanced resolution of colonic inflammation, as indicated by a greater decrease in fecal calprotectin in the group receiving the probiotic compared with that extensively receiving hydrolyzed formula alone. (36) Conversely, a randomized controlled trial of breastfed infants with allergic proctocolitis failed to show the benefit of adding lactobacillus GG to CM elimination diet versus placebo. (37)

FPIES caused by CM typically presents between 3 and 6 months of age and is most commonly incited by CM and soy. FPIES is characterized by exposure to a food antigen followed by repetitive emesis within 1 to 4 hours. Up to 15% of reactions will include hypotension. This may be accompanied by lethargy, pallor, dehydration, or diarrhea. (35) Acute FPIES reactions are managed with supportive care, including intravenous fluid, if appropriate, and ondansetron (if >6 months old). (38) Continued exposure to the culprit food, including infant formula or CM solids in the case of CM FPIES, can lead to chronic diarrhea, failure to thrive, and intermittent-to-frequent emesis as seen in chronic forms of FPIES. Infants may appear to have sepsis and occasionally may have bloody diarrhea, which can appear similar to neonatal NEC. (39)

FPIES evaluation includes a medical history, and when appropriate, an oral food challenge, to establish the diagnosis. If the history includes a severe reaction with hypotension or is indicative of several reactions consistently with 1 food, a history alone may be sufficient to make the diagnosis. In addition, the symptoms should resolve with elimination of the offending food. (23) Given that this is a non–IgE-mediated process, serum-specific IgE and skin prick testing is not indicated. Laboratory studies after an FPIES episode show leukocytosis with neutrophilia and thrombocytosis. (23)(31) The diarrhea may contain occult blood, and a stool smear may show eosinophils and leukocytes. (40)

FPIES is considered to be the result of enhanced activation of food antigen–specific T cells. (38)(41) Activated T cells secrete cytokines that damage the gut barrier, increasing exposure to the antigen. The increased intestinal permeability then leads to more antigen uptake in the gut mucosa and creation of more antigen-specific T cells. (42) More recent studies have further investigated a role for innate immune activation, cellular and antibody responses in FPIES. (43)(44) FPIES is rare in breastfed infants, (38) which is considered to be because of the small amount of CM protein in breast milk or other protective factors in breast milk. The prognosis of FPIES is favorable because the vast majority of children outgrow FPIES by school age. (45)(46)

In contrast to FPIAP, and similar to FPIES, FPE is rarely symptomatic in exclusively breastfed infants. (47) FPE is a less common form of non–IgE-mediated food allergy caused by CM and typically presents at 1 to 2 months of age within weeks of CM formula introduction. (35)(48) The prevalence of FPE induced by CM is not clear, and currently it appears to be relatively rare, which may be because of the common use of hypoallergenic infant formulas. Risk factors for FPE include infective enteritis, immaturity, and history of abdominal surgery. (48) FPE is a malabsorptive syndrome characterized by vomiting, chronic diarrhea, failure to thrive, and steatorrhea. (49) FPE is more insidious than the acute reaction that occurs in FPIES. Intestinal mucosal changes include severe mucosal damage that is similarly seen in celiac disease, including villous atrophy. (49)(50) It has been shown that intestinal permeability is increased in patients with untreated FPE. The underlying mechanism for FPE is thought to be a T cell–mediated immune reaction causing jejunal damage after stimulation from CM antigens. (50) The best mode of prevention of FPE is exclusive breastfeeding. (48) The prognosis is favorable, because FPE will resolve in most children by 1 to 2 years of age. (35)

No specific test is used in the diagnosis of FPIAP, FPE, and FPIES. A careful history and physical examination are vital in establishing the diagnosis. With regard to FPIAP, the diet should be closely examined, including the type of infant formula consumed and maternal diet if the infant is breastfed. This includes maternal intake of CM and CM-containing products (including baked milk), soy, and eggs. FPE and FPIES rarely occur as a result of food proteins in the maternal diet; therefore, a maternal elimination diet is rarely necessary. Skin prick tests and serum-specific IgE antibody levels are often negative, and in the absence of other allergic comorbidities, are not recommended in the initial evaluation of non–IgE-mediated CMA. (35) The diagnostic criteria include resolution of symptoms with removal of the offending food antigen and recurrence of the symptoms upon reintroduction of the food protein during an oral food challenge. However, it is not always necessary to do the oral food challenge to confirm the diagnosis if there is a convincing history and/or the reaction was severe. (23) Infants with FPIES should be offered an observed oral food challenge 12 to 18 months from the most recent reaction. Because of the risk of developing IgE-mediated food allergy during the period of food avoidance, skin prick testing or serum-specific IgE is often performed before the oral food challenge. (38) In infants with FPIAP or FPE, in the absence of symptoms suggestive of IgE-mediated reactions or other allergic comorbidities (such as severe atopic dermatitis), CM can gradually be reintroduced at home after 12 months of age.

Non–IgE-mediated CMA can rarely present with pulmonary manifestations. Heiner syndrome is a rare disorder in which food hypersensitivity causes recurrent pulmonary infiltrates, failure to thrive, hemosiderosis, and anemia. (23)(51) Nakamura et al described a case of a preterm male infant with a family history of atopy who recurrently developed wheezing and apnea episodes after exposure to human milk fortifier and associated eosinophilia with negative infectious evaluation. His symptoms improved with discontinuation of the milk fortifier, and he tolerated breast milk alone. By 18 months of age, he was able to tolerate CM products. (52) Tekgündüz et al describe a case of a preterm infant with bronchopulmonary dysplasia who developed hematochezia and worsening respiratory distress with pulmonary infiltrates seen on radiography. (53) Symptoms improved with transition to a completely hydrolyzed formula. He returned to baseline and 2 weeks later restarted typical formula. Within 1 week, he again developed hematochezia and an increased oxygen requirement. His diagnosis of CMA was confirmed, and hydrolyzed formula was reinitiated. By 5 months of age, he no longer required supplemental oxygen or had hematochezia. (53)

Atopic dermatitis is a common manifestation of CMA, especially in infants. (4)(9) One-third of children with refractory moderate to severe atopic dermatitis can be attributed to food allergy. (54) Earlier onset and increased severity of atopic dermatitis increases the risk for food allergy. Before an evaluation for food allergy, skin care should be optimized. (55) Atopic dermatitis caused by a food allergy can be difficult to distinguish given the delayed onset of symptoms and often the lack of other IgE-mediated symptoms. Serum-specific IgE and skin prick tests can be helpful in identifying contributing foods. However, although these results may be positive, indicating sensitization, a positive result does not always correlate with clinical food allergy. To mitigate the risk of nutritional deficiencies with extensive elimination diets, it is imperative that a period of reintroduction occurs after foods are eliminated from an infant or maternal diet to confirm whether the symptoms were conclusively caused by the food.

EoE is an antigen-driven inflammatory condition characterized by eosinophilic infiltrate of the esophagus. (56) The inflammation of EoE is a separate entity from gastroesophageal reflux and likely will not respond to proton pump inhibitors alone. (57) Symptoms of EoE include vomiting, food refusal, failure to thrive, and in older children or adults, include food impaction and dysphagia. (56)(58)(59) Infants with reflux, feeding difficulties, prematurity, eczema, or diagnosis of milk protein allergy are at increased risk for EoE. (60) The diagnosis is made based on symptoms of esophageal dysfunction, when significant eosinophilia is seen on esophageal biopsy, and after excluding other causes such as proton pump inhibitor–responsive esophageal eosinophilia. (61) Endoscopic findings also include linear furrowing, concentric rings, white exudates, esophageal strictures, and superficial tears. (59) Given the invasive nature of endoscopy, it is not typically performed in infants who can be medically treated; thus the incidence of EoE in infancy is not well characterized because esophageal eosinophilia is a required diagnostic criterion. Food allergy can play a role in EoE, because some patients are responsive to elimination of certain foods, particularly CM. (62)(63)

Reported cases of CMA in preterm infants are non–IgE-mediated, and the symptoms are broad. The prevalence of CMA in preterm infants has not been well-established. In a study of 2,116 neonates in a NICU, Morita et al found that the incidence of CMA with systemic symptoms was 1.1%. (64) For infants with CMA, the mean time to onset was 23 days in preterm infants and 3.5 days in full-term infants. The delay of onset in preterm infants may be because of the time it takes for the infants to reach a postmenstrual age of 32 weeks when the immune mechanisms have matured enough to produce an adverse immune response. (64) Contrary to rectal bleeding in an otherwise healthy full-term infant with FPIAP or singular episodes of FPIES reactions in older infants and young children, preterm infants may present with more acute illness that may mimic NEC, sepsis, or shock, or may have nonspecific symptoms, including abdominal distention and fever. (64)(65)

Many case reports have described preterm infants with symptoms mimicking NEC who were then diagnosed with FPIAP. (33)(39)(65)(66) Coviello et al (33) described preterm twins with different presentations of CMA. One infant developed vomiting, abdominal distention, bloody stools, and lethargy with a normal abdominal radiograph and complete blood cell count. She improved after pausing enteral feeding; however, symptoms recurred several days after restarting enteral feeding. She developed eosinophilia and a positive radioallergosorbent test (RAST) result to BLG and α-lactoglobulin. Her symptoms resolved with transition to amino acid–based formula. Her twin presented with a positive result on fecal occult blood test, dilated bowel loops, and abdominal pain that resolved after stopping enteral feedings. Seven days after restarting enteral feedings, she developed lethargy, bradycardia, abdominal distention, positive result on fecal occult blood test, and pneumatosis of bowel wall seen on an abdominal radiography. She was medically treated for stage 2A NEC. She had eosinophilia and a positive RAST result to BLG; therefore, she was transitioned to an amino acid–based formula, which led to resolution of her symptoms. Both these infants had been introduced to milk fortifiers in the days leading up to their symptoms. (33) Srinivasan et al highlighted a case of a 26-week preterm infant with 3 episodes of allergic enterocolitis that closely resembled NEC. (65) The infant had pneumatosis each time, and the episodes correlated to eosinophilia. The symptoms fully resolved with elimination of CM antigen exposure. (65) Powell described 2 preterm infants who developed milk and soy enterocolitis. Their clinical presentations were similar to that of NEC with severe vomiting, a septic appearance, bloody stools, and abdominal distention. They improved after elimination of milk and soy. (39) In a retrospective study of 348 infants in a NICU who required parental nutrition, 5% were ultimately diagnosed with CMA as manifested by feeding intolerance or “late-onset or recurring NEC-like illness.” (67) These cases highlight the importance of considering CMA when caring for preterm infants with NEC-like illness or feeding intolerance. Chuang et al demonstrated an upregulation of proinflammatory cytokines after peripheral blood mononuclear cells from infants with NEC were stimulated by CM proteins (BLG and casein), whereas the controls exhibited no or negligible response. These findings provide evidence of CM protein sensitization and suggest the importance of CMA in NEC. (68)

Preterm and very-low-birthweight infants have additional caloric needs beyond human milk. These needs are met through the use of human milk fortifiers. Traditionally, human milk fortifiers have been CM-based; however, a commercial human milk–based human milk fortifier is now available. In a group of preterm infants with challenge-proven gastrointestinal intolerance to bovine-based milk fortifier, the resolution of symptoms was demonstrated with transition to human milk–based human milk fortifier. (69) Sullivan et al (70) found lower rates of NEC in extremely preterm infants who were fed an exclusive human milk–based diet (including human milk–based fortifier) compared with infants fed bovine-based milk fortifier. (70)

Preterm and very-low-birthweight infants have a greater risk of developing NEC when given formula compared with donor breast milk when maternal breast milk is not available. (71) There has been controversy regarding the use of CM formula versus hydrolyzed formula in NICUs. Evidence regarding the effect of hydrolyzed formula in place of standard CM formula when human milk is unavailable on NEC and feeding intolerance is of low quality. (72) This is an area in need of further study.

CMA can present with a wide spectrum of clinical manifestations ranging from rectal bleeding in an otherwise well term infant to FPIAP or urticaria in the setting of IgE-mediated CMA. The diagnostic tools for CMA differ depending on whether the allergy is IgE- or non–IgE-mediated. IgE-mediated CMA can be diagnosed with clinical history and evidence of sensitization (skin prick test and/or serum IgE). Diagnosis of non–IgE-mediated CMA is made based on clinical history, improvement of symptoms with removal of the offending food antigen, and recurrence of symptoms with a trial of reintroduction of the food antigen (if clinically possible).

Infants with CMA should avoid direct consumption of CM and CM-based formula. Mothers of symptomatic infants who are exclusively breastfed should pursue a CM elimination diet only if symptoms are exacerbated by maternal ingestion of CM. Occasionally, maternal elimination of CM may be required as the first step to remove CM from the infants’ and mothers’ diet, especially if symptoms are severe, to achieve quick symptom resolution; however, reintroduction of CM in a mother’s diet may subsequently be tolerated and should be attempted. A period of reintroduction of CM into the maternal diet is, however, critical for confirming the diagnosis of non–IgE-mediated CMA in an exclusively breastfed infant who is not critically ill and likewise important to see whether maternal elimination is required even if direct introduction of CM to the infant is not tolerated. Infants with CMA who are formula fed should be transitioned to an extensively hydrolyzed formula, and if no response is seen, then an amino acid formula should be started. The prognosis of CMA in infants is favorable. Most infants with non–IgE-mediated CMA are able to tolerate CM by 12 months of age, and children with IgE-mediated CM often reach tolerance by early school age. (22)(73)

BLG

β-lactoglobulin

CI

confidence interval

CMA

cow’s milk allergy

EoE

Eosinophilic esophagitis

FPE

food protein–induced enteropathy

FPIAP

food protein–induced allergic proctocolitis

FPIES

food protein–induced enterocolitis syndrome

IgE

immunoglobulin E

NEC

necrotizing enterocolitis

OR

odds ratio

RAST

radioallergosorbent test

RR

risk ratio