Allergic and Eosinophilic Gastrointestinal Disease

IgE-Mediated Allergy and Anaphylaxis

Type I (IgE-mediated) immediate hypersensitivity reactions to foods are most common in young children, with 50% of these reactions occurring in the first year of life. The majority are reactions to cow’s milk or to soy protein from infant formulas. Other food allergies begin to predominate in older children, including egg, fish, peanut, and wheat. Together with milk and soy, these account for more than 90% of food allergy in children.

Blinded food challenges have shown that symptoms referable to the gastrointestinal tract in IgE-mediated allergy typically begin within minutes of the ingestion, although occasionally may be delayed for up to 2 hours. They tend to be short-lived, lasting 1 to 2 hours. ^, Symptoms include nausea, vomiting, abdominal pain, and diarrhea; they may also include oral symptoms, skin manifestations, wheezing, or airway edema.

Anaphylactic reactions to food ingestion are life-threatening. For the purpose of this discussion, the term anaphylaxis is limited to those IgE-mediated reactions causing upper airway obstruction, hypotension, and circulatory collapse. Although any protein may be implicated, certain foods have a propensity to cause severe reactions in susceptible persons: cow’s milk, egg, peanut, and shellfish. The altering of food protein antigens through cooking or prior hydrolysis does not preclude a type I allergic reaction because some proteins are relatively resistant to denaturation.

Meanwhile, the oral allergy syndrome (OAS) is another IgE-mediated allergy, where the reaction is typically oral itching or tingling. Occasionally, facial or throat swelling may occur as well; it is unclear whether OAS causes other gastrointestinal symptoms such as pain or diarrhea. With IgE-mediated food allergy, the rapid onset of gastrointestinal (GI) symptoms after food ingestion correlates highly with positive Immunoglobulin E-Radioallergosorbent Test (IgE-RAST) or skin prick tests to the offending antigen, demonstrating that these reactions are related to typical type-1 hypersensitivity. On the other hand, in patients with OAS, symptoms relate to cross-reaction between similar epitopes on certain pollens and certain fruits and vegetables.

Patients with a history of a significant reaction to one or more foods should be tested by skin prick against those foods and against a limited battery of common food allergens (milk, soy, egg, peanut, fish, wheat). Skin testing has a sensitivity of 90% to 100%, depending on the antigen, so patients with negative skin testing are very unlikely to have IgE-mediated disease and should be challenged openly with the food in question. Non-IgE-mediated hypersensitivity may still cause symptoms on challenge, but these may be delayed for hours or days. IgE-RAST testing for specific foods does not have greater positive or negative predictive value than skin prick testing, and combining the two does not improve the diagnostic yield. The specificity of skin prick testing for food allergies ranges from 40% to 80%, which implies that the possibility of a false positive skin test is not inconsequential. Therefore, the use of skin testing in the evaluation of food allergy should be limited to cases where there is a high clinical suspicion of allergies as the etiology of symptoms, which increases the positive predictive value of the test.

If IgE-RAST or skin prick tests are positive, the food should be avoided for at least 3 weeks; if symptoms improve, the elimination diet should be continued. If there is no improvement, then open or single-blind challenges with the food can be given to try to elicit a response. These challenges should be performed in a setting in which access to emergency treatment of allergic reactions is available, and this is generally best handled by an allergist.

Positive challenges should lead to consultation with a dietitian to educate the patient and family concerning avoidance of the food and to ensure that adequate nutrition is maintained. Groups such as the Food Allergy and Anaphylaxis Network and the American Partnership for Eosinophilic Disorders provide support and educational materials for families. Patients with a history of serious reactions to foods should be provided with an epinephrine kit for home use, proper instruction on how the device is used, and a MedicAlert bracelet.

Food Protein–Induced Enterocolitis Syndrome

The food protein–induced enterocolitis syndrome, often abbreviated as FPIES, is a non-IgE mediated gut reaction to ingested food(s). FPIES is most commonly characterized by a severe reaction that includes profuse vomiting and diarrhea, with an onset between 1 and 6 hours after ingestion of an inciting food. The most common food antigens to cause FPIES are milk and soy, although the reaction has been seen with numerous foods. Rice may be the most common solid food, although once one is sensitized to one grain, it is common to react to various grains.

FPIES occurs most frequently in infancy, but somewhat later than the onset of typical milk-protein intolerance. The reaction with FPIES can be severe, often confused with an episode of gastroenteritis. Dehydration is the most common serious consequence of an FPIES reaction, and the patient may present with significant lethargy or shock, as well as bloody stool, anemia, and hypoalbuminemia. Transient methemoglobinemia has been reported as a result of an FPIES reaction as well. It may take several episodes before the trigger of the episodes is uncovered, or symptoms may become chronic if the inciting food is a common component of the diet. It can be difficult to isolate the exact food(s) responsible for the reaction.

Because FPIES is not IgE-mediated, traditional allergy testing (i.e., skin prick or RAST) is not useful to identify the trigger foods. Atopy patch testing (APT) has been reported to predict foods that cause FPIES, confirmed by oral food challenge.

Etiology

Eosinophilic esophagitis (EoE) is caused by an abnormal immunologic response to specific antigens. In the vast majority of cases, the antigens responsible are food antigens, although there appears to be a contribution from other environmental antigens in certain individuals. While several studies have documented resolution of EoE with the strict avoidance of food antigens, in 1995, Kelly published the seminal paper on EoE. Because the suspected etiology was an abnormal immunologic response to specific unidentifiable food antigens, each patient was treated with a strict elimination diet that included an amino acid–based formula. Patients were also allowed clear liquids, corn, and apples. Seventeen patients were initially offered a dietary elimination trial with 10 patients adhering to the protocol. The initial trial was determined by a history of anaphylaxis to specific foods and abnormal skin testing. These patients were subsequently placed on a strict diet consisting of an amino acid–based formula for a median of 17 weeks. Symptomatic improvement was seen within an average of 3 weeks after the introduction of the elemental diet (resolution in eight patients, improvement in two). In addition, all 10 patients demonstrated a significant improvement in esophageal eosinophilia. All patients reverted to previous symptoms upon the reintroduction of foods.

While an exact explanation for this type of response was not determined, Kelly suggested an immunologic basis, secondary to a delayed hypersensitivity or a cell-mediated hypersensitivity response, as the cause for EoE. More recently, Spergel demonstrated that foods that cause EoE do not do so through immediate hypersensitivity reactions. By using a combination of traditional skin prick testing and a lesser used technique of APT, he established that a delayed cellular-mediated allergic response may be responsible for many cases of EoE. Further supporting a delayed type response, CD8+ lymphocytes have been identified as the predominant T cell within the squamous epithelium of patients diagnosed with EoE.

A link between eosinophilic esophagitis and atopy has been established. ^, It is these links between atopy and EoE that originally suggested food allergies play a role in the pathogenesis of this disease. The role of food allergy was confirmed as patients improved on elemental diets. Elimination of the responsible food usually does not lead to immediate resolution of the symptoms. Rather, improvement of symptoms occurs approximately 1 to 2 weeks after the removal of the causative antigen. Also, in patients with EoE, symptoms do not always occur immediately after reintroduction to the foods. It may take several days for symptoms to develop suggesting either a mixed IgE and T-cell mediated allergic response or strictly a T-cell delayed mechanism in the pathogenesis of this disease. While both IgE and T-cell mediated reactions have been identified as possible causative factors, T-cell mediated reactions seem to be the predominant mechanism of disease.

Several authors have suggested that aeroallergens may contribute to the development of EoE. Mishra and Rothenberg used a mouse model to show that the inhalation of Aspergillus caused EoE. They found that the allergen-challenged mice developed elevated levels of esophageal eosinophils and features of epithelial cell hyperplasia that mimic EoE. In addition, Spergel reported a case of a 21-year-old female with asthma and allergic rhino-conjunctivitis who also had EoE. The patient’s EoE became prominent during pollen seasons, followed by resolution during winter months.

Familial clustering of cases of EoE has led to the assumption that there may be a genetic predisposition to the disease. One twin study showed 41% disease concordance in monozygotic twins compared with 22% in dizygotic twins and 2.4% in nontwin siblings of EoE patients. In recent years, several candidate genes have been identified as risk variants for the development of EoE. Among those are the genes that code for Eotaxin-3, thymic stromal lymphoprotein (TSLP), and filaggrin. A genome-wide association approach to identify EoE risk variants was undertaken in two independent EoE and control populations, revealing a single susceptibility locus in both cohorts that corresponded to locus 5q22.1, in which 11 single nucleotide polymorphisms (SNPs) resided within a single haplotype block spanning the TSLP gene. Interestingly, TSLP has been implicated in the development of atopic disease previously. ^, In addition, the EoE transcriptome, a set of esophagus-specific epithelial genes dysregulated in patients with EoE, is being investigated for disease pathogenesis. Most notably, decreased expression of the serine protease inhibitor Kazal-type 7 (SPINK7) has been demonstrated to unleash proteolytic activity and proinflammatory cytokines (including TSLP) in patients with EoE.

However, genetic factors alone do not account for the development of EoE. As mentioned previously, environmental exposure to allergens in the diet, and possibly aeroallergens, are a necessary factor for developing disease. It remains unclear why genetically susceptible individuals develop EoE in the presence of certain allergens. Atopic diseases such as asthma seem to be on the rise, and it is possible that EoE is just another manifestation of this phenomenon. Recent evidence suggests that early exposures may also influence risk of developing EoE. In a small cohort study, Jensen found that antibiotic exposure in infancy conferred a sixfold increased risk of developing EoE in childhood, while other factors such as Caesarian section, preterm birth, and formula exposure also had trends toward increased risk. The role of microbiota in epithelial barrier function and induction of EoE is also being explored.

Clinical Manifestations

Eosinophilic esophagitis can occur in all ages but was first extensively demonstrated in younger patients with a male to female ratio of about 3:1. However, with increased awareness of the disorder among internist-gastroenterologists, there has also been increased recognition of the disorder in adults. Estimates of prevalence are approximately 50 patients per 100,000 population, but it appears to be more prevalent in certain populations. Based on insurance claims, men 35 to 39 years of age in the United States have a prevalence of over 100 per 100,000. And, among patients with airway complaints such as cough, referred to an aerodigestive referral center staffed by otolaryngologists, 3.6% were found to have EoE.

Patients typically present with one or more of the following symptoms: vomiting, regurgitation, nausea, epigastric or chest pain, water brash, globus, and decreased appetite. ^, Less common symptoms include growth failure and hematemesis. Esophageal dysmotility and dysphagia are less common in younger children, but become increasingly prevalent in adolescents and adults. Symptoms can be frequent and severe in some patients while extremely intermittent and mild in others. Most patients may experience daily dysphagia or chronic nausea or regurgitation, while others may have infrequent or rare episodes of dysphagia. Up to 50% of patients manifest additional allergy-related symptoms such as asthma, eczema, or rhinitis. Furthermore, more than 50% of patients have one or more parents with a history of allergy ( Table 37.2 ).

Children with EoE have been studied in comparison to those with GER. ^, While the symptoms of vomiting and abdominal pain occurred similarly in both groups, dysphagia, diarrhea, and growth failure are more common in those with EoE. In addition, patients with EoE are more likely to have had allergic symptoms, another family member with an allergic history, and peripheral eosinophilia. Patients with EoE have higher numbers of eosinophils in their esophagus and also tend to have relatively normal pH-metry when compared with those with GER ( Table 37.3 ).

Evaluation and Diagnosis

Patients with chronic refractory symptoms of GER disease or dysphagia should undergo evaluation for EoE. While laboratory and radiologic assessment may be appropriate, the majority of these patients should undergo an upper endoscopy with biopsy. Historically, the diagnosis of EoE was often given when an isolated severe histologic esophagitis unresponsive to aggressive acid blockade, associated with symptoms similar to those seen in GER disease, was seen. The diagnosis is further supported if the patient responds both clinically and histologically to the elimination of a specific food(s). In the past, a 24 hour pH probe was used to demonstrate that the esophageal disease was not acid induced; however, more recent guidelines allow for diagnosis in the setting of appropriate clinical and histologic findings. According to consensus guidelines, the threshold of esophageal eosinophilia should be 15 or more eosinophils per high-power field (HPF) on esophageal biopsies.

While eosinophil density is the mainstay for assessing EoE status, there now exists a scoring system that takes other histologic findings into account. The EoE Histologic Scoring System (EoEHSS) examines eosinophil density in addition to basal zone hyperplasia, the presence of eosinophilic abscesses, eosinophil layering on the mucosal surface, the presence of dilated intercellular spaces, the presence of surface epithelial alteration, dyskeratotic epithelial cells, and lamina propria fibrosis and converts these findings into a score that demonstrates near perfect intra- and interrater reliability.

Very high levels of esophageal eosinophilia have been demonstrated with GER alone, emphasizing that failure of appropriate medical therapy is an important feature of the diagnosis. Adult studies performed after EoE became recognized in adults found that a large proportion had resolution of both symptoms and histologic findings after treatment with proton pump inhibitors, without dietary manipulation. This phenomenon became known as PPI-responsive esophageal eosinophilia (PPI-Ree) and led to some controversy as to whether EoE was the same in children and adults. In a prospective study of 173 adults undergoing upper endoscopy for dysphagia, 66 patients were found to have greater than 15 eosinophils per HPF on biopsy. Of this group, 40 had persistent eosinophilia and 24 had improvement consistent with PPI-Ree.

However, with further study, it has become apparent that PPI-Ree is likely just a subset of EoE that is responsive to proton pump inhibitors (PPIs). The first evidence that PPI-Ree eventually reverts to a more classical EoE phenotype was demonstrated in four pediatric patients with PPI-Ree who remained on PPI yet had recurrence of eosinophilis in the esophagus. With the recognition of responsiveness of esophageal eosinophilia to PPIs, a multidisciplinary conference was convened to better define the role of PPI in both the diagnosis and treatment of EoE. In 2018, the results of the A Working Group on PPI-REE (AGREE) conference were published, with refinement of previous consensus recommendations to state that EoE should only be diagnosed in the setting of esophageal dysfunction and eosinophilic density in esophageal biopsies of 15 per HPF regardless of whether a trial of PPI had been given before diagnosis. Treatment options to be considered after diagnosis include PPI as first line therapy.

In most pediatric patients, it is preferable to defer endoscopy until after a course of aggressive acid suppression with proton pump inhibitors. However, some have adjusted their thinking to perform endoscopy prior to treatment with PPI in order to definitively diagnose EoE that may potentially resolve on PPI. At this point, it is a matter of clinical judgment without a definitive consensus. It is the authors’ practice to treat typical reflux symptoms with PPI prior to endoscopy and only proceed to endoscopy if symptoms are refractory to treatment or relapse repeatedly upon discontinuation of therapy. In practice, if EoE is present, it will eventually present itself regardless of approach above. It should be noted that in adult patients, the predominant presenting feature of dysphagia is considered a warning symptom that warrants urgent endoscopy, which usually results in diagnosis before a trial of PPI. This likely accounts for the recognition of the PPI responsiveness of some EoE patients in the adult population first.

Once EoE is diagnosed, patients should be encouraged to seek an allergy consultation. Skin prick testing and serum RAST tests may provide some clues to possible food allergens. Unfortunately, these tests are most useful in determining IgE-based allergic disorders. Since EoE is considered to be either a T-cell mediated disease or a mixed IgE and T-cell mediated disorder, the sensitivity and specificity of skin prick tests alone are low. APT (the placement of an antigen on the skin for several days followed by assessment for localized skin reaction) may be useful in determining additional antigens responsible for causing esophageal eosinophilia and should be considered, if available. If no specific antigen(s) are found through allergy testing, a trial of an elimination diet, consisting of removal of the antigens that most commonly cause EoE, can be attempted. The most common foods identified as causing EoE in the United States are milk, soy, egg, wheat, nuts, fish, and shellfish. However, there may be regional differences in which foods trigger EoE in individual patients. If diet manipulation based on testing or empiric removal fails, an elemental diet utilizing an amino acid based formula should be considered. The assessment of success should be based on both the improvement of clinical symptoms and histologic improvement.

Once EoE has responded to diet elimination, foods should be reintroduced in a systematic manner. Because of the possibility of delayed reactions, it is advisable to wait several days to 1 week between each new food introduction. This time period is usually sufficient to see a recurrence of symptoms; if symptoms develop, the food should be discontinued. However, in some cases symptoms do not occur, despite recurrence of eosinophilic infiltration. A repeat endoscopy with biopsy is required in order to evaluate for the presence of esophageal mucosal injury. Because clinical symptoms often occur sporadically, biopsy remains the most important way to accurately determine the presence or resolution of EoE.

While upper endoscopy with biopsy can precisely determine the diagnosis, noninvasive diagnostic tests have proven to be less useful, yet remain an area of focus. These include the evaluation of serum IgE levels and quantitative peripheral eosinophils, radiographic upper gastrointestinal series (UGI), pH probe and manometry, RAST testing, and skin prick and patch testing. When used alone, serum IgE levels and serum eosinophils have been found to be unreliable, as these tests usually respond to environmental allergens as well as ingested or inhaled allergens. While radiographs determine anatomic abnormalities, these tests cannot identify tissue eosinophilia. Patients with EoE usually have normal or borderline pH-metry. These patients may have mild gastroesophageal reflux disease (GERD) secondary to abnormalities in esophageal motility due to tissue eosinophilic infiltration. A promising new modality is the esophageal string test, where a string is swallowed but then anchored at the mouth so it can be removed and analyzed for the presence of eosinophil-derived proteins. Initial reports show that results from this test correlate strongly with histologic findings of eosinophilia. An additional noninvasive method to obtain esophageal cells is the Cytosponge, which is swallowed while encapsulated but expands in the stomach and then is withdrawn by a string across the esophageal mucosa, providing ample material to assess for EoE-specific cytokines. Another modality being explored for its ability to be used without sedation is the unsedated transnasal endoscopy (TNE). In a prospective study with 21 EoE children, there were no serious adverse events and a high degree of satisfaction over esophagogastroduodenoscopy (EGD) among both parents (85.7%) and patients (52.4%), while maintaining adequate biopsy specimen size.

EoE should be considered only when the eosinophilia is isolated strictly to the esophageal mucosa. In the past, early reports suggested that eosinophilia in EoE was more characteristic in the mid-esophagus than the distal (which is more likely to have eosinophils relating to acid reflux); however, it has now been demonstrated that a severe mucosal eosinophilia can occur in either the distal or proximal esophagus. To make an accurate diagnosis, the remainder of the gastrointestinal tract must be normal. When EoE is suspected, the sensitivity for detecting the disease is increased when more biopsies are obtained from the esophagus. Sensitivity seems to be highest when at least five biopsies are obtained.

EoE has been associated with visual findings on endoscopy: concentric ring formation called “trachealization” or a “feline esophagus,” longitudinal linear furrows and patches of small, white papules on the esophageal surface. Most investigators believe that the esophageal rings and furrows are a response to full thickness esophageal tissue inflammation. The white papules appear to represent the formation of eosinophilic micro-abscesses ( Figs. 37.1 and 37.2 ).

In 1999, Ruchelli evaluated 102 patients presenting with GERD symptoms who also were found to have at least one esophageal eosinophil without any other gastrointestinal abnormalities. Patients were subsequently treated with aggressive acid blockade. It was demonstrated that the treatment response could be classified into three categories. Patients who improved had on average 1.1 eosinophils per HPF, patients who relapsed upon completion of therapy had 6.4 eosinophils per HPF, and patients who remained symptomatic had on average 24.5 eosinophils per HPF.

In 2000, Fox utilized high resolution probe endosonography in patients with EoE in order to determine the extent of tissue involvement. He compared eight patients identified with EoE to four control patients without esophagitis. He discovered that the layers of the esophageal wall were thicker in EoE patients than in the control group (2.8 to 2.2 mm). In addition, the mucosa to submucosa ratio was greater in EoE patients (1.6 to 1.1 mm), and the muscularis propria thickness was greater in EoE patients (1.3 to 1.0 mm). These findings suggested EoE patients had more than just surface involvement of eosinophils.