In this article, the epidemiology of eosinophilic esophagitis (EoE) is reviewed. Demographic features and natural history are described, the prevalence and incidence of EoE are highlighted, and risk factors for EoE are discussed. EoE can occur at any age, there is a male predominance, it is more common in whites, and there is a strong association with atopic diseases. EoE is chronic, relapses are frequent, and persistent inflammation increases the risk of fibrostenotic complications. The prevalence is currently estimated at 0.5–1 in 1000, and EoE is now the most common cause of food impaction. The incidence of EoE is approximately 1/10,000 new cases per year, and the increase in incidence is outpacing increases in recognition and endoscopy volume, but the reasons for this evolving epidemiology are not yet fully delineated.
Key points
- •
Eosinophilic esophagitis (EoE) affects patients of all ages, is more commonly seen in males, and whites, and is strongly associated with atopy.
- •
EoE is chronic, relapses are frequent after treatment is stopped, and diagnostic delay with persistent inflammation increases the risk of esophageal strictures and fibrostenotic complications.
- •
The prevalence of EoE is 0.5 to 1 cases/1000 persons.
- •
The incidence of EoE is approximately 10 cases/10,000 persons per year.
- •
Elucidating the reasons for the rapid increase in incidence and prevalence of EoE is an active area of research, and possibilities include changes in food allergens and aeroallergens, other environmental factors, the decrease of Helicobacter pylori , and early life exposures.
Introduction
Epidemiology is the study of patterns and causes of diseases in defined populations. It uses observational study designs, such as case control or cohort studies, to make inferences about causes and risk factors for diseases. These results generate hypotheses that can subsequently be tested in experimental study designs, such as clinical trials. Over the past 2 decades, great strides have been made in understanding the epidemiology of eosinophilic esophagitis (EoE). Case reports and case series provided initial descriptions of clinical characteristics of patients with EoE, and then, as larger cohorts were reported and natural history data accumulated, formal disease definitions and guidelines were put forth. Epidemiologic techniques have been used to estimate the incidence and prevalence of EoE at the single-center, regional, and national levels to provide an understanding of trends in the number of cases as well as the burden of disease attributable to EoE. The clear result of these studies is that EoE is rapidly increasing both in incidence and in prevalence. Investigations have also begun to determine potential causes that might explain this evolving epidemiology. In this article, the current knowledge base related to the epidemiology of EoE is reviewed. In particular, demographic features and natural history are described, data summarizing the prevalence and incidence of EoE throughout the world are highlighted, and risk factors for EoE are discussed.
Introduction
Epidemiology is the study of patterns and causes of diseases in defined populations. It uses observational study designs, such as case control or cohort studies, to make inferences about causes and risk factors for diseases. These results generate hypotheses that can subsequently be tested in experimental study designs, such as clinical trials. Over the past 2 decades, great strides have been made in understanding the epidemiology of eosinophilic esophagitis (EoE). Case reports and case series provided initial descriptions of clinical characteristics of patients with EoE, and then, as larger cohorts were reported and natural history data accumulated, formal disease definitions and guidelines were put forth. Epidemiologic techniques have been used to estimate the incidence and prevalence of EoE at the single-center, regional, and national levels to provide an understanding of trends in the number of cases as well as the burden of disease attributable to EoE. The clear result of these studies is that EoE is rapidly increasing both in incidence and in prevalence. Investigations have also begun to determine potential causes that might explain this evolving epidemiology. In this article, the current knowledge base related to the epidemiology of EoE is reviewed. In particular, demographic features and natural history are described, data summarizing the prevalence and incidence of EoE throughout the world are highlighted, and risk factors for EoE are discussed.
Demographic features
EoE has been reported throughout the life span, from infancy to almost 100 years of age. However, most cases are in children, adolescents, and adults younger than 50 years. There is a consistent gender discrepancy, with males affected 3 to 4 times more commonly than females, and EoE is also more frequently reported in whites compared with other races/ethnicities. The reason for the male predominance is not known, and as more data accrue from larger centers with more diverse patient populations, increasing numbers of racial minorities have been identified with EoE. Although most clinical, endoscopic, and histologic features are shared by patients with EoE of different races, some data suggest that African Americans with EoE may be diagnosed at earlier ages and are less likely to have typical endoscopic findings of EoE. EoE is also strongly associated with atopic disease and can run in families; both of these topics, as well as the clinical features and differences in the presentation between adults and children, are addressed in detail in other articles elsewhere in this issue.
Natural history
EoE is considered a chronic disease. Data from the placebo arms of randomized clinical trials and prospective and retrospective cohort studies show that EoE does not tend to spontaneously resolve or burn out. Specifically, the endoscopic signs and esophageal eosinophilia persist in the absence of treatment. Moreover, if treatment is stopped, symptoms, endoscopic signs, and esophageal eosinophilia recur in most patients over a period of several months. There are no published reports of EoE transforming into hypereosinophilic syndrome, extending to involve other areas of the gastrointestinal (GI) tract, or causing a malignancy.
Recent data have supported the concept that EoE may progress from an inflammatory-predominant phenotype (primarily seen in children) to a fibrosis-predominant one (seen in adults). These data help to explain the clinical differences between adults and children, and are consistent with basic science work showing that esophageal remodeling and deposition of fibrosis are key pathogenic features of EoE. The hypothesis is that in children, eosinophilic inflammation in the esophagus manifests with white plaques or exudates, decreased vascularity or mucosa edema, and linear furrows, but without esophageal rings or strictures, and causes symptoms such as pain, heartburn, and failure to thrive. However, with ongoing inflammation, subepithelial collagen is deposited, esophageal rings, narrowing, and strictures develop, and symptoms become dysphagia predominant.
A retrospective cohort study of patients in Switzerland explored this issue by assessing diagnostic delay (how long patients experienced symptoms before the EoE diagnosis and before any treatment) as a proxy for ongoing inflammation. There was a strong association between increasing diagnostic delay and the prevalence of strictures at diagnosis. For example, only 17% of those with less than 2 years of symptoms before diagnosis had strictures compared with 71% with more than 20 years of symptoms. These investigators calculated that for every increased decade of untreated EoE, there was a doubling of odds of having an esophageal stricture. Another study assessed the same issue by examining endoscopically defined phenotypes of EoE and noted nearly identical results, with the odds of having a fibrostenotic EoE phenotype doubling with each increasing decade of age. These natural history data provide important information for patients, show long-term consequences of EoE, and support the need for treatment of this condition.
Prevalence
The prevalence of a condition is defined by how many total cases exist during a given time frame in a specified location and is a useful measure of the burden of that disease. EoE has been described in many places throughout the world, including North America, Europe, South America, Australia, Asia, and the Middle East, but the prevalence seems to be highest in the United States, Western Europe, and Australia compared with Japan or China. There are no reported EoE cohorts in sub-Saharan Africa or India.
The prevalence of EoE depends on the population that is studied, the definition of EoE that was used (and whether the study was conducted before the recognition of proton pump inhibitor [PPI]-responsive esophageal eosinophilia [REE]), and the study methodology (prospective vs retrospective), and this contributes to variation in the range of prevalence estimates ( Tables 1 and 2 ). Although most prevalence estimates are derived from single centers with defined catchment areas, there are some studies that have used either population-based techniques or national databases to attempt to generate more accurate or generalizable estimates. Because EoE is chronic and nonfatal, studies universally report an increasing prevalence of EoE, regardless of the geographic location.
| Author | Location | Population | Time Frame | Prevalence (per 100,000) |
|---|---|---|---|---|
| Noel | Hamilton County, OH | Pediatric | 2000–2003 | 43.0 |
| Buckmeier | 2000–2006 | 90.7 | ||
| Cherian | Perth, Australia | Pediatric | 1995, 1999, 2004 | 89.0 |
| Straumann | Olten County, Switzerland | Adult | 1989–2004 | 23.0 |
| Hruz | 1989–2009 | 42.8 | ||
| Ronkainen | Northern Sweden | Adult | 1998–2001 | 400 |
| Prasad | Olmstead County, MN | Adult and pediatric | 1976–2005 | 55.0 |
| Gill | Huntington region, WV | Pediatric | 1995–2004 | 73.0 |
| Dalby | Southern Denmark | Pediatric | 2005–2007 | 2.3 |
| Spergel | United States | Adult and pediatric | 2010 | 52.2 |
| Ally | United States (military) | Adult and pediatric | 2009 | 9.7 |
| van Rhijn | Netherlands | Adult and pediatric | 1996–2010 | 4.1 |
| Syed | Calgary, Canada | Adult and pediatric | 2004–2008 | 33.7 |
| Arias | Castilla-La Mancha region, Spain | Adult and pediatric | 2005–2011 | 44.6 |
| Dellon | United States | Adult and pediatric | 2009–2011 | 56.7 |
| Author | Population | Time Frame | Prevalence (per 100) |
|---|---|---|---|
| Patients undergoing endoscopy for any reason | |||
| Veerappan | Adults | 2007 | 6.5 |
| Joo | Adults | 2009 | 6.6 |
| Sealock | Adults (VA population) | n/a | 2.4 |
| Patients undergoing endoscopy for dysphagia | |||
| Prasad | Adults | 2005–2006 | 15 |
| Mackenzie | Adults | 2005–2007 | 12 |
| Ricker | Adults (nonobstructive dysphagia only) | 2007–2009 | 22 |
| Dellon | Adults | 2009–2011 | 23 |
| Patients undergoing endoscopy for food bolus impaction | |||
| Desai | Adults | 2000–2003 | 55 |
| Kerlin | Adults | n/a | 50 |
| Sperry | Adults and pediatric | 2002–2009 | 46 |
| Hurtado | Pediatric | 2005–2009 | 63 |
| Patients undergoing endoscopy for refractory reflux | |||
| Liacouras | Pediatric | 1993–1995 | 3 |
| Rodrigro | Adult | 2002–2005 | 0.2 |
| Veerappan | Adult | 2007 | 8 |
| Sa | Adult | 2006–2008 | 1 |
| Poh | Adult | n/a | 1 |
| Foroutan | Adult | 2006 | 8 |
| Garcia-Campean | Adult | 2007–2009 | 4 |
| Dellon | Adult | 2009–2011 | 2 |
| Patients undergoing endoscopy for noncardiac chest pain | |||
| Achem | Adult | 2006–2007 | 6 |
| Patients undergoing endoscopy for abdominal pain | |||
| Thakkar | Pediatric | 2007–2010 | 4 |
| Patients with refractory aerodigestive symptoms | |||
| Hill | Pediatric | 2003–2013 | 4 |
Prevalence of EoE in the General Population
How common is EoE? In reviewing data assessing the prevalence of EoE in general populations (see Table 1 ), a reasonable answer is 0.5 to 1 cases/1000 persons. Most estimates in the United States range from 40 to 90 cases/100,000 persons. The largest study in the United States examined administrative health claims data and found the prevalence to be 57/100,000, or approximately 152,000 cases, with a range from 39 to 153/100,000 (106,000–411,000 cases), based on the case definition that was used. These findings are consistent with the most recent estimates from Australia (89/100,000), Switzerland (43/100,000), Spain (45/100,000), and Canada (34/100,000). In addition, these prevalence estimates are of the same order of magnitude of pediatric Crohn disease and ulcerative colitis, and are beginning to approach the prevalence of inflammatory bowel disease overall, a remarkable observation, given that EoE was essentially unknown 2 decades ago.
However, several studies have provided estimates that are outside this range. In Northern Sweden, data from a population-based study assessing the prevalence of Barrett esophagus were reanalyzed, and a prevalence of 400/100,000 was found for EoE. These patients were largely asymptomatic and, although they had esophageal eosinophilia, they may not have met current diagnostic criteria for EoE. A study in the US military found a prevalence of 10/100,000, but this might have been an underestimate because of the specialized population assessed or because of low sensitivity of the administrative code for EoE. Studies from the Netherlands and the southern region of Denmark have also reported lower prevalences of EoE. It is unknown if these results are because of practice patterns in those countries or a true difference in the prevalence of disease.
Prevalence of EoE in Patients Undergoing Endoscopy
Although these previous studies assessed the overall prevalence of EoE, depending on the specific patient population seen in health care settings, the prevalence may be orders of magnitude higher (see Table 2 ). For example, a prospective study, which collected esophageal biopsies on consecutive patients undergoing upper endoscopy for any indication, found that 6.5% of patients were diagnosed with EoE. Other studies with a similar design found comparable rates, from 2.4% to 6.6%. Given that approximately 6.9 million upper endoscopies are performed annually in the United States, endoscopists should expect to commonly encounter EoE in the procedure suite.
EoE is even more common in patients undergoing upper endoscopy for symptoms of dysphagia. In 3 prospective studies conducted in this focused patient population, the prevalence of EoE ranged from 12% to 22%. However, not every patient in these studies with a finding of esophageal eosinophilia would meet current diagnostic guidelines with exclusion of PPI-REE, because patients were enrolled before the recognition of this phenomenon. A recent study conducted in an esophageal referral center reported an EoE prevalence of 23% for patients undergoing endoscopy for dysphagia, after excluding those patients with PPI-REE. In that study, esophageal eosinophilia itself was seen in 40% of those undergoing upper endoscopy for dysphagia, before a PPI trial was conducted. Given that approximately 20% of all upper endoscopies performed are for an indication of dysphagia, EoE must be highly suspected in this population.
In the most extreme form of dysphagia (esophageal food bolus impaction requiring a visit to an emergency department and urgent endoscopy), EoE is now the most frequent condition identified. In this setting, between 46% and 63% of patients with food impaction have EoE. However, these data are limited because less than half of patients with food or foreign body impactions have esophageal biopsies to evaluate for esophageal eosinophilia, so EoE in this setting may still be underdiagnosed. If patients have a bolus impaction and no biopsies are obtained, it is reasonable to perform a follow-up procedure to identify the underlying cause.
The prevalence of EoE in patients undergoing endoscopy for other upper GI symptoms has also been described. For patients with heartburn-predominant symptom who do not respond clinically to PPI therapy (PPI-refractory reflux), EoE has been identified as the cause 1% to 8% of the time. In adults with noncardiac chest pain, 1 study found that EoE was the cause in 6%. In children undergoing upper endoscopy for abdominal pain, the frequency of EoE was 4%, and in children undergoing multispecialty evaluation for refractory aerodigestive symptoms, it was also 4%.
Although these epidemiologic data are useful for understanding how common EoE is, building a differential diagnosis for patients undergoing endoscopy, and measuring the burden of disease attributable to EoE, they also imply a practical point. Because EoE is common in patients undergoing endoscopic procedures, practitioners should strongly consider the diagnosis and obtain esophageal biopsies to evaluate for it. Currently, guidelines recommend obtaining biopsies for all patients presenting with dysphagia, regardless of the endoscopic appearance. There is also a recent analysis that suggests if the clinical probability of EoE is at least 8% to 10%, obtaining biopsies in patients with PPI-refractory reflux is also cost-effective.
Incidence
The incidence of a condition is how many new cases occur during a given time frame in a specified location, and is a measure of the number of people newly affected by a disease. The incidence of a condition can be low (a few new cases), but the prevalence can be relatively high if the condition is chronic and does not affect longevity. In addition, the incidence of a condition can be approximately equivalent to the prevalence for a condition that is either short-lived or highly morbid.
To study the incidence of a condition, it is necessary not only to detect all new (incident) cases but to ensure that existing (prevalent) cases are not falsely counted as new cases. In EoE, this goal has been accomplished primarily at referral centers in regions in which the patient population and catchment area are well defined, but there are some emerging regional and national data as well. Most of the reports estimate the incidence of EoE from 6 to 13 cases/100,000 persons ( Table 3 ). The 2 studies that have reported lower incidence, from the Netherlands and southern Denmark, were the same ones that reported lower prevalence.
| Author | Location | Population | Time Frame | Incidence (per 100,000) |
|---|---|---|---|---|
| Noel | Hamilton County, OH | Pediatric | 2003 | 12.8 |
| Hruz | Olten County, Switzerland | Adult | 2007–2009 | 7.4 |
| Prasad | Olmstead County, MN | Adult and pediatric | 2001–2005 | 9.5 |
| Dalby | Southern Denmark | Pediatric | 2005–2007 | 1.6 |
| van Rhijn | Netherlands | Adult and pediatric | 2010 | 1.3 |
| Syed | Calgary, Canada | Adult and pediatric | 2004–2008 | 11 |
| Arias | Castilla-La Mancha region, Spain | Adult and pediatric | 2005–2011 | 6.4 |
Increasing Incidence and Risk Factors
Studies are consistent in showing that the incidence of EoE has been increasing rapidly. In a report from Hamilton County, Ohio, the incidence of EoE was noted to increase from 9 to 12.8/100,000 over a 3-year period. In a report from Olmstead County, Minnesota, no EoE cases were seen before 1990, but the incidence increased from 0.35 to 9.5/100,000 over a 15-year period. Similar increases have been reported in Switzerland (1.2–7.4/100,000 over a 20-year period) and in the Netherlands (0.01–1.3/100,000 over 14 years). This increasing incidence has also been reflected in temporal trends in the relative prevalence of dysphagia causes, with EoE becoming a more frequent cause of dysphagia over time.
Why is EoE Increasing?
Although these marked changes in EoE incidence account for its increasing prevalence, they also beg the question of why the incidence is itself increasing. Although the answer is not known, there are several hypotheses. One explanation is simply that EoE is increasing because of increasing recognition; the condition was always there but given the research and clinical interest in the condition, practitioners are more aware of EoE, have a higher degree of suspicion when performing endoscopy, take more esophageal biopsies, and are therefore more likely to make the diagnosis. The exponential increase in the number of publications related to EoE is a testament to the increased interest in EoE (from 0 to 1 publication per year before 2000, to >200 publications in 2013). Some data that support this theory show that the increase in incidence rates relatively closely matches the increase in endoscopy volume or biopsy rates. However, there are other studies in which biopsy rates do increase 2 to 3 times over the study period, but the incidence of EoE outpaces that increase by several-fold, indicating that increased recognition is not the only explanation. In addition, in studies in which archived pathology samples have been retrieved and reanalyzed, cases of probable EoE identified in the early 1980s and 1990s occur at lower rates than are currently being observed.
There are other epidemiologic clues that provide insight. Recent data suggest that the prevalence of EoE steadily increases as age increases, peaks in the 35-year to 45-year age range, and then decreases ( Fig. 1 ). This is a counterintuitive observation for a chronic and nonfatal disease (in which prevalence would be expected to continue to increase as age increases) and suggests a possible cohort effect. It is interesting to speculate whether something might have changed in the environment 40 to 50 years ago that began to affect children born after that time, but did not affect older individuals. In contrast, genetic changes would not be expected to affect a new disease over a relatively short time frame. This time frame is also consistent with decades-long symptom duration before EoE diagnosis in some older patients with EoE.
