A practical approach to small bowel biopsy interpretation: Celiac disease and its mimics

Abstract

While celiac disease is the most common cause of villous remodeling and intraepithelial lymphocytosis in the proximal small bowel, there are many entities that can mimic its histologic appearance. The purpose of this review is to discuss normal small bowel histology and the differential diagnosis of celiac disease. Approaches to evaluate increased intraepithelial lymphocytes are presented, followed by a detailed discussion of the pathology of celiac disease. Particular emphasis is given to those conditions that cause intraepithelial lymphocytosis in the setting of preserved villous architecture, although other important entities, such as peptic injury, idiopathic inflammatory bowel disease, medication injury, eosinophilic (allergic) gastroenteritis, autoimmune enteropathy, common variable immunodeficiency, and infections are also reviewed.

Introduction

Celiac disease (i.e., gluten-sensitive enteropathy or celiac sprue) is the most common cause of small intestinal mucosal remodeling. It usually causes variable villous blunting, crypt hyperplasia, plasma cell-rich inflammation, and increased intraepithelial lymphocytes, although some cases show subtle abnormalities, such as increased intraepithelial lymphocytes in the presence of normal villous architecture. Unfortunately, the diagnosis of celiac disease is not straightforward in all cases, as many other conditions can produce similar histologic features (Table 1). The purpose of this review is to provide the reader with a general approach to small bowel biopsy evaluation and discuss entities that can cause villous blunting and/or increased intraepithelial lymphocytes.

The duodenum, jejunum, and ileum show slight histologic differences and are prone to artifactual distortion that poses diagnostic challenges. Villi at all sites are lined by absorptive cells with admixed goblet cells, although the ratio of goblet cells to absorptive cells increases in the distal bowel compared to the duodenum. Paneth cells and endocrine cells are confined to the deeper crypt region where one may also encounter occasional mitotic figures. The normal villus-to-crypt ratio in the duodenum is approximately 3–4:1 (Fig. 1A), but villi may be broader and shorter, overlying gastric heterotopias and Brunner glands in the duodenal bulb. In addition, duodenal bulb biopsies frequently show changes of peptic injury that simulate celiac disease, as described in subsequent sections, so biopsies to exclude celiac disease should always include samples distal to the duodenal bulb. Villi of the jejunum and ileum tend to be slightly taller (Fig. 1B) except in areas overlying lymphoid aggregates where they are often broad-based or flat. Tangentially sectioned villi and those that have been stripped of surface epithelium can also appear blunted and, thus, these areas should be avoided when assessing villous architecture. An adequate tissue sample consists of at least three or four consecutive, intact villi that are well-oriented in the plane of section.

The small bowel lamina propria normally houses abundant gut-associated lymphoid tissue comprising lymphocytes, plasma cells, and macrophages, as well as scattered eosinophils and neutrophils. Granulocytes are considered normal components of the lamina propria, provided they are not present in crypt or surface epithelium, whereas scattered intraepithelial lymphocytes can be normally present in villous epithelium (Fig. 1C). Although the acceptable number of “normal” intraepithelial lymphocytes has been reported to be as high as 40/100 enterocytes, more recent analyses have defined the upper limit of normal to be 20–25/100 enterocytes in hematoxylin and eosin (H&E)-stained sections and slightly more (25–35/100 enterocytes) when using an anti-CD3 immunohistochemical stain.1, 2, 3 Use of immunohistochemistry to detect intraepithelial lymphocytes is not recommended, as these stains are not superior to evaluation of H&E-stained sections for diagnosing celiac disease and may, in fact, lead to overdiagnosis of gluten sensitivity.⁴ Intraepithelial lymphocytes are more numerous on the lateral aspects of villi and diminish in number toward the villous tips. However, diseases characterized by increased intraepithelial lymphocytes, particularly celiac disease, show more intraepithelial lymphocytes in the villous tips.5, 6 Thus, analyzing villous tips for intraepithelial lymphocytes is a convenient way to determine whether or not they are increased. A count of 6 or more intraepithelial lymphocytes per 20 enterocytes at the villous tip is considered abnormal.6, 7, 8 Formal counting of intraepithelial lymphocytes is not necessary in most cases, as any increase is usually readily apparent upon careful observation. Of note, increased numbers of intraepithelial lymphocytes are normally present in epithelium overlying and adjacent to lymphoid aggregates, so care must be taken to avoid over-interpreting this finding.

Celiac disease is a systemic disease with primarily gastrointestinal manifestations, particularly in the small intestine where it causes malabsorption owing to an abnormal immune response to dietary gluten. Gluten activates both the adaptive and the innate immune systems in susceptible patients.⁹ Gliadin, a degradation product of gluten, directly interacts with intestinal epithelium, resulting in upregulation of heat shock proteins and secretion of several cytokines, including interleukin-15, which promote recruitment of intraepithelial lymphocytes. Gliadin also enters the lamina propria where it is modified by tissue transglutaminase (tTG) and presented to CD4⁺ T-cells via human leukocyte antigen (HLA) complexes. This action results in T-cell activation and secretion of various cytokines that stimulate CD8⁺ intraepithelial lymphocytes to injure enterocytes. These T-cells also stimulate B-cells to produce autoantibodies to tissue transglutaminase, gliadin, and other proteins.⁹

Celiac disease is very common in Western populations and is now diagnosed four times more frequently than it was 60 years ago. It affects approximately 1% of individuals in North America and Europe, although recent estimates suggest that only 20% of patients with celiac disease are actually diagnosed with the disorder.10, 11 The disease usually presents in the third and fourth decades, but is increasingly detected in pediatric and older patients. The disorder is more common among women than among men and is heavily influenced by genetic factors. Patients with specific HLA complexes, namely HLA-DQ2 and HLA-DQ8 alleles, are highly susceptible to celiac disease.¹⁰

Celiac disease may cause diverse clinical symptoms and laboratory abnormalities. Adults typically present with steatorrhea, abdominal cramps, distension, diarrhea, and vomiting. Non-gastrointestinal symptoms such as iron deficiency anemia, osteoporosis, neurologic symptoms, infertility, and increased liver function tests are increasingly recognized as presenting manifestations of disease.10, 12, 13 Infants may suffer from failure to thrive, vomiting, anorexia, and diarrhea; whereas older children may have short stature, neurologic symptoms, and anemia.¹⁴ Dermatitis herpetiformis, autoimmune hepatitis, cystic fibrosis, type 1 diabetes mellitus, Sjögren syndrome, primary biliary cirrhosis, and many other immune-mediated disorders show associations with celiac disease and, thus, diagnosis of these entities should prompt exclusion of gluten sensitivity.¹⁰ Patients with Down syndrome are also particularly at risk for celiac disease.¹⁵

The diagnosis of celiac disease can be problematic because no single test shows 100% sensitivity and specificity in every patient.¹³ A presumptive diagnosis of celiac disease is usually made in the context of positive serologic studies and characteristic histologic changes present in duodenal biopsies, although one may argue that suggestive symptoms and markedly elevated celiac antibody serologies alone may be adequate to establish a diagnosis.¹⁴ Importantly, a patient must be on a gluten-containing diet for serologic and histologic test results to be meaningful. The two best serologic studies measure IgA anti-tissue transglutaminase (tTG) antibodies and IgA anti-endomysial antibodies. Of these, anti-tTG is the preferred screening test. It is easy to perform and shows both high sensitivity (94%) and specificity (97%) for celiac disease.¹⁶ However, this marker has several limitations. False-negative results may be encountered among patients with IgA deficiency, which is seen in approximately 10% of patients with celiac disease. These individuals require testing for the IgG isotype of anti-tTG. False-positive results may be encountered in patients with inflammatory bowel disease, primary biliary cirrhosis, heart disease, autoimmune enteropathy, and other immune-mediated disorders. In addition, serum anti-tTG antibody titers correlate with severity of histologic damage and their positive predictive value can be quite low in patients with minimal histologic abnormalities.17, 18 Newer tests that measure antibodies against deamidated gliadin peptides may also be helpful in some settings.¹⁹

Patients with a positive serologic test should undergo endoscopic examination with biopsies of the duodenum.¹³ The endoscopic findings are neither sensitive nor specific. The duodenum may appear entirely normal or display innumerable indentations on mucosal folds, which have been described as showing a scalloped appearance. While evaluating villous architecture in the duodenal bulb can be challenging, recent studies have shown that biopsies from the duodenal bulb improve diagnostic yield in both pediatric and adult patients with celiac disease due to the patchy nature of small bowel involvement.20, 21 For this reason, most authorities suggest obtaining biopsies from multiple sites within the duodenum including the duodenal bulb, proximal duodenum, and distal duodenum.²²

Classic features of celiac disease include total villous blunting, hyperplastic crypts, expansion of the lamina propria by plasma cell-rich inflammatory infiltrates, and increased intraepithelial lymphocytes (Fig. 2A and B). Indeed, biopsies with severe villous abnormalities bear a close resemblance to colonic mucosa. Notably, the absence of villi is accompanied by regeneration and hyperplasia of crypts and, thus, the findings are more appropriate classified as “villous remodeling” than “villous atrophy.” Surface enterocytes are damaged and may display loss of polarity with an attenuated brush border. Mucosal erosions and ulcerations are quite rare and, if present, may herald the onset of refractory sprue or lymphoma. Scattered neutrophils are occasionally encountered in crypts and some patients even have crypt or surface microabscesses. However, these findings are generally mild and focal in nature; more substantial neutrophilic infiltrates should raise the possibility of an alternative diagnosis.²³ The severity of histologic damage can be quantified using the Marsh classification scheme. This grading system, which was later modified by Oberhuber et al.,²⁵ assesses the degree of villous blunting, intraepithelial lymphocytes, and crypt hyperplasia (Table 2).²⁴ While the Marsh classification is an important research tool that can be used to determine responses to treatment, it is not routinely used in daily practice.

The diagnosis of celiac disease is straightforward when the clinical findings, serologic studies, and biopsy results are all in agreement. However, there are many instances in which one or more of these tests are negative. Some patients may have normal duodenal biopsies in the presence of positive serologies. These patients may be provisionally labeled with latent celiac disease, which may be managed with a combination of repeat biopsies, a trial of gluten withdrawal, or, more commonly, watchful waiting.10, 13 If, however, intraepithelial lymphocytes are encountered in the absence of other supportive evidence of celiac disease, one should consider alternative explanations for the histologic findings. A strong clinical suspicion in the absence of supportive histologic or laboratory findings may prompt evaluation for high-susceptibility alleles HLA-DQ2 or DQ8 or repeat testing prior to initiation of gluten withdrawal.¹⁰

The histologic abnormalities and clinical symptoms of celiac disease usually regress rapidly upon adherence to a gluten-free diet. Incomplete resolution, or recurrence, of symptoms despite at least one year of adherence to a gluten-free diet is classified as refractory celiac disease. Possible etiologies include non-compliance to gluten withdrawal, collagenous sprue, or concomitant lymphocytosis elsewhere in the gastrointestinal tract. Once these possibilities are excluded, refractory celiac disease is classified based upon the predominant phenotype of intraepithelial T-cells.26, 27 Type I refractory celiac disease is characterized by a normal T-cell phenotype (CD3⁺/CD8⁺), whereas type II refractory celiac disease displays loss of CD8 expression and clonality by molecular studies. Type II disease may progress to overt enteropathy-associated T-cell lymphoma. This lethal complication of celiac disease consists of an infiltrate of medium-to-large cells with round nuclei, prominent nucleoli, and moderate amounts of pale cytoplasm, which is accompanied by a reactive inflammatory cell infiltrate rich in macrophages and eosinophils.²⁸

Collagenous sprue was originally described by Weinstein et al.²⁹ in 1970. It shares many features with celiac disease, including severe villous architectural abnormalities, crypt hyperplasia, and intraepithelial lymphocytosis, but also shows an irregularly thickened layer of type 1 collagen subjacent to the surface epithelium (Fig. 3). Collagenous sprue shares many features with collagenous colitis. The lamina propria is expanded by a mixed inflammatory cell infiltrate and increased eosinophils and the subepithelial collagen layer contains entrapped capillaries and inflammatory cells. Most cases show some degree of surface epithelial cell damage, including cytoplasmic vacuolization, loss of columnar shape, and detachment from the basement membrane. Occasional neutrophils may be present, but are not a prominent feature.

The etiology of collagenous sprue is often unknown, although some cases are probably related to gluten sensitivity. Several reports describe a well-documented progression of celiac disease to collagenous sprue, whereas other de novo cases of collagenous sprue respond to a gluten-free diet.30, 31, 32 Some cases of collagenous sprue have been attributed to medication injury or immunodeficiency. The disorder is commonly associated with other immune-mediated diseases and may be seen in combination with collagenous gastritis and collagenous colitis, raising the possibility that it represents an abnormal immune response to dietary antigens including, but not limited to, gluten. First-line treatment consists of adherence to a gluten-free diet as well as initiation of immunomodulatory medications if symptoms do not improve. Unfortunately, histologic features of collagenous sprue often persist despite therapy and some patients develop serious complications, including severe malnutrition and T-cell lymphoma.³²

Increased intraepithelial lymphocytosis may be seen in nearly 3% of duodenal biopsies that show normal villous architecture. Although early reports indicated that biopsies displaying increased intraepithelial lymphocytes in the setting of preserved villous architecture may reflect celiac disease, more recent data suggest that this finding is fairly non-specific.6, 33, 34 Kakar et al.³⁵ assessed the clinical features of 43 patients with increased intraepithelial lymphocytes and normal duodenal villous architecture and found that only 9% proved to have celiac disease. Results of larger studies also confirm that only 20% of biopsies with increased intraepithelial lymphocytes and normal villous architecture reflect celiac disease, although it is possible that some patients will develop more characteristic clinical and histologic features of celiac disease following a gluten challenge.33, 36

The differential diagnosis of increased intraepithelial lymphocytes in the duodenum is broad. There is a strong association between Helicobacter pylori gastritis and increased duodenal intraepithelial lymphocytes, the latter of which resolves upon successful H. pylori eradication (Fig. 4).37, 38 Several immune-mediated diseases, including Hashimoto thyroiditis, Grave disease, rheumatoid arthritis, psoriasis, multiple sclerosis, systemic lupus erythematosis, Crohn disease, and ulcerative colitis may also cause intraepithelial lymphocytosis, although care must be taken to exclude celiac disease in these patients since they all may have concomitant gluten sensitivity.5, 33, 35, 39 Other etiologies include autism, lactose intolerance, peptic injury, medications, infection, obesity, and non-gluten food allergies.40, 41 Medications are increasingly recognized for their effects on small bowel inflammation. Kakar et al.³⁵ found that 14% of patients with increased intraepithelial lymphocytes and normal villous architecture were taking non-steroidal anti-inflammatory drugs (NSAIDs) compared to only 2.2% of those without this histologic finding. Some infections such as those by Giardia lamblia, Cryptosporidium, and Cystoisospora belli may also elicit intraepithelial lymphocytes as described in the manuscript by Lai et al. in this issue.

The duodenal bulb is susceptible to symptomatic or subclinical acid-induced injury due to exposure to acidic gastric contents. Peptic injury is characterized by variable villous blunting and intraepithelial lymphocytosis that mimic features of celiac disease. However, the lamina propria usually contains increased neutrophils and the surface epithelium displays gastric mucous cell (foveolar) metaplasia with, or without, associated neutrophils (Fig. 5). Brunner gland hyperplasia has been described as a feature of peptic duodenitis, although it may be seen in the normal duodenum as well. The features of peptic injury are quite non-specific and can be seen in the setting of medication injury, particularly NSAIDs, Crohn disease, and some infections.

Patients with established ulcerative colitis or Crohn disease may have duodenal biopsies that demonstrate increased intraepithelial lymphocytes. While there are conflicting reports regarding the association between idiopathic inflammatory disease and celiac disease, one may consider the possibility of coexistent gluten sensitivity in these individuals.42, 43 However, both ulcerative colitis and Crohn disease may have manifestations in the duodenum. Indeed, 10–22% of ulcerative colitis patients have features of duodenitis upon endoscopic examination and biopsy, including variable villous blunting, expansion of the lamina propria by plasma cell-rich inflammation, and active inflammation.39, 44, 45 Several recent reports also describe patients with well-documented ulcerative colitis who developed severe, diffuse inflammatory changes in the upper gastrointestinal tract characterized by shallow ulcers, marked villous shortening, and basal lymphoplasmacytosis.46, 47

Crohn disease is a well-known cause of small intestinal inflammation and villous distortion. Endoscopically apparent strictures, aphthous erosions, and linear ulcerations are commonly present in the ileum, biopsies of which show active inflammation, villous architectural distortion, basal lymphoplasmacytosis, and pseudopyloric gland metaplasia. Granulomata are helpful in confirming the diagnosis, but they are present in less than half of cases. Clinically apparent duodenal Crohn disease is less frequent, but a substantial number of patients without upper gastrointestinal tract symptoms have microscopic disease.⁴⁸ Biopsies show variable active inflammation, villous architectural distortion, basal lymphoplasmacytosis, foveolar metaplasia, and occasional granulomata (Fig. 6). In contrast to celiac disease, most cases of Crohn disease that elicit villous shortening are accompanied by neutrophil-rich inflammation, crypt abscesses, and edema. The primary differential diagnosis includes peptic or medication-related injury. Correlation with biopsy findings from other sites is also helpful, as isolated duodenal Crohn disease is extremely rare.⁴⁹

Medication-related gastrointestinal injury is increasingly common. Perhaps the most common is injury by NSAIDs, many of which are available without a prescription and cause a wide variety of colonic and small intestinal injuries. Mucosal injury may not elicit symptoms and manifest only as occult bleeding and/or anemia. The stomach and duodenum are most commonly affected, but distal small bowel and colonic injury are increasingly recognized. Endoscopic findings include diffuse erythema and well demarcated, punched out ulcers and erosions. Jejunal and ileal changes are similar to those of the duodenum, although diaphragm disease shows a predilection for the distal small bowel.⁵⁰ Diaphragm disease is caused by prolonged use of non-steroidal anti-inflammatory drugs and is characterized by numerous thin, web-like mucosal septa that cause luminal narrowing and obstruction. The villi in these areas are often blunted and pseudopyloric gland metaplasia is common. These agents can also be responsible for increased intraepithelial lymphocytes in small bowel biopsies either with, or without, villous architectural abnormalities.

Recent reports describe severe gastrointestinal injury related to the use of the angiotensin-II receptor antagonist olmesartan.51, 52 Rubio-Tapia et al.⁵² evaluated the findings in 22 patients with olmesartan-induced enteropathy and found that all had at least partial villous blunting and 15 showed complete villous shortening in duodenal biopsies (Fig. 7). Most (68%) cases showed some degree of neutrophilic inflammation, 64% displayed increased intraepithelial lymphocytes in the duodenum, and 32% had thickened subepithelial collagen that raised the possibility of collagenous sprue. Indeed, the clinical diagnosis for many of these patients was non-responsive celiac disease or unclassified sprue. Importantly, none of the patients had positive celiac serologies or responded to a gluten-free diet. Abnormal histologic findings were also seen in the stomach and colon. Discontinuation of olmesartan resulted in reversion of duodenal histology to normal in most cases.

Other medications that can cause villous abnormalities include colchicine, mycophenolate mofetil, ipilimumab (anti-CTLA4), and various chemotherapy agents. Colchicine causes mitotic arrest, producing an abundance of mitotic figures, including ring mitoses.⁵³ Mycophenolate mofetil, a medication commonly used in solid organ transplantation, induces villous abnormalities, inflammation, and increased epithelial cell apoptosis.54, 55, 56, 57, 58 Ipilimumab is a monoclonal antibody that is increasingly being used to treat advanced melanoma, renal cell, ovarian, and prostate cancer. It causes increased crypt epithelial cell apoptosis, active inflammation, and villous blunting resembling autoimmune enteropathy.⁵⁹ Celiac disease may rarely develop in patients receiving ipilimumab, raising the possibility that this medication can unmask latent celiac sprue.⁶⁰ Numerous chemotherapy medications and radioactive materials can also injure the small bowel. A complete discussion of medication-related gastrointestinal tract injury is provided in the article by Panarelli, which is included in this issue.

Eosinophil-rich inflammatory infiltrates may be encountered in small bowel biopsies obtained in a diverse range of settings, including medication injury, parasitic infection, Crohn disease, collagen vascular disease, hematologic malignancies, and idiopathic eosinophilic (allergic) gastroenteritis.⁶¹ Idiopathic eosinophilic gastroenteritis is a diagnosis of exclusion and, thus, should only be entertained after other potential causes of increased eosinophils are excluded. Mucosal involvement by idiopathic eosinophilic gastroenteritis may result in variable villous shortening with crypt hyperplasia and intraepithelial lymphocytosis that simulates celiac disease, although it is also characterized by a prominent eosinophil-rich inflammatory infiltrate, often in combination with other abnormalities, including degranulated eosinophils, intraepithelial eosinophils, eosinophilic crypt abscesses, epithelial degenerative changes, villous blunting, and deep eosinophils in the muscularis mucosa and submucosa.

Autoimmune enteropathy was first described by McCarthy et al.⁶² who reported a patient with severe enteropathy associated with total villous blunting and gut autoantibodies. A few years later, Walker-Smith et al.⁶³ coined the term “autoimmune enteropathy” to describe findings in a 15-month-old child with severe enteropathy and antibodies directed against gut epithelium. Powell et al.⁶⁴ later identified a genetic basis for the disease when they studied a family with eight boys afflicted with severe diarrhea, insulin-dependent diabetes, and other endocrinopathies. This pediatric X-linked syndrome of immune dysregulation polyendocrinopathy autoimmune enteropathy (IPEX) proved to result from mutations in FOXP3, a gene that controls regulatory T-cells.⁶⁵ Recent data suggest that autoimmune enteropathy can also occur in young girls without features of IPEX syndrome, as well as adults.65, 66, 67

Autoimmune enteropathy shows a predilection to affect the small bowel and colon. It induces variable, but often severe, villous shortening and crypt hyperplasia, which is accompanied by dense mononuclear cell rich inflammation of the lamina propria, similar to celiac disease (Fig. 8). These entities can usually be distinguished owing to the presence of neutrophilic inflammation and prominent crypt epithelial cell apoptosis typically present in cases of autoimmune enteropathy. Autoimmune enteropathy also shows decreased, or absent, specialized intestinal cells such as goblet, Paneth, and endocrine cells. It tends to display more severe intraepithelial lymphocytosis affecting crypts compared to the surface epithelium, whereas celiac disease shows similar degrees of lymphocytosis in crypt and surface epithelium. Importantly, some patients with autoimmune enteropathy have small bowel biopsies that show histologic features indistinguishable from those of celiac disease and these patients may even have anti-tTG antibodies in addition to anti-enterocyte antibodies.⁶⁷ However, patients with autoimmune enteropathy do not respond to a gluten-free diet.⁶⁸

Common variable immunodeficiency (CVID) is the second most common primary immunodeficiency with an estimated prevalence between 1:10,000 and 1:50,000.69, 70 The diagnostic criteria include exclusion of other causes of immunodeficiency, recurrent infections, decreased serum IgG levels at least two standard deviations below normal accompanied by decreased levels of least one other immunoglobulin subclass, and a failure to mount a response to vaccination.69, 70 The disease is equally prevalent among males and females and can occur at any age, although patients most commonly present between the first and third decades of life. Common variable immunodeficiency likely represents a heterogeneous group of disorders, the genetic basis of which is unknown in the majority of cases.

Most patients with common variable immunodeficiency suffer from recurrent respiratory tract infections as well as symptoms related to the gastrointestinal tract. The two sites most commonly affected are the small bowel and colon, followed in frequency by changes in the stomach and esophagus.71, 72, 73, 74 Nearly two-thirds of small bowel biopsies from patients with common variable immunodeficiency show intraepithelial lymphocytosis, which is often accompanied by villous architectural abnormalities that simulate the features of celiac disease.⁷⁴ Findings that help distinguish these two entities include prominent crypt apoptosis and a paucity, or absence, of plasma cells in the former (Fig. 9), whereas plasma cells are quite prominent in the latter. However, data from a recent study suggest that up to 30% of common variable immunodeficiency cases contain normal numbers of plasma cells, so this feature may prove to be of limited diagnostic utility.⁷⁴ Other findings typically present in biopsies from patients with common variable immunodeficiency include intraepithelial neutrophils, granulomata, and lymphoid aggregates that may raise concern for Crohn disease or autoimmune enteropathy. Patients with common variable immunodeficiency are prone to gastrointestinal infections, particularly giardiasis and cytomegalovirus.72, 74 Thus detection of either of these diseases should prompt the pathologist to exclude the possibility of common variable immunodeficiency.

Numerous infections, including tropical sprue, viral enteritis, bacterial overgrowth, Whipple disease, and Mycobacterium avium complex (MAC), can produce villous abnormalities either with or without increased intraepithelial lymphocytes. Tropical sprue affects residents of tropical areas, such as India, Southeast Asia, and the Caribbean. It likely results from an infectious agent, as epidemics and seasonal variations are well described. Small bowel aspirates from affected patients grow a variety of bacteria, although no one species is implicated in the pathogenesis of this disorder.⁷⁵ Symptoms and histologic features are indistinguishable from those of celiac disease, but treatment consists of antibiotic administration as well as supplementation with folate and vitamin B12. Of note, the ileum tends to be affected to the same degree as the duodenum in patients with tropical sprue.

Viral enteritis also elicits intraepithelial lymphocytosis with variable villous blunting, although infected patients are rarely biopsied because their symptoms spontaneously resolve. Detection of adenovirus is facilitated by recognition of smudgy intranuclear viral inclusions in epithelial cells that can be enhanced using immunohistochemistry. Patients infected with the human immunodeficiency virus (HIV) may develop chronic diarrhea (>1 month), malabsorption, and weight loss, yet they lack other detectable pathogens. These patients are presumed to have HIV enteropathy.76, 77, 78 Features of this disorder are diverse and include focal active inflammation, erosions, villous blunting, crypt epithelial cell apoptosis, and intraepithelial lymphocytosis.

Bacterial infections, such as Whipple disease and infection by Mycobacterium avium complex, expand the lamina propria with numerous foamy macrophages. Whipple disease is caused by failed phagocytosis of Tropheryma whipplei, which can be detected in lysosomes ultrastructurally and appears as chunky PAS-D-positive material in macrophages. Whipple disease is often accompanied by dilated lacteals and extracellular lipid deposits that can be a helpful diagnostic feature (Fig. 10).⁷⁹ Polymerase chain reaction for T. whipplei can be used to confirm the infection.⁸⁰ In contrast, Mycobacterium avium complex appears as long, thin cytoplasmic rods reminiscent of sheaves of wheat. These organisms show acid-fast and PAS-D positivity (Fig. 10C and D). Small bowel bacterial overgrowth is another potential cause of villous blunting.⁸¹ Overgrowth with anaerobic bacteria occurs in patients with decreased motility resulting from neural disorders, amyloidosis, or diabetes mellitus; structural defects due to strictures or surgical manipulation; and immune deficiency states. Patients with bacterial overgrowth present with diarrhea and malabsorptive symptoms. Histologic features are variable. Some symptomatic patients have entirely normal small bowel biopsies, whereas approximately 25% show some degree of villous blunting. Intraepithelial lymphocytosis has been reported in some studies, whereas others have failed to identify this as a consistent finding.81, 82