Context.—An increased intraepithelial lymphocyte density in an architecturally normal proximal small intestinal mucosal biopsy is a common finding facing surgical pathologists dealing with gastrointestinal biopsy specimens. Approximately 1% to 2% of all proximal small intestinal biopsies will show this change. It is increasingly recognized by surgical pathologists that gluten-sensitive enteropathy is an important cause of this pattern; however, gluten-sensitive enteropathy accounts for the minority of all cases. A wide variety of immunologic stimuli can raise intraepithelial lymphocyte numbers. Among the other common associations are enteric infection, autoimmune disease, drugs, and gastric Helicobacter infection.
Objective.—To outline the causes of intraepithelial lymphocytosis, to highlight the importance and the difficulties faced in establishing gluten-sensitive enteropathy as the cause, and to aid the surgical pathologist in the routine sign out of these cases.
Data Sources.—A review of the literature detailing the causes or associations of proximal small intestinal intraepithelial lymphocytosis is presented.
Conclusions.—Increased lymphocyte numbers in the epithelium of architecturally preserved proximal small intestinal biopsies is a morphologic feature associated with a broad differential diagnosis.
Intraepithelial lymphocytes (IELs) are a normal constituent of the small intestinal mucosa where they play a significant role in immune surveillance and activation.1 The vast majority of IELs are of T-cell type and express an α/β T-cell receptor on their surface.2 Only approximately 5% have surface γ/δ T-cell receptors (a population that is expanded in the setting of gluten sensitivity).2–4
The upper normal limit of IELs in the proximal small intestine has long been considered to be 40 IELs per 100 epithelial cells,1,5 a ratio used as a criterion of the Marsh classification for gluten sensitivity.6 This figure was derived by counts performed on thickly sectioned (7-μm) hematoxylin-eosin sections.5 However, recent studies using hematoxylin-eosin sections cut at 3 μm7 and 4 μm8 and CD3 immunohistochemistry7 have shown the upper limit of normal in the proximal small intestine (especially in the duodenum) to be 25 IEL per 100 epithelial cells. Theoretically up to 2.5% (<2 standard deviations above the mean) of duodenal biopsies could have an IEL density above this level. Of interest with regard to this review, elevated IEL counts have been observed in 1.3%9 to 2.2%10 of patients undergoing upper gastrointestinal endoscopy and small intestinal biopsy.
Given that the villi in proximal small intestinal biopsies are seldom perfectly orientated and may appear bent or are tangentially cut,11 it has been suggested that 4 normal fingerlike villi in a row with a villus-to-crypt ratio between 3:1 and 5:1 be seen to consider the biopsy architecturally normal.11,12 In the normal villi, IELs decrease in number from the base toward the tip13 and counting IELs at the tips of the villi is recommended as the most useful method of confirming an increase.13–15
In practice, it is wise to avoid counting IELs in villi that are not well oriented because tangential sections of the villus base can produce an artificial impression of an increased number.
CD3 immunohistochemistry (Figure 1) can aid the detection of IELs particularly because some lymphocytes can have irregular nuclear outline mimicking granulocytes and others could be mistaken for epithelial cell nuclei.14,16 However, not all pathologists have found this additional stain helpful, cautioning that it decreases the sensitivity derived from appreciating a loss of the normal IEL decrescendo pattern.17
CAUSES OF INCREASED IELs IN ARCHITECTURALLY NORMAL SMALL INTESTINAL BIOPSIES
Because IELs are a functioning component of the intestinal immune system, it is not unexpected that immunologic stimulation whether by an ingested antigen, endogenous antigen, or autoimmune causation can lead to increased IELs. In many cases there will be associated epithelial cell injury and death (by apoptosis) leading in time to villus atrophy. Conditions associated with normal villus architecture would seem to represent a mild or early manifestation of this process.
Gluten-Sensitive Enteropathy With Normal Villus Architecture
That gluten sensitivity can be morphologically manifest by intraepithelial lymphocytosis without villus atrophy is increasingly being appreciated by pathologists (Figure 2). The reported prevalence of gluten-sensitive enteropathy (GSE) as the cause of increased IELs in architecturally normal small intestinal biopsies has ranged from 9% to 40%,9,10,18 although for reasons mentioned subsequently the exact prevalence is difficult to determine.
It is, nonetheless, important to try to recognize the gluten-sensitive patients with mild morphologic changes because these people can develop significant nutritional deficiencies leading to anemia (usually iron deficiency type) and osteoporosis.19,20 Furthermore, there is a small risk for the subsequent development of malignancy (particularly intestinal T-cell lymphoma, but also upper gastrointestinal tract carcinomas) in these patients as well as in their first-degree relatives.19,20 Terms such as gluten sensitivity with mild enteropathy, latent celiac disease, and potential coeliac disease have been applied to this condition.17,21 Despite mild changes on small bowel biopsy, these patients can have significant clinical symptoms and signs because these correlate better with the length of involvement of the small intestine than with the severity of villus architectural change.22 Also because the architectural change in GSE can be patchy, biopsies showing normal architecture may not reflect villus change elsewhere in the small intestine.
A major problem confronting clinicians is to establish which cases of intraepithelial lymphocytosis represent the Marsh type 1 pattern of gluten sensitivity from the other possible causes of this morphologic pattern. The magnitude of this problem is highlighted when one considers the severe dietary alterations and lifelong clinical follow-up required in patients diagnosed with gluten sensitivity, while failure to establish the diagnosis puts patients at risk of the complications listed previously. Although small intestinal biopsy remains the gold standard for the diagnosis of GSE, it is usually supplemented by serologic tests. The latter includes immunoglobulin (Ig) A antiendomysial antibodies, IgA antitissue transglutaminase (which has largely replaced the IgA antiendomysial antibody test), and IgA and IgG antigliadin antibodies. The first 2 have optimal sensitivity and specificity as well as high positive predictive value.23,24 A positive IgA antiendomysial antibody test or a significantly raised anti–tissue transglutaminase titer helps confirm the diagnosis of GSE even when villus architecture is normal.25 Unfortunately, the likelihood of a positive test correlates with the degree of mucosal injury and GSE cases with intraepithelial lymphocytosis alone can often have antibody titers in the normal range.25–28 Thus, negative serology does not exclude a diagnosis of GSE with a Marsh type 1 pattern.
Where available, testing for human leukocyte antigen DQ haplotype is being increasingly used. However, this is only helpful in excluding gluten sensitivity (or making it less likely)29 because celiac disease associated DQ haplotypes (DQ2 and DQ8) are seen in many of the normal population.30 This leaves many clinicians to resort to a trial of gluten-free diet or sometimes a gluten challenge followed by rebiopsy.
Importantly, the surgical pathologist may be able to assist in this situation. It is recognized that biopsies displaying an even distribution of IELs along the villi or a villus tip accentuation with a consequent loss of the normal decrescendo pattern (decrease in IEL numbers from base to tip of villi) are more likely to be associated with GSE than cases without this alteration (Figure 3).9,13,14,15 This is, however, not specific to GSE and not all cases of gluten sensitivity show this feature.9,13 However if present, it should be highlighted in the surgical pathology report.
Hypersensitivity to Other Alimentary Proteins
Hypersensitivity to nongluten components of foods including cereals, cow's milk, soy products, fish, rice, and chicken are also associated with increased IELs in affected patients, although villus atrophy is usually present.31–37
We and others have observed intraepithelial lymphocytosis in children with autism, whether or not it is associated with lactase deficiency frequently recorded in these patients.38
Parasitic infections such as Giardia lamblia and Cryptosporidium can lead to increased lymphocytic exocytosis. Elevated IEL counts can also be seen as a result of viral enteritis and in the setting of tropical sprue,39,40 although some degree of villus atrophy is usually recognized (Figure 4). Goldstein17 has also noted increased IELs in the duodenal bulb in patients with Helicobacter pylori gastritis. We and others41 have observed this association with gastric H pylori infection not infrequently (Figure 5). The intraepithelial lymphocytosis can also be seen beyond the duodenal bulb and is usually not associated with gastric intraepithelial lymphocytosis.
Bacterial overgrowth whether related to gastric hypochlorhydria or intestinal dysmotility has been associated with an increase of IELs (in the absence of villus atrophy) when compared with control biopsies.42 However, only some cases in that study had an IEL density greater than what is currently considered the normal range. In the study by Kakar et al,9 2 cases of intraepithelial lymphocytosis were related to bacterial overgrowth.
Various drugs have been associated with colonic intraepithelial lymphocytosis (ie, lymphocytic colitis),43,44 so it is intuitively expected that drugs could be responsible for similar change in proximal intestine. However, at present there is a paucity of literature supporting this presumption. Kakar et al9 noted nonsteroidal anti-inflammatory drug use in 14% of patients in their series. Similarly, in our experience 15% of patients with increased IELs were on nonsteroidal anti-inflammatory drugs at the time of biopsy. We also found 20% of patients were taking a proton pump inhibitor (I. Brown, unpublished data, 2004).
Disorders of immune regulation are frequently associated with elevated IELs and various degrees of villus atrophy. Fourteen percent of patients in the series by Kakar et al9 had an associated immune dysregulation condition. Reported conditions include Hashimoto thyroiditis, Graves disease, rheumatoid arthritis, psoriasis, multiple sclerosis, and systemic lupus erythematosus (Figure 6).1,9,45 We have also noted an association in 2 patients with ankylosing spondylitis and in 1 patient with progressive systemic sclerosis. Intestinal intraepithelial lymphocytosis in the setting of glomerulonephritis particularly when presenting with nephrotic syndrome is also reported.46
Autoimmune enteropathy, an ostensibly uncommon condition usually seen in a pediatric setting, can mimic GSE clinically and histologically.47,48 Villus atrophy is normally observed but morphologically milder cases could enter the differential diagnosis under discussion. In addition, this condition is reported in adults48 and may be underrecognized.
Patients with hypogammaglobulinemia whether due to IgA deficiency or common variable immunodeficiency often have increased IELs in the proximal small intestine. Despite the defective humoral immune system in these patients, this finding has a cellular immunologic basis pathogenically analogous to GSE. Villus atrophy is usually seen but can be mild. Of diagnostic value are associated features including reduced numbers of plasma cells, frequent parasitic infection (eg, Giardia) and lymphoid follicle formation.49–51
Recently, it has been reported that chronic inflammatory conditions of the colon such as ulcerative colitis,52 Crohn disease (Figure 7),9,17,52 and collagenous and lymphocytic colitis9 can be concurrently associated with proximal small intestinal intraepithelial lymphocytosis. In both lymphocytic and collagenous colitis, the association appears to be exclusive of gluten sensitivity in many cases despite the recognized association of both conditions with celiac disease.44,53,54
Interestingly, the finding of small intestinal intraepithelial lymphocytosis in Crohn disease can be seen at the time of active disease but may also precede active Crohn enteritis,17 possibly in a fashion analogous to focal colonic intraepithelial lymphocytosis preceding Crohn colitis.55
An enigmatic condition termed diffuse lymphocytic gastroenteropathy 56 characterized by intraepithelial lymphocytosis throughout the upper gastrointestinal tract has been reported. Whether this is a distinct morphologic entity, a variant of gluten sensitivity, or an unusual immune reaction pattern to various (and unrecognizable) stimuli is unclear.
Finally with respect to immune causation, proximal small intestinal biopsies in patients with graft versus host disease can, although uncommonly, display prominent numbers of IELs.11 The finding of epithelial cell apoptosis and some degree of architectural disturbance57 together with the clinical setting allows for the correct diagnosis.
The preinfiltrative (or cryptic) form of intestinal T-cell lymphoma (enteropathy-associated T-cell lymphoma) in which the neoplastic cells (often surprisingly nonatypical) can present within a minimally or nonatrophic small intestine mucosa (Figure 8).58 Failure to respond to a gluten-free diet and a significant malabsorption problem are frequent accompaniments. In addition to immunohistochemical evaluation, flow cytometric studies or polymerase chain reaction for T-cell receptor gene rearrangement are usually necessary to establish a firm diagnosis.
Intraepithelial Lymphocytosis Without Definitive Diagnosis
Finally there are a significant number of patients displaying intraepithelial lymphocytosis in their proximal small intestine who manifest ongoing often nonspecific gastrointestinal symptoms but for whom a specific diagnosis cannot be made. Some of these patients meet clinical criteria for irritable bowel syndrome and may respond symptomatically to a gluten-free diet. Whether these patients have a subtle form of gluten sensitivity remains uncertain although there is some experimental support for this suggestion.59
Up to 2.5% of proximal small intestinal mucosal biopsies display increased IELs (>25 IELs per 100 epithelial cells) in the absence of villus architectural change. In most cases this is due to immunologic activation of the lymphocytes that are normally resident in the epithelium. The causes for this increase in numbers of IELs are multiple and include reactions to intraluminal antigens and small intestinal manifestations of autoimmune or other allied diseases (see Table). Gluten sensitivity is a common cause, accounting for 9% to 40% of cases.9,10,18 Other common associations include gastric H. pylori infection, a drug reaction, and autoimmune lymphocyte stimulation. In a significant number of cases, a cause is never established. It is the practice of one of us (I.B.) to add a comment to the surgical pathology report in cases in which intraepithelial lymphocytosis is the diagnosis. This comment states “the finding of intraepithelial lymphocytosis with preserved villus architecture is a non specific immunological phenomenon that has a large number of possible causes. A mild histological manifestation of gluten sensitivity is one of these causes. Unfortunately, celiac serology may be negative even in patients that subsequently prove to be gluten sensitive. Other common causes of this appearance include infective enteritis, H. pylori infection, a drug effect eg NSAIDs, and autoimmune disease. In many cases a specific cause is not identified.”
The establishment of a diagnosis of gluten sensitivity can be clinically difficult and relies on a weighted assessment of clinical, serologic, and histopathologic data. Surgical pathologists can help by highlighting cases that have a uniform distribution of IELs over the villus length rather than those showing increased numbers but in a persisting decrescendo pattern because the latter is an unusual feature for GSE.
Awareness of the various conditions associated with increased IELs in architecturally preserved proximal small intestinal mucosa is important to guide the clinician toward a correct diagnosis.
The authors have no relevant financial interest in the products or companies described in this article.
Reprints: Ian Brown, MBBS, FRCPA, Sullivan Nicolaides Pathology, Taringa, Queensland, Australia (firstname.lastname@example.org)