Objectives: Small bowel intraepithelial lymphocytosis (IELs) with normal villous architecture is a relatively common finding, often of uncertain significance. The aims of our study were to determine the prevalence of this finding, its clinical associations, its specificity for celiac disease (CD), and whether histologic clues exist that increase the specificity for CD in this setting, all in the pediatric population.
Methods: The Mayo Clinic electronic pathology database was searched between January 1, 2000 and December 31, 2009 for patients younger than 18 years who had the terms “normal villi” and “increased intraepithelial lymphocytes” in their small bowel biopsy reports. All of the slides were reviewed to confirm the histologic findings. Demographic, serologic, pathologic, and clinical informations were obtained.
Results: Among 1290 duodenal biopsies obtained from children during the years 2000 and 2009, 56 (4.3%) were noted to have “normal villous architecture with increased intraepithelial lymphocytes.” In the 54 patients not known to have CD before biopsy, 48 (89%) had serologic testing for CD. Of these 48 patients, 9 were labeled with CD, although only 5 of 9 met the definite criteria for the diagnosis, based on a combination of serologic markers, human leukocyte antigen haplotyping, and response to gluten-free diet. No clinical features pointed to a diagnosis of CD. There was no correlation between CD and number of IELs, but patients with newly diagnosed CD were more likely to have a tip-heavy lymphocyte distribution. Other diagnoses made during the study period and in follow-up were inflammatory bowel disease (5), Helicobacter pylori infection (3), medication-related injury (10), and systemic autoimmune disorders (2).
Conclusions: Increased IELs with normal villous architecture in small bowel biopsies are clinically important in children, and are associated with a new definite diagnosis of CD in 9% of pediatric patients. Even at this low sensitivity, clinical work-up for CD is mandated in all of the patients with this finding.
Increased intraepithelial lymphocytosis (IELs) with preserved villous architecture can represent celiac disease (CD), corresponding to grade 1 of the Marsh classification (1). In addition to symptomatic, latent, or partially treated CD, this histologic finding has been reported in those with dermatitis herpetiformis and in 38% of first-degree relatives of patients with CD (2). Conversely, this histologic finding alone is far from specific for CD; other reported associations in adults include systemic autoimmune disorders, Helicobacter pylori gastritis, enteric viral infection, Crohn disease, and exposure to nonsteroidal anti-inflammatory drugs (NSAIDs) (3–5). The significance of this histologic finding has not been elucidated in children.
The goal of our study was to assess the clinical characteristics and long-term outcomes of children with small bowel biopsies demonstrating increased IELs with normal villous architecture. We aimed to determine the prevalence of this histologic finding, its clinical associations, its specificity for CD, and whether there are histologic clues that increase the specificity for CD in this patient population.
We searched the Mayo Clinic electronic pathology database from January 1, 2000 to December 31, 2009 for patients younger than 18 years who had the terms “normal villi” and “increased IELs” in their small bowel biopsy reports. From the total of 1290 duodenal biopsies obtained during that study period, 56 patients fulfilled search criteria. One pathologist with an interest in gastrointestinal pathology reviewed all of the slides to confirm the histologic findings.
Demographic and clinical information was collected, including age at biopsy; sex; race; medical history; family history of CD; use of NSAIDs, aspirin, and cyclooxygenase-2 inhibitors within 3 months before biopsy; celiac serological studies including gliadin antibodies (IgA and IgG), endomysial antibodies (EMA), and tissue transglutaminase (TTG) antibodies (IgA and IgG); human leukocyte antigen-DQ haplotype; indication for biopsy; final clinicopathological diagnosis; length of follow-up; and additional diagnosis during follow-up.
In patients who were labeled with a new diagnosis of CD by their provider, we categorized the likelihood of that being the correct diagnosis in the following way. “Definite” CD was used to define patients with elevation in 2 different serologic markers (TTG and EMA), or those with elevation in 1 serologic marker (TTG) and a documented clinical response to a gluten-free diet (GFD). “Possible” CD described patients who did not have a serologic marker above the normal range, but had both a decrease in the serologic titer and a documented clinical response to a GFD. “Unlikely” CD categorized patients with 2 different serologic markers that were both negative for CD, despite the fact that there was compatible human leukocyte antigen haplotyping and clinical improvement on a GFD. These definitions were adapted from clinical guidelines, but may not fully represent all of the clinical categories (6).
All of the biopsy specimens had been fixed in formalin, embedded in paraffin, and cut at 4 μm for staining with hematoxylin and eosin (H&E). The routine sectioning protocol calls for 2 ribbons of tissue on each of the 2 slides, giving approximately 16 levels of tissue available for review on each patient, optimizing the potential for finding well-oriented pieces. Immunohistochemical stains for T-cell subsets are not used in routine clinical practice, but for this retrospective review, antibody to CD3 (Novocastra Laboratories, Newcastle, UK) was applied to paraffin sections using standard immunohistochemical technique.
All of the slides were reviewed independently by a single pathologist to confirm that villous architecture was preserved and that there were at least 30 IELs per 100 epithelial cells (Fig. 1). IELs were counted independently on H&E- and CD3-stained slides, without knowledge of the result from the complementary technique. The distribution of IELs was characterized as even, predominantly at the tips of villi, or predominantly on the sides of villi. The present study was approved by the Mayo Clinic institutional review board.
Among 1290 duodenal biopsies obtained from children during the years 2000–2009, 56 (4.3%) had “normal villous architecture with increased IELs” (Fig. 2). No diagnosis was changed after re-review of histology. IEL counts using CD3 were within 10% of those obtained on H&E, and were not consistently either higher or lower (data not shown). The mean age of patients with this histologic pattern was 7.8 years, with the majority being girls (64.3%), of white race (60.7%), who presented with abdominal cramps/pain (60.7%) (Table 1).
Two patients carried a diagnosis of CD before the present biopsies were obtained. The first was a 15-year-old girl who remained symptomatic 1 year after receiving a GFD; TTG was persistently elevated. The finding of increased IELs on biopsy prompted strict dietary compliance. Follow-up biopsies 4 years later were normal. The second patient was a 9-year-old girl diagnosed as having CD as a toddler. She was asymptomatic on a GFD and hoped to relax her dietary restrictions. Repeat biopsies showed increased IELs; she remained on a GFD.
Among 54 patients not known to have CD before biopsy, 48 (89%) had serologic testing for CD. In 9 of these patients (18.8% of the 48 tested for CD), a new diagnosis of CD was made by the provider, despite differing levels of support for the diagnosis. Five of the 9 patients met “definite” criteria for CD, representing 9% of those with isolated increased IELs on duodenal biopsy, whereas 2 patients met “possible” criteria and 2 met “unlikely” criteria, each representing 4% of those with the histology of interest. No clinical feature pointed specifically to a diagnosis of CD. There was no correlation between CD and numbers of IELs. Table 2 shows the demographic, serologic, and histologic features for this group.
Table 3 shows clinical and histologic information for the patients without an earlier diagnosis of CD, broken down by final diagnosis. The average length of follow-up was 24 months (range 0–100 months). During this period, 12 patients had follow-up gastroduodenal biopsies. A total of 3 patients had H pylori infection (2 were preexisting diagnoses, 1 was newly diagnosed); 5 patients had inflammatory bowel disease, including 4 with Crohn disease and 1 with ulcerative colitis (3 of the Crohn cases were preexisting diagnoses); 10 patients had exposure to aspirin, NSAIDs, or COX-2 inhibitors within the 3 months before biopsy. Two patients were noted to have autoimmune conditions, and each was discovered to have lactase deficiency, fructose malabsorption, and selective IgA deficiency. No patients were diagnosed as having CD during the follow-up period.
In 22 patients, the finding of increased IELs remained idiopathic, despite clinical workup and focused retrospective chart review. The group comprised 4 boys, ages 1 to 4 years, and 16 girls, mostly teens. Two of the boys were thought to have postinfectious gastroenteritis, an impression supported by the fact that follow-up duodenal biopsies 6 months later were normal. One boy continued to be troubled by fevers and joint pain 5 years after the abnormal duodenal biopsy, but no specific systemic illness has been diagnosed. One girl had recently been treated for Hodgkin disease and 1 had undergone multiple surgeries for gastroschisis, providing at least an intuitive explanation for duodenal inflammation. Ten girls in their teen years were given upper gastrointestinal diagnoses of varying specificity (reflux, functional dyspepsia, abdominal pain). None of those patients had a follow-up biopsy; 4 have been followed into young adulthood and their gastrointestinal symptoms have resolved.
Between January 1, 2000, and December 31, 2009, we obtained duodenal biopsies from 1290 pediatric patients. Of these, 56 (4.3%) were diagnosed as “normal villous architecture with increased IELs.” The majority of these patients were worked up for CD with serologic studies. Although 9 of the 56 patients were labeled with a new diagnosis of CD in the medical record, only 5 (9%) met our “definite” criteria, 2 (4%) met “possible” criteria, and 2 (4%) were “unlikely” to have CD based on weak supporting data. Two additional patients had confirmation of a previous CD diagnosis. Although 23 of the 56 patients (41%) had a possible explanation for the histologic finding, 22 (39%) had no definite clinical explanation, despite clinical follow-up. There was no difference in clinical presentation between patients who had CD and those who did not. Patients with CD were more likely to have a tip-heavy IEL pattern on the duodenal villi.
Possible causes we found for IELs other than CD included medication exposure (n = 10), inflammatory bowel disease (n = 5), H pylori infection (n = 3), autoimmune conditions (n = 2), lactase deficiency, fructose malabsorption, and selective IgA deficiency (n = 1, each). In 22 patients (39%), the histologic finding was unexplained; 16 of those 22 had a serologic work-up for CD and 6 did not. This case mix is similar to that described for adult patients (4), although an association between this histologic finding and inflammatory bowel disease has been controversial in the adult literature (7).
The only comparable study in children comes from Lähdeaho et al (8) who evaluated clinical outcomes in 236 children with suspected CD who had minor changes on small bowel biopsies. Inclusion criteria were decreased villous/crypt ratio and/or increased IELs. The study reported 5 new CD diagnoses, but only 76 patients were available for follow-up. Additionally, the average age at biopsy of cases in that study was 3.1 versus 7.8 years in our study, which may affect the comparability of the results.
According to a 2008 study by Hill and Holmes (9) using data from 2002 to 2003, the positive predictive value for CD was 100% when IgA-TTG was >10 times the upper limit of normal. The same study reported a 58% positive predictive value using data from 2004 to 2005 (9). As demonstrated in Table 2, only 2 of our eligible patients met this criterion. This may indicate a correlation between antibody titer and the degree of intestinal mucosal damage.
Small bowel biopsy is a critical component of the diagnostic workup for CD. But even the most floridly abnormal duodenum—completely flat mucosa with an attenuated surface epithelium overrun by IELs—is not specific for CD, thus the adage that “all that flattens is not sprue.” An irreducible lack of specificity also accompanies the histologic finding of increased IELs with normal villous architecture. Larger numbers of IELs do not make CD more likely. A tip-heavy distribution of IELs is characteristic of CD, as previously reported by Goldstein (10). Four of 5 patients with “definite” CD and both patients with “possible” CD had IELs concentrated at the villus tip, compared with neither in the “unlikely category and 2 of 45 of the others. Although this finding does not appear to be sufficiently robust to be considered diagnostic of CD, it should heighten the awareness of this association, and prompt a thorough evaluation for CD in patients when it is noted.
We noted 2 patients that were labeled by their care provider as having CD, although they met “unlikely” criteria based on our definition. Although both had compatible celiac gene pairs and clinical improvement on a GFD, the fact that 2 different serologic markers (TTG and EMA) were negative makes the diagnosis less likely. The diagnosis of CD was found persistently in their medical record from that point forward, which may have lifelong implications as related to dietary restriction, future testing/procedures, and insurability.
The present study has several limitations. First, the retrospective nature of the present study makes it susceptible to sampling bias. Next, our tertiary care center may alter the mix of patients who come to biopsy. Additionally, there were 6 patients who had increased IELs with normal architecture who were not tested for CD, so it is possible that some cases of CD may have been missed, although this was not apparent in the follow-up period. Finally, the biopsy protocol in effect during the course of the study called only for biopsies from the second part of the duodenum, meaning that patchy disease, particularly disease involving only the duodenal bulb or more distal duodenum, may have been missed (11,12).
We conclude that the histologic finding of increased IELs with normal villous architecture in small bowel biopsies is clinically important. It is associated with a new “definite” diagnosis of CD in 9% of pediatric patients. Even at this low sensitivity, clinical workup for CD is mandated in all of the children with this finding. As in adults, there are other conditions that can produce this histologic change. Although a tip-heavy distribution of IELs favors CD, there do not appear to be histologic features that can stand alone as diagnostic of CD in this setting. As in adults, the etiology for increased IELs remains idiopathic in a substantial subset of patients.
1. Marsh MN. Gluten, major histocompatibility complex, and the small intestine. A molecular and immunobiologic approach to the spectrum of gluten sensitivity (‘celiac sprue’). Gastroenterology 1992; 102:330–354.
2. Marsh MN, Bjarnason I, Shaw J, et al. Studies of intestinal lymphoid tissue. XIV-HLA status, mucosal morphology, permeability and epithelial lymphocyte populations in first degree relatives of patients with coeliac disease. Gut 1990; 31:32–36.
3. Memeo L, Jhang J, Hibshoosh H, et al. Duodenal intraepithelial lymphocytosis with normal villous architecture: common occurrence in H pylori gastritis. Mod Pathol 2005; 18:1134–1144.
4. Carmack SW, Lash RH, Gulizia JM, et al. Lymphocytic disorders of the gastrointestinal tract; a review for the practicing pathologist. Adv Anat Pathol 2009; 16:290–306.
5. Kakar] S, Nehra V, Murray HA, et al. Significance of intraepithelial lymphocytosis in small bowel biopsy samples with normal mucosal architecture. Am J Gastroenterol 2003; 98:2027–2033.
6. Hill ID, Dirks MH, Liptak GS, et al. Guideline for the diagnosis and treatment of celiac disease in children: recommendations of the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr 2005; 40:1–19.
7. Wright CL, Riddell RH. Histology of the stomach and duodenum in Crohn's disease. Am J Surg Pathol 1998; 22:383–390.
8. Lähdeaho ML, Kaukinen K, Collin P, et al. Celiac disease: from inflammation to atrophy: a long-term follow-up study. J Pediatr Gastroenterol Nutr 2005; 41:44–48.
9. Hill PG, Holmes GKT. Coeliac disease: a biopsy is not always necessary for diagnosis. Aliment Pharmacol Ther 2008;27:572–77.
10. Goldstein NS, Underhill J. Morphologic features suggestive of gluten sensitivity in architecturally normal duodenal biopsy specimens. Am J Clin Pathol 2001; 116:63–71.
11. Bonamico M, Mariani P, Thanasi E, et al. Patchy villous atrophy of the duodenum in childhood celiac disease. J Pediatr Gastroenterol Nutr 2004; 38:204–207.
12. Bonamico M, Thanasi E, Mariani P, et al. Duodenal bulb biopsies in celiac disease: a multicenter study. J Pediatr Gastroenterol Nutr 2008;47:5:618–22.