"Give Us This Day Our Daily Bread"-Evolving Concepts in Celiac Sprue

Celiac sprue (CS) is a complex multigenic disease. Genetically susceptible individuals may develop small intestinal injury and malabsorption from dietary exposure to "gluten" proteins in wheat, barley, and rye. Confirmation of diagnosis is 2-fold: first, biopsy evidence of a characteristic but nonspecific pattern of injury, including villous blunting or flattening, surface enterocyte damage, and increased intraepithelial lymphocytes; and second, dramatic clinical response to a gluten-free diet (GFD). Complete gluten removal from the diet is effective treatment for patients with symptoms of malabsorption (steatorrhea is the hallmark); however, lifelong adherence to the diet is expensive, socially limiting, and nearly impossible on a contemporary diet with manufactured foodstuffs. Therefore, pathologists should avoid overdiagnosis of celiac disease based on minimal, nonspecific histologic changes. Biopsy artifact can be confused with true villous flattening. Diseases with identical histology include kwashiorkor, tropical sprue, pseudo-obstruction, cow or soy milk injury, acid-peptic injury in Zollinger-Ellison syndrome, and infectious injury. These do not have a clinical response to a GFD and usually respond to other specific therapies. Dermatitis herpetiformis (DH) shares CS genes and is characterized by a pruritic skin rash, but affected patients rarely have overt steatorrhea. Nearly all DH patients have intestinal abnormalities if eating gluten, but lesions are focal and minimal. Approximately 8% of CS patients develop small intestinal T-cell lymphomas, mostly in elderly patients after many years of a flat celiac intestinal lesion responsive to a GFD. There is no convincing evidence that a GFD can prevent lymphoma.

WHAT IS THE IMPACT OF A DIAGNOSIS OF CS?

Civilization as group living in a civitas (Latin, city or town) did not occur prior to the cultivation of grain in sufficient quantities to support a stationary human habitat. The earliest archeologic evidence shows that the first cultivated grains were wheat and barley. The word companionship derives from the Latin word companionem (literally, bread fellow) from the roots con (with) and panis (bread). The breaking of bread, the sharing of grain-based foods, is such a fundamental part of participation in many societies that it forms the substrate for important rituals, including the Jewish use of challah at Shabbat and matzoh at Passover, and the Christian rite of communion with bread as the host. There are numerous secular rituals in which grains or grain-based products are also central in the form of birthday cakes, wedding cakes, dumplings, piroshki, flat breads, noodles, cookies, gravies, sauces, and condiments. Lifelong removal of gluten from a patient's diet is almost a form of excommunication. Pathologists who render a diagnosis of CS must remember this and the potential impact on the patient's social life that results from dietary elimination of gluten, especially in children. We must not render this diagnosis on minimal clinical or histologic evidence.

WHAT IS CS?

Celiac sprue is a complex multigenic disease in which certain polypeptide sequences in wheat, barley, and rye trigger an interleukin 15 response in genetically susceptible individuals. When these sequences are deamidated by tissue transglutaminase (tTG), they bind to DQ2 and DQ8 antigen-presenting cells. These, in turn, activate CD4+ helper T lymphocytes and induce their proliferation. The CD4+ helper cells injure the small bowel mucosa, causing intestinal injury and malabsorption.1-7

There are several factors required for CS to manifest as a disease process. First, the patient must have hereditary susceptibility to injury by gluten, which comprises gliadens, found in wheat, and homologous food grain proteins hordeins and secalins in rye and barley, respectively. And the patient must be exposed to adequate amounts of gluten in wheat, barley, and rye to initiate injury. The clinical symptoms and signs of CS are related to the extent and location of intestinal damage. The first line of treatment is the elimination of dietary exposure to gluten.

WHAT ARE THE SYMPTOMS OF CS? WHAT IS MEANT BY CLASSIC SPRUE, ATYPICAL SPRUE, AND LATENT SPRUE?

Classic CS is characterized by gastrointestinal symptoms that bring the patient to medical attention. Malabsorption is the main symptom, and steatorrhea is the hallmark. The classic childhood presentation is failure to thrive, which may be associated with iron deficiency anemia, and diarrhea with steatorrhea, abdominal bloating, and flatulence. The affected child eats more than siblings or friends but may not gain weight or grow to the expected height.

Atypical CS is classic CS without overt steatorrhea. Patients may present with a wide range of other symptoms, all caused by malabsorption, including growth arrest, iron deficiency anemia, infertility, and symptoms such as tetany due to calcium deficiency. Other manifestations may include depression, psychosis, weight loss, fatigue, amenorrhea, arthralgias and, rarely, neurologic symptoms. A number of CS patients also manifest autoimmune disorders, including autoimmune thyroiditis and other immune-mediated endocrinopathies.1-7 There is also an association of CS with selective immunoglobulin A (IgA) deficiency, Turner syndrome, and Down syndrome. Diabetes mellitus type 1 may coexist with celiac disease, which should be ruled out in every type 1 diabetic, especially if diarrhea or steatorrhea develops.1-7

Another important group of atypical sprue patients are patients with dermatitis herpetiformis (DH), all of whom carry genes for celiac disease but only rarely manifest overt steatorrhea.8,9 Approximately 75% of patients with DH have some flat intestinal lesions, although they may be delimited or focal in extent. Patients with DH respond to gluten restriction because their intestinal lesions heal quickly; however, their skin lesions take months to resolve by diet therapy alone. The most efficient and effective treatment of the skin rash and symptoms is medical therapy with drugs, such as dapsone. The DH patient without severe intestinal injury is not likely to adhere to dietary elimination of grain and gluten-containing foods, and treatment by a GFD is too slow to be reasonable as a firstline treatment for the skin rash. Only those patients with moderate to severe mucosal flattening characteristic of classic CS should be given a trial of a GFD, along with monitoring of symptomatic response.8-10 Asymptomatic patients without histologic alterations do not have overt CS, should not be treated as such, and should not be labeled with the wrong diagnosis. Few asymptomatic patients will adhere voluntarily to gluten restriction.

Latent sprue is DH with no histologic evidence of intestinal lesions, despite many proximal intestinal biopsies; thus, the celiac genes may be present but not penetrating. This clinical syndrome is only rarely seen. These patients, if overdosed with high gluten for 4 to 5 months, will develop a flat intestinal lesion, but intestinal histology on a usual diet is normal.11 This intestinal injury is an iatrogenic disease and would not develop on its own, nor should it be provoked. Since the intestine may be histologically normal, there may be no elevation in tTG, which requires intestinal injury to be detectable. Moreover, unless the patients have clinical symptoms other than skin rash, they are unlikely to adhere to a GFD, and there is no clear evidence that such a regimen will provide any benefit.

DO CS PATIENTS DEVELOP MALIGNANCIES?

Celiac sprue patients are at an increased risk for developing certain malignancies when compared with controls, including an increased relative risk (2- to 6-fold increase compared with nonceliacs) for non-Hodgkin lymphoma in any primary site.12 Within the category of lymphomas, CS is linked particularly with enteropathy-associated T-cell lymphoma, which is reported to affect 8% of elderly celiac patients. There is a mild increased risk for small intestinal adenocarcinoma, which is, nonetheless, quite rare. Most cases of lymphoma occur in CS patients older than 70 years. Moreover, the efficacy of a GFD to reduce or to eliminate the risk for the development of malignancy has not been proven. It is important to emphasize that fear of potential malignancy, which is a low risk event overall in patients with celiac disease, is not a good reason to subject patients to a lifelong limited diet, unless there is sufficient other symptomatic or histologic evidence to warrant treatment, or until there is good evidence that a GFD effectively reduces risk for malignancy.12-15

HOW IS CS DIAGNOSED?

Long experience has shown that CS requires the fulfillment of 2 criteria: First, there is a characteristic, but not specific or diagnostic, flat lesion of the small intestine seen on histologic examination of a small intestinal biopsy, especially in the proximal small intestine, from a patient who is eating gluten. Second, the patient with proven CS experiences a dramatic clinical symptomatic response to a GFD, including the decrease and eventual disappearance of malabsorption (steatorrhea) and improvement, or decrease, in psychiatric symptoms.

WHEN SHOULD PATIENTS UNDERGO BIOPSY?

Patients with symptoms of classical or atypical CS should undergo serologic screening for IgA antibodies to tTG while eating a diet containing gluten.2 Patients with tTG elevation should undergo small intestinal biopsy. Following establishment of a diagnosis of CS, and following symptomatic improvement on a GFD, patients who relapse or become refractory to GFD should undergo rebiopsy. Workup should include careful nutritional evaluation to exclude inadvertent gluten ingestion, which is the most common cause of loss of response to a GFD. For patients who have relapsed, multiple biopsies of duodenum and jejunum are indicated, with submission of samples for flow cytometry to search for cell markers of lymphoma, in all patients who have lost their previous normal villous architecture while on a GFD.

Biopsy alone is not a good way to assess the efficacy of GFD. There is a sampling error; proximal small intestinal biopsies may demonstrate injury and underestimate improvement as they do not sample more distal areas that have healed with the restoration of intestinal absorption. The more distal jejunum is less injured than the proximal small bowel and heals more rapidly on a GFD.16

WHAT IS THE HISTOLOGY OF CLASSIC CS?

The normal small intestinal mucosa contains long villi, varying in length depending on biopsy orientation and depth of biopsy (Figure 1). Shallow biopsies that do not include the muscularis mucosae may normally have shorter and broader villi than biopsies that are "tethered" by the muscle, and normal villi may be short or absent compared with Brunner glands. The normal surface enterocytes are columnar in shape, with basally located nuclei and intact brush border at their surface. Intraepithelial lymphocytes at the surface are variable in number, and excessive numbers are obvious but may be patchy. The normal lamina propria contains a scant mixture of lymphocytes and plasma cells.17

In classic CS there is severe small intestinal mucosal epithelial injury with nonspecific pathologic findings, including severe villous injury or destruction (to flat mucosa), increased mononuclear and especially plasma cells in lamina propria, intraepithelial lymphocytosis, most prominent at the villous surface, histologic evidence of surface cell damage, and variable extent of involvement of mucosal damage beyond the duodenum (Figure 2, A). The duodenum, especially the proximal duodenum, is most severely injured, and the severity of malabsorption is related to the length of jejunum, and even ileum, injured. In milder lesions, mucosal damage is focal. In classic CS, loss of villous architecture may be severe, yielding a flat small intestinal mucosa with marked intraepithelial lymphocytosis. If less severely affected areas are sampled, especially distally, the villi may be short and stubby, or even normal in length (Figure 2, B). The histology may appear moderately or focally abnormal if the most severely affected areas are not biopsied, or if the patient is on an incomplete GFD.17 In occasional patients, the only site of disease is the duodenal bulb, and this site should always be included as one of the biopsy samples for evaluation.18

There is no need to rebiopsy the small intestine after initiation of GFD if the patient has a clinical response to dietary gluten exclusion. Rebiopsy may lead to misunderstanding or misdiagnosis, since the proximal small intestinal morphologic improvement, in the region most commonly biopsied, may occasionally lag considerably behind the more rapid healing beyond the ligament of Treitz (ie, jejunum).16 The only reasons to consider rebiopsy are (1) failure to respond clinically and symptomatically to a GFD, or (2) relapse of symptoms while on a GFD and after a long interval of symptomatic improvement on the diet.

HOW SHOULD THE PATHOLOGIST COMMUNICATE THE DIAGNOSIS OF BIOPSIES WITH HISTOLOGIC FEATURES OF CS?

One of the common themes encountered in our consultation practice at the University of Washington (Seattle) is the overdiagnosis of celiac disease, based on histologic features that are minimal or nonspecific, or rendered without reference to the clinical context. It is our experience that the vast majority (more than 75%) of cases of classic CS have a characteristic flat small intestinal biopsy with a dramatic clinical response to a GFD. Unfortunately, clinical information provided to the pathologist may be sparse or absent, and the pathologist may not know if the patient is ingesting gluten or has started a GFD. Therefore, when we encounter histologic features characteristic of but not specific for CS, a suggested wording of the diagnosis in such cases is "Duodenal (or jejunal) mucosa with a moderate (or severe) villous abnormality, with increased intraepithelial lymphocytes and surface epithelial injury, consistent with CS, if the patient is eating a diet that contains gluten; please see Comment." The comment would include the statement that "the diagnosis of CS is a twopronged process. In addition to the characteristic biopsy of moderately or severely injured small intestinal mucosal injury, the diagnosis requires confirmation by a clinical and symptomatic response to a gluten free diet."

With respect to biopsies that exhibit villi of normal length with increased intraepithelial lymphocytes and no evidence of surface enterocyte damage, the appropriate diagnosis is either "No diagnostic alteration" or "Small intestinal mucosa with nonspecific increase in intraepithelial lymphocytes without enterocyte injury." We do not raise the diagnostic consideration of CS unless a biopsy exhibits moderate villous blunting or severe flattening. Minimally increased intraepithelial lymphocytes without surface injury or marked villous blunting are nonspecific, and they are seen in many patients who do not respond to a GFD and who have a disorder other than CS. The correct diagnosis is often made by the patient's response to another form of specific therapy.

WHAT IS THE ROLE OF ANTI-tTG SEROLOGIC SCREENING IN CS?

Immunoglobulin A anti-tTG is the most specific and accurate serologic test available for screening patients with CS, and it is reproducible and easy to perform.19,20 It is not infallible, however, and not all commercial laboratories are reliable in their test methodology or quality. Anti-tTG screening should be used before small bowel biopsy, but it is useful only in patients who are eating a diet that contains grains and gluten, and in patients who do not have IgA deficiency.1,19,20 Currently, small bowel biopsy remains central to the diagnosis of CS, despite advances in serologic testing and in the understanding of the genetics of the disease. Finally, CS is not a consideration if a patient lacks HLA-DQ2 or HLA-DQ8.

Although the advent of sensitive serologic assays for tTG has led some authors to question the need for histologic confirmation of the diagnosis, long experience has shown that a flat proximal small bowel biopsy when the patient is on a regular diet, followed by a dramatic clinical response to a GFD, is the only reliable way to diagnose CS.

WHAT WERE THE CLASSIC RUBIN EXPERIMENTS IN CS?

In experiments using serial hydraulic biopsies of both CS patients on GFDs and healthy control subjects, Cyrus E. Rubin, MD, and his colleagues at the University of Washington (Seattle) proved histologically in real time that wheat, barley, or rye instilled into the ileum of celiac patients flattened the villi in 10 to 12 hours, and that the villi regenerated in approximately 48 hours.16,21-24 The effect was local, rather than systemic, because injury was confined to the downstream small bowel exposed to the grain. These studies also confirmed that childhood celiac disease and idiopathic sprue in adults were the same permanent genetic sensitivity to gluten, sharing similar proximal small intestinal histology, leading to the unified term celiac sprue.23 Instillation of oats did not cause a flat ileum but did cause some edema and neutrophilic infiltration of the mucosae of both CS patients and healthy controls. Based on this evidence, oats have been permitted in the diet of CS patients at the University of Washington for the last 50 years with no adverse effect.

Although the pathology literature is replete with the use of the descriptor villous atrophy that carries over from the early descriptions of celiac histology,25-27 the Rubin experiments provided evidence that celiac changes are not atrophy, but that they are rapidly reversible injury by removing gluten.

An additional important observation was the persistence of the histologic injury months to years after the patient was placed on a GFD. This led to the mistaken belief that there was a "lack of correlation between the histological severity of the proximal lesion and the patient's clinical status. While the former improves slowly, or not at all, the latter usually improves dramatically." 21 Rubin and colleagues demonstrated that histologic recovery begins distally and progresses proximally, that the distal jejunum and ileum recover far more rapidly than the proximal jejunum and duodenum, and that the extreme proximal and most severe injury may, rarely, never regress. Thus, the severity of the patient's clinical symptomatology correlates best with the length of intestine affected and by the extent of the injury rather than with the degree of severity of the injury in a proximal small intestinal sample.

IS THE HISTOLOGY OF CS SPECIFIC?

No, not all patients with sprue have mucosal changes on histology, and patients with sprue can manifest histologic changes of variable severity. In addition, histologic changes, including villous blunting and shortening and increased intraepithelial lymphocytes with surface epithelial injury, can be seen in other disorders.

WHEN CAN NORMAL MUCOSA HAVE FEATURES THAT MIMIC CS-LIKE HISTOLOGY?

The most commonly encountered "mimic" of CS is biopsy artifact. Villi may be amputated or inapparent due to poor orientation, denudation, or tangential sectioning. Assessment of small intestinal mucosal architecture requires the presence of at least 3 well-oriented crypts, with intact overlying mucosal surface and villi, if present. In addition to marked villous blunting or flattening, CS is characterized by surface enterocyte injury and increased intraepithelial lymphocytes, and these will not be seen in artifactually flattened biopsy material from healthy patients. Another common normal histologic finding that may be misinterpreted as CS is the reduced villous height that may occur over intramucosal Brunner glands (Figure 3); again, careful inspection confirms the absence of enterocyte injury or increased lymphocytes in biopsies from these sites in patients without CS.17

HOW SHOULD BIOPSIES BE HANDLED? HOW SHOULD OPTIMAL SPECIMENS BE OBTAINED FOR HISTOLOGY?

Always label the sites of the biopsies (duodenum, jejunum) and place them into separate containers to distinguish proximal duodenal from distal duodenal and duodenal-jejunal junctional biopsies. Although much has been written about the adequacy of small biopsies for diagnosis, one of the major causes for the overdiagnosis of CS is excessively small biopsies. It is much harder to orient small biopsies, and tangential artifact is common, leading to spurious shortening or absence of villi. Jumbo biopsies are both easier for the histology technicians to orient during embedding and easier for the pathologist to interpret than smaller biopsies. Jumbo biopsies are long enough to compensate for focal areas of tangential sectioning. The best way for the biopsies to be handled to minimize crush distortion and mangling is for the endoscopist or the endoscopic assistant to drop them from the forceps directly into liquid fixative. Modified formalin enhances histology; however, neutral-buffered formalin is adequate to permit identification of surface epithelial injury, intraepithelial lymphocytes, and architectural features.16

WHERE SHOULD PATIENTS BE BIOPSIED?

In CS, the most severe injury occurs proximally, at the sites of first intestinal exposure to the undigested and concentrated slurry of food that contains gluten-based proteins. Therefore, the duodenum is the region of the small intestine that is most frequently injured, and it may be the only site of histologic injury. The biopsy strategy should always include at least 1 biopsy sample of the duodenal bulb, because in some patients it is the only injured area.18 Additional biopsies should include at least 3 samples from the crests of the valvulae conniventes (primary and secondary mucosal folds) in the descending duodenum. Push enteroscopy into the jejunum is not necessary if numerous duodenal biopsy samples are taken, but endoscopic biopsies should ideally extend to the duodenal-jejunal junction.16

WHAT INTESTINAL DISORDERS CAN PRESENT WITH A CS-LIKE HISTOLOGY BUT ARE NOT RESPONSIVE TO A GFD AND ARE NOT CS?

By the mid-1960s, Rubin and his Seattle colleagues recognized other conditions that may present a histologic picture identical to sprue but that do not respond to a GFD. These can be distinguished clinically. For example, kwashiorkor improves after an adequate protein diet. Tropical sprue improves after broad-spectrum antibiotics and folic acid. Pseudo-obstruction may respond to broad-spectrum antibiotics or to release of the obstruction, if it is mechanical. Cow's milk or soy milk protein injury responds to removal of the injurious protein from the diet, and these 2 diseases are confined to children and disappear when the child matures. Acid-peptic injury, as in Zollinger-Ellison syndrome, responds to protein pump inhibitors or surgical removal of the gastrin-producing neoplasm. Flat infectious lesions in the small intestine are seen in children and disappear when the infection clears.16,17

WHEN DOES DIETARY TREATMENT FAIL IN CELIAC DISEASE? WHAT SHOULD WE CONSIDER WHEN A PATIENT ON A GFD REMAINS SYMPTOMATIC OR GETS SICKER?

If the diagnosis of sprue is correct, the major cause of treatment failure and of recurrent or persistent symptoms and signs is dietary lapse or inadvertent consumption of gluten. If a patient has been successfully treated in the past by a GFD, the recurrence or worsening of the patient's symptoms may be a dire development, heralding the development of a malignancy, the most common being malignant lymphoma, particularly the enteropathy-associated T-cell lymphoma of refractory CS type 2, a CS variant that develops resistance to a GFD. An alternate consideration is the evolution of collagenous sprue.28 A subset of patients never respond to dietary elimination of gluten, and these rare patients may have true refractory sprue, or they may need to have additional workup to exclude another disease.28-30

WHAT IS REFRACTORY OR UNCLASSIFIED SPRUE?

Refractory CS is defined as continuing flat intestinal injury pattern despite a prolonged completely GFD, or persistent symptoms despite partial response to GFD. In some patients there is an initial improvement with subsequent relapse while on GFD. Other patients have experienced no symptomatic, clinical, or histologic response to GFD, and these patients may be considered unclassified, or nonceliac, sprue necessitating consideration of other causes of the injury.

Recently, refractory sprue has been subclassified as refractory CS type 1, without aberrant T cells, and refractory CS type 2, with aberrant T cells. Patients with refractory CS type 1 do not respond to a GFD, but they typically respond to immunosuppressive therapy, such as prednisone and/or azathioprine, with successful treatment in 95% of patients. Patients with this disorder do not have evidence of clonally altered lymphocytes and do not progress to develop enteropathy-associated T-cell lymphoma, ulcerative jejunoileitis, or refractory CS type 2. Refractory CS type 1 may be an immunopathy.

Refractory CS type 2 affects approximately 75% of patients with refractory CS, and it appears to be more common in a subset of patients with HLA-DQ2 homozygosity. Patients with CS type 2 have had a positive anti-tTG serologic test with documented clinical response and negative serum tTG on a GFD, and over time their symptoms have recurred while on the diet. Refractory CS type 2 is a cause for concern and for a workup to exclude evolution to a monoclonal T-cell neoplasia. Until recently, enteropathic T-cell lymphoma was uniformly fatal, but there are recent reports of successful treatment with autologous hematopoietic stem cell transplantation.29,30

WHEN SHOULD PATIENTS BE TREATED WITH A GFD?

There is a recent trend to overtreat patients who do not have biopsy-proven CS, as well as a trend to diagnose CS in patients with minimal and nonspecific histologic changes. Patients with biopsy-proven classic sprue require a GFD for the rest of their lives. They respond clinically to a GFD and therefore strive to follow it completely.

Patients with atypical CS who have biopsy evidence of moderately severe or severe small intestinal injury are also encouraged to try a GFD for symptoms. Some respond; however, patients who do not experience symptomatic improvement are unlikely to adhere long-term to a GFD. For diet to be effective, it is essential for patients to work with skilled nutritionists who can provide information about the myriad occult ways that gluten is incorporated into apparently "nongrain" foodstuffs, including condiments and vitamins, and particularly as introduced in the refinement and manufacturing processes of canned and bottled foods. Patients with latent CS and DH do not require a GFD unless they develop malabsorption and have biopsy evidence of more than focal intestinal injury.

With deepest gratitude to Cyrus Rubin, MD, of the Division of Gastroenterology, Department of Medicine, University of Washington, Seattle. As my teacher, mentor, colleague, and friend, Dr Rubin shared his personal wisdom, his life work, and his slides, and he kindly reviewed this manuscript and advised me during its preparation. Dr Rubin is preparing 2 extensive lectures based on his experience and knowledge, the first on CS and the second on biopsy diagnosis of intestinal diseases that can be confused with it. They will be available online at the University of Washington Gastroenterology Division Web site. The reason for Dr Rubin's effort to report his more than 50 years of work on CS and to correlate his early observations and findings with more recent ones is that parts of the history are pertinent but forgotten. As George Santayana said, "Those who do not learn from history are doomed to repeat it."

© 2008 College of American Pathologists Provided by ProQuest LLC. All Rights Reserved.