Duodenal bulb biopsies remain relevant in the diagnosis of adult celiac disease

Celiac disease is an immune mediated enteropathy that occurs in some genetically susceptible individuals upon exposure to dietary gluten. The prevalence of celiac disease has increased due in part to better diagnostic and screening practices^1,2. Undiagnosed or under-treated celiac disease places these individuals at significant risk for complications, thereby necessitating accurate and early diagnosis^2–4.

Current estimates place the prevalence of celiac disease to be approximately 1% in many parts of the world. Prevalence is known to be higher among individuals with other autoimmune disorders, or a family history of celiac disease⁸ and some studies have suggested it to be higher among individuals with certain gastrointestinal complaints⁵, irritable bowel syndrome⁶, or refractory gastroesophageal reflux⁷. The diagnosis of celiac disease is based on positive serological tests, characteristic histologic changes in the small intestinal mucosa and response to a gluten free diet^1,9,10. It is currently understood that celiac disease can have a patchy distribution¹¹ throughout the small intestine and therefore several studies in adult patients have documented the utility of combining mucosal biopsy sampling from the first and distal portions of the duodenum to increase diagnostic yield and aid in identification of the most severe lesion^11–14. The most recent celiac disease guidelines published by the American College of Gastroenterology and British Society of Gastroenterology both recommend obtaining mucosal biopsy samples from the first and second portions of the duodenum when a diagnosis of celiac disease is being considered^9,10.

In the current issue of Clinical Gastroenterology and Hepatology, Stoven et al.¹⁵ from the Mayo Clinic report on their retrospective evaluation of the diagnostic utility of mucosal biopsies from both the first and second portions of the duodenum in a population with a low pretest probability for celiac disease. From across all the Mayo Clinic sites they retrospectively identified 737 patients who had underwent a clinically indicated upper endoscopy between January 1, 2011 and December 31, 2011, and also had mucosal biopsies obtained from both the first and second portions of the duodenum sent separately for histologic analysis. All abnormal histology reports were retrospectively reviewed by a single expert gastroenterology pathologist. The major analysis of this cohort excluded 33 individuals with known celiac disease and/or dermatitis herpetiformis and 25 individuals with positive serological tests prior to endoscopy in an effort to focus on patients with a low pretest probability of celiac disease. This is in contrast to prior studies that have reported on the enhanced diagnostic utility of combining mucosal biopsy samples from both the first and second portions of the duodenum to increase diagnostic yield in high pretest probability cohorts (Table 1)^11–14,16.

Table 1.

Summary of Available Adult Cohort Studies Evaluating the Diagnostic Utility of Duodenal Bulb Biopsies for the Diagnosis of CD

Study	Year published	Newly diagnosed CD cohort (% with villous atrophy limited to the first portion)	Established CD cohort (% with villous atrophy limited to the first portion)	CD cohort with villous atrophy limited to the first portion (% of total CD cohort)	Location of most severe lesion identified in the first portion of the duodenum, %	Entire cohort (% of CD with villous atrophy limited to the first portion)	Biopsy samples orientated at endoscopy or embedding stage
Evans et al	2011	126 (8.7)	85 (14.1)	23 (10.9)	76.9	461 (5.0)	Embedding
Hopper et al	2007	53 (1.9)	0	1 (1.9)	55	53 (1.9)	Embedding
Cammarota et al	2007	49 (8.2)	18 (38.9)	11 (16.4)	N/A	67 (16.4)	Endoscopy
Nenna et al	2012	43 (2.2)	0	1 (2.2)	16.3	43 (2.2)	Endoscopy
Gonzalez et al	2010	15 (26.7)	25 (4)	5 (12.5)	N/A	80 (6.3)	Embedding
Kurien et al	2012	28 (17.9)	0	5 (17.9)	N/A	77 (6.5)	Embedding
Vogelsang et al	2001	12	9 (22.2)	2 (9.5)	N/A	51 (3.9)	No information
Brocchi et al	2005	1	0	1	N/A	N/A	No information
Stoven et al	2016	16 (6.25)	28 (7.1)	3 (6.8)	N/A	737 (0.4)	No information

CD, celiac disease; N/A; not applicable.

Within the low pretest probability for celiac disease cohort as defined by Stoven et al. they report a new diagnosis of celiac disease in 16 individuals, which represents 2.4% of the total cohort. The additional 1.4% diagnosis rate above the national average, likely demonstrates the effects of sampling bias of large tertiary referral centers and increased prevalence of celiac disease in individuals with gastrointestinal symptoms.

To our knowledge the Stoven at al. study represents the largest adult low pretest probability cohort that was evaluated in an attempt to determine the diagnostic utility of combining biopsies from both the first and distal portions of the duodenum. However, due to the studies’ design the celiac disease diagnosis cohort was significantly smaller than that of the Evans et al. study which included 126 newly diagnosed and 85 established cases of celiac disease from a total cohort of 461 patients¹². The dramatic difference in the number of newly diagnosed celiac disease cases between these two studies serve to highlight the clinical utility of serologic based screening methods for celiac disease. However, serologic screening remains imperfect, given that approximately 10–15% of all individuals with celiac disease are seronegative¹⁷.

Of the 16 newly diagnosed celiac disease cases from the Stoven at al. study, only a single individual had villous atrophy limited to the first portion of the duodenum, which represents 0.15% of the total cohort, but 6.25% of the newly diagnosed celiac disease cohort. In the high pretest probability cohort of the Evans et al. study, 8.7% of the newly diagnosed and 14.1% of the established celiac disease cases had villous atrophy limited to the first portion of the duodenum; if biopsies had not been obtained from the first portion of the duodenum, 11 newly diagnosed cases and 12 cases of ongoing mucosal injury in established celiac disease would have been missed. In the Stoven et al. low pretest probability cohort, 1 case of newly diagnosed and 2 cases of ongoing mucosal injury in established celiac disease patients would have been missed.

How rare is villous atrophy limited to the first portion of the duodenum in individuals with celiac disease? The answer depends on the population under consideration. For example, if the denominator includes all individuals who undergo a clinically indicated endoscopy for some gastrointestinal complaint, the number is small (i.e. 0.15%) but not zero as demonstrated by the Stoven et al. study. However, if the denominator is all individuals newly diagnosed with celiac disease, the number appears to range from ~ 2 to 27% with current data, refer to Table 1. Therefore, within the cohort of 16 individuals with newly diagnosed celiac disease reported by Stoven et al., the single individual with villous atrophy limited to the first portion of the duodenum, represents 6.25% of all newly diagnosed individuals. Furthermore, 2 individuals of the 29 with known celiac disease had villous atrophy limited to the first portion of the duodenum, representing 6.9% of the established celiac disease cohort. In total, 3 individuals with celiac disease of the total cohort of 737 had villous atrophy limited to the first portion of the duodenum, representing 0.4% of the total cohort or 6.8% of the celiac disease cohort (44 individuals), which includes both new and established cases of celiac disease. When we consider the “cost”, whether it be health expectancy, life expectancy or monetary costs, of an inaccurate, a missed, or delayed diagnosis, it could be argued the small incremental benefit from combining first and second portion duodenal biopsies should not be overlooked^2,3.

In the prior studies evaluating the diagnostic utility of combining duodenal bulb and distal biopsies, the numbers are reported relative to the number of celiac disease cases rather than the total cohort, thereby providing the percentage of celiac diagnoses that depend on identifying villous atrophy in biopsy samples from only the first portion of the duodenum. It is not clear why the authors chose to evaluate the cases of villous atrophy limited to the first portion of the duodenum to their entire cohort rather than the more valuable comparison to the celiac disease cohort as this dramatically alters the interpretation of their results. In fact their findings of an enhanced diagnostic yield specifically for the diagnosis of celiac disease is consistent with the prior published studies from adult cohorts, refer to Table 1.

However, what Stoven et al. demonstrates by analyzing this low pretest probability cohort, is that there is only a very small incremental enhanced diagnostic yield for this biopsy strategy in this low risk cohort. The authors argue that given the additional cost of this biopsy collection strategy, biopsies should not be obtained from the first portion of the duodenum of low pretest probability individuals. The validity of this conclusion is dependent on whether there is a necessity to separate biopsies from the first and second portions of the duodenum prior to sending for histologic analysis. It does seem that in centers with expert gastroenterology pathologists there is no longer a need to separate biopsies obtained from the duodenal bulb and more distal areas as the unique histology of the bulb is no longer a source of diagnostic confusion¹⁸. Furthermore, there is no longer a need for endoscopy technicians to orientate the samples prior to formalin fixation^11–14, thereby increasing the speed at which samples can be collected during the endoscopy procedure. In our center, there is a flat rate pathology charge for each bottle containing six gastrointestinal mucosal biopsy samples, therefore at least in our center there is not an increased cost to collecting samples from both the duodenal bulb and more distal areas.

How much do additional mucosal biopsies from the first portion of the duodenum cost? The answer depends on whether one is considering the cost of endoscopy and pathology alone or the much greater total costs of an inaccurate, missed, or delayed diagnosis. Individuals who continue to have symptoms will continue to seek medical care, many becoming desperate enough to seek care by multiple physicians and through emergency services, driving up their costs for care. These individuals often undergo more diagnostic testing and treatments in search of a diagnosis and resolution of their symptoms. It has been reported that 5% – 13.6% of individuals eventually diagnosed with CD have undergone previous non-diagnostic endoscopy for evaluation of their symptoms^19,20. While we acknowledge the limited incremental benefit of this biopsy strategy in low pretest probability individuals, it is often difficult to discern an individual’s actual risk at the time of endoscopy, especially in open access endoscopy units. In our opinion, it remains prudent to continue to obtain biopsies from the first and distal portions of the duodenum until a better pre-procedural risk assessment can be determined.

It would have been of benefit for Stoven et al. to report the clinical indication for each of the newly diagnosed celiac disease cases in order to help us better understand the clinical features that define increased risk for celiac disease from a low pretest probability cohort. As this information may then allow for the development of better pre-procedural screening modalities that could help us determine which patients will truly benefit from this biopsy strategy.

The diagnostic benefit of this biopsy strategy is well documented within the pediatric celiac disease literature^21–23. Furthermore, previous work has demonstrated that simply increasing the number of duodenal biopsies increases the diagnosis of celiac disease^11,24, especially when duodenal bulb biopsies are included¹¹. Therefore, at the present time, we believe that there is clinical utility in obtaining at least 4–6 mucosal biopsy samples from both the first and distal portions of the duodenum in a single formalin bottle not just to enhance the diagnosis of celiac disease but also for the potential to diagnose other diseases and provide the best care for our patients. We applaud Stoven et al. for this intriguing study and continuing this important discussion related to the diagnosis of celiac disease.

Abbreviations

CD

Celiac disease