Article Text

Download PDFPDF

Invasive or non-invasive methods for the diagnosis of subclinical coeliac disease?
  1. A Tursi1
  1. 1Department of Emergency, “L Bonomo” Hospital, Andria (BA), Italy
  1. Correspondence to:
    Dr A Tursi, Galleria Pisani, 4, 70031 Andria (BA), Italy;
  1. C B Pearce2,
  2. H D Duncan3,
  3. D Sinclair3,
  4. D N Poller3
  1. 2Department of Gastroenterology, Queen Alexandra Hospital, Portsmouth, UK
  2. 3Queen Alexandra Hospital, Portsmouth, UK

Statistics from

I read with great interest the letters published in Gut on the role of small bowel biopsies and antiendomysial antibodies (EMA) as methods for diagnosing coeliac disease in patients with iron deficiency anaemia (IDA) (Gut 2001;49:595). In particular, Pearce et al seem to prefer EMA instead of histological evaluation because of the considerable resource implications for histology departments, and Scott seems to agree with Pearce's affirmation even if he takes into consideration the important role of histological evaluation. However, I do not completely agree with the conclusions of these authors.

Firstly, I would debate the role of non-invasive methods in diagnosing subclinical/silent coeliac disease. In my experience, IDA is the main indication of subclinical forms of coeliac disease1 and thus it should be taken into consideration in clinical practice. Although EMA are a well known hallmark of coeliac disease and the prevalence is more than 90% in classical forms of coeliac disease, our study and other recent studies clearly showed that the prevalence of EMA (as well as AGA) is lower than expected in clinical practice,2–4 probably due to the high prevalence of slight histological lesions in these patients (Marsh I-IIIa lesions according to the Marsh classification5). In contrast, the sorbitol H2-breath test (H2-BT) seems to be more effective than EMA in diagnosing this form of coeliac disease, probably because of a better correlation with slight histological lesions.6 In the light of these experiences, sorbitol H2-BT may be a good alternative to small bowel biopsy in identifying coeliac disease in patients with IDA but unfortunately this is not always true. In fact, in clinical practice, it is easy to observe patients with IDA EMA− and a negative sorbitol H2-BT test who show slight histological lesions (Marsh I-II type lesions) with disappearance of IDA and improvement in histology after a gluten free diet (GFD). In these cases the use of non-invasive methods (such as EMA) may be a serious mistake as we may run the risk of not identifying hidden coeliac disease. These experiences are very important and should be considered in the cost/benefit ratio of diagnosing coeliac disease.

Other important points are the patchiness of the disease, difficulties for pathologists in obtaining biopsies orientated sufficiently, and the cost of small bowel biopsy. Pearce et al are not in favour of biopsy. Firstly, many patients suspected of having coeliac disease have upper gastrointestinal endoscopy as an initial investigation which provides an opportunity to perform a biopsy on the second part of the duodenum. Although routine biopsies in all patients undergoing endoscopy would have significant resource implications, endoscopic abnormalities of the second portion of the duodenum associated with coeliac disease have been described,7 and these may be used to select patients for biopsy, even if recent studies have re-evaluated the accuracy of endoscopic markers of the disease.8 Secondly, multiple biopsy samples obtained from the second portion of the duodenum overcome the problem of the patchiness of the histological lesions (and we routinely take at least six endoscopic biopsies from the descending duodenum). Also, the pathologist's expertise in the Marsh classification of histological lesions in coeliac disease may certainly overcome the problem of incorrectly orientated biopsies. Thirdly, I disagree about the excessive expensive of histological evaluation. In Italy the cost of histological evaluations from a single seat (in this case descending duodenum) is about $12.40: I do not believe that this is an excessive additional cost to a routine upper gastrointestinal endoscopy.

In light of these consideration, the final question is: should we always perform small bowel biopsies in patients with IDA or other pathologies hiding a subclinical/silent form of coeliac disease? I believe that small bowel biopsy remains the gold standard in diagnosing subclinical forms of coeliac disease (such as IDA), even if the sorbitol H2-BT test is promising as a non-invasive method9: the sorbitol H2-BT test seems to be more promising in the follow up of the disease after a GFD (unpublished data). It remains to be determined whether serological testing for antibodies to antitissue transglutaminase improves the diagnosis in cases of mild mucosal lesions. I think that patients at high risk for coeliac disease (such as those with unexplained IDA) should always undergo duodenal biopsy. The costs could be quite high due to the high number of endoscopies that need to be performed but is cost/effectiveness if we consider the significant proportion of patients with coeliac disease who may be missed if screened by serology alone.


Authors' reply

We would like to thank Dr Tursi for his comments. We suggest that Dr Tursi's and our opinions on the use of antiendomysial antibodies (anti-EMA) and duodenal biopsy (DDB) in the management of coeliac disease are convergent.

We agree that at present anti-EMA can be a flawed diagnostic test and its results must be treated with caution. However, we believe that there is potential for close to 100% sensitivity and specificity, as has been obtained in some laboratories.1,2 This highlights the fact that there is a difference in accuracy between centres due to the nature of the test, as discussed in our first letter (Gut 2001;49:595). The reasons for these differences are not always easily addressed and may be related not just to the laboratory itself but also to the local population. In centres that have 100% accuracy in anti-EMA testing for coeliac disease, it is clear that its use in the management of coeliac disease will be different in centres that achieve much lower accuracy, some of which are cited in Dr Tursi's letter.

In our letter we were attempting to raise the question as to whether DDB should be obtained in all patients with suspected coeliac disease, and to suggest that if anti-EMA is as accurate as pooled published data, suggesting sensitivity and specificity of 94% and 98%, respectively,3,4 theoretically it could be used as a firstline investigation. To put this in context, at present in our biochemistry laboratory, in a hospital serving 550 000, 400 anti-EMA tests are performed per month. Currently at our hospital there is not the capacity to perform 400 extra duodenoscopies each month, and if there were a clinical need, the department would have to be reorganised to do so. Anecdotally, other regions of the UK carry out similar numbers of serological tests for coeliac disease diagnosis.

At present in our patient population, anti-EMA is therefore used as a screening tool, and if positive, patients are offered DDB. This is likely to be the case across the country. Our regional teaching centre has formalised this: tissue transglutaminase (tTG) is used as a screening tool, and anti-EMA and subsequently duodenal biopsy are offered to selected patients.5 We believe that this approach has merit, although the routine use of tTG lacks confirmation in the clinical setting.

We agree that currently DDB is the gold standard investigation and that at this moment in time, on the basis of current clinical evidence, should be obtained in all patients with positive serology.

Although we are aware of the benefits of other diagnostic methods, such as sorbitol H2 breath testing (H2 BT), this is still a “resource intensive” test; a quick, cheap, and accurate screening test is required. If sorbitol H2 BT, when thoroughly evaluated, was found to be more effective than DDB in the diagnosis of coeliac disease, we would welcome its introduction.

Better education of primary care physicians in anti-EMA testing may also be required. We have become aware that in our region only a minority of patients with positive anti-EMA are biopsied. This situation should be corrected; it has arisen as a result of anti-EMA being available on a direct access basis to primary care physicians. We wonder if the difference between published guidelines, and what is currently practicable, has contributed to some of the confusion.

On the subject of duodenal biopsies, we believe, as discussed in our letter, that at some point in the future its validity as a gold standard should be reviewed. However, we accept it must remain as such until other methods are thoroughly evaluated, hopefully in well designed controlled trials. If performing an endoscopy for iron deficiency, we would certainly accept that it makes sense to biopsy the duodenum.

In summary, the numbers of patients who are tested at present using anti-EMA means that it has effectively become a screening tool for suspected coeliac disease. However, as the gold standard, we believe that DDB should be offered to all anti-EMA positive patients to confirm the diagnosis and other patients in whom the diagnosis is in doubt, which could include patients who are iron deficient but with negative anti-EMA. We suggest that guidelines for the management of coeliac disease should reflect this.


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.