Introduction

The beginning of the 21st Century brought with it a profound change in knowledge about celiac disease, thus shaking the apparent certainties that those knowledgeable of this enigmatic disease had compiled over past decades. Just as Catassi had correctly foreseen a few years earlier, only a minority of the world population of celiac disease sufferers were clinically recognized, while the vast majority of those who suffer from the disease remained underdiagnosed [1], [2].

Studies conducted in various parts of the world (albeit with a few geographical exceptions) very faithfully reproduced the figures Catassi published in 1994, and which were obtained from a study he had performed in Italy [1], [2], [3]. Thus, it was possible to establish that the prevalence of celiac disease is considerably higher than what had been previously thought, especially due to the high proportion of clinically atypical cases, along with clinically silent forms of the disease [3]. Today it is believed that between 0.6 and 1.0% of the population in the studied geographical zones suffer from celiac disease [2], [3], [4]. The prevalence is almost twice as high in females than in men, and it is higher among people who have a first degree relative who suffers from the disease (between 10 to 15%) or who suffer from type 1 diabetes (between 3 to 16%) as well as certain autoimmune diseases such as Hashimoto’s thyroiditis, Sjögren’s syndrome, IgA deficiency and liver autoimmune diseases [3], [4], [5]. A corollary of these results has been the publication of guidelines bearing suggestions with respect to which individuals and risk groups should be assessed for possible celiac disease [4], [5].

Predisposing factors

Gluten consumption and consumption of similar proteins contained in wheat, rye and barley are necessary for developing clinically active celiac disease, along with possessing the DQ2 and DQ8 haplotypes (histocompatibility antigens). However, as factors they are not sufficient by their own right when it comes to disease development [3]. At least 40 human non-leukocyte antigen genes appear to confer a predisposition toward the disease. These genes are mostly involved in immune response or in disease predisposition towards developing inflammatory states [3]. Therefore, celiac disease is considered as a genetically complex disease.

There is evidence suggesting that the frequency of celiac disease is growing in many countries, perhaps due to the increase in wheat flour and its derivatives, the so-called “westernization” of diets, along with other factors which have not been fully identified. Of course, the number of celiac disease sufferers is increasing, owing to greater disease awareness. This has led to the development of active search strategies [2].  There are significant variations in celiac disease prevalence rates among different European countries, albeit without a clear explanation. It is very likely that certain environmental factors are responsible for these differences [3].

Clinical presentation

Accordingly, up until a few decades ago, celiac disease was a condition that was considered to afflict the pediatric and Caucasian population. It is now being more and more frequently diagnosed in individuals from all over the world and increasingly among adult patients [2], such as the case in Chile [6]. Furthermore, celiac disease was seen as an exclusively gastrointestinal affection of various presentations [7]. Today it is seen as a multifaceted entity in which gastrointestinal symptoms can vary from mild to severe or even be non-existent, and include a wide array of extraintestinal manifestations [7].

The reason behind the remarkable variations in the clinical presentation and impact of celiac disease from one patient to another is not completely clear. These variations can range from disease devoid of symptoms to full-blown clinical manifestations. Apart from the aforementioned genetic factors, it is necessary to keep in mind variables such as quantity and nature of the gluten which is consumed, breastfeeding period duration and the age at which children are introduced to gluten [2], [8], the composition of the gut flora [9], certain external immune response polarization  determiners (Th1 vs. Th2) such as infections, drugs, biological products, pregnancy, and surgeries, among others [10], [11], [12], [13]. 

Diagnostic principles

Diagnostic confirmation of celiac disease is based on the presence of a typical enteropathy which is dependent on gluten consumption, as well as a histological exam on the intestinal biopsy [2], [3], [4]. Serological tests are useful for identifying the population of patients for whom a biopsy is recommended [4], [5], [8]. The diagnostic process is not without its flaws. This process starts at clinical suspicion of celiac disease in certain patients, then conducting the proper screening tests and subsequently interpreting them in order to select those who will undergo an intestinal biopsy and finally, to properly characterize the histological findings that will lead to a final verdict.

Some authors have suggested that there is a certain lack of critical mass among clinical practitioners, pathologists, and laboratory personnel who are experienced and familiarized enough with the diagnostic complexities of the disease [14]. Diagnostic errors stemming from this situation, such as under and overdiagnosing, can have a profound and lasting impact on patients as well as their families. Since celiac disease constitutes a chronic entity, lasting the entirety of a patient’s life, diagnostic error in one sense or another are clearly harmful to the patient [13],[14].

Differential diagnosis

In a broad sense, celiac disease shares specific symptoms or groups of symptoms with many other illnesses, on account of its multifaceted characteristics. However, upon analyzing the patient in depth, several of these other diseases can be ruled out with relative ease. Among the differential diagnoses [13], [14], [15], a few chronic intestinal inflammatory diseases can be mentioned, as well as food allergies, irritable bowel syndrome , various causes of anemia, enteroparasitosis, fermented oligosaccharide/disaccharide/monosaccharide intolerance, HIV infection, tuberculosis, lymphoma and other forms of cancer, graft versus host disease, radiation damage, anorexia/bulimia nervosa, fictitious disorders, as well as others. Moreover, with pediatric patients it is necessary to consider infectious diseases (viral, bacterial, or parasitic enteritis), metabolic diseases, rare congenital absorption defects, congenital immunodeficiencies, cystic fibrosis, structural malformations of the digestive tract, eosinophilic gastroenteritis, malnutrition due to insufficient intake, among others.

Several entities can develop presenting variable degrees of villous atrophy and intestinal wall inflammatory manifestations [15]. Table 1 offers a list of the most representative causes.

Diagnostic error possibilities

Among the factors contributing to difficulties and errors in diagnosing celiac disease, are the following:

Misinformation regarding celiac disease
There is abundant misinformation that persists regarding the distinct characteristics of celiac disease and its diagnostic criteria in vast sectors of medical practice. Overall, this situation has led to a low threshold in diagnostic suspicions, and thus a low threshold of disease recognition. This situation occurs particularly at the primary healthcare level, yet also –though to a lower degree – at levels of greater medical complexity.

It is clear that patients with less typical forms of celiac disease are the most affected by this misinformation, seeing as they are for the most part undiagnosed [17]. For different reasons, the opposite situation can occur [16]. In an environment with insufficient expertise, public pressure and the growing media portrayal of the disease through non-specialized press outlets, this can lead those who handle limited specific information concerning the problem to systematically lean towards making diagnoses based on trends and media pressure, even when the required criteria are absent.

Problems with the serological tests

1. General problems. Albeit serological screening tests for celiac disease have considerably improved the ability to identify those who may suffer from the illness, they still are not sufficiently reliable for confidently diagnosing a condition which requires a life-time of adhering to a strictly gluten-free diet[4],[18],[19].
   These serological tests attain their maximum quality and efficiency as they appear in renowned research laboratories. Nevertheless, with the gradual transfer of technology toward other laboratories, they may suffer a drop in their performance characteristics [20]. For a test to continue being totally valid, it is necessary to conduct a thorough and constant follow-up of its performance, by monitoring the behavior of the internal control mechanisms over time. Otherwise, the test’s performance characteristics begin to progressively deteriorate. This is often the case with commercial celiac disease serological tests, which by not having been subject to standardization and by exhibiting a marked variability in its results from one laboratory to another, have almost always suffered a decrease compared to their original sensitivity and specificity characteristics. [3],[4],[13],[16],[20].
   The population which is selected for the study on initial test validation may be markedly different from that which is studied in the community using clinical methods. This can lead to non-representative results [4], [5]. Furthermore, there is quite limited data regarding the serological profile of celiac disease in children under the age of 5. Thus, these cases cannot necessarily be extrapolated based on the compiled experiences of subjects from older age groups [4].
   The performance of an antibody test within a certain geographical medium depends on the characteristics of the evaluated patients, such as age, genetic predisposition, pre-test probability of the disease which is being researched, the phase of the disease, and the amount of ingested gluten. These factors therefore must be kept in mind when interpreting antibody test results, be they positive or negative. It is equally important to set the best possible cut-off limits, which are also highly probable causes for error in either direction [22], [23].
2. Conventional Anti-gliadin antibody test. After nearly two decades of regular use with the conventional anti-gliadin antibody test in celiac disease serological tracing, the balance has been disappointing. This test has a relatively low performance, both in sensitivity and specificity, which can lead both to under and overdiagnosing celiac disease. For example, there have been cases of false positives in subjects who suffer from a variety of different clinical entities[4],[5],[12],[19]. Therefore, based on current evidence, the pertinent international societies no longer recommend this test. The most novel situation regarding antibodies directed against deamided gliadin peptides[13],[17] has been different, owing to its greater performance [17] though without exceeding the anti-transglutaminase antibodies [24].
3. Anti-endomysial antibodies. The anti-endomysial antibodies test, which is generally considered to be of a higher performance than the aforementioned test, is nevertheless affected by practical problems that can limit its application or interpretation. On the one hand, its relative complexity –coupled with being operator-dependent– make it less feasible in less sophisticated mediums, apart from making it more expensive [2], [25]. It also loses accuracy in children under the age of 2 [4],[5]. Furthermore, if the protocols for running the test are not strictly adhered to, ambiguous results can be obtained due to systematic detection of false positives. For example, this has been known to occur if the 1/5 dilution [26] is not used as the starting point during initial preparation for the sample[26], with a more concentrated solution being used instead. This has often been the case in Chilean laboratories.
4. Anti-transglutaminase antibodies. Measuring anti-transglutaminase antibodies is perhaps the most reliable serological test for celiac disease currently available, due to its excellent level of standardization, along with its performance characteristics such as sensitivity and specificity.  However, general problems regarding test application detailed hereinabove in item 1 must be kept in mind upon analyzing a given result.
   A special situation which can also lead to error are those who have selective IgA deficiency (more common among celiac disease sufferers than in the general population), wherein both the measurement of anti-transglutaminase antibodies and other usual serological testing can lead to false negatives, since they are based on IgA class antibodies [2], [6]. In this case, there are various strategies that can be followed in order to overcome the potential problem. One of them consists of simultaneously measuring serum IgA levels in each individual who is subjected to serological testing for celiac disease. If IgA is undetectable, the next step is to directly proceed to performing an intestinal biopsy, when there is a sound suspicion of celiac disease. 
   An additional instance of error occurs when screening tests are conducted using a commercial reactant on a drop of blood. These tests are often performed by the patients themselves [27], which can lead to false negatives [2].
5. Progression in the serological behavior of celiac disease. There is evidence that in some children suffering from type 1 diabetes or Down syndrome, as well in first degree relatives of celiac disease sufferers who initially obtained negative results for their celiac disease serology tests, there can be a change over the years, and their biopsies may turn out with the typical profile for celiac disease [4]. If the natural progressive nature of some celiac disease sufferers is not taken into account, these patients may be overlooked. It is therefore recommended that subjects falling into any of these categories be subject to new studies later on [4].

Errors in the histopathological interpretation of celiac disease
Another fertile ground for error can be found in the transcendental area of histological interpretation. In other words, nothing less than within the realm of diagnostic confirmation of celiac disease, it’s Gold Standard. Histopathology for the disease is characterized by the increase in intra-epithelial lymphocytes, crypt hyperplasia, and small intestine villous atrophy.

1. Obtaining and processing the biopsy samples. The changes described above may be parceled out in terms of distribution, and can vary in intensity throughout the intestine. Faced with a scenario which does not appear in a black and white manner, and where the alterations flow across a broad gradient, the probability of diagnostic as well as technical and conceptual error is high. With all this in mind, it is vital that the processing of the samples be extremely thorough and bound to protocol, and that the biopsy interpretation be precise and reproducible both among different pathologists and in diverse clinical practice scenarios [28].
   Due to the sometimes parceled out nature of the lesions, if the person who conducts the endoscopy and draws the biopsy sample  randomly picks out an area of the intestine where the lesions are not as pronounced, there is a high probability for diagnostic error [5],[7],[13],[16],[18],[25],[28]. It is therefore important to take multiple endoscopic biopsy samples –ideally six– from the small intestine, including both the duodenal bulb and the second duodenal portion. The samples must also be large enough to make it possible for proper orientation, so that the cuts will not be tangential. In the case of the latter, misinterpretations can occur on the part of the pathologist [25]. Evidently, diagnostic studies for possible celiac disease must be performed while patient is consuming gluten, otherwise the obtained results will be erroneous and difficult to interpret.
2. Marsh classification. The Marsh score system, originally for purely research-related purposes [29], has been widely adapted for semi-quantitative evaluation of intestinal histology for celiac disease, and therefore as a diagnostic tool [29]. The Marsh criteria, which were later modified by Oberhüber [30], and that classify intestinal alterations in five stages, are useful for conceptualizing the existence of a histological anomaly gradient which reveals the remodeling of the small intestine mucosa in celiac disease. This evaluative system has broadened the histopathological view over this disease for the treating physician. Nevertheless, extreme caution should be exercised towards not labeling as indicative of celiac disease any non-specific change detected in the biopsy, particularly if the appearance of the sample falls under Marsh category 1 [28],[31],[32],[33],[34].
   It is well known that changes consistent with category Marsh 1 may be present due to a variety of causes, the most noteworthy of which are Giardialamblia or Helicobater pylori infections, non-steroidal anti-inflammatory drug use, and several other situations [13]. In fact, the majority of biopsies where such changes are detected belong to patients with causes of intra-epithelial lymphocyte increases  other than from celiac disease [13],[31],[32],[33]. On the other hand, 5% of the general population exhibits this histological characteristic. Regrettably, not all of the medical community is aware of this fact, and the diagnosis for celiac disease is automatically reached on a relatively regular basis in patients whose biopsies showed Marsh score of 1, lacking the due considerations towards the pertinent differential diagnoses.
3. Systematic variability among different pathologists. Systematic variability has been reported concerning the diagnosis of celiac disease among different pathology departments, which might explain the diagnostic errors which have come up, particularly in underdiagnosing this disease in some places [28]. These errors appear to be occurring more frequently in public hospitals and private centers rather than in academic institutions [28].

An Italian university center reported that 20% of the patients who were referred to tertiary care for a previously diagnosed case of celiac disease had come back negative upon proper re-evaluation [33]. The main errors causing this overdiagnosis were wrongly attributing minimal changes to celiac disease, along with intra-epithelial lymphocyte increases (this topic has been discussed hereinabove), and coming to a diagnosis based solely on serological tests, and without the subsequent biopsy for confirmation [33].

Researchers from Argentina have also reported a considerable degree of overdiagnosis owing to problems with biopsy assessment [16]. Discrepancies often occur among the diagnoses made by expert pathologists and non-experts. The aforementioned study detected a divergent level of diagnoses between these 2 practical scenarios, reaching 46% of the studied cases, and leading to significant overdiagnosing of celiac disease in patients assessed in the community [16]. The poor quality of the histological samples and their impact on diagnostic error has been a particularly repeated and worrying situation. Twenty-eight percent of the serological results had been inconsistent with the histological findings.

Conclusion

When faced with a patient exhibiting the characteristic clinical profile for celiac disease, with some of its less typical manifestations, a combination thereof, or if the patient is part of the at-risk groups for celiac disease, it is prudent to start by measuring tissue anti-transglutaminase antibodies or, failing this, anti-endomysial antibodies. When conducted in reliable laboratories, these tests hold reasonably high levels of sensitivity and specificity. If the results are positive, it is currently a formal indication that an endoscopy be performed, together with a small intestine biopsy. If the sample is characteristic of celiac disease, a gluten-free diet should be permanently prescribed for the patient. Several professional and scientific societies have published guidelines for diagnosing and managing this disease.

Diagnostic error for celiac disease –both under as well as overdiagnosis– can have profoundly negative effects for patients and their families. They also contribute toward undermining public policies created in favor of chronic patients such as these. There are numerous areas which are prone to errors in the deductive process itself, due to relative lack of knowledge concerning the distinctive characteristics of celiac disease on the part of the medical community. Furthermore, in the serological testing, there are several causes for mishaps, the majority of which are due to technical reasons. Finally, in the histological confirmation phase, inappropriate biopsy sample processing –coupled with the variability in the histopathological report from one pathologist to another– are other causes for diagnostic error. Everything expressed herein highlights the need to keep working toward obtaining a higher degree of reliability from the laboratories that process the serological screening tests, as well as the need for standardization among pathologists regarding the histopathological report.

Notes

Potential conflicts of interest
The author has completed the ICMJE uniform disclosure form translated into Spanish by Medwave and declares no relationships or activities that could appear to have influenced the submitted work. He also declares that until 2011 he was honorary vicepresident of the Convivir Foundation (charity) for celiac disease.

Esta obra de Medwave está bajo una licencia Creative Commons Atribución-NoComercial 3.0 Unported. Esta licencia permite el uso, distribución y reproducción del artículo en cualquier medio, siempre y cuando se otorgue el crédito correspondiente al autor del artículo y al medio en que se publica, en este caso, Medwave.