Statistics from Altmetric.com
Although mortality of patients with coeliac disease has being studied for >25 years, it is not yet clear whether it is higher than in the general population. In the first years after the diagnosis of coeliac disease, some studies showed a definitely increased mortality rate that, however, decreases over time. On the other hand, other studies showed that this was not always the case (reviewed in ref 1). All these studies are difficult to compare because they differ widely from each other for epidemiological and clinical reasons. To explain these apparently contrasting results, we suggested that mortality rates were different because patients with coeliac disease were different, especially regarding the type of clinical presentation and age at diagnosis of coeliac disease. We also hypothesised that the quantity of gluten consumed before the diagnosis of coeliac disease could have a role in inducing complications and thus influencing mortality rate.1
In the last few years, mortality of patients with coeliac disease was the focus of studies that were not based on clinically collected patients with coeliac disease but on data from patients with coeliac disease recorded in massive databases. The study by Sultan et al2 is one of them. This and other studies show that overall mortality of patients with coeliac disease is not increased compared with the general population.2–5 Does this mean that the previous studies showing an increased mortality rate for patients with coeliac disease were actually all wrong? As usual, the truth is likely somewhere in the middle.
Although these database studies benefit from a huge statistical power, we cannot be sure that they represent the ideal tool for studying this issue. First of all, the accuracy of these diagnoses is a reason for concern. Nowadays, misdiagnoses of coeliac disease are very common and confirmation of the initial diagnosis of coeliac disease is required before making a diagnosis of refractory coeliac disease.6 ,7 On the other hand, patients with coeliac disease are well known to have a significant diagnostic delay, most often because they were labelled as patients affected by irritable bowel syndrome.8 Since all these database studies take into account thousands of patients with coeliac disease and controls, is it not possible that a considerable proportion of patients with coeliac disease were included among the controls and non-coeliac subjects were included among patients with coeliac disease? And is it not possible that mortality rates of patients with coeliac disease were levelled and flattened by these 'misplacements'?
A confirmation of this concern could be the observation that some of the results obtained by these studies are contrasting and difficult to explain. For example, the highest risk of developing enteropathy-associated T-cell lymphoma, a ferocious complication of coeliac disease with a dismal prognosis, was found in patients with a Marsh type 3 lesion, while the risk was lower in patients with a Marsh type 2 lesion.4 However, a similar study, performed in the same population, showed that mortality of patients with coeliac disease diagnosed with a Marsh type 2 lesion was higher than that of patients with coeliac disease with a Marsh type 3 lesion.3
In the last few years, quite a few studies investigated the mortality of undiagnosed, undetected coeliac disease. The most accurately performed ones suggested that mortality is not increased in these patients.4 ,9 This further supports the observation that the type of clinical presentation is of paramount importance in determining the risk of complications and thus mortality rates.10 So, if we assume that only patients with coeliac disease with frank symptoms of malabsorption at diagnosis are affected by an increased mortality rate, is it not possible that the database studies did not show an overall increased mortality because silent and non-classical/minor patients, nowadays the majority of patients with coeliac disease, 'diluted' the patients with classical/major symptoms? It must be very difficult, if not impossible, to evaluate the clinical picture through a database. Again, some evidence seems to suggest that mortality of patients with coeliac disease on a strict gluten-free diet is reduced compared with the general population.10 If this should be the case, is it possible to suggest that database studies did not show an increased mortality because the reduced mortality of patients with coeliac disease on a strict gluten-free diet ‘counter-balanced’ the increased mortality of patients on a poor gluten-free diet?
To conclude, although the study of Sultan et al2 provides evidence that mortality of patients with coeliac disease is similar to that of the general population, in our opinion this conclusion should be taken with care. In view of the comments raised in this commentary, we are not sure that mortality is not increased in patients with coeliac disease. However, we feel that in patients with coeliac disease mortality rate varies in different types of patients with coeliac disease, adherence to a gluten-free diet, type of clinical presentation and age at diagnosis being the most important discriminant factors. Since we do not know the natural history of untreated silent coeliac disease, this cannot be considered a good reason for not suggesting a strict gluten-free diet to silent patients with coeliac disease too. However, it can be a good point to start tailoring follow-up modalities of patients with coeliac disease according to the specific clinical characteristics of each individual patient. So, patients with coeliac disease diagnosed in the first two or three decades of life or because of familiarity or minor symptoms could be followed up on a clinical and serological basis only. On the other hand, in order to diagnose complications as soon as they arise, patients with coeliac disease diagnosed later in life and because of frank symptoms of malabsorption could be followed up on a regular basis with duodenal biopsy and other instrumental procedure, analogously to what is nowadays performed in patients affected by atrophic gastritis or colonic polyps.
We are grateful to Susan West for reading and correcting the manuscript.
Contributors FB and GRC wrote the paper.
Competing interests None.
Provenance and peer review Commissioned; internally peer reviewed.
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.