Gluten-Free Diets in Celiac Disease: Does Life Mean Life?

Abstract

Matysiak-Budnik T, Malamut G, de Serre NP, et al. (INSERM, U793, Faculté de Médecine René Descartes, Paris, France). Long-term evaluation of 61 coeliac patients diagnosed in childhood: evolution towards latency is possible on a normal diet. Gut 2007;56:1379–1386.When seeing children with recently confirmed celiac disease, most pediatricians dictate that this is a permanent condition and that a gluten-free diet (GFD) is required for life. Many will not advise gluten challenge except in a small proportion of patients. This dogmatic stance has now been challenged by a group in France who report that up to 10% develop long-term latency of their celiac disease when returning to a gluten-containing diet (Gut 2007;56:1379–1386).The study is a retrospective review of adult patients who had celiac disease diagnosed in childhood. The presence of celiac disease was characterized as the presence of duodenal villous atrophy, clinical symptoms of gluten intolerance with improvement after initiating the GFD, and/or positivity of specific celiac disease serum antibodies. At the time, it was the accepted management to subsequently challenge the patients after a period of 1–7 years. Although most clinicians would have reinstituted a GFD in the face of histologic abnormalities in the follow-up biopsies, in this unit they took the approach that in the absence of symptoms, even in the presence of histologic relapse, the patients could continue to ingest gluten in their diet. This characterized the patient population evaluated in this study—patients who had no evidence of clinical relapse after the first gluten challenge who were adults and had been on a normal diet for ≥2 years. The ability to tolerate the presence of gluten in their diet was reported to be present in approximately 50% of all the subjects diagnosed with celiac disease.All patients were assessed by means of clinical history, anthropometric measures, bone mineral density (BMD), blood investigations (including celiac antibodies and human leukocyte antigen [HLA] typing), and small intestinal biopsy. The Marsh classification was used to assess the degree of villous atrophy, intraepithelial lymphocytes were counted, and immunohistochemistry was examined to evaluate CD3/CD8 and T-cell receptors.Sixty-one adult patients with celiac disease confirmed in childhood were identified (38 women and 23 men with a median age at evaluation of 26 years) who were taking gluten-containing diets. Of these 61 patients who had small bowel biopsies, 48 had the presence of villous atrophy (now referred to as “silent” disease) and 13 had no villous atrophy (termed “latent” disease).The subjects had been on a normal diet for a median of 10 years before evaluation; before this, they had had a GFD for a median of 7 years in total. Thus, the median length of time from diagnosis was 17 years. Median age at diagnosis of celiac disease was 17 months; only 4 participants were diagnosed beyond the age of 8 years, the oldest being 16 years. The groups with latent disease were slightly older at diagnosis than those with silent disease.Almost half of the patients did have mild clinical symptoms (bloating, episodic abdominal pain) that were insufficient to prompt the patients to return to a GFD. The proportions of patients with these symptoms did not significantly differ between the silent and the latent groups.The main differences found in the silent compared with the latent group were reduced BMD, higher positivity to celiac antibodies, elevated serum aminotransferase levels, elevated CD3+ and CD8+ levels, and reduced TCRαβ+ intraepithelial lymphocytes (TCRγδ was not significantly different). There were no statistically significant differences between the HLA types in the 2 groups.The authors subsequently compared the subjects with latent celiac disease to 7 subjects who had been on GFD diets since diagnosis. The only differences they found was a higher incidence of celiac antibody positivity. In both silent and latent groups, elevated CD3+, CD8+, TCRαβ+, and TCRγδ+ IELs were compared with nonceliac controls.Finally, 2 (of 4 biopsied) in the latent group have since developed histologic relapse. One patient noted that they had ingested greatly increased amounts of gluten-containing foods; the other reported increased professional stress.CommentThis paper confirms that the majority (>90%) of subjects diagnosed with celiac disease continue to have disease with histologic abnormalities even if apparently tolerating gluten-containing products. Although the data were not presented in detail, it is perhaps surprising that as many 50% of all subjects diagnosed as having celiac disease as children reported tolerating gluten in their diet when challenged later in life.Other presented data challenge the dogma that once diagnosed celiac disease is always permanent and requires a GFD for life. In this study, up to 20% of the subjects tolerating normal diets developed latent celiac disease and thus now had normal histology. This would relate to approximately 10% of all the patients diagnosed with celiac disease during the period studied. These figures are similar to previous reports (J Pediatr Gastroenterol Nutr 1986;5:565–569; Arch Dis Child 1989;64:1604–1607).A critical question from the standpoint of whether these patients have latent celiac disease is whether they were accurately diagnosed initially. In European countries, these patients span 2 periods of time wherein different criteria for the diagnosis of celiac disease were present. The first criteria were published in 1974 and required 3 biopsies to confirm the diagnosis of celiac disease; 1 with typical changes in the first biopsy, the second a normal biopsy when on a GFD, and the third a rebiopsy after a gluten challenge again showing typical features of celiac disease (Proceedings of the Second International Coeliac Disease Symposium 1974:10–16). In 1990, the revised European Society of Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) criteria were published advising that the only mandatory items required to confirm a diagnosis of celiac disease were typical histologic changes in the biopsy and the presence of clinical remission on a strict GFD (Arch Dis Child 1990;65:909–911). Under the 1990 criteria, further biopsies before and after gluten challenge are not required, unless the patients had been asymptomatic at the time of the initial screening, had histologic features not typical of celiac disease, had inadequate specimens, or where other causes of enteropathy should be considered (especially in younger subjects <2 years of age). Although this paper straddles these 2 periods, the presence of celiac disease is defined as the presence of duodenal villous atrophy plus a clinical response to the GFD with or without celiac disease serum antibodies, that is, requiring a single biopsy in keeping with the more recent criteria. What they then did differently was to challenge the asymptomatic subjects (presumably all the patients who started and responded to GFD) with gluten and allowed those with good clinical tolerance to the gluten to continue on a normal diet, even if biopsies continued to show histologic abnormalities. What is not clear in the paper is whether all of the 13 subjects with latent disease had ≥1 abnormal biopsy after their gluten challenge (and subsequent normal diet); if this was not the case, then under the diagnostic criteria they should not be labeled as having celiac disease at all. It also may be of some relevance to the accuracy of the initial diagnosis that the group with latent disease were younger than those with silent disease; it is well recognized that enteropathies not due to celiac disease are more common in the younger age group; indeed the ESPGHAN 1990 criteria advise considering the full 3 biopsies specifically for this age group. Thus, it is possible that at least some of the 13 latent subjects did not have celiac disease but an enteropathy due to other causes. All individuals that were included in the latent group had DQ2 or DQ8 HLA haplotypes, and 5 of 13 had ≥1 serum CD antibody (compared with 41 of 44 in the silent group) T-cell and TCRγδ-positive cells were elevated compared with controls (and similar to the silent group), all found in subjects with celiac disease.Assuming an accurate diagnosis for at least the majority of the 13 subjects, how else could they have developed normal mucosa? An obvious answer is that they were not taking sufficient gluten to produce the villous atrophy. Although as a group the mean gluten consumption was the same as the silent group, the ranges were not reported (particularly the lowest levels) and it is possible that 1 or 2 could have been consuming insufficient quantities to cause the mucosal changes, effectively being on a GFD although not deliberately so. With 2 of the 13 later developing overt disease histologically and 1 admitting markedly increased gluten consumption, this could explain some of the supposed development of latency.The other major question raised by this paper is what to do with patients with known celiac disease who are taking gluten without clinical symptoms but have continuing histologic abnormalities. This is particularly pertinent given the increasing identification of asymptomatic celiac disease via antibody screening. Although >40% of the subjects on gluten with silent disease had mild gastrointestinal symptoms, the proportion is no higher than those with latent disease and cannot be directly attributed to histologic abnormalities. However, despite this lack of sufficient symptoms to encourage the patients to return to a GFD, there were clinical abnormalities that are important, the major one being reduced BMD with the associated risks of increased fractures.Another cause for concern in the long term is the risk of developing lymphomas. Although there is a good body of evidence that celiac disease can predispose one to developing lymphoma, the increased risk is not as great as the earliest reports and in terms of individuals is almost certainly very small (Gastroenterology 2005;128:S79–S86). Population data have suggested that if celiac disease is diagnosed in childhood (Gastroenterology 2002;123:1428–1435) or at least at a younger age (BMC Gastroenterol 2007;7:8) (and presumably treated with GFD) there is no increased risk of lymphoma. This is information I not infrequently discuss with my more recalcitrant teenage patients. Recent data suggest that those with silent disease do not have an increased risk of lymphoma (Eur J Gastroenterol Hepatol 2006;18:187–194). Whether this would be true of subjects who initially were diagnosed with clinical symptoms and who later develop silent disease is not known. Clearly knowing and allowing a continuing histologic abnormality requires some discussion of these risks so the patients can decide whether the benefits and ease of a normal diet can outweigh the risks.In summary, this paper reports that after gluten challenge ≥50% of subjects develop overt symptoms and histologic abnormalities requiring a return to a GFD. As many as 40%–45% of subjects with celiac disease may clinically tolerate a gluten-containing diet, but continue to have mucosal abnormalities and thus run the risk of complications such as fractures and lymphoma in the long term. Returning to a GFD diet would be advisable; however, at the very least open and honest discussion of the risks of continuing a normal diet must be made. A small proportion (probably <10%) of all celiac patients diagnosed in childhood may develop tolerance to gluten, which itself may not be permanent, so even if asymptomatic and on gluten they need continuing and long-term follow-up to identify serologic or histologic abnormalities. This small number of patients may be the lucky ones where the advice of life-long adherence to GFD does not really mean “life.” Matysiak-Budnik T, Malamut G, de Serre NP, et al. (INSERM, U793, Faculté de Médecine René Descartes, Paris, France). Long-term evaluation of 61 coeliac patients diagnosed in childhood: evolution towards latency is possible on a normal diet. Gut 2007;56:1379–1386. When seeing children with recently confirmed celiac disease, most pediatricians dictate that this is a permanent condition and that a gluten-free diet (GFD) is required for life. Many will not advise gluten challenge except in a small proportion of patients. This dogmatic stance has now been challenged by a group in France who report that up to 10% develop long-term latency of their celiac disease when returning to a gluten-containing diet (Gut 2007;56:1379–1386). The study is a retrospective review of adult patients who had celiac disease diagnosed in childhood. The presence of celiac disease was characterized as the presence of duodenal villous atrophy, clinical symptoms of gluten intolerance with improvement after initiating the GFD, and/or positivity of specific celiac disease serum antibodies. At the time, it was the accepted management to subsequently challenge the patients after a period of 1–7 years. Although most clinicians would have reinstituted a GFD in the face of histologic abnormalities in the follow-up biopsies, in this unit they took the approach that in the absence of symptoms, even in the presence of histologic relapse, the patients could continue to ingest gluten in their diet. This characterized the patient population evaluated in this study—patients who had no evidence of clinical relapse after the first gluten challenge who were adults and had been on a normal diet for ≥2 years. The ability to tolerate the presence of gluten in their diet was reported to be present in approximately 50% of all the subjects diagnosed with celiac disease. All patients were assessed by means of clinical history, anthropometric measures, bone mineral density (BMD), blood investigations (including celiac antibodies and human leukocyte antigen [HLA] typing), and small intestinal biopsy. The Marsh classification was used to assess the degree of villous atrophy, intraepithelial lymphocytes were counted, and immunohistochemistry was examined to evaluate CD3/CD8 and T-cell receptors. Sixty-one adult patients with celiac disease confirmed in childhood were identified (38 women and 23 men with a median age at evaluation of 26 years) who were taking gluten-containing diets. Of these 61 patients who had small bowel biopsies, 48 had the presence of villous atrophy (now referred to as “silent” disease) and 13 had no villous atrophy (termed “latent” disease). The subjects had been on a normal diet for a median of 10 years before evaluation; before this, they had had a GFD for a median of 7 years in total. Thus, the median length of time from diagnosis was 17 years. Median age at diagnosis of celiac disease was 17 months; only 4 participants were diagnosed beyond the age of 8 years, the oldest being 16 years. The groups with latent disease were slightly older at diagnosis than those with silent disease. Almost half of the patients did have mild clinical symptoms (bloating, episodic abdominal pain) that were insufficient to prompt the patients to return to a GFD. The proportions of patients with these symptoms did not significantly differ between the silent and the latent groups. The main differences found in the silent compared with the latent group were reduced BMD, higher positivity to celiac antibodies, elevated serum aminotransferase levels, elevated CD3+ and CD8+ levels, and reduced TCRαβ+ intraepithelial lymphocytes (TCRγδ was not significantly different). There were no statistically significant differences between the HLA types in the 2 groups. The authors subsequently compared the subjects with latent celiac disease to 7 subjects who had been on GFD diets since diagnosis. The only differences they found was a higher incidence of celiac antibody positivity. In both silent and latent groups, elevated CD3+, CD8+, TCRαβ+, and TCRγδ+ IELs were compared with nonceliac controls. Finally, 2 (of 4 biopsied) in the latent group have since developed histologic relapse. One patient noted that they had ingested greatly increased amounts of gluten-containing foods; the other reported increased professional stress. CommentThis paper confirms that the majority (>90%) of subjects diagnosed with celiac disease continue to have disease with histologic abnormalities even if apparently tolerating gluten-containing products. Although the data were not presented in detail, it is perhaps surprising that as many 50% of all subjects diagnosed as having celiac disease as children reported tolerating gluten in their diet when challenged later in life.Other presented data challenge the dogma that once diagnosed celiac disease is always permanent and requires a GFD for life. In this study, up to 20% of the subjects tolerating normal diets developed latent celiac disease and thus now had normal histology. This would relate to approximately 10% of all the patients diagnosed with celiac disease during the period studied. These figures are similar to previous reports (J Pediatr Gastroenterol Nutr 1986;5:565–569; Arch Dis Child 1989;64:1604–1607).A critical question from the standpoint of whether these patients have latent celiac disease is whether they were accurately diagnosed initially. In European countries, these patients span 2 periods of time wherein different criteria for the diagnosis of celiac disease were present. The first criteria were published in 1974 and required 3 biopsies to confirm the diagnosis of celiac disease; 1 with typical changes in the first biopsy, the second a normal biopsy when on a GFD, and the third a rebiopsy after a gluten challenge again showing typical features of celiac disease (Proceedings of the Second International Coeliac Disease Symposium 1974:10–16). In 1990, the revised European Society of Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) criteria were published advising that the only mandatory items required to confirm a diagnosis of celiac disease were typical histologic changes in the biopsy and the presence of clinical remission on a strict GFD (Arch Dis Child 1990;65:909–911). Under the 1990 criteria, further biopsies before and after gluten challenge are not required, unless the patients had been asymptomatic at the time of the initial screening, had histologic features not typical of celiac disease, had inadequate specimens, or where other causes of enteropathy should be considered (especially in younger subjects <2 years of age). Although this paper straddles these 2 periods, the presence of celiac disease is defined as the presence of duodenal villous atrophy plus a clinical response to the GFD with or without celiac disease serum antibodies, that is, requiring a single biopsy in keeping with the more recent criteria. What they then did differently was to challenge the asymptomatic subjects (presumably all the patients who started and responded to GFD) with gluten and allowed those with good clinical tolerance to the gluten to continue on a normal diet, even if biopsies continued to show histologic abnormalities. What is not clear in the paper is whether all of the 13 subjects with latent disease had ≥1 abnormal biopsy after their gluten challenge (and subsequent normal diet); if this was not the case, then under the diagnostic criteria they should not be labeled as having celiac disease at all. It also may be of some relevance to the accuracy of the initial diagnosis that the group with latent disease were younger than those with silent disease; it is well recognized that enteropathies not due to celiac disease are more common in the younger age group; indeed the ESPGHAN 1990 criteria advise considering the full 3 biopsies specifically for this age group. Thus, it is possible that at least some of the 13 latent subjects did not have celiac disease but an enteropathy due to other causes. All individuals that were included in the latent group had DQ2 or DQ8 HLA haplotypes, and 5 of 13 had ≥1 serum CD antibody (compared with 41 of 44 in the silent group) T-cell and TCRγδ-positive cells were elevated compared with controls (and similar to the silent group), all found in subjects with celiac disease.Assuming an accurate diagnosis for at least the majority of the 13 subjects, how else could they have developed normal mucosa? An obvious answer is that they were not taking sufficient gluten to produce the villous atrophy. Although as a group the mean gluten consumption was the same as the silent group, the ranges were not reported (particularly the lowest levels) and it is possible that 1 or 2 could have been consuming insufficient quantities to cause the mucosal changes, effectively being on a GFD although not deliberately so. With 2 of the 13 later developing overt disease histologically and 1 admitting markedly increased gluten consumption, this could explain some of the supposed development of latency.The other major question raised by this paper is what to do with patients with known celiac disease who are taking gluten without clinical symptoms but have continuing histologic abnormalities. This is particularly pertinent given the increasing identification of asymptomatic celiac disease via antibody screening. Although >40% of the subjects on gluten with silent disease had mild gastrointestinal symptoms, the proportion is no higher than those with latent disease and cannot be directly attributed to histologic abnormalities. However, despite this lack of sufficient symptoms to encourage the patients to return to a GFD, there were clinical abnormalities that are important, the major one being reduced BMD with the associated risks of increased fractures.Another cause for concern in the long term is the risk of developing lymphomas. Although there is a good body of evidence that celiac disease can predispose one to developing lymphoma, the increased risk is not as great as the earliest reports and in terms of individuals is almost certainly very small (Gastroenterology 2005;128:S79–S86). Population data have suggested that if celiac disease is diagnosed in childhood (Gastroenterology 2002;123:1428–1435) or at least at a younger age (BMC Gastroenterol 2007;7:8) (and presumably treated with GFD) there is no increased risk of lymphoma. This is information I not infrequently discuss with my more recalcitrant teenage patients. Recent data suggest that those with silent disease do not have an increased risk of lymphoma (Eur J Gastroenterol Hepatol 2006;18:187–194). Whether this would be true of subjects who initially were diagnosed with clinical symptoms and who later develop silent disease is not known. Clearly knowing and allowing a continuing histologic abnormality requires some discussion of these risks so the patients can decide whether the benefits and ease of a normal diet can outweigh the risks.In summary, this paper reports that after gluten challenge ≥50% of subjects develop overt symptoms and histologic abnormalities requiring a return to a GFD. As many as 40%–45% of subjects with celiac disease may clinically tolerate a gluten-containing diet, but continue to have mucosal abnormalities and thus run the risk of complications such as fractures and lymphoma in the long term. Returning to a GFD diet would be advisable; however, at the very least open and honest discussion of the risks of continuing a normal diet must be made. A small proportion (probably <10%) of all celiac patients diagnosed in childhood may develop tolerance to gluten, which itself may not be permanent, so even if asymptomatic and on gluten they need continuing and long-term follow-up to identify serologic or histologic abnormalities. This small number of patients may be the lucky ones where the advice of life-long adherence to GFD does not really mean “life.” This paper confirms that the majority (>90%) of subjects diagnosed with celiac disease continue to have disease with histologic abnormalities even if apparently tolerating gluten-containing products. Although the data were not presented in detail, it is perhaps surprising that as many 50% of all subjects diagnosed as having celiac disease as children reported tolerating gluten in their diet when challenged later in life. Other presented data challenge the dogma that once diagnosed celiac disease is always permanent and requires a GFD for life. In this study, up to 20% of the subjects tolerating normal diets developed latent celiac disease and thus now had normal histology. This would relate to approximately 10% of all the patients diagnosed with celiac disease during the period studied. These figures are similar to previous reports (J Pediatr Gastroenterol Nutr 1986;5:565–569; Arch Dis Child 1989;64:1604–1607). A critical question from the standpoint of whether these patients have latent celiac disease is whether they were accurately diagnosed initially. In European countries, these patients span 2 periods of time wherein different criteria for the diagnosis of celiac disease were present. The first criteria were published in 1974 and required 3 biopsies to confirm the diagnosis of celiac disease; 1 with typical changes in the first biopsy, the second a normal biopsy when on a GFD, and the third a rebiopsy after a gluten challenge again showing typical features of celiac disease (Proceedings of the Second International Coeliac Disease Symposium 1974:10–16). In 1990, the revised European Society of Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) criteria were published advising that the only mandatory items required to confirm a diagnosis of celiac disease were typical histologic changes in the biopsy and the presence of clinical remission on a strict GFD (Arch Dis Child 1990;65:909–911). Under the 1990 criteria, further biopsies before and after gluten challenge are not required, unless the patients had been asymptomatic at the time of the initial screening, had histologic features not typical of celiac disease, had inadequate specimens, or where other causes of enteropathy should be considered (especially in younger subjects <2 years of age). Although this paper straddles these 2 periods, the presence of celiac disease is defined as the presence of duodenal villous atrophy plus a clinical response to the GFD with or without celiac disease serum antibodies, that is, requiring a single biopsy in keeping with the more recent criteria. What they then did differently was to challenge the asymptomatic subjects (presumably all the patients who started and responded to GFD) with gluten and allowed those with good clinical tolerance to the gluten to continue on a normal diet, even if biopsies continued to show histologic abnormalities. What is not clear in the paper is whether all of the 13 subjects with latent disease had ≥1 abnormal biopsy after their gluten challenge (and subsequent normal diet); if this was not the case, then under the diagnostic criteria they should not be labeled as having celiac disease at all. It also may be of some relevance to the accuracy of the initial diagnosis that the group with latent disease were younger than those with silent disease; it is well recognized that enteropathies not due to celiac disease are more common in the younger age group; indeed the ESPGHAN 1990 criteria advise considering the full 3 biopsies specifically for this age group. Thus, it is possible that at least some of the 13 latent subjects did not have celiac disease but an enteropathy due to other causes. All individuals that were included in the latent group had DQ2 or DQ8 HLA haplotypes, and 5 of 13 had ≥1 serum CD antibody (compared with 41 of 44 in the silent group) T-cell and TCRγδ-positive cells were elevated compared with controls (and similar to the silent group), all found in subjects with celiac disease. Assuming an accurate diagnosis for at least the majority of the 13 subjects, how else could they have developed normal mucosa? An obvious answer is that they were not taking sufficient gluten to produce the villous atrophy. Although as a group the mean gluten consumption was the same as the silent group, the ranges were not reported (particularly the lowest levels) and it is possible that 1 or 2 could have been consuming insufficient quantities to cause the mucosal changes, effectively being on a GFD although not deliberately so. With 2 of the 13 later developing overt disease histologically and 1 admitting markedly increased gluten consumption, this could explain some of the supposed development of latency. The other major question raised by this paper is what to do with patients with known celiac disease who are taking gluten without clinical symptoms but have continuing histologic abnormalities. This is particularly pertinent given the increasing identification of asymptomatic celiac disease via antibody screening. Although >40% of the subjects on gluten with silent disease had mild gastrointestinal symptoms, the proportion is no higher than those with latent disease and cannot be directly attributed to histologic abnormalities. However, despite this lack of sufficient symptoms to encourage the patients to return to a GFD, there were clinical abnormalities that are important, the major one being reduced BMD with the associated risks of increased fractures. Another cause for concern in the long term is the risk of developing lymphomas. Although there is a good body of evidence that celiac disease can predispose one to developing lymphoma, the increased risk is not as great as the earliest reports and in terms of individuals is almost certainly very small (Gastroenterology 2005;128:S79–S86). Population data have suggested that if celiac disease is diagnosed in childhood (Gastroenterology 2002;123:1428–1435) or at least at a younger age (BMC Gastroenterol 2007;7:8) (and presumably treated with GFD) there is no increased risk of lymphoma. This is information I not infrequently discuss with my more recalcitrant teenage patients. Recent data suggest that those with silent disease do not have an increased risk of lymphoma (Eur J Gastroenterol Hepatol 2006;18:187–194). Whether this would be true of subjects who initially were diagnosed with clinical symptoms and who later develop silent disease is not known. Clearly knowing and allowing a continuing histologic abnormality requires some discussion of these risks so the patients can decide whether the benefits and ease of a normal diet can outweigh the risks. In summary, this paper reports that after gluten challenge ≥50% of subjects develop overt symptoms and histologic abnormalities requiring a return to a GFD. As many as 40%–45% of subjects with celiac disease may clinically tolerate a gluten-containing diet, but continue to have mucosal abnormalities and thus run the risk of complications such as fractures and lymphoma in the long term. Returning to a GFD diet would be advisable; however, at the very least open and honest discussion of the risks of continuing a normal diet must be made. A small proportion (probably <10%) of all celiac patients diagnosed in childhood may develop tolerance to gluten, which itself may not be permanent, so even if asymptomatic and on gluten they need continuing and long-term follow-up to identify serologic or histologic abnormalities. This small number of patients may be the lucky ones where the advice of life-long adherence to GFD does not really mean “life.” ReplyGastroenterologyVol. 134Issue 3PreviewWe thank Dr Croft for his comments and his interest in our study. However, some of the issues highlighted by Dr Croft deserve explanations. Full-Text PDF