Long-term response to gluten-free diet as evidence for non-celiac wheat sensitivity in one third of patients with diarrhoea- dominant and mixed-type irritable bowel syndrome
Barmeyer C, Schumann M, Meyer T et al.
Int J Colorectal Dis. 2016. DOI 10/1007/s00384-016-2663-x
Irritable bowel syndrome (IBS) is a common functional bowel disorder with a recognised social and economic impact due to the associated reduced quality of life and high rates of IBS-related non-productive time.
Int J Colorectal Dis. 2016. DOI 10/1007/s00384-016-2663-x
Irritable bowel syndrome (IBS) is a common functional bowel disorder with a recognised social and economic impact due to the associated reduced quality of life and high rates of IBS-related non-productive time.
Patients often suspect food but classical food allergy is rarely the reason. In recent years, evidence has arisen that subgroups of IBS patients may be ‘gluten sensitive’ and could therefore respond to a gluten-free diet (GFD), even in the absence of the histological abnormalities characteristic of coeliac disease (CD). This led to the concept of a new disorder called ‘non-coeliac gluten sensitivity’, or more recently termed ‘non-coeliac wheat sensitivity’ due to the fact that wheat components other than gluten might be involved.
The authors previously reported a benefit for patients with IBS-D, carrying the HLA-DQ2 allele, after being on a GFD for 6 months in an open-uncontrolled trial. The authors therefore speculated that HLA-DQ2 could be a useful marker to identify a subgroup of gluten-sensitive IBS patients who benefit from a GFD.
The aim of the study was to analyse if HLA-DQ2 and DQ8 are suitable markers for the diagnosis of wheat sensitivity (WS) and to evaluate the long-term clinical response to a GFD in a defined cohort of patients with IBS-D and IBS-M. Patients were recruited from secondary care, primary care and through public advertising in newspapers and public transportation. Upon study entry other medical conditions presenting with similar symptoms to IBS were excluded.
91 subjects (>18 years of age) fulfilled the primary inclusion criteria (ROME III criteria, self-reported weekly symptoms) for IBS-D or IBS-M and were eligible to enter the 4-week observation period without any intervention (as constipation is not a classical feature of CD, it was not expected that WS would exist in IBS-C patients and they were excluded). At the end of each week the ‘Subject’s Global Assessment’ (SGA) of relief was determined. The SGA asks the subject to rate the relief from their symptoms during the past week compared to the way they felt prior to entering the study. Five different answers were available and subjects who answered ‘unchanged’ or ‘worse’ at all four timepoints, underwent diagnostic tests to exclude other medical conditions before being included in the study. 45 subjects did not finish the observation period; 15 of these were due to ‘significant improvement’ or ‘symptoms less than once a week’; 30 participants due to other reasons and 11 dropped out due to a new diagnosis: CD (n=3), lactose intolerance (n=3), pregnancy (n=1), erosive duodenitis (n=1), giardiasis (n=1) and mastocytosis (n=1).
Thirty-five subjects (15 IBS-D and 20 IBS-M) entered the study and received two dietary consultations from registered dietitians at an interval of two weeks; the first to educate subjects on a GFD and the second to ascertain adherence to a GFD. Subjects were then monitored for 4 months by weekly assessment of the SGA of relief via phone calls. For the SGA, those who answered ‘considerably relieved’ or ‘completely relieved’ on at least 75% of the weeks over the 4 month dietary intervention were defined as responders (R75), taking into account both magnitude and persistence of effect. The stricter requirement of 75% of time (vs 50% of time commonly used) was used as it was assumed symptoms would disappear with a GFD when WS was their primary cause. However, the weaker response criterion of 50% (R50) of the time was used as a secondary outcome measure. Other secondary outcomes measures included changes in the score of the IBS-QOL, IBS-SSS and EQ-5D questionnaires and changes in pain intensity, stool consistency and stool frequency before and at the end of the study. Prior to the dietary intervention, blood was collected for genotyping; determination of HLA-DQ2/8 status followed dietary intervention to avoid bias. At the end of the study period, subjects were followed up at 1 year to determine their long-term dietary habits and ongoing symptom relief.
Patient Characteristics
The study found that 40% of the study population was positive for HLA-DQ2 or DQ8; a slightly higher proportion than in the general population. A significantly higher proportion of HLA-DQ2/8 positive subjects was observed in the IBS-M group compared with the IBS-D group (n=10 (50%) vs n=4 (27%). 94% (33/35 subjects) adhered to a GFD for the study period; two IBS-M subjects dropped out between 4-6 weeks; one due to increased symptom severity (HLA-DQ2 +ve) and one due to commencement of treatment for another condition (HLA-DQ 2/8 –ve). Both were were classed as non-responders.
Wheat Sensitivity and Response to a GFD
12/35 subjects (34%) reported considerable or complete relief from symptoms on at least 75% of weeks over the 4 month dietary intervention. These were classed as responders having WS. The proportion of responders was similar in both IBS subgroups. This figure increased to 51% when applying the less strict criterion of relief for at least 50% of the time (R50 group).
Association of HLA-DQ2/8 status with wheat sensitivity
HLA-DQ2 or DQ8 expression was not found to be associated with WS; the proportion of responders to a GFD was in fact higher in HLA-DQ2/8 –ve subjects. The sensitivity and specificity of HLA-DQ2 or DQ8 expression for WS were 25% and 52% respectively and therefore did not identify IBS patients with and without WS in this cohort.
QOL measures in responders and non-responders
In the R75 group, five out of 9 patients reported an overall IBS-SSS score of below 75 after 4 months of a GFD which reflects that these patients went into ‘remission’ whilst on the diet. This was in contrast to only 4 out of 17 in the non-responder group (NR75). Similarly, the overall score of the IBS-QOL and EQ-5D results showed a tendency towards improvement. On comparison of two relevant IBS parameters, pain and distention/tightness, only the latter showed a significant improvement in the R75 group versus the NR75 group. As a parameter, pain only showed a tendency towards improvement and the same observation was made in relation to ‘satisfaction with bowel habits’ and ‘interference with life in general’.
In the R50 group, an IBS-SSS score of <75 was reported in 9 out of 15 patients after 4 months on a GFD, with an increase in both the IBS-QOL and EQ-5D scores, consistent with the R75 group but the differences were more pronounced. In fact, a significant symptom improvement could be detected after the diet in comparison to the non-responders group (NR50) with corresponding results for the occurrence of pain, abdominal distention/tightness, satisfaction with bowel habits and interference with life in general. Stool frequency did not change in either the IBS-D or IBS-M responders and only a tendency to decrease in IBS-D was observed. In summary, when comparing the two different response criteria and the corresponding secondary outcomes, in both groups, a significant symptom and QOL improvement were observed after 4 months of a GFD but the results obtained from the 50% of time criterion were more pronounced than for the 75% of time.
GFD duration until clinical response
A considerable number of subjects 6 IBS-D (1 DQ2+ve) and 7 IBS-M (3 DQ2 +ve/1 DQ8 +ve) presented with a delayed response to the GFD. A response (‘considerably relieved or completely relieved’) was seen after approximately two months and was ongoing. These patients would have been classed as ‘non-responders’ when applying the 75% criterion and 7 patients (4 IBS-D; 0 DQ2/8+ve and 3 IBS-M; 0 DQ2/8+ve) were even non-responders when applying the 50% criterion because of the delayed response.
One year follow up
In the R75 group, 7/11 patients (64%) still followed a strict GFD with the majority still reporting complete or considerable relief from symptoms. The remaining four patients reported to feel considerably relieved. In fact, all continued to follow a GFD with some exceptions. In the non-responder group, six patients (33%) continued to follow a GFD with the majority still reporting some degree of relief. The authors found a consistent adherence to a GFD in the responder group with long-term improvement of symptoms.
In summary, the observations in this study show that WS is common within the group of IBS-D and IBS-M patients. WS is not associated with the HLA-DQ2/8 status and therefore this cannot be used as a predictive marker. There is evidence that a GFD offers relief for a group of non-constipated IBS patients, however, a long-term GFD appears to be necessary to identify all responders as two response characteristics were seen in the responder group: i) early responders who immediately benefitted from a GRD and ii) late responders who only benefitted after 2 months. This indicates that two different mechanisms may underlie the improvement seen in gastrointestinal symptoms.
Link to full study
The authors previously reported a benefit for patients with IBS-D, carrying the HLA-DQ2 allele, after being on a GFD for 6 months in an open-uncontrolled trial. The authors therefore speculated that HLA-DQ2 could be a useful marker to identify a subgroup of gluten-sensitive IBS patients who benefit from a GFD.
The aim of the study was to analyse if HLA-DQ2 and DQ8 are suitable markers for the diagnosis of wheat sensitivity (WS) and to evaluate the long-term clinical response to a GFD in a defined cohort of patients with IBS-D and IBS-M. Patients were recruited from secondary care, primary care and through public advertising in newspapers and public transportation. Upon study entry other medical conditions presenting with similar symptoms to IBS were excluded.
91 subjects (>18 years of age) fulfilled the primary inclusion criteria (ROME III criteria, self-reported weekly symptoms) for IBS-D or IBS-M and were eligible to enter the 4-week observation period without any intervention (as constipation is not a classical feature of CD, it was not expected that WS would exist in IBS-C patients and they were excluded). At the end of each week the ‘Subject’s Global Assessment’ (SGA) of relief was determined. The SGA asks the subject to rate the relief from their symptoms during the past week compared to the way they felt prior to entering the study. Five different answers were available and subjects who answered ‘unchanged’ or ‘worse’ at all four timepoints, underwent diagnostic tests to exclude other medical conditions before being included in the study. 45 subjects did not finish the observation period; 15 of these were due to ‘significant improvement’ or ‘symptoms less than once a week’; 30 participants due to other reasons and 11 dropped out due to a new diagnosis: CD (n=3), lactose intolerance (n=3), pregnancy (n=1), erosive duodenitis (n=1), giardiasis (n=1) and mastocytosis (n=1).
Thirty-five subjects (15 IBS-D and 20 IBS-M) entered the study and received two dietary consultations from registered dietitians at an interval of two weeks; the first to educate subjects on a GFD and the second to ascertain adherence to a GFD. Subjects were then monitored for 4 months by weekly assessment of the SGA of relief via phone calls. For the SGA, those who answered ‘considerably relieved’ or ‘completely relieved’ on at least 75% of the weeks over the 4 month dietary intervention were defined as responders (R75), taking into account both magnitude and persistence of effect. The stricter requirement of 75% of time (vs 50% of time commonly used) was used as it was assumed symptoms would disappear with a GFD when WS was their primary cause. However, the weaker response criterion of 50% (R50) of the time was used as a secondary outcome measure. Other secondary outcomes measures included changes in the score of the IBS-QOL, IBS-SSS and EQ-5D questionnaires and changes in pain intensity, stool consistency and stool frequency before and at the end of the study. Prior to the dietary intervention, blood was collected for genotyping; determination of HLA-DQ2/8 status followed dietary intervention to avoid bias. At the end of the study period, subjects were followed up at 1 year to determine their long-term dietary habits and ongoing symptom relief.
Patient Characteristics
The study found that 40% of the study population was positive for HLA-DQ2 or DQ8; a slightly higher proportion than in the general population. A significantly higher proportion of HLA-DQ2/8 positive subjects was observed in the IBS-M group compared with the IBS-D group (n=10 (50%) vs n=4 (27%). 94% (33/35 subjects) adhered to a GFD for the study period; two IBS-M subjects dropped out between 4-6 weeks; one due to increased symptom severity (HLA-DQ2 +ve) and one due to commencement of treatment for another condition (HLA-DQ 2/8 –ve). Both were were classed as non-responders.
Wheat Sensitivity and Response to a GFD
12/35 subjects (34%) reported considerable or complete relief from symptoms on at least 75% of weeks over the 4 month dietary intervention. These were classed as responders having WS. The proportion of responders was similar in both IBS subgroups. This figure increased to 51% when applying the less strict criterion of relief for at least 50% of the time (R50 group).
Association of HLA-DQ2/8 status with wheat sensitivity
HLA-DQ2 or DQ8 expression was not found to be associated with WS; the proportion of responders to a GFD was in fact higher in HLA-DQ2/8 –ve subjects. The sensitivity and specificity of HLA-DQ2 or DQ8 expression for WS were 25% and 52% respectively and therefore did not identify IBS patients with and without WS in this cohort.
QOL measures in responders and non-responders
In the R75 group, five out of 9 patients reported an overall IBS-SSS score of below 75 after 4 months of a GFD which reflects that these patients went into ‘remission’ whilst on the diet. This was in contrast to only 4 out of 17 in the non-responder group (NR75). Similarly, the overall score of the IBS-QOL and EQ-5D results showed a tendency towards improvement. On comparison of two relevant IBS parameters, pain and distention/tightness, only the latter showed a significant improvement in the R75 group versus the NR75 group. As a parameter, pain only showed a tendency towards improvement and the same observation was made in relation to ‘satisfaction with bowel habits’ and ‘interference with life in general’.
In the R50 group, an IBS-SSS score of <75 was reported in 9 out of 15 patients after 4 months on a GFD, with an increase in both the IBS-QOL and EQ-5D scores, consistent with the R75 group but the differences were more pronounced. In fact, a significant symptom improvement could be detected after the diet in comparison to the non-responders group (NR50) with corresponding results for the occurrence of pain, abdominal distention/tightness, satisfaction with bowel habits and interference with life in general. Stool frequency did not change in either the IBS-D or IBS-M responders and only a tendency to decrease in IBS-D was observed. In summary, when comparing the two different response criteria and the corresponding secondary outcomes, in both groups, a significant symptom and QOL improvement were observed after 4 months of a GFD but the results obtained from the 50% of time criterion were more pronounced than for the 75% of time.
GFD duration until clinical response
A considerable number of subjects 6 IBS-D (1 DQ2+ve) and 7 IBS-M (3 DQ2 +ve/1 DQ8 +ve) presented with a delayed response to the GFD. A response (‘considerably relieved or completely relieved’) was seen after approximately two months and was ongoing. These patients would have been classed as ‘non-responders’ when applying the 75% criterion and 7 patients (4 IBS-D; 0 DQ2/8+ve and 3 IBS-M; 0 DQ2/8+ve) were even non-responders when applying the 50% criterion because of the delayed response.
One year follow up
In the R75 group, 7/11 patients (64%) still followed a strict GFD with the majority still reporting complete or considerable relief from symptoms. The remaining four patients reported to feel considerably relieved. In fact, all continued to follow a GFD with some exceptions. In the non-responder group, six patients (33%) continued to follow a GFD with the majority still reporting some degree of relief. The authors found a consistent adherence to a GFD in the responder group with long-term improvement of symptoms.
In summary, the observations in this study show that WS is common within the group of IBS-D and IBS-M patients. WS is not associated with the HLA-DQ2/8 status and therefore this cannot be used as a predictive marker. There is evidence that a GFD offers relief for a group of non-constipated IBS patients, however, a long-term GFD appears to be necessary to identify all responders as two response characteristics were seen in the responder group: i) early responders who immediately benefitted from a GRD and ii) late responders who only benefitted after 2 months. This indicates that two different mechanisms may underlie the improvement seen in gastrointestinal symptoms.
Link to full study
www.drschaer-institute.com