Introduction Gluten has been shown to alter bowel barrier function in patients with diarrhoea-predominant irritable bowel syndrome (D-IBS), particularly those who are HLA-DQ positive.

Aims To assess the clinical response to a gluten-free diet (GFD) in D-IBS patients previously naïve to the effects of gluten and blinded to HLA-DQ status.

Method Between May 2012 to December 2014 we enrolled 48 D-IBS patients (24 HLA-DQ positive and 24 HLA-DQ negative) to undertake a six-week GFD. Patients commenced a GFD after seeing an expert dietitian. Both patients and dietitian were blinded to HLA-DQ status. Questionnaires were self-completed at baseline and at week 6. The primary end-point was change in IBS-symptom severity score (IBS-SSS), with a 50 point reduction conferring clinical benefit. Secondary end-points were change in IBS-quality of life (IBS-QOL), short form-36 (SF-36), hospital anxiety and depression score (HADS) and fatigue impact score (FIS).

Results Data from 41 patients (74.4% women, mean-age 40.3 yrs) was available for per protocol analysis. Of these, 21 were HLA-DQ positive and 20 HLA-DQ negative; baseline characteristics were similar other than worse physical-fatigue (0.04) and vitality (0.05) in the HLA-DQ positive group. Overall, a GFD reduced IBS-SSS ≥50 points in 76% (n = 31). In fact, the mean-total IBS-SSS decreased from 286 to 131 (change –155, p < 0.001); this was seen similarly across both HLA-DQ groups. There was also a significant improvement in IBS-QOL, SF-36, HADS, and FIS amongst both groups. However, HLA-DQ positive subjects showed a greater response to depression (p = 0.02), fatigue (physical, p = 0.04; cognitive, p = 0.07) and vitality (p = 0.05) compared to HLA-DQ negative subjects. On study completion, 90% (n = 37) were discharged and 61% (n = 25) continued with a GFD.

Conclusion A GFD is a therapeutic option for the management of D-IBS. A clinical improvement was seen in 76% of patients undertaking a six-week GFD, with 61% opting to continue with a GFD for the foreseeable future. The pathophysiological mechanism may differ according to HLA-DQ status.

Disclosure of interest None Declared.