WHO ARE THE PLACEBO RESPONDERS IN IBS CLINICAL TRIALS?: A PROSPECTIVE STUDY OF A PLACEBO-CONTROLLED MULTISITE NIH TRIAL

Background: Irritable bowel syndrome (IBS) is a common disorder with complex pathophysiological mechanisms. Prior studies suggest that 25-50% of patients with diarrhea-predominant IBS (IBS-D) have evidence of bile acid (BA) diarrhea (BAD). Studies of candidate genes involved in the BA pathway have identified single nucleotide variants (SNVs) in the synthesis of farnesoid X receptor (NR1H4, MALRD1/DIET1), proteins involved in hepatic re-uptake of BA (KLB, FGFR4), the rate-limiting enzyme of BA synthesis (CYP7A1), and BA receptors (GPBAR1/TGR5) that may be implicated in the pathophysiology of IBS-D and BAD. Aim: To explore the burden of allelic variants among known SNVs across these genes in IBS-D. Methods: Design: Case-control analysis of the genetic differences between those with diagnosis of IBS-D compared to age- and sex-matched controls. Genomic data were provided by the Mayo Clinic Tapestry Study for the target genes listed above along with links exomic database which was linked to the electronic medical record (EMR). Cases of IBS-D were defined according to ICD-10 code, and controls were those without documentation of the same ICD-10 diagnosis. SNVs with either low quality, read depth, or probability of a functional impact (e.g. synonymous and non-splice site intronic variants) were disregarded. The total burden of non-reference alleles between IBS-D and control patients was compared using ANOVA with α=0.05. Results: After data cleaning, our final sample included 526 patients with IBS-D and 2,950 controls. The mean (SD) age of the entire sample was 53.1 (15.8) years, and majority (78.9%) were female (Table 1). Three thousand and eighteen (95.5%) participants self-identified as White. Only ethnicity was statistically significantly different between the groups, with 4.6% and 9.4% identifying as Hispanic or Latino in the IBS-D and control groups, respectively. The summary of the genetic findings is shown in Table 2. Participants in the IBS-D group had, on average, a significantly greater genetic burden of homozygous alternate alleles in MALRD1 compared to controls (p=0.006). There was a greater proportion of heterozygote genotypes for key SNVs in CYP7A1 in controls compared to those with IBS-D. Other SNVs in candidate genes did not demonstrate a statistically significant difference in frequency of alternate alleles between IBS-D and controls. Conclusions: MALRD1 (previously DIET1), a gene that codes for an intestinal protein involved in enterohepatic circulation by regulating FGF-19 and CYP7A1 (BA synthesis), may contribute to pathophysiological mechanisms in IBS-D. Further studies are needed to explore specific genetic variants and their functional impact in MALRD1 and CYP7A1 in IBS-D that may elucidate disease mechanisms and potential therapeutic targets, such as with FGF-19 analogs or FXR agonists.

Background: Patients with irritable bowel syndrome (IBS) often report gastrointestinal symptoms after consuming wheat/gluten-containing foods. It is, however, unclear whether gluten is the main driver of symptoms, as other immunogenic peptides, such as amylase trypsin inhibitors (ATI), poorly digestible fiber (inulin, part of FODMAP) or even the nocebo effect may contribute to symptom generation.

Aim: To evaluate whether whole wheat (containing ATIs) or purified gluten alone triggers symptoms compared to nocebo in patients with IBS adopting a gluten-free diet (GFD).

Method: We conducted a double-blind, randomized, nocebo-controlled, crossover study in adult IBS patients (Rome IV criteria) who previously perceived symptomatic improvement while on a GFD. Participants were challenged for 7 days with whole wheat, purified gluten, and nocebo (gluten-free flour) added to low-FODMAP cereal bars. Each challenge was followed by a 2-week washout. Patients remained on a GFD throughout the study. Diet adherence was assessed by a dietitian and stool gluten immunogenic peptides (GIP; Biomedal, Spain). Symptoms were assessed by IBS Symptom Severity Score (IBS-SSS); an increase >50 points was considered significantly worse. Blood samples were collected to assess immune markers and celiac genotype. Statistical comparisons used Friedman rank sum tests and paired Wilcoxon signed rank tests.

Results: Twenty-nine IBS patients (27 female, mean age 42 years) were enrolled in the study; 1 dropped. Similar proportions of patients reacted symptomatically to wheat (11/28, 39.3%), gluten (10/28, 35.7%), and nocebo (8/28, 28.6%). However, there was an overall increase in IBS symptoms after wheat (+39.5 on IBS-SSS; p=0.030) but not after gluten (+27.5; p=0.051) or nocebo (+5.5; p=0.236) challenges (Figure 1). Ten participants experienced moderate symptoms (IBS-SSS >175) during baseline and did not worsen during the challenges. Median GIP levels were 0.58 after wheat, 0.43 after gluten, and 0.09 µg/g after nocebo; p<0.001. Four out of 28 patients had GIP levels 0.10 µg/g after the gluten/wheat challenges. TNF-α, IL-8, and IFN-γ levels were low and did not increase after challenges. Celiac genes were present in 19/24 patients (HLA DQ2 n=10, DQ8 n=2, DQ7 n=9).

Conclusions: IBS patients self-reporting wheat/gluten sensitivity had worse symptoms after whole wheat, but not after purified gluten or nocebo challenges. However, similar proportions of patients reacted to each challenge, suggesting that central mechanisms play an important role in symptom genesis. One third of patients had persistently high symptoms during a GFD and challenges, suggesting that other mechanisms are underlying their IBS. Furthermore, negative GIP in some patients suggests a lack of adherence to the study protocol, highlighting the complexity and challenges of dietary intervention studies.

BACKGROUND. Challenges in developing medical therapies for IBS include regulatory hurdles, unknown biological mechanisms and biomarkers, poor consensus around what constitutes clinically meaningful endpoints, heterogeneity of patient population, and lack of a proof-of-concept conceptual model with predictive value in Phase IIB and III trials (Camilleri & Chang 2008, Camilleri 2017). One relatively underexplored factor involves high (40%) placebo rates (Enck & Klosterhalfen 2020) which raises bar for demonstrating a “true drug” effect, rendering many trials time consuming and expensive, delays delivery of new treatments into market, and reinforces negative stigma of IBS treatments. Detailed patient phenotyping of a broader range of variables that potentiate placebo response is needed. Patient-related data readily obtained in placebo-controlled behavioral trials are likely to have more explanatory value than general characteristics drug trials collect. AIM: To identify factors that predict treatment response of placebo-treated patients. METHOD. 145 Rome III-diagnosed IBS patients (79% F, M age = 42 yrs.) with moderate to severe GI symptoms (sx) completed baseline (BL) assessment of an NIH trial before randomized to a state-of-the-art psychological placebo (EDUcation) that mimicked elements common across therapies (e.g., alliance between patient and treating doctor) including pharmacotherapy without containing their technical properties. BL testing included measures of motivation, GI sxs, psychiatric comorbidity, distress. Clinician expectancy of treatment effectiveness and quality of working alliance by patient (pt.) were measured at end of session 1. We used dichotomous (CGI-IBS) and continuous (IBS Symptom Severity Scale) indices of global treatment response. RESULTS. For CGI, 17% were categorized as early (wk. 4) and post treatment (PT, wk. 12) responders (≧ moderate improvement). For IBS-SSS, 31% of EDU pts qualified as treatment responders (≥ 50 unit reduction) at wk. 4 and 12 (PT). IBS-SSS improvement was predicted by shorter sx duration, r (127) = -.27, p = .002, self-rated importance of achieving primary treatment goal, r (127) = .21, p = .020, as well as 2 alliance dimensions, doctor-patient bond r (124) = .19, p = .034, and their agreement around treatment goals, r (127) = .19, p = .040. Posttest CGI responder status related to clinician expectancy ratings, d = 0.38 (95% CI = 0.02 / 0.72). CONCLUSION. This study identifies a set of previously unexplored variables (alliance, motivation, clinician expectancy of sx relief) that prospectively predicted placebo response at PT. IBS pts who regarded their primary treatment goal as a more important change goal in their life to address in a supportive and collaborative relationship with a doctor with a positive treatment expectancy are more likely to be placebo responders. Funding: NIH/NIDDK 77738