Effects of FGF19 Analogue Aldafermin in Patients With Bile Acid Diarrhea: A Randomized, Placebo-Control Trial

Irritable bowel syndrome (IBS) is a functional gastrointestinal disorder characterized by recurrent abdominal pain and altered bowel movements.1Mearin F. et al.Gastroenterology. 2016; 150: 1393-1407.e5Abstract Full Text Full Text PDF Scopus (1723) Google Scholar Approximately 25%–50% of patients with diarrhea-predominant IBS (IBS-D) have evidence of bile acid (BA) diarrhea (BAD).2Valentin N. et al.Gut. 2016; 65: 1951-1959Crossref PubMed Scopus (94) Google Scholar In patients with idiopathic BAD with IBS-D, there is decreased fibroblast growth factor 19 (FGF19) production3Johnston I.M. et al.Am J Gastroenterol. 2016; 111: 423-432Crossref PubMed Scopus (44) Google Scholar,4Walters J.R. et al.Clin Gastroenterol Hepatol. 2009; 7: 1189-1194Abstract Full Text Full Text PDF PubMed Scopus (255) Google Scholar and decreased negative feedback on BA synthesis, as manifested by increased serum 7α-hydroxy-4-cholesten-3-one (7αC4). BA sequestrants are approved for the treatment of hyperlipidemia and are used off label to treat BAD.5Camilleri M. Nurko S. Neurogastroenterol Motil. 2022; 34e14287Crossref Scopus (10) Google Scholar However, they might interfere with the absorption of other drugs. Aldafermin (NGM282), a 190–amino acid peptide, is an engineered analogue6Harrison S.A. et al.Lancet Gastroenterol Hepatol. 2022; 7: 603-616Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar of recombinant human FGF19 with a 95.4% homology. The aim of this study was to examine the effects of aldafermin on BA synthesis, BA excretion, and bowel function in patients with IBS-D and BAD. We performed a 28-day, randomized, double-blinded, placebo-controlled (1:1 ratio) trial of aldafermin vs placebo in patients with BAD (Supplemental Figure 1, upper panel). This study was conducted at the Mayo Clinic between December 2021 and November 2022. Block randomization ensured that equal numbers were randomized to placebo and aldafermin in 2 blocks: patients on BA sequestrants and patients off BA sequestrants. Patients were required to have a clinical diagnosis of IBS-D, age of 18–75 years, and body mass index of 18.0–45.0 kg/m2. Diagnosis of BAD was based on at least 1 of these criteria: (1) serum 7αC4 of ≥52 ng/mL; (2) fecal BA of >2337 μmol/48 h; (3) total fecal BA >1000 μmol/48 h + 4% primary BA; and (4) fecal primary BA of >10%/48 h. The coprimary endpoints were the fasting serum 7αC4 from baseline on day 28 and the stool consistency averaged over the 28-day study. The latter was originally listed as a secondary outcome in the original protocol. Further details regarding experimental design, ethical approval, eligibility criteria, assessment of endpoints and statistical analysis (based on 2-group comparisons using analysis of covariance) are provided in the Supplementary Material. Among 35 patients enrolled, 31 were randomized. One patient withdrew after 1 dose of aldafermin because of severe nausea unresponsive to antiemetics. Three patients in the aldafermin group and 2 in the placebo group remained on a stable dose of BA sequestrants during the study. The demographics, baseline and on-treatment bowel functions, and BA synthesis and excretion are reported in Table 1.Table 1Demographics, Baseline Characteristics, and Effects of Aldafermin and Placebo in 30 Patients With Bile Acid DiarrheaDataAldafermin (n = 15)Placebo (n = 15)P valueDemographics Age, y, median (IQR)49.0 (40.0 to 53.0)43.0 (39.0 to 58.0).633 Sex, % female80.093.3.282 BMI, kg/m2, median (IQR)31.5 (25.4 to 36.7)30.1 (23.2 to 35.7).561 Prior cholecystectomy, n (%)6 (40.0)5 (33.3).705 Off BA sequestrants, n (%)12 (80.0)13 (86.7).624Baseline bowel function Stool consistency, median (IQR)aBristol Stool Form Scale: 1–7.5.3 (4.1 to 5.6)4.5 (4.2 to 5.3).407 Abdominal pain, median (IQR)bNumerical scale: 0–10.0.9 (0.0 to 3.6)0.4 (0 to 1.0).377 Stool frequency, BM/day (IQR)2.3 (1.9 to 3.2)2.3 (1.8 to 3.0).901Baseline BA synthesis and excretion, median (IQR) Serum 7αC4, ng/mL40.7 (17.5 to 102.0)53.0 (18.7 to 82.4).804 Serum FGF19, pg/mL66.80 (34.2 to 80.8)94.30 (48.5 to 108.8).294 Random stool total BA, μmol/g3.1 (1.9 to 4.0)3.5 (1.9 to 4.3).913 % Primary BA, CA + CDCA5.4 (1.9 to 14.6)5.2 (2.0 to 12.9).896 % Secretory BA, DCA + CDCA63.8 (58.3 to 66.7)63.2 (57.4 to 66.6).777On treatment BA synthesis and excretion,cP value obtained using analysis of covariance, adjusted for the baseline measurement of the outcome, BMI, sex, and baseline serum 7αC4. median (IQR) Serum 7αC4, day 14, ng/mL2.6 (1.6 to 3.8)50.5 (19.4 to 122.0)<.001 Serum 7αC4, day 28, ng/mL2.2 (1.6 to 3.3)50.4 (25.8 to 106.0)<.001 Random stool total BA, day 14, μmol/g0.5 (0.3 to 0.7)2.7 (1.2 to 5.1).002 Random stool total BA, day 28, μmol/g0.5 (0.3 to 0.8)3.6 (1.0 to 7.3)<.001 % Primary BA, day 14, CA + CDCA10.3 (3.3 to 15.4)2.8 (0.7 to 9.8).738 % Primary BA, day 28, CA + CDCA9.9 (3.0 to 20.0)14.3 (7.4 to 27.4).417 % Secretory BA, day 14, DCA + CDCA43.2 (32.5 to 53.5)59.2 (53.1 to 63.9)<.001 % Secretory BA, day 28, DCA + CDCA45.5 (39.8 to 55.6)60.0 (46.3 to 65.1).033On treatment bowel function,cP value obtained using analysis of covariance, adjusted for the baseline measurement of the outcome, BMI, sex, and baseline serum 7αC4. median (IQR) Stool consistency, days 1–284.7 (4.3 to 5.1)4.6 (4.4 to 5.3).224 Individual change in stool consistency, days 1–28 minus baselinedP value obtained using Mann-Whitney rank sum test.–0.4 (–1.2 to 0.1)–0.02 (–0.4 to 0.3).125 Stool consistency, week 34.6 (4.3 to 5.0)4.8 (4.0 to 5.3).247 Stool consistency, week 44.4 (4.2 to 5.0)5.1 (4.4 to 5.5).047 Abdominal pain, days 1–280.6 (0.0 to 1.3)0.5 (0.1 to 1.1).649 Stool frequency, days 1–282.3 (2.1 to 3.6)2.1 (1.7 to 2.7).261NOTE. Stool consistency and frequency represent medians and IQRs of the individual averages over the specified time periods. Bolded P values are statistically significant.BM, bowel movements; BMI, body mass index; CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; IQR, interquartile range.a Bristol Stool Form Scale: 1–7.b Numerical scale: 0–10.c P value obtained using analysis of covariance, adjusted for the baseline measurement of the outcome, BMI, sex, and baseline serum 7αC4.d P value obtained using Mann-Whitney rank sum test. Open table in a new tab NOTE. Stool consistency and frequency represent medians and IQRs of the individual averages over the specified time periods. Bolded P values are statistically significant. BM, bowel movements; BMI, body mass index; CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; IQR, interquartile range. Patients on aldafermin had significantly decreased serum 7αC4 at days 14 and 28 (Supplementary Figure 1, lower panel) but no significant change in stool consistency during days 1–28, 1–14, or 15–28. The aldafermin-treated group also had significantly decreased fecal total BA and percent secretory BA at days 14 and 28. There were no significant differences in abdominal pain or stool frequency between the 2 groups during days 1–28, 1–14, or 15–28. However, there was numerically improved stool consistency in patients on aldafermin during days 15–28 (Bristol Stool Form Scale, 4.6 for aldafermin vs 4.7 for placebo; P = .082), particularly in week 4 of treatment (P = .047). Total IBS quality of life score and scores on each of the 8 quality of life domains were not significantly different on day 28 in patients on aldafermin or placebo. The aldafermin group had a greater increase in low-density lipoprotein cholesterol (LDL-C) from baseline compared to the placebo group (P = .052). Four patients on aldafermin were considered to have a clinically significant increase in LDL-C at day 28, which was normalized 1 month after cessation in 3 of the 4 patients. There were more adverse events in the group receiving aldafermin compared to those receiving placebo (Supplementary Table 1). One patient who was randomized to aldafermin was hospitalized with myocardial infarction with normal coronary angiogram, an event deemed unrelated to the study drug. Over 28 days, rescue medication with loperamide was received by 2 participants on placebo (totals: 6 mg and 16 mg) and 3 on aldafermin (totals: 20 mg, 20 mg, and 12 mg). In this study, we showed that aldafermin, 1 mg subcutaneous daily, led to a significant change in the biochemical measurements of BAD in patients with IBS-D. Improved stool consistency with aldafermin was only observed during week 4 of treatment. Although the biochemical changes associated with aldafermin have been previously demonstrated in other patient populations,6Harrison S.A. et al.Lancet Gastroenterol Hepatol. 2022; 7: 603-616Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar increased frequency of bowel movements with higher doses has been reported as a side effect of aldafermin,7Hirschfield G.M. et al.J Hepat. 2019; 70: 483-493Abstract Full Text Full Text PDF PubMed Scopus (111) Google Scholar,8Harrison S. et al.J Hepatol. 2017; 1: S92-S93Abstract Full Text PDF Google Scholar and acceleration of colonic transit documented with a higher dose of 6 mg NGM282 (aldafermin) was associated with increased frequency in contrast to the 1-mg dose used in the current study.9Oduyebo I. et al.Am J Gastroenterol. 2018; 113: 725-734Crossref PubMed Scopus (25) Google Scholar Overall, aldafermin had an acceptable side effect and tolerability profile. More patients on aldafermin had a clinically but not statistically significant increase in LDL-C levels at study completion (Supplementary Table 1). Aldafermin has been associated with increased LDL-C in prior studies8Harrison S. et al.J Hepatol. 2017; 1: S92-S93Abstract Full Text PDF Google Scholar secondary to decreased conversion of cholesterol to BA, and this could be reversed by coadministration of rosuvastatin.10Rinella M.E. et al.J Hepatol. 2019; 70: 735-744Abstract Full Text Full Text PDF PubMed Scopus (55) Google Scholar The strengths of this study include assessment of the effects of aldafermin on both biochemical endpoints and patient-reported outcomes; the primary biochemical endpoint was met in this small study sample. Although the small sample size in our study had sufficient power to detect >1 or 1.5 points on the 7-point Bristol Stool Form Scale, such a change may not be observed with most agents approved for IBS-D, and a larger sample size would have been needed to detect a significant effect on symptoms. Limitations include enrollment of 5 of 30 patients on BA sequestrants, although block randomization ensured equal distribution of patients across both arms and the inclusion of patients with stool frequency and consistency that were less severe than patients in other IBS-D trials used for approval of medications. Although this trial was conducted at a tertiary referral center, all participants were residents of the tristate area of Minnesota, Wisconsin, and Iowa seeking care in their communities for gastroenterologic disorders. Overall, these data suggest that this FGF19 analogue is a promising approach to the treatment of BAD. (Clinical Trial Registration: ClinicalTrials.gov, Number NCT05130047.) The authors thank Wassel Sannaa and Ting Zheng for the recruitment and medical care of participants during the study, Sanna McKinzie and Irene Busciglio for assistance in design of the protocol, Paula Carlson for biospecimen preparation and biochemical measurements, and Cindy Stanislav for secretarial assistance. Joelle BouSaba, MD (Investigation: Equal; Writing – original draft: Lead; Writing – review & editing: Equal). Monique Torres, (Investigation: Equal; Writing – review & editing: Supporting). Saam Dilmaghani, MD (Investigation: Equal; Writing – review & editing: Equal). W. Scott Harmsen, MS (Formal analysis: Lead; Writing – review & editing: Supporting). Lei Ling, MD (Resources: Equal; Writing – review & editing: Supporting). Michael Camilleri, MD (Conceptualization: Lead; Funding acquisition: Lead; Methodology: Lead; Resources: Lead; Writing – original draft: Equal; Writing – review & editing: Equal). The study was approved by Mayo Clinic institutional review board (IRB no. 21-009348) and registered at ClinicalTrials.gov (NCT05130047). All authors had access to the study data and reviewed and approved the final manuscript. The Consolidated Standards of Reporting Trials (CONSORT) checklist is included in the Supplementary Material. The Mayo Clinic Research Pharmacy received from the study statistician (W. Scott Harmsen) a computer-generated randomization schedule, using blocks of size 2 and 4. There was concealed allocation, and both study participants and staff were blinded to the assignments. All patients fulfilled Rome III or IV criteria for IBS-D based on clinical documentation in the electronic medical records. Patients on antidepressants or BA sequestrants were included only if they were on a stable dose of medication for 30 days before enrollment. Inclusion of patients on BA sequestrants was a modification of the original protocol. All females of child-bearing potential were required to be on a highly effective method of contraception for 30 days before, throughout study, and 90 days after study completion for male participants and through 30 days after study completion for female participants. The main exclusion criteria were structural, surgical, inflammatory, neoplastic or metabolic conditions affecting the gastrointestinal system; clinically relevant changes in dietary, lifestyle, or exercise regimen within 30 days or in medications that might affect the gastrointestinal system within 14 days of screening; history of cerebrovascular or cardiac disease or specific findings at screening: electrocardiography (heart rate of <40 or >100 beats/min, QTc interval of > 470 ms, QRS interval of ≥110 ms, PR interval of ≥220 ms); hepatic dysfunction (alanine aminotransferase or aspartate aminotransferase of >3× the upper limit of normal); total direct bilirubin or alkaline phosphatase of >2× the upper limit of normal); or renal insufficiency (creatinine of >2.5 mg/dL). Other exclusions were history of suicide attempt within 1 year; alcohol or substance use disorder within 2 years; and significant medical history that would prevent participation, alter the participant’s well-being, or affect the scientific validity of study assessments. Study treatments were 1 mg aldafermin and a matched placebo, both being clear liquids provided in clear syringes identical in appearance. Pharmacy personnel provided randomly assigned medication to blinded study personnel, and a nurse observed the patient properly self-administering the first dose into the abdomen. Patients self-administered the injection into the abdomen at home at a similar time each day for 28 days, preferably in the evening after dinner and before bed. Biochemical endpoints included fasting serum 7αC4 (measured by liquid chromatography–tandem mass spectrometry1Camilleri M. et al.Neurogastroenterol Motil. 2009; 21734–e43Crossref Scopus (5) Google Scholar,2Donato L.J. et al.Clin Biochem. 2018; 52: 106-111Crossref PubMed Scopus (18) Google Scholar), serum FGF-19 (by commercial enzyme-linked immunosorbent assay3Odunsi-Shiyanbade S.T. et al.Clin Gastroenterol Hepatol. 2010; 8: 159-165Abstract Full Text Full Text PDF PubMed Scopus (152) Google Scholar), and bile acids in stool (by liquid chromatography–tandem mass spectrometry, as detailed elsewhere4Wong B.S. et al.Clin Gastroenterol Hepatol. 2012; 101009–15.e3Google Scholar). Blood samples were collected before 9 AM (after a 12-hour fast) on days 1, 14, and 28. Concentration of total BA, individual BA (cholic acid [CA], chenodeoxycholic acid [CDCA], deoxycholic acid [DCA], lithocholic acid [LCA], ursodeoxycholic acid), percent primary BA (CA, CDCA), and percent secretory BA (DCA, CDCA) were measured in a single random stool sample provided by the participant on days 1, 14, and 28. Patients were requested to fill out a bowel pattern diary5Zinsmeister A.R. et al.Dig Dis Sci. 2013; 58: 509-518PubMed Google Scholar for 7 days before medication initiation and for 28 days while on aldafermin/placebo. Participants documented daily the number of bowel movements/day, the consistency of each bowel movement (as measured by the Bristol Stool Form Scale scale,6Heaton K. et al.Gut. 1992; 33: 818-824Crossref PubMed Scopus (574) Google Scholar 1–7), ease of passage of the bowel movement, sense of completeness of evacuation, and most severe abdominal pain (on a scale of 0–10, with 10 being most severe). The stool consistency was averaged across each day, each week, and each study period (days 1–14 and days 15–28). Participants filled out an IBS-specific, validated quality of life (QOL) questionnaire7Andrae D.A. et al.Health Qual Life Outcomes. 2013; 11: 208Crossref PubMed Scopus (60) Google Scholar at the beginning and at the end of the 28-day study period. The questionnaire consists of 34 questions related to 8 different domains (dysphoria, interference with activity, body image, health worry, food avoidance, social reaction, sexual, and relationship). The primary endpoints are detailed in the main body of the text. Secondary endpoints included absolute change in stool consistency from baseline to day 14 and to day 28 of treatment, average daily stool consistency at days 14 and 28, maximum abdominal pain score, absolute change in abdominal pain from baseline to days 14 and 28, stool frequency, total fecal BA concentration, change in total fecal BA concentration between baseline and day 28, proportion of fecal secretory BA (CDCA + DCA), and proportion of primary BA (CDCA + CA) in a single random stool sample. Usage of loperamide as a rescue medication was allowed, with a maximum dose of 4 mg in 24 hours except within 48 hours of a stool collection. Use of loperamide was recorded in the daily bowel diary. An electrocardiogram and safety laboratory tests (complete blood count, basic metabolic panel, liver function tests, creatinine, lipids, and urine analysis) were performed at study entry and on day 28. Adverse events reported by patients were evaluated, and study team members followed up with study participants. When appropriate, laboratory tests were repeated 1 month after study completion. Based on the endpoints of interest in patients with BAD,8Camilleri M. et al.Aliment Pharmacol Ther. 2015; 41: 438-448Crossref PubMed Scopus (92) Google Scholar the proposed sample sizes (n = 15 in each treatment arm, with anticipation of 6 patients on BA sequestrants in each group) were estimated to have ∼80% power (2-sided α = .05) to detect clinically relevant detectable difference in serum 7αC4 and stool form between the 2 groups (Table 1). Descriptive statistics are reported as number (percentage) for discrete variables and as median (interquartile range) for continuous variables. Wilcoxon and chi-square tests were used to compare baseline demographics, biochemical features, bowel function, lipid level, and QOL. Effects of treatment on bile acid synthesis and excretion and bowel functions were analyzed following intent-to-treat principles using analysis of covariance, with body mass index, sex, baseline serum 7αC4, and baseline measurements of the outcome of interest as covariates. Effects of treatment on lipid levels were similarly analyzed without baseline serum 7αC4 as a covariate. Effects of treatment on QOL were evaluated using analysis of covariance, with baseline measurements of outcome as a covariate. The Fisher exact test was used to compare the prevalence of adverse events in the aldafermin and placebo groups.Supplementary Table 1Adverse events and lipid levels compared in 2 treatment groups using the Fisher exact testAdverse events and lipid levelsAldafermin (n = 15)Placebo (n = 15)P valueAdverse events, n (%) Treatment-emergent adverse events14 (93.3)7 (46.7).014Grade 111 (73.3)6 (40.0).272Grade 25 (33.3)2 (13.3).389Grade 34 (26.7)3 (20.0).999Grade 40 (0)0 (0).999Grade 50 (0)0 (0).999 Serious adverse events1 (6.7)0 (0).999Most common treatment-emergent adverse events (≥10%), n (%) GastrointestinalIncreased diarrhea2 (13.3)1 (6.7).999Abdominal pain1 (6.7)2 (13.3).999Other gastrointestinal symptoms3 (20.0)3 (20.0).999 Increased LDL-C4 (26.7)0 (0).100 Skin disordersIrritation at injection site3 (20)1 (6.7).598 Infectious5 (33.3)2 (13.3).390 NeurologicHeadache2 (13.3)2 (13.3).999 Musculoskeletal3 (20)2 (13.3).999 Metabolism and nutrition2 (13.3)0 (0).483 Renal and urinary2 (13.3)2 (13.3).999Baseline lipids, median (IQR) Total cholesterol, mg/dL192.0 (175.0–211.0)210.0 (177.0–220.0).418 LDL-C, mg/dL114.0 (98.0–130.0)127.0 (103.0–137.0).534 HDL-C, mg/dL54.0 (42.0–67.0)57.0 (42.0–66.0).771 Triglycerides, mg/dL91.0 (68.0–226.0)120.0 (46.0–149.0).999On-treatment lipid levels,aP value obtained using analysis of covariance, adjusted for the baseline measurement of the outcome, body mass index, and sex. median (IQR) Total cholesterol, day 28, mg/dL198.0 (188.0–243.0)193.0 (174.0–226.0).145 LDL-C, day 28, mg/dL130.0 (106.0–153.0)117.0 (96.0–142.0).052 HDL-C, day 28, mg/dL53.0 (41.0–78.0)53.0 (41.0–69.0).039 Triglycerides, day 28, mg/dL88.0 (60.0–133.0)104.0 (74.0–176.0).007NOTE. Bolded P values are statistically significant.HDL-C, high-density lipoprotein cholesterol; IQR, interquartile range.a P value obtained using analysis of covariance, adjusted for the baseline measurement of the outcome, body mass index, and sex. Open table in a new tab NOTE. Bolded P values are statistically significant. HDL-C, high-density lipoprotein cholesterol; IQR, interquartile range. Download .pdf (.07 MB) Help with pdf files Supplementary Data