Liraglutide and Colesevelam Change Serum and Fecal Bile Acid Levels in a Randomized Trial With Patients With Bile Acid Diarrhea

Abstract

Introduction: Both liraglutide and colesevelam improve bile acid diarrhea symptoms. Colesevelam binds excess amounts of diarrhea-causing bile acids in the colon, whereas the mode of action for liraglutide remains elusive. In this article, we examined the impact of colesevelam and liraglutide treatment on the concentrations of bile acids in serum and feces and the fecal microbiota composition to better understand the 2 drugs' modes of action.

Methods: Bile acid species were analyzed in serum and fecal samples from a randomized, double-blind, double-dummy trial at baseline and after 3 and 6 weeks of orally administered colesevelam (1,875 mg twice daily, n = 26) or subcutaneously administered liraglutide (uptitrated by weekly increments of 0.6 mg from 0.6 to 1.8 mg daily, n = 26) in patients with 75 selenium-homotaurocholic acid test-verified, idiopathic, or postcholecystectomy bile acid diarrhea. Fecal microbiota composition was analyzed by 16S rRNA gene amplicon sequencing at the same time points.

Results: Colesevelam increased the fecal concentrations of all bile acid species, whereas it decreased serum concentrations of secondary bile acids. Liraglutide induced a small increase in serum unconjugated bile acid concentrations without affecting fecal bile acid concentrations. No changes in fecal microbiota composition were observed with either treatment.

Discussion: Colesevelam and liraglutide exhibit distinct effects on serum and fecal bile acid concentrations with colesevelam reducing serum concentrations of secondary bile acids and promoting fecal bile acid excretion, whereas liraglutide enhances serum concentrations of unconjugated bile acids, potentially through deceleration of small intestinal transit time allowing more time for passive absorption of bile acids.

Graphical abstract

Figure 1.

Total, primary, and secondary bile acids. Effect of 3 and 6 weeks of treatment with liraglutide (green circles) or colesevelam (blue triangles) on the levels of total bile acids (a, b), primary bile acids (c, d), and secondary bile acids (e, f) in serum (a, c, e) and in feces (b, d, f). Data are presented as log-transformed geometric means with 95% confidence intervals; P values are adjusted by the Benjamini and Hochberg procedure; asterisks represent statistical difference between data points at week 3 or 6 compared with baseline, daggers represent statistical significance in the estimated treatment difference; */†P < 0.05; **/††P < 0.01; ***/†††P < 0.001.

Figure 2.

Ratios of serum primary/secondary and unconjugated/conjugated bile acids. Effect of 3 and 6 weeks of treatment with liraglutide (green circles) or colesevelam (blue triangles) on the ratios between primary and secondary bile acids (a) and unconjugated and conjugated bile acids (b) in serum. Data are presented as log-transformed geometric means with 95% confidence intervals; P values are adjusted by the Benjamini and Hochberg procedure; asterisks represent statistical difference between data points at week 3 or 6 compared with baseline, daggers represent statistical significance in the estimated treatment difference; †P < 0.05; **/††P < 0.01; ***/†††P < 0.001.

Figure 3.

Unconjugated and conjugated bile acids. Effect of 3 and 6 weeks of treatment with liraglutide (green circles) or colesevelam (blue triangles) on the levels of unconjugated bile acids (a, b), conjugated bile acids (c, d), taurine-conjugated (T-conj.) bile acids (e, f), and glycine-conjugated (G-conj.) bile acids (g, h) in serum (a, c, e, g) and in feces (b, d, f, h). Data are presented as log-transformed geometric means with 95% confidence intervals; P values are adjusted by the Benjamini and Hochberg procedure; asterisks represent statistical difference between data points at week 3 or 6 compared with baseline, daggers represent statistical significance in the estimated treatment difference; */†P < 0.05; **/††P < 0.01; ***/†††P < 0.001. BA, bile acid; NA, not available.

Figure 4.

Cholic acid. Effect of 3 and 6 weeks of treatment with liraglutide (green circles) or colesevelam (blue triangles) on the levels of cholic acid (a, b), unconjugated cholic acid (c, d), taurine-conjugated (T-conj.) cholic acid (e, f), and glycine-conjugated (G-conj.) cholic acid (g, h) in serum (a, c, e, g) and in feces (b, d, f, h). Data are presented as log-transformed geometric means with 95% confidence intervals; P values are adjusted by the Benjamini and Hochberg procedure; asterisks represent statistical difference between data points at week 3 or 6 compared with baseline, daggers represent statistical significance in the estimated treatment difference; */†P < 0.05; **/††P < 0.01; ***/†††P < 0.001. CA, cholic acid; NA, not available.

Figure 5.

Chenodeoxycholic acid. Effect of 3 and 6 weeks of treatment with liraglutide (green circles) or colesevelam (blue triangles) on the levels of chenodeoxycholic acid (a, b), unconjugated chenodeoxycholic acid (c, d), and glycine-conjugated (G-conj.) chenodeoxycholic acid (e, f) in serum (a, c, e) and in feces (b, d, f). Taurin-conjugated chenodeoxycholic acid was below detection range in both serum and feces. Data are presented as log-transformed geometric means with 95% confidence intervals; P values are adjusted by the Benjamini and Hochberg procedure; asterisks represent statistical difference between data points at week 3 or 6 compared with baseline, daggers represent statistical significance in the estimated treatment difference; */†P < 0.05; **/††P < 0.01; ***/†††P < 0.001. CDCA, chenodeoxycholic acid; NA, not available.

Figure 6.

Deoxycholic acid. Effect of 3 and 6 weeks of treatment with liraglutide (green circles) or colesevelam (blue triangles) on the levels of deoxycholic acid (a, b), unconjugated deoxycholic acid (c, d), taurine-conjugated (T-conj.) deoxycholic acid (e, f), and glycine-conjugated (G-conj.) deoxycholic acid (g, h) in serum (a, c, e, g) and in feces (b, d, f, h). Data are presented as log-transformed geometric means with 95% confidence intervals; P values are adjusted by the Benjamini and Hochberg procedure; asterisks represent statistical difference between data points at week 3 or 6 compared with baseline, daggers represent statistical significance in the estimated treatment difference; */†P < 0.05; **/††P < 0.01; ***/†††P < 0.001. DCA, deoxycholic acid; NA, not available.

Figure 7.

Microbial beta and alpha diversity analysis. Principal coordinate analysis plot based on the Bray-Curtis dissimilarity metric (a), indicating no significant differences in relation to time or treatment (liraglutide and colesevelam). Differences were tested using a distance-based redundancy analysis, with results presented in the plot (highlighted in red font). Bacterial species richness and alpha diversity (b) based on Chao1 and Shannon indices in patients treated with liraglutide or colesevelam. (c) colesevelam; dbRDA, distance-based redundancy analysis; L, liraglutide; PCo, principal coordinate.