Fracture risk under selective and non-selective PPAR agonists in primary biliary cholangitis

Individuals with primary biliary cholangitis (PBC) have an increased risk of fractures and higher post-fracture mortality than the general population.1 We read with great interest—and some concern—the article by Dr Vierling, Dr Hirschfield, and colleagues reporting a potential imbalance in fracture risk in the randomized, placebo-controlled phase III RESPONSE trial evaluating the selective peroxisome proliferator-activated receptor δ (PPARδ) agonist seladelpar in patients with PBC and an incomplete response to ursodeoxycholic acid (UDCA).2 In this post hoc analysis, 5 bone fractures occurred in the seladelpar group (n=128), corresponding to an incidence rate of 4.1 fractures per 100 person-years, whereas no fractures were observed in the placebo group (n=65) during the 1-year study period.2 A similar finding was reported in the randomized, placebo-controlled phase III ELATIVE trial evaluating the dual PPARα/δ agonist elafibranor in a comparable PBC population, where 7 patients (6.5%) in the treatment group (n=108) developed fractures, compared with none in the placebo group (n=53).3 These data suggest that fracture incidence rate under elafibranor may reach up to 6.5 per 100 person-years. These observations are unexpected, as adverse effects on bone metabolism have been reported with selective PPARγ agonists, but not with PPARα or PPARδ agonists. Although interpretation is limited by small sample sizes and potential baseline differences in bone health, these findings prompted us to assess whether the non-selective pan-PPAR agonist bezafibrate is associated with a similar fracture risk. Bezafibrate, a pan-PPAR (α, δ, γ) agonist, is a well-tolerated and affordable second-line therapy for PBC patients with incomplete UDCA response, widely used in Eastern Asia, Europe, and elsewhere. In combination with UDCA, bezafibrate improved biochemical markers of cholestasis and noninvasive fibrosis parameters in the BEZURSO trial, reduced cholestatic pruritus in the FITCH trial, and, in real-world studies, was associated with improved 5-year histology and a 68% lower risk of mortality or liver transplantation over 15 years.4 We performed a post hoc analysis of fracture risk in the phase III BEZURSO trial cohort (n=100), the longest-duration (2-year) randomized, placebo-controlled study of a PPAR agonist in PBC.5 Results are summarized in Table 1. Serum vitamin D levels and vitamin D deficiency rates were comparable between groups at baseline and end of study (EOS). In the placebo group (n=50), 3 participants each sustained a single post-traumatic fracture (toe, thumb, rib). In the bezafibrate group (n=50), 3 fractures occurred in 2 participants: one sustained post-traumatic ankle and elbow fractures, and one had a vertebral compression fracture. Fracture incidence rates per 100 person-years (95% confidence interval) were 2.2 (0.6–8.8) in the placebo group and 3.1 (1.0–9.5) in the bezafibrate group, indicating no meaningful between-group difference. TABLE 1 - Vitamin D status and incidence of bone fractures during the 2-year BEZURSO trial Total (n=100) Placebo (n=50) Bezafibrate (n=50) 25-OH vitamin D level (ng/mL) Baseline 27.8 ± 17.3 29.2 ± 21.0 26.4 ± 12.8 EOS 28.4 ± 16.3 28.3 ± 13.4 28.6 ± 18.4 25-OH vitamin D <20 ng/mL Baseline 32 (32%) 15 (30%) 17 (34%) EOS 25 (25%) 11 (22%) 14 (28%) 25-OH vitamin D <10 ng/mL Baseline 8 (8%) 4 (8%) 4 (8%) EOS 4 (4%) 1 (2%) 3 (6%) Cumulative fractures Number of fractures 6 (6%) 3 (6%) 3 (6%) Patients with fractures 5 (5%) 3 (6%) 2 (4%) Fracture incidence rate (per 100 person-years) 2.6 (1.1–6.4) 2.2 (0.6–8.8) 3.1 (1.0–9.5) Note: Data are presented as mean ± standard deviation for 25-OH vitamin D serum levels; number (percentage) for vitamin D deficiency and cumulative fractures; and mean (95% confidence interval) for fracture incidence rates. Fracture incidence rates were estimated using a Poisson regression model with an offset term.Abbreviations: EOS, end of study; PBC, primary biliary cholangitis. In conclusion, these data suggest—albeit cautiously, given the limited number of events—that selective PPARδ and PPARα/δ agonists (seladelpar and elafibranor), but not the pan-PPAR agonist bezafibrate, may be associated with an increased fracture risk during treatment. These preliminary findings warrant confirmation in larger, long-term studies across independent cohorts of patients with cholestatic disease treated with selective and non-selective PPAR agonists. Further research is needed to clarify the effects of these agents on bone metabolism in cholestatic conditions.