2025 年 32 巻 11 号 p. 1400-1415
Aim: Recently, we reported that a pemafibrate extended-release (XR) formulation lowered low-density lipoprotein cholesterol (LDL-C) and cholesterol synthesis and absorption markers in a phase 2 clinical pharmacology study. Here we describe our post-hoc analysis of that study, discuss the mechanism by which pemafibrate lowers LDL-C, and suggest which patients may respond favorably to pemafibrate treatment.
Methods: In the phase 2 study, patients with hypertriglyceridemia received treatment with pemafibrate immediate-release (IR) 0.2 mg/day or XR 0.4 mg/day or 0.8 mg/day. This post-hoc subgroup analysis examined the percentage change in LDL-C, apolipoprotein B (ApoB), non-HDL-C, and cholesterol synthesis and absorption markers, in subgroups by baseline LDL-C, and then determined the correlation between the percentage change in LDL-C and the percentage change in cholesterol synthesis and absorption markers.
Results: Our analysis included 60 patients who received two of three formulations of the drug. A total of 78.3% (47/60) were male, 16.7% (10/60) had type 2 diabetes mellitus, and 10% (6/60) received concomitant statins. The percentage of LDL-C lowering was greater in the population with high baseline LDL-C, and similar trends were noted for the ApoB, non-HDL-C, and cholesterol synthesis and absorption markers. The percentage change in LDL-C was positively correlated with the percentage change in lathosterol, β-sitosterol, and campesterol.
Conclusions: In patients with hypertriglyceridemia, results suggested that pemafibrate lowered LDL-C by inhibiting cholesterol synthesis in the liver and cholesterol absorption from the intestinal tract. This lowering effect was greater in populations with higher baseline LDL-C.
See editorial vol. 32: 1372-1374
Numerous epidemiological studies have shown that elevated low-density lipoprotein cholesterol (LDL-C) is associated with a high incidence of atherosclerotic cardiovascular disease (ASCVD)1). The use of statins to lower LDL-C can reduce the risk of ASCVD2), and the percentage of LDL-C lowering correlates with that risk reduction3). Notably, the significant decline in coronary artery disease mortality in the United States over the last several decades has been attributed to the widespread use of evidence-based treatment for hypertension and dyslipidemia (e.g., renin-angiotensin-aldosterone system inhibitors and statins)4). To further reduce the risk of ASCVD, statins, ezetimibe, and proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors are now being recommended for hyper LDL-cholesterolemia5). In addition, fibrates, which are widely used in the treatment of lipid disorders in patients with type 2 diabetes, primarily target hypertriglyceridemia. These drugs have been reported to lower LDL-C6), and a recent meta-analysis has demonstrated a positive correlation between the extent of LDL-C lowering achieved by fibrates and the corresponding decrease in ASCVD risk7).
Pemafibrate, a selective peroxisome proliferator-activated receptor modulator alpha (SPPARMα), has higher peroxisome proliferator-activated receptor alpha (PPARα) selectivity than conventional fibrates, with a favorable balance of efficacy and safety8). It primarily lowers triglyceride (TG) and increases high-density lipoprotein cholesterol (HDL-C), with few renal and hepatic adverse events. It is widely used in Asia for the treatment of hypertriglyceridemia. Regarding the effects of pemafibrate on LDL-C, results have been inconsistent: many past trials have reported that the effects were negligible9), while the PROMINENT (Pemafibrate to Reduce Cardiovascular Outcomes by Reducing Triglycerides in Patients with Diabetes) trial found that pemafibrate increased LDL-C10), and at least one trial conducted on patients with MASLD has linked pemafibrate to decreased LDL-C11, 12).
We recently reported the findings from a phase 2 clinical pharmacology study of an extended-release (XR) formulation of pemafibrate. Results showed a 3–11% reduction in LDL-C at four weeks of pemafibrate treatment, with similar reductions in markers for cholesterol synthesis and absorption13). To further understand the implications of these findings, we subjected that data to post-hoc analysis.
We performed a post-hoc analysis of the phase 2 clinical pharmacology study data of pemafibrate XR to assess the mechanism by which pemafibrate lowers LDL-C and to identify populations in which this reduction is likely to be achieved.
We conducted a post-hoc subgroup analysis of a phase 2 clinical pharmacology trial on pemafibrate XR tablets from September to December 2019 at three sites in Japan as a multicenter, randomized, single-blind, active-controlled, 12-arm, two-phase, crossover study. Details of the study design and criteria have been described previously13). Pemafibrate doses used in this study were immediate-release (IR) 0.2 mg/day, XR 0.4 mg/day, and XR 0.8 mg/day. Sixty subjects with hypertriglyceridemia were randomly assigned to 12 groups and received different formulations and doses of pemafibrate during each four-week treatment period (Supplementary Fig.1).
Modified from a Figure in Reference 13).
§A washout period is implemented if subjects have been using drugs that may affect the efficacy and pharmacokinetic evaluation of pemafibrate.
†Randomization of subjects eligible for study participation based on screening test findings.
IR, immediate-release; XR, extended-release; PK, pharmacokinetics; PD, pharmacodynamics.Disposition of patients
TG, triglycerides; IR, immediate-release; XR, extended-release.
All laboratory tests used in this post-hoc analysis were prespecified according to the study protocol. Blood samples were collected after patients had fasted for at least 10 hours. Tests at screening and during treatment periods 1 and 2 were performed every two weeks from Week 0 to Week 8. The prespecified laboratory assessments were centrally performed primarily by LSI Medience Corporation (Tokyo, Japan) and some tests (ApoB48, lathosterol, campesterol, β-sitosterol) conducted by SRL Inc. (Tokyo, Japan). LDL-C was measured using a homogeneous assay (direct method). Lathosterol, campesterol, and β-sitosterol were measured by gas chromatography used in previous studies14, 15).
EndpointsEndpoints were the percentage change from baseline to four weeks for each treatment period in total cholesterol, LDL-C, apolipoprotein B (ApoB), non-HDL-C, lathosterol, β-sitosterol, and campesterol, as well as the changes in lathosterol, β-sitosterol, and campesterol when normalized for total cholesterol levels. Essentially, baseline values were calculated as the mean of measurements that were taken one day before drug administration in period 1 and at Week 0 of the treatment period.
Statistical AnalysisThe analysis population was the full analysis set (FAS), defined as the population of randomly assigned subjects who had at least one dose of the study drug and had baseline and post-baseline measurements for the endpoints. In this study, patients were divided into groups based on the timing of drug administration (before meals and after meals). For the present analysis, the before meals and after meals were combined into a single group. We evaluated the percentage change from baseline to four weeks for each treatment period in LDL-C, ApoB, non-HDL-C, lathosterol, β-sitosterol, and campesterol in subgroups of baseline LDL-C ≥ 100 mg/dL, ≥ 120 mg/dL, ≥ 140 mg/dL, and ≥ 160 mg/dL. The LDL-C cutoff levels were based on the target values for lipid management by risk category, obtained from the Japan Atherosclerosis Society (JAS) Guidelines for Prevention of Atherosclerotic Cardiovascular Diseases 2022 16). The two-sided significance level was 5%, and the confidence coefficient was 95%. Differences in least square means (LS means) and two-sided 95% confidence intervals (CI) were calculated for each XR group in relation to the IR group, using a marginal model and assuming compound symmetry in the error variance-covariance matrix for each subject. In this calculation, fixed effects were allocation groups pooled by the timing of drug administration (before meals, after meals), treatment period (period 1, period 2), treatment (IR 0.2 mg/day, XR 0.4 mg/day, XR 0.8 mg/day), baseline values, site, sex, and concomitant statin use or non-use. A t-test was used for the comparative superiority of the XR formulation over the IR formulation. The correlation coefficient (r) and 95% prediction ellipse were calculated by plotting the percentage change in LDL-C versus the percentage change in lathosterol, β-sitosterol, and campesterol at four weeks of treatment for IR 0.2 mg/day, XR 0.4 mg/day, and XR 0.8 mg/day, respectively, for each treatment period. SAS ver. 9.4 or higher was used for analysis.
Data were analyzed for 60 patients who received the treatment13). Briefly, 87 patients gave written consent, of whom 63 eligible patients were randomly assigned to each group, and three of those patients were withdrawn from the study after enrollment but before receiving the study drug.
Patient characteristics for this population have been reported previously13). Age was 57.5±9.8 years (mean±standard deviation), body weight was 71.5±13.6 kg, and body mass index (BMI) was 25.5±3.7 kg/m2. A total of 78.3% (47/60) were male, 16.7% (10/60) had type 2 diabetes mellitus, 53.3% (32/60) had metabolic syndrome, 10.0% (6/60) were taking concomitant statins, and no patient was taking concomitant ezetimibe. Baseline fasting TG was 221.3±68.1 mg/dL, LDL-C (direct method) was 134.5±31.2 mg/dL, non-HDL-C was 174.1±33.7 mg/dL, and ApoB was 104.5±16.8 mg/dL (Supplementary Table 1).
|
Total (n = 60) |
IR 0.2 mg/day (n = 40) |
XR 0.4 mg/day (n = 40) |
XR 0.8 mg/day (n = 40) |
|
|---|---|---|---|---|
| Age, years | 57.5±9.8 | 57.4±9.6 | 56.8±10.1 | 58.4±9.7 |
| Male | 47 (78.3) | 33 (82.5) | 30 (75.0) | 31 (77.5) |
| Weight, kg | 71.5±13.6 | 73.0±13.7 | 69.9±13.0 | 71.5±14.0 |
| BMI, kg/m2 | 25.5±3.7 | 25.9±3.8 | 25.2±3.4 | 25.5±4.0 |
| TG, mg/dL | 221.3±68.1 | 220.9±76.3 | 229.0±68.3 | 214.0±58.6 |
| LDL-C (direct), mg/dL | 134.5±31.2 | 132.3±29.8 | 131.2±30.6 | 140.0±32.8 |
| HDL-C, mg/dL | 42.6±9.3 | 42.4±10.5 | 43.0±9.1 | 42.5±8.4 |
| non-HDL-C, mg/dL | 174.1±33.7 | 171.7±30.3 | 172.2±35.4 | 178.4±35.2 |
| Creatinine, mg/dL | 0.85±0.18 | 0.86±0.19 | 0.82±0.15 | 0.85±0.19 |
| eGFR, mL/min/1.73m2 | 71.3±13.1 | 70.9±13.7 | 72.8±12.9 | 70.2±12.9 |
| HbA1c, % | 5.9±0.5 | 5.9±0.5 | 5.8±0.4 | 5.9±0.5 |
| Concomitant disease | ||||
| Hypertension | 17 (28.3) | 11 (27.5) | 11 (27.5) | 12 (30.0) |
| Type 2 diabetes mellitus | 10 (16.7) | 8 (20.0) | 3 (7.5) | 9 (22.5) |
| Metabolic syndrome | 32 (53.3) | 22 (55.0) | 20 (50.0) | 22 (55.0) |
| Concomitant statin use | 6 (10.0) | 4 (10.0) | 5 (12.5) | 3 (7.5) |
Data are presented as mean±standard deviation for continuous parameters and n (%) for categorical parameters. Reprint from a Table in Reference 13).
IR, immediate-release; XR, extended-release; BMI, body mass index; TG, triglycerides; LDL-C, low-density lipoprotein cholesterol; HDL-C, high- density lipoprotein cholesterol; eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c.
Fig.1 shows the LS means of the percentage change from baseline to four weeks for LDL-C, ApoB, non-HDL-C, lathosterol, β-sitosterol, and campesterol in the overall population for this post-hoc analysis. For lathosterol, the LS means of the percentage change showed statistically significant decreases in all three groups (−15.8% for IR 0.2 mg/day, −19.9% for XR 0.4 mg/day, and −24.7% for XR 0.8 mg/day). The percentage of lathosterol lowering was significantly greater for XR 0.8 mg/day than IR 0.2 mg/day, with a difference in the LS means (95% CI) of −4.1% (−10.4%, 2.2%) in the XR 0.4 mg/day group and −8.9% (−15.3%, −2.6%) in the XR 0.8 mg/day group (Supplementary Table 2). The LS means of the percentage change from baseline to four weeks for β-sitosterol and for campesterol showed statistically significant lowering in all groups, but no differences in the percentage of lowering between treatment groups (β-sitosterol: −26.4% to −24.8%; campesterol: −23.7% to −21.2%, respectively). The LS means of the percentage change from baseline to four weeks for lathosterol and β-sitosterol showed statistically significant lowering in all groups even after values were normalized for total cholesterol levels (Supplementary Fig.2).
Data are presented as (number of patients) [mean (mg/dL)] or <mean (mg/L)> for baseline levels and as LS mean (95%CI) for change in level. ＊p＜0.05, ＊＊p＜0.01, ＊＊＊p＜0.001 (vs. IR 0.2 mg/day).
IR, immediate-release; XR, extended-release; LDL-C, low-density lipoprotein cholesterol; Apo, apolipoproteins; HDL-C, high-density lipoprotein cholesterol; LS mean, least square mean; CI, confidence interval
| (A) LDL-C (direct method), mg/dL | ||||
| IR 0.2 mg/day | XR 0.4 mg/day | XR 0.8 mg/day | ||
| Total | n | 40 | 40 | 39 |
| Baseline | 132.3±29.8 | 131.2±30.6 | 140.3±33.2 | |
| Week 4 | 124.3±30.5 | 121.8±29.4 | 117.1±23.6 | |
| % Change | −3.5 (−8.4, 1.4) | −7.0 (−11.9, −2.1) | −11.3 (−16.2, −6.4) | |
| Differences (vs. IR0.2) | −3.5 (−8.0, 0.9) | −7.8 (−12.3, −3.3) ＊＊ | ||
| Baseline LDL-C levels | ||||
| ≥ 100 mg/dL | n | 35 | 33 | 36† |
| Baseline | 139.6±23.8 | 141.3±22.4 | 146.3±28.0 | |
| Week 4 | 129.4±28.1 | 128.5±27.2 | 120.1±22.7 | |
| % Change | −5.8 (−10.9, −0.8) | −9.7 (−14.8, −4.5) | −14.7 (−19.7, −9.6) | |
| Differences (vs. IR0.2) | −3.8 (−8.9, 1.2) | −8.8 (−13.7, −3.9) ＊＊＊ | ||
| ≥ 120 mg/dL | n | 27 | 28 | 29‡ |
| Baseline | 147.8±20.8 | 146.6±20.0 | 155.1±23.6 | |
| Week 4 | 131.8±27.2 | 129.4±27.8 | 123.0±22.9 | |
| % Change | −9.3 (−14.6, −4.1) | −13.2 (−18.4, −8.1) | −18.6 (−23.8, −13.5) | |
| Differences (vs. IR0.2) | −3.9 (−10.0, 2.2) | −9.3 (−15.3, −3.3) ＊＊ | ||
| ≥ 140 mg/dL | n | 13 | 14 | 19 |
| Baseline | 164.2±18.6 | 161.1±18.7 | 167.5±19.7 | |
| Week 4 | 136.8±28.1 | 132.6±29.4 | 125.5±22.6 | |
| % Change | −14.1 (−20.9, −7.3) | −18.4 (−25.0, −11.7) | −25.4 (−31.6, −19.1) | |
| Differences (vs. IR0.2) | −4.3 (−10.9, 2.3) | −11.3 (−16.9, −5.7) ＊＊＊ | ||
| ≥ 160 mg/dL | n | 6 | 5 | 11 |
| Baseline | 180.5±14.2 | 180.4±17.6 | 180.5±14.9 | |
| Week 4 | 142.8±23.8 | 134.8±34.2 | 125.7±20.8 | |
| % Change | −20.5 (−34.2, −6.9) | −26.1 (−39.9, −12.3) | −30.4 (−44.4, −16.3) | |
| Differences (vs. IR0.2) | −5.6 (−19.0, –7.9) | −9.9 (−18.9, −0.8) ＊ | ||
| (B) ApoB, mg/dL | ||||
| IR 0.2 mg/day | XR 0.4 mg/day | XR 0.8 mg/day | ||
| Total | n | 40 | 40 | 39 |
| Baseline | 103.2±15.7 | 103.0±17.0 | 107.4±17.7 | |
| Week 4 | 92.9±17.6 | 91.5±18.2 | 90.4±14.6 | |
| % Change | −9.8 (−13.0, −6.6) | −11.9 (−15.1, −8.7) | −13.4 (−16.7, −10.2) | |
| Differences (vs. IR0.2) | −2.1 (−5.3, 1.1) | −3.7 (−6.9, −0.4) ＊ | ||
| Baseline LDL-C levels | ||||
| ≥ 100 mg/dL | n | 35 | 33 | 36† |
| Baseline | 106.6±13.5 | 108.1±13.8 | 110.8±14.6 | |
| Week 4 | 95.2±16.7 | 95.1±17.4 | 92.3±14.1 | |
| % Change | −11.2 (−14.6, −7.7) | −12.7 (−16.2, −9.2) | −14.7 (−18.1, −11.3) | |
| Differences (vs. IR0.2) | −1.5 (−5.2, 2.1) | −3.5 (−7.1, 0.0) | ||
| ≥ 120 mg/dL | n | 27 | 28 | 29‡ |
| Baseline | 110.1±12.6 | 110.1±13.5 | 115.3±12.2 | |
| Week 4 | 96.6±15.9 | 95.2±17.6 | 94.0±14.6 | |
| % Change | −12.5 (−16.3, −8.7) | −14.5 (−18.3, −10.8) | −16.9 (−20.7, −13.2) | |
| Differences (vs. IR0.2) | −2.0 (−6.3, 2.2) | −4.4 (−8.7, −0.2) ＊ | ||
| ≥ 140 mg/dL | n | 13 | 14 | 19 |
| Baseline | 117.8±11.1 | 119.1±11.6 | 120.9±11.0 | |
| Week 4 | 99.2±17.3 | 97.9±19.2 | 95.0±14.8 | |
| % Change | −15.4 (−20.3, −10.5) | −17.8 (−22.6, −12.9) | −22.0 (−26.6, −17.4) | |
| Differences (vs. IR0.2) | −2.3 (−6.8, 2.2) | −6.5 (−10.4, −2.7) ＊＊ | ||
| ≥ 160 mg/dL | n | 6 | 5 | 11 |
| Baseline | 124.2±10.2 | 128.2±11.7 | 126.0±10.6 | |
| Week 4 | 100.7±12.9 | 95.8±21.7 | 93.6±14.1 | |
| % Change | −21.0 (−29.5, −12.5) | −23.2 (−31.9, −14.5) | −25.9 (−34.5, −17.3) | |
| Differences (vs. IR0.2) | −2.2 (−11.6, 7.1) | −4.9 (−11.1, 1.4) | ||
| (C) non-HDL-C, mg/dL | ||||
| IR 0.2 mg/day | XR 0.4 mg/day | XR 0.8 mg/day | ||
| Total | n | 40 | 40 | 39 |
| Baseline | 171.7±30.3 | 172.2±35.4 | 178.7±35.6 | |
| Week 4 | 151.9±34.2 | 150.6±34.5 | 146.0±27.3 | |
| % Change | −10.7 (−14.6, −6.9) | −12.8 (−16.6, −9.0) | −15.8 (−19.6, −12.0) | |
| Differences (vs. IR0.2) | −2.1 (−5.9, 1.7) | −5.1 (−8.9, −1.2) ＊ | ||
| Baseline LDL-C levels | ||||
| ≥ 100 mg/dL | n | 35 | 33 | 36† |
| Baseline | 178.6±24.5 | 182.7±28.2 | 185.6±28.8 | |
| Week 4 | 157.1±31.3 | 157.8±32.0 | 149.7±25.9 | |
| % Change | −11.8 (−15.8, −7.8) | −14.1 (−18.2, −10.0) | −17.6 (−21.5, −13.6) | |
| Differences (vs. IR0.2) | −2.3 (−6.7, 2.1) | −5.7 (−10.0, −1.5) ＊＊ | ||
| ≥ 120 mg/dL | n | 27 | 28 | 29‡ |
| Baseline | 185.1±23.2 | 187.9±27.0 | 193.7±25.3 | |
| Week 4 | 159.6±29.4 | 157.9±33.1 | 152.8±27.0 | |
| % Change | −13.7 (−18.0, −9.5) | −16.6 (−20.7, −12.4) | −19.9 (−24.0, −15.7) | |
| Differences (vs. IR0.2) | −2.9 (−7.9, 2.1) | −6.1 (−11.1, −1.2) ＊ | ||
| ≥ 140 mg/dL | n | 13 | 14 | 19 |
| Baseline | 197.3±18.8 | 202.4±24.2 | 206.6±20.7 | |
| Week 4 | 163.3±31.6 | 161.4±34.7 | 155.9±26.4 | |
| % Change | −17.0 (−22.6, −11.4) | −19.7 (−25.1, −14.2) | −24.9 (−30.1, −19.7) | |
| Differences (vs. IR0.2) | −2.7 (−7.7, 2.4) | −7.9 (−12.2, −3.6) ＊＊ | ||
| ≥ 160 mg/dL | n | 6 | 5 | 11 |
| Baseline | 210.7±16.6 | 227.2±20.8 | 218.2±19.6 | |
| Week 4 | 167.8±26.4 | 166.0±43.5 | 154.5±25.7 | |
| % Change | −22.3 (−33.5, −11.0) | −25.0 (−36.6, −13.5) | −29.2 (−40.4, −18.0) | |
| Differences (vs. IR0.2) | −2.8 (−15.6, 10.1) | −6.9 (−15.6, 1.7) | ||
| (D) Lathosterol, mg/L | ||||
| IR 0.2 mg/day | XR 0.4 mg/day | XR 0.8 mg/day | ||
| Total | n | 40 | 40 | 40 |
| Baseline | 3.7±1.7 | 3.6±1.8 | 3.7±1.6 | |
| Week 4 | 2.9±1.3 | 2.8±1.2 | 2.6±1.1 | |
| % Change | −15.8 (−21.4, −10.1) | −19.9 (−25.6, −14.2) | −24.7 (−30.4, −19.0) | |
| Differences (vs. IR0.2) | −4.1 (−10.4, 2.2) | −8.9 (−15.3, −2.6) ＊＊ | ||
| Baseline LDL-C levels | ||||
| ≥ 100 mg/dL | n | 35 | 33 | 36 |
| Baseline | 3.9±1.7 | 3.8±1.7 | 3.9±1.5 | |
| Week 4 | 3.0±1.2 | 3.0±1.1 | 2.7±1.0 | |
| % Change | −15.8 (−22.0, −9.7) | −20.2 (−26.5, −13.9) | −26.4 (−32.5–−20.3) | |
| Differences (vs. IR0.2) | −4.4 (−11.1, 2.4) | −10.5 (−17.0, −4.1) ＊＊ | ||
| ≥ 120 mg/dL | n | 27 | 28 | 29 |
| Baseline | 3.9±1.8 | 4.0±1.7 | 4.0±1.6 | |
| Week 4 | 3.0±1.2 | 3.1±1.1 | 2.8±1.1 | |
| % Change | −17.5 (−24.5, −10.5) | −21.3 (−28.2, −14.5) | −26.4 (−33.2, −19.6) | |
| Differences (vs. IR0.2) | −3.8 (−11.8, 4.1) | −8.8 (−16.6, −1.1) ＊ | ||
| ≥ 140 mg/dL | n | 13 | 14 | 19 |
| Baseline | 4.3±1.7 | 4.4±1.6 | 4.5±1.6 | |
| Week 4 | 3.2±1.1 | 3.3±1.1 | 3.1±1.0 | |
| % Change | −17.4 (−28.2, −6.6) | −23.9 (−34.4, −13.3) | −28.4 (−38.1, −18.7) | |
| Differences (vs. IR0.2) | −6.5 (−18.1, 5.2) | −11.0 (−21.0, −1.1) ＊ | ||
| ≥ 160 mg/dL | n | 6 | 5 | 11 |
| Baseline | 4.8±1.7 | 4.3±1.0 | 4.5±1.4 | |
| Week 4 | 3.4±1.3 | 2.8±1.1 | 2.7±0.8 | |
| % Change | −24.6 (−47.6, −1.6) | −36.5 (−60.2, −12.8) | −40.3 (−63.0, −17.6) | |
| Differences (vs. IR0.2) | −11.9 (−39.2, 15.4) | −15.7 (−34.0, 2.7) | ||
| (E) β-Sitosterol, mg/L | ||||
| IR 0.2 mg/day | XR 0.4 mg/day | XR 0.8 mg/day | ||
| Total | n | 40 | 40 | 40 |
| Baseline | 2.2±0.8 | 2.1±0.6 | 2.2±0.8 | |
| Week 4 | 1.6±0.6 | 1.5±0.5 | 1.6±0.5 | |
| % Change | −24.8 (−28.6, −20.9) | −26.4 (−30.3, −22.5) | −26.4 (−30.3, −22.5) | |
| Differences (vs. IR0.2) | −1.6 (−6.0, 2.7) | −1.6 (−6.0, 2.7) | ||
| Baseline LDL-C levels | ||||
| ≥ 100 mg/dL | n | 35 | 33 | 36 |
| Baseline | 2.3±0.8 | 2.1±0.7 | 2.2±0.8 | |
| Week 4 | 1.7±0.6 | 1.5±0.5 | 1.6±0.5 | |
| % Change | −26.4 (−30.5, −22.3) | −26.9 (−31.0, −22.7) | −27.5 (−31.5, −23.4) | |
| Differences (vs. IR0.2) | −0.5 (−5.2, 4.3) | −1.1 (−5.6, 3.5) | ||
| ≥ 120 mg/dL | n | 27 | 28 | 29 |
| Baseline | 2.4±0.8 | 2.1±0.6 | 2.2±0.9 | |
| Week 4 | 1.7±0.6 | 1.5±0.5 | 1.5±0.5 | |
| % Change | −28.3 (−33.0, −23.7) | −28.5 (−33.1, −23.9) | −29.6 (−34.1, −25.1) | |
| Differences (vs. IR0.2) | −0.1 (−5.8, 5.5) | −1.2 (−6.8, 4.3) | ||
| ≥ 140 mg/dL | n | 13 | 14 | 19 |
| Baseline | 2.3±0.6 | 2.0±0.7 | 2.0±0.5 | |
| Week 4 | 1.6±0.4 | 1.4±0.4 | 1.4±0.5 | |
| % Change | −30.4 (−37.7, −23.2) | −27.0 (−34.2, −19.9) | −30.8 (−37.4, −24.1) | |
| Differences (vs. IR0.2) | 3.4 (−4.1, 10.8) | −0.4 (−6.7, 6.0) | ||
| ≥ 160 mg/dL | n | 6 | 5 | 11 |
| Baseline | 2.3±0.5 | 2.1±0.5 | 2.2±0.5 | |
| Week 4 | 1.6±0.3 | 1.4±0.3 | 1.4±0.4 | |
| % Change | −32.6 (−49.2, −16.0) | −27.0 (−43.7, −10.3) | −34.9 (−52.1, −17.7) | |
| Differences (vs. IR0.2) | 5.6 (−9.8, 21.1) | −2.3 (−12.6, 8.1) | ||
| (F) Campesterol, mg/L | ||||
| IR 0.2 mg/day | XR 0.4 mg/day | XR 0.8 mg/day | ||
| Total | n | 40 | 40 | 40 |
| Baseline | 3.9±1.4 | 3.8±1.1 | 4.0±1.5 | |
| Week 4 | 3.1±1.1 | 2.9±1.0 | 2.9±0.9 | |
| % Change | −21.2 (−25.1, −17.3) | −23.3 (−27.2, −19.5) | −23.7 (−27.6, −19.8) | |
| Differences (vs. IR0.2) | −2.1 (−6.8, 2.6) | −2.5 (−7.2, 2.3) | ||
| Baseline LDL-C levels | ||||
| ≥ 100 mg/dL | n | 35 | 33 | 36 |
| Baseline | 4.1±1.4 | 3.9±1.1 | 4.0±1.6 | |
| Week 4 | 3.2±1.1 | 3.0±1.0 | 2.9±0.9 | |
| % Change | −22.2 (−26.3, −18.1) | −24.2 (−28.4, −20.0) | −25.0 (−29.0, −21.0) | |
| Differences (vs. IR0.2) | −2.0 (−7.2, 3.2) | −2.8 (−7.8, 2.1) | ||
| ≥ 120 mg/dL | n | 27 | 28 | 29 |
| Baseline | 4.3±1.5 | 3.8±1.0 | 4.0±1.5 | |
| Week 4 | 3.2±1.2 | 2.8±0.9 | 2.8±0.8 | |
| % Change | −24.7 (−29.7, −19.8) | −25.9 (−30.7, −21.0) | −26.8 (−31.6, −22.0) | |
| Differences (vs. IR0.2) | −1.1 (−7.3, 5.1) | −2.1 (−8.1, 4.0) | ||
| ≥ 140 mg/dL | n | 13 | 14 | 19 |
| Baseline | 4.1±0.9 | 3.7±1.0 | 3.8±0.9 | |
| Week 4 | 2.9±0.7 | 2.6±0.7 | 2.7±0.8 | |
| % Change | −27.9 (−35.2, −20.5) | −26.1 (−33.3, −18.9) | −28.9 (−35.5, −22.4) | |
| Differences (vs. IR0.2) | 1.8 (−6.4, 9.9) | −1.0 (−8.0, 5.9) | ||
| ≥ 160 mg/dL | n | 6 | 5 | 11 |
| Baseline | 3.9±1.1 | 3.9±0.9 | 3.9±1.0 | |
| Week 4 | 2.8±0.5 | 2.7±0.6 | 2.6±0.6 | |
| % Change | −28.0 (−43.6, −12.3) | −25.3 (−41.4, −9.2) | −32.5 (−47.9, −17.1) | |
| Differences (vs. IR0.2) | 2.7 (−16.0, 21.4) | −4.6 (−17.1, 8.0) | ||
Data are presented as mean±SD for values at baseline and Week 4, and LS mean (95%CI) for % change and differences. Differences are vs. IR 0.2 mg/day. ＊p＜0.05, ＊＊p＜0.01, ＊＊＊p＜0.001 (vs. IR 0.2 mg/day). †n = 35, ‡n = 28 for week 4.
LDL-C, low-density lipoprotein cholesterol; IR, immediate-release; XR, extended-release; Apo, apolipoproteins; SD, standard deviation; LS mean, least square mean; CI, confidence interval.
Data are presented as (number of patients) [mean (mg/dL)] or <mean (mg/dL / mg/dL)> for baseline levels and as LS mean (95%CI) for change in level. ＊＊p＜0.01 (vs. IR 0.2 mg/day). IR, immediate-release; XR, extended-release; LS mean, least square mean; CI, confidence interval.
Fig.2 shows the LS means of the percentage change from baseline to four weeks of treatment for LDL-C, ApoB, non-HDL-C, lathosterol, β-sitosterol, and campesterol in subgroups by baseline LDL-C. All four of the subgroups by baseline LDL-C showed statistically significant lowering of LDL-C in each treatment group, and this lowering was significantly greater for XR 0.8 mg/day than for IR 0.2 mg/day in each subgroup. In the XR 0.8 mg/day group, the LS means of the percentage change in LDL-C were greater for the groups with higher LDL-C at baseline (−14.7%, −18.6%, −25.4%, and −30.4% for LDL-C subgroups of baseline ≥ 100 mg/dL, ≥ 120 mg/dL, ≥ 140 mg/dL, and ≥ 160 mg/dL, respectively). ApoB and non-HDL-C showed statistically significant lowering in all subgroups for each treatment group, and as seen for LDL-C, the LS means of percentage change in each parameter were greater for the patients with higher baseline LDL-C. Statistically significant lowering was also observed in lathosterol, β-sitosterol, and campesterol in all subgroups and in each treatment group. Lathosterol showed a trend toward greater lowering in patients with higher LDL-C at baseline, and the XR 0.8 mg/day group was associated with significantly greater lathosterol lowering than the IR 0.2 mg/day group, with the exception of the LDL-C ≥ 160 mg/dL subgroup. β-Sitosterol and campesterol showed a trend toward greater lowering in patients with higher LDL-C at baseline, but there was no statistical difference in the percentage change among the treatment groups.
Data are presented as (number of patients) [mean (mg/dL)] or <mean (mg/L)> for baseline levels and as LS mean (95%CI) for change in level. ＊p＜0.05, ＊＊p＜0.01, ＊＊＊p＜0.001 (vs. IR 0.2 mg/day). †n = 35, ‡n = 28; LDL-C, ApoB and non-HDL-C.
LDL-C, low-density lipoprotein cholesterol; Apo, apolipoproteins; IR, immediate-release; XR, extended-release; LS mean, least square mean; CI, confidence interval.
Plots and 95% prediction ellipses were determined for the percentage change in LDL-C versus the percentage change in lathosterol, β-sitosterol, and campesterol from baseline to four weeks of treatment in the IR 0.2 mg/day, XR 0.4 mg/day, and XR 0.8 mg/day groups, shown in Fig.3. In all groups, the percentage change in cholesterol absorption and synthesis markers was positively correlated with the percentage change in LDL-C (r = 0.352–0.631).
IR, immediate-release; XR, extended-release; LDL-C, low-density lipoprotein cholesterol
To our knowledge, this study is the first to evaluate LDL-C changes related to the cholesterol synthesis marker lathosterol and the cholesterol absorption markers β-sitosterol and campesterol for pemafibrate in this patient population. Just as with cholesterol, the sterols used as markers of cholesterol intestinal synthesis and hepatic absorption are contained in lipoproteins. If pemafibrate lowers lipoproteins and total cholesterol, we would expect to see these markers of cholesterol synthesis and absorption lowered as well. Notably, in our study pemafibrate lowered these synthesis and absorption markers even when corrected for total cholesterol level. In addition, there is a trade-off between intestinal absorption and hepatic synthesis: conventional statins suppress the hepatic synthesis of cholesterol, which increases intestinal absorption, while ezetimibe suppresses intestinal absorption of cholesterol, which increases hepatic synthesis. Notably, pemafibrate is the first drug proven in humans to inhibit both intestinal absorption and hepatic synthesis. Our study showed a positive correlation between the percentage changes in LDL-C and in cholesterol synthesis and absorption markers, and patients with higher baseline LDL-C showed more pronounced lowering of LDL-C and also tended to experience greater reduction of markers for both cholesterol synthesis and absorption. A similar trend was observed in a study involving patients with MASLD12), indicating that the effects of pemafibrate have become more evident in patients with higher baseline LDL-C. Additionally, in the present study, higher doses of pemafibrate exhibited a more pronounced LDL-C-lowering effect compared to lower doses.
In the present study, pemafibrate provided significant lowering of lathosterol that tended to be dose-dependent. Previously, in patients with metabolic syndrome, fenofibrate was found to lower lathosterol17). Notably, basic research has shown that liver cholesterol content was lowered in a mouse model of steatohepatitis following the administration of pemafibrate18) and that 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase activity was lowered in the liver of rats following the administration of fenofibrate or bezafibrate19). These findings suggest that PPARα agonists may inhibit cholesterol synthesis in the liver. Since fenofibrate did not suppress HMG-CoA reductase activity in vitro19), this suppression by PPARα agonists may be an indirect effect. However, following pemafibrate administration, no reduction of HMG-CoA reductase mRNA was observed in the liver of a mouse model of bile duct ligation20), and no studies have reported HMG-CoA reductase activity.
The cholesterol absorption markers β-sitosterol and campesterol were significantly reduced following pemafibrate administration in this study, as is also seen with conventional fibrates. Previous papers have reported that, in dyslipidemic patients, gemfibrozil significantly lowered β-sitosterol21) and fenofibrate significantly lowered β-sitosterol and campesterol22). Meanwhile, in basic research using an LDL receptor-deficient mouse model and a high-fat diet-loaded mouse model, pemafibrate suppressed expression of NPC1L1 mRNA in the small intestine, and this suppression appeared to be PPARα-dependent23).
We also considered several other possible mechanisms of LDL-C lowering by pemafibrate. First, pemafibrate administration increased cholesterol excretion in rat feces20), potentially inducing ATP binding cassette subfamily G member 8 (ABCG8) gene expression in the jejunum and accelerating intestinal excretion of cholesterol. The same study indicated that pemafibrate upregulated ABCG5 gene expression in the liver, and potentially also in the bile duct, to promote biliary excretion of cholesterol, and also upregulated cholesterol 7α-hydroxylase gene expression in the liver to promote catabolism of cholesterol into bile acids20). These findings suggest that pemafibrate increases both biliary and transintestinal cholesterol excretion.
Second, fenofibrate administration increases the amount of LDL receptors and the binding affinity of lipoproteins to hepatocytes in the mouse liver by inducing the maturation of the sterol regulatory element-binding protein 2 (SREBP2)24). Pemafibrate administration tended to decrease hepatic cholesterol in a mouse model of fatty liver23), which would be expected to increase the number of LDL receptors. However, the effect of pemafibrate on the expression of hepatic SREBP2 remains unclear, emphasizing the need for further investigation of the role of pemafibrate in cholesterol metabolism and receptor regulation.
Third, tracer studies have shown that small dense LDL (sdLDL) particles remain in the blood approximately 1.5 times longer than normal LDL particles25) and that these smaller LDL particles have a lower affinity for LDL receptors in the liver than normal LDL particles26). Pemafibrate may act to increase LDL particle size12, 27, 28), resulting in the uptake of LDL particles into the liver and thereby decreasing the number of those particles in the blood. This mechanism is supported by the reduction in ApoB12, 29).
Fourth, pemafibrate promotes the metabolism of TG-rich lipoproteins, which may lead to a transient increase in LDL-C27) and would be expected to increase LDL-C and ApoB; however, this increase is not seen in all clinical studies. Our current study also showed decreases in LDL-C and ApoB, and those findings may reflect the LDL-C-lowering effect of pemafibrate, achieved by inhibiting the intestinal absorption and hepatic synthesis of cholesterol. In clinical practice, the balance between these two opposing effects of pemafibrate on LDL-C levels may result in a phenotypical increase or decrease in LDL-C, which could explain some of the inconsistencies.
Finally, there are two factors to affect this balance, especially with regard to masking the LDL-C-lowering effect of pemafibrate. The first factor is high baseline TG. Previous studies of pemafibrate have shown that baseline TG correlates inversely with the percentage change in LDL-C27, 28). It is likely that TG-rich lipoproteins are elevated in patients who have high TG, and that pemafibrate administration in such patients would cause more TG-rich lipoproteins to be catabolized to LDL, thus increasing LDL-C in hypertriglyceridemic patients. In the J-BENEFIT, bezafibrate lowered LDL-C in a population with high baseline LDL-C30), similar to the present study findings, and the population with high baseline TG tended to have low baseline LDL-C and vice versa. This association between low baseline LDL-C and relatively high TG suggests more effective PPARα-induced promotion of TG-rich lipoprotein catabolism. The second factor is that LDL-C can be strictly controlled with statins, in which case cholesterol synthesis is already strongly inhibited, and pemafibrate may be unable to provide further inhibition. In such patients, LDL receptor expression is already maximized, with no room for further enhancement of LDL uptake by pemafibrate. Statin treatment also increases the expression of PCSK9 31), potentially negating any pemafibrate-induced increase in LDL receptors. It should be noted that the expression of PCSK9 under pemafibrate administration is not clearly understood, and previous findings for fenofibrate are inconsistent, showing both increases32) and decreases33) in PCSK9. Pemafibrate would thus be expected to decrease LDL-C most effectively in a population with high baseline LDL-C and low or normal TG. The possible effects of pemafibrate on lipoprotein metabolism are summarized in Fig.4.
A. Patients with low LDL-C and high TG who are taking statins (a) and receiving additional treatment with pemafibrate (b). B. Patients with high LDL-C and high TG who are not taking statins (a) but are receiving treatment with pemafibrate (b). Thickness and size of arrows indicates intensity of action. Dashed line and light color of arrows indicates masking of medication effect. Circle size indicates size of particle.
LDL-C, low-density lipoprotein cholesterol; TG, triglycerides; Cho, cholesterol; VLDL, very low-density lipoprotein; LPL, lipoprotein lipase; HTGL, hepatic triglyceride lipase; PCSK9, proprotein convertase subtilisin kexin 9.
To our knowledge, there have been no reported effects on ASCVD events among populations who decreased TG, LDL-C, and ApoB under pemafibrate administration, but a previous report has noted the importance of lowering LDL-C and ApoB when managing hypertriglyceridemia as a residual risk factor for ASCVD34). Based on findings from our present study, some populations may experience lowering of LDL-C and ApoB as well as TG, and information on baseline TG and LDL-C may help to identify those populations. Even in cases where LDL-C levels increase with pemafibrate treatment, it redeuces not only TG-rich lipoproteins but also sdLDL particles27, 28), which are particularly atherogenic35). The impact on ASCVD following sdLDL-C lowering by pemafibrate is a subject for future study.
Although statins are commonly used to manage LDL-C, it is estimated that 7–29% of patients with hyper-LDL cholesterolemia are completely statin-intolerant or cannot take the recommended dose due to adverse events including statin-associated muscle symptoms36). ASCVD risk is reportedly increased in patients who have difficulty taking statins37), suggesting that statin-intolerant patients with limited treatment options include a subset who have insufficient risk reduction for ASCVD. Comprehensive management, including of residual risk, is especially desirable for such patients. A placebo-controlled phase 3 study is currently underway in Japan to evaluate the efficacy and safety of pemafibrate XR 0.2 mg/day or 0.4 mg/day for 12 weeks in statin-intolerant hypercholesterolemic patients (jRCT2051230055).
First, it included only 60 patients who participated in a phase 2 clinical pharmacology study, and the proportion of those patients who were taking statins was very small (10%). The presence of statins may have influenced results when examining the effects on cholesterol synthesis and absorption markers. In particular, we could reach no conclusions on whether pemafibrate reduces cholesterol synthesis and absorption markers under high-intensity statin use. Second, this study was a post-hoc analysis, making it an exploratory rather than a hypothesis-testing study. However, synthetic and absorption markers, which were collected prospectively, may provide some insight into the mechanism of action of pemafibrate. Third, the treatment period was short (four weeks), and the design was a crossover comparative study without a washout period. Previous pemafibrate clinical trials28) have confirmed that the TG-lowering effect of pemafibrate reaches steady state two to four weeks after administration and returns to pre-treatment levels four weeks after completion of treatment. We thus assumed that the TG-lowering effect would return to baseline in four weeks without the need for a washout period. However, we have not fully confirmed whether this holds true for indices other than TG. Fourth, this study was limited to Japanese patients and did not include any patients with severe obesity, which is a common comorbidity of hypertriglyceridemia in western populations. In the present study, no differences in the effects of pemafibrate on LDL-C or cholesterol synthesis and absorption markers were observed between the subpopulations with a BMI of 25 kg/m2 or more and those with a BMI of less than 25 kg/m2 (data not shown), but the influence of body size and race was not fully examined in this study. Fifth, because of the small sample size of 60 patients, subgroup analysis based on patient characteristics could not be performed. With regard to TG-lowering effect, previous clinical trials on IR formulations showed no differences based on patient characteristics38), and similar results were observed in the Phase 3 trial of the XR formulation39).
In patients with hypertriglyceridemia, results suggested that pemafibrate lowered LDL-C by inhibiting cholesterol synthesis in the liver and cholesterol absorption from the intestinal tract. This lowering effect was greater in populations with higher baseline LDL-C.
Authors acknowledge the investigators and patients who participated in this study. Medical writing support for the first draft was provided by EDIT, Inc. (Tokyo, Japan) and was funded by Kowa Company, Ltd.
This study has not been the recipient of grants from any funding agency in the public, commercial, or not-for-profit sectors.
This study was funded by Kowa Company, Ltd. The study sponsor had a role in the study design; data collection, analysis, and interpretation; and writing of the report.
Yamashita S has received personal fees from Kowa Company, Ltd., Novartis Pharma K.K., Otsuka Pharmaceutical Co., Ltd., and Skylight Biotech, Inc. Araki E has received personal fees from Sumitomo Pharma Co., Ltd., Novo Nordisk Pharma Ltd., MSD K.K., and Kowa Company, Ltd.; and grants from Manpei Suzuki Diabetes Foundation. Arai H has received personal fees from Kowa Company, Ltd., and Astellas Pharma Inc. Yokote K has received personal fees and grants from Kowa Company, Ltd., Sumitomo Pharma Co., Ltd., Mitsubishi Tanabe Pharma Corporation, and Nippon Boehringer Ingelheim Co., Ltd.; personal fees from Astellas Pharma Inc., Sanofi K.K., Daiichi Sankyo Company, Limited, Taisho Pharmaceutical Co., Ltd., Novartis Pharma K.K., Novo Nordisk Pharma Ltd., and Bayer Yakuhin, Ltd.; and grants from Takeda Pharmaceutical Company Limited. Tanigawa R, Saito A, Suganami H, and Minamikawa S are employees of Kowa Company, Ltd.
Yamashita S collected data and wrote the original draft of the manuscript. Araki E, Arai H, Yokote K and Ishibashi S collected data, edited the manuscript, and contributed to the interpretation of data and writing. Tanigawa R and Saito A conducted the clinical trial. Suganami H analyzed data. Minamikawa S edited the manuscript. All authors reviewed the final manuscript.
Data sharing including the protocol are not applicable in this study.
As a post-hoc analysis of a phase 2 clinical trial (NCT0407953), this study did not require independent registration.
