Population Pharmacokinetic and Exposure–Response Analysis of Finerenone: Insights Based on Phase IIb Data and Simulations to Support Dose Selection for Pivotal Trials in Type 2 Diabetes with Chronic Kidney Disease

Mineralocorticoid receptor antagonists (MRAs) reduce albuminuria but their use is limited by hyperkalemia and kidney dysfunction risks, especially in chronic kidney disease (CKD) patients receiving renin–angiotensin system blockade. Finerenone is a novel, selective, non-steroidal MRA with higher receptor selectivity/affinity versus steroidal MRAs and showed favorable safety with albuminuria reduction in earlier studies (ARTS, ARTS-DN, ARTS-DN Japan). The research aim was to characterize finerenone population pharmacokinetics (PK) and pharmacodynamics (PD) in adults with type 2 diabetes (T2D) and CKD using sparse phase IIb ARTS-DN and ARTS-DN Japan data, quantify covariate effects, define exposure–response relationships for efficacy (UACR) and safety (serum potassium, eGFR), and simulate dose–exposure–response to support dose selection for pivotal trials.

Prior evidence shows steroidal MRAs reduce albuminuria/proteinuria but increase hyperkalemia risk, particularly in CKD on ACEI/ARB therapy. Finerenone demonstrated smaller potassium increases than spironolactone and reduced albuminuria in CHF with CKD (ARTS) and in T2D with CKD (ARTS-DN; ARTS-DN Japan). Phase I studies in healthy volunteers showed dose-linear PK (1–80 mg), rapid absorption/elimination, minimal renal elimination, increased exposure with moderate/severe renal impairment, and higher AUC with moderate hepatic impairment without Cmax change. Drug–drug interaction and biotransformation studies were conducted. Albuminuria change (UACR) is a validated prognostic surrogate associated with ESRD and mortality risk, and both MDRD and CKD-EPI equations are used to estimate eGFR; CKD-EPI is more specific in CKD. These findings motivate detailed popPK/PD modeling in the target T2D-CKD population.

Data sources: Phase IIb ARTS-DN (global, non-Japanese) and ARTS-DN Japan (Japanese) randomized studies in T2D with CKD on background RAS blockade; once-daily finerenone doses 1.25–20 mg and placebo. Primary efficacy endpoint: change in UACR at day 90 vs baseline; safety markers: serum potassium and eGFR. Datasets included 893 subjects for PK/PD and 787 for PK; multiple dose groups and rich PD sampling across visits. Modeling approach: PopPK model based on prior Phase IIa structure, refined with ARTS-DN and combined ARTS-DN+JP data. Structural PK model: two-compartment disposition with first-order elimination, absorption via transit compartments and a fixed lag time. Covariates: eGFR-MDRD and body weight assessed on PK parameters (CL/F, V/F, F); ethnicity effect evaluated. PK linearity over dose and time assessed. PD models: - UACR: turnover model with inhibitory effect on production (kin), concentration–effect described by a maximum effect (Imax) relationship; separate baseline UACR estimated for very high albuminuria (screening UACR >300 mg/g). - Serum potassium: turnover model with a log-linear concentration–effect relationship; outcomes included proportions exceeding 5.5 and 6.0 mmol/L at any visit. - eGFR (CKD-EPI): power concentration–effect model for relative change from baseline; proportions with declines ≥25%, ≥30%, ≥40%, and ≥57% evaluated. Estimation and evaluation: Models refit first to ARTS-DN, then to ARTS-DN+JP; statistical comparisons used minimum value of objective function and likelihood ratio tests; 90% confidence intervals reported; precision via relative standard errors (<50%). Exposure metric: AUC at steady state (AUCss) derived from individual EBEs of PK parameters. eGFR equation usage: CKD-EPI used for PD analyses; MDRD used as PK covariate for consistency with historical PK models. Simulations: Hypothetical phase III-like scenario with inclusion criteria (e.g., serum K+ <4.8 mmol/L at run-in/screening, UACR ≥30 mg/g, eGFR 25–90 mL/min/1.73 m²), OD dosing, full compliance, 180-day treatment, assessments on days 1, 30, 60, 90, and 180. Doses simulated: placebo, 5, 7.5, 10, 15, 20, and extrapolated 30 mg. Event derivations for serum K+ thresholds independent of study discontinuation.

Pharmacokinetics: - PK described adequately by a two-compartment model with transit absorption and first-order elimination; dose- and time-linear over 1.25–20 mg OD. - Covariates: Lower eGFR-MDRD increased exposure (CL/F decreased and F increased as eGFR decreased); higher body weight increased apparent volume. Example: eGFR-MDRD 30 vs 90 mL/min/1.73 m² led to +32.9% AUCss and +17.5% Cmax; BW 50 vs 100 kg led to +43.1% Cmax. - Steady-state PK reached after 2 days; half-life ~2–3 h. - Japanese vs global: dose-normalized AUCss +0.4% and Cmax,ss +7.4% in Japanese; no structural ethnicity effect required; minor differences attributable to slightly lower BW and eGFR distribution. Pharmacodynamics/exposure–response: - UACR: Decreased with increasing exposure; maximum-effect (Imax) model indicated saturation at high exposures; effect proportional to baseline UACR (ratio UACRday90/UACRbaseline independent of baseline). Time to 99% steady-state effect ~138 days. - Serum potassium: Log-linear concentration–effect; attenuation of effect gains at higher exposures; effect proportional to baseline K+. Time to 99% steady-state effect ~20 days. Predicted proportion with any K+ >5.5 mmol/L increased with dose and baseline K+. - eGFR (CKD-EPI): Power concentration–effect describing modest sustained decreases during treatment; time to 99% steady-state effect ~85 days. Threshold exceedance probabilities increased with dose but remained low. Ethnicity: No apparent differences in PK/PD relationships for UACR, serum potassium, or eGFR between Japanese and non-Japanese subjects. Simulations to 180 days: - UACR targets at day 180 for 20 mg OD: ≤0.8: 66.4%; ≤0.7: 58.6%; ≤0.6: 49.0%; ≤0.5: 39.5%. Increasing to 30 mg added ~5–7 percentage points across targets. - Serum potassium: Proportion with any K+ >5.5 mmol/L increased from 1.3% (90% CI 1.1–1.7) at 20 mg to 1.7% (1.5–2.1) at 30 mg; none expected >6.0 mmol/L. - eGFR declines at day 180: ≥25% from 5.0% (4.3–6.1) at 20 mg to 5.5% (4.5–6.8) at 30 mg; ≥30% from 1.8% (1.6–2.2) to 2.0% (1.7–2.6); ≥40% and ≥57% <0.1% at 20 and 30 mg. Overall: Efficacy and safety effects approach saturation or attenuate around 20 mg OD; doses 10 and 20 mg OD predicted safe and efficacious for albuminuria reduction.

The modeling results address the key question of how finerenone exposure relates to efficacy (albuminuria reduction) and safety (serum potassium increase, eGFR decline) in T2D with CKD. PK was predictable and linear, enabling reliable individual exposure estimates. Exposure–response analyses revealed timescale separation: rapid PK steady state (2 days) versus slower PD equilibration, particularly for UACR (≈138 days) and eGFR (≈85 days), aligning with underlying biology of MR antagonism (electrolyte effects within weeks; anti-inflammatory/anti-fibrotic effects requiring months). The UACR Imax model indicates diminishing returns at higher exposures, suggesting limited added benefit beyond 20 mg OD, while serum potassium (log-linear) and eGFR (power) relationships similarly show attenuating increases in risk, which remained low in simulations. Absence of ethnic effects supports generalizability across global and Japanese populations after accounting for body weight and renal function. Collectively, the findings support selecting 10 and 20 mg OD doses for pivotal trials to balance efficacy and safety, with limited incremental gains predicted at 30 mg.

PopPK and PK/PD models adequately described finerenone PK and its effects on UACR, serum potassium, and eGFR in ARTS-DN and ARTS-DN Japan. Pharmacological effects on UACR reduction and on safety parameters (serum potassium increase, eGFR decrease) approach saturation or attenuate by 20 mg once daily. Simulations indicate only small additional benefits and small increases in safety events when increasing dose to 30 mg. Doses of 10 and 20 mg once daily appear safe and efficacious for reducing albuminuria, providing a rationale for dose selection in phase III. The established models enable further investigation of dose–exposure–response relationships and linkage to clinical outcomes in pivotal trials.

- Structural PK simplification (equal central and peripheral volumes) may limit identification of covariate effects on distribution and may inadequately predict terminal elimination beyond 24 hours, though AUCss and Cmax are likely unaffected. - Simulations did not capture disease progression; horizon limited to 180 days. - Hyperkalemia event estimations based purely on modeled serum potassium values irrespective of study discontinuation or reassessment, differing from trial-defined hyperkalemia endpoints. - Ethnic comparisons were not powered specifically to detect population differences; absence of detected differences may reflect limited sample sizes. - Dependence on baseline covariate distributions (e.g., baseline K+, eGFR) affects threshold-crossing probabilities; inclusion/exclusion criteria in simulations (e.g., no reassessment if K+ ≥4.8 mmol/L) influence predicted rates.