Exposure-Response Analysis of the Macitentan Effects on NT-proBNP and WHO functional class in adult patients with Pulmonary Arterial Hypertension
Roberta Bartolucci(1), Dénes Csonka(2), Juan José Pérez Ruixo(3), Navin Goyal(4), Alberto Russu(1)
(1) Janssen-Cilag SpA, Cologno Monzese, Italy (2) Janssen Research and Development, Allschwil, Switzerland (3) Janssen Research and Development, Beerse, Belgium (4) Janssen Research and Development, Raritan, NJ
Introduction: Pulmonary arterial hypertension (PAH) is a rare, chronic disease characterized by increased pulmonary vasoconstriction and pulmonary artery smooth muscle cell proliferation, which leads to ventricular hypertrophy and a progressive worsening of pulmonary and cardiac complications. N-terminal pro-brain natriuretic peptide (NT-proBNP, secreted in response to ventricular hypertrophy) and World Health Organization functional class (WHO FC, a 4-level classification of symptoms from I, least severe, to IV, most severe) are two relevant measures routinely evaluated to monitor disease severity [1]. Macitentan is an oral, non-peptide, potent endothelin receptor antagonist approved in adults for the treatment of PAH, at the dose of 10 mg/day [2]. This work assessed the exposure-response (E-R) relationships of NT-proBNP and WHO FC with respect to macitentan after 6 months of treatment.
Methods: Data for NT-proBNP and WHO FC data were obtained from SERAPHIN (NCT00660179), a Phase 3 study in which adult PAH patients were randomly assigned to receive placebo, macitentan 3 mg or macitentan 10 mg QD. The individual macitentan exposure (area under the concentration-time curve, AUC0-24h,ss) estimated via population PK analysis [3] was used to guide the E-R analysis.
For NT-proBNP, the final dataset contained 399 subjects (164, 114, and 121 in placebo, 3 mg and 10 mg treatment arms, respectively). First, the correlation between relevant variables (sex, race, etiology, age, weight, height, creatinine clearance, BMI, baseline NT-proBNP, dose and exposure) and NT-proBNP change from baseline, defined as the log-transformed ratio of month 6 and baseline levels, was assessed as an exploratory analysis. Then, the E-R relationship was assessed with linear regressions models.
For WHO FC, the final dataset contained 468 subjects (136 in placebo, 139 in 3 mg and 193 in 10 mg treatment arms). Given the ordered categorical nature of WHO FC, a first-order Markov model was developed to describe the effect of AUC0-24h,ss on the probability to change class from baseline to month 6. The model was then used to predict the probability of improvement (i.e., any class lower than the baseline class).
Results: For NT-proBNP change from baseline, a significant difference was observed between the treatment arms, with a 62.8% of subjects worsening in the placebo group and a majority of subjects improving in both the macitentan groups (58.1% and 63.5% for 3 mg and 10 mg, respectively). Among the tested covariates, baseline NT-proBNP showed a correlation of 30% (larger NT-proBNP decrease from baseline for larger baseline value) and was therefore tested using the base regression model. Other covariates showed relatively small correlations (<10%) and were therefore not tested. The final regression model was characterized by a positive NT-proBNP change from baseline in the placebo group (intercept of NT-proBNP log-ratio of 8.6e-02, p-value<0.01), and by a negative linear relationship with macitentan AUC0-24h,ss (slope of -2.8e-05 (ng.h/mL)-1, p-value<0.01). Therefore, compared to placebo, the drug effect at 7081 ng.h/mL (median AUC0-24h,ss for the macitentan 10 mg group) resulted in a 24% reduction in NT-proBNP change from baseline. A significant effect of baseline (log-transformed and centered on the median) was obtained on the intercept (-1.7e-01, p-value<0.01), demonstrating that patients who were more severe at baseline had a larger improvement.
For WHO FC, the Markov model was defined by 3 logit functions for each baseline class, with linear placebo and exposure effects. Since insufficient data were available for subjects with baseline class I or IV (only 1 for class I and 3 for class IV), the effect of macitentan exposure was estimated only on the logits related to baseline classes II and III. The final model was characterized by piecewise linear functions of AUC0-24h,ss on the probability to change class when the baseline class is II or III. The model showed that patients with a higher exposure had a higher probability of improvement. In particular, for an exposure of 7180 ng.h/mL (median AUC0-24h,ss for the macitentan 10 mg group) the predicted probability of improvement was 17% higher than in the placebo group.
Conclusions: E-R analyses on NT-proBNP and WHO FC demonstrated a statistically significant drug effect linked to macitentan AUC0-24h,ss after 6 months of treatment, when accounting for the disease progression observed in the placebo group.
References:
[1] Galiè, N et al. “2015 ESC/ERS Guidelines for the diagnosis and treatment of pulmonary hypertension: The Joint Task Force for the Diagnosis and Treatment of Pulmonary Hypertension of the European Society of Cardiology (ESC) and the European Respiratory Society (ERS): Endorsed by: Association for European Paediatric and Congenital Cardiology (AEPC), International Society for Heart and Lung Transplantation (ISHLT).” European heart journal vol. 37,1 (2016): 67-119. doi:10.1093/eurheartj/ehv317
[2] US Food and Drug Administration. Opsumit® (macitentan) prescribing information. 2013. https://www.accessdata.fda.gov/drugsatfda_docs/label/2019/204410s017lbl.pdf. Accessed 17 Feb 2023.
[3] Bartolucci, R et al. “A Population Pharmacokinetic Model of Macitentan and Its Active Metabolite Aprocitentan in Healthy Volunteers and Patients with Pulmonary Arterial Hypertension.” Clinical Pharmacokinetics vol. 60,12 (2021): 1605-1619. doi:10.1007/s40262-021-01049-3