concentration concentration concentration

certara.com
  • No tags were found...

(OM), Amlodipine (AML), and Hydrochlorothiazide (HCTZ) - Certara

T2416A Fixed-Dose Combination of Olmesartan Medoxomil (OM), Amlodipine (AML), and Hydrochlorothiazide (HCTZ): Use of Modeling and Simulationto Support an Understanding of the Dose Response of Intermediate Dose Combinations not Included in the Pivotal Phase 3 StudyTimothy J Carrothers, ScD 1 ; Michelle Green, PhD 1 ; Helen Moore, PhD 1 ; Shashank Rohatagi, PhD 2 ; SaeHeum Song, PhD 2 ; James Lee, PhD 2 ; Antonia Wang, PhD 2 ; Reinilde Heyrman, MD 2 ; Daniel E. Salazar, PhD 21Pharsight Corporation, Mountain View, CA; 2 Daiichi Sankyo Pharma Development, Edison, NJObjectivesPrimary Study for AnalysisDevelop population pharmacokinetic (PopPK) models of olmesartan medoxomil (OM), amlodipine (AML), andhydrochlorothiazide (HCTZ) for application to the hypertensive patient population using relevant Phase 1 andPhase 3 datasets from the clinical development programs for CS-866 (OM+HCTZ), CS-8663 (OM+AML), andCS-8635 (OM+AML+HCTZ).Characterize and quantify the effects of covariates on the oral clearances of the compounds, OM, AML, andHCTZ, including demographics (age, weight, gender, creatinine clearance, alanine aminotransferase and aspartateaminotransferase) and disease status (hypertension with and without diabetes).Develop exposure-response models that characterize the effect of the drug on seated trough diastolic bloodpressure (SeDBP) and seated trough systolic blood pressure (SeSBP) based on the data in the three relevantPhase 3 studies: CS866-318 (OM+HCTZ), CS8663-A-U301 (OM+AML), and CS8635-A-U301 (OM+AML+HCTZ).Characterize and quantify the effects of covariates (age, race, weight, sex, creatinine clearance, and baselineSeDBP or SeSBP) on the exposure-response models.Simulate the systolic and diastolic blood pressure lowering effects of clinically unevaluated market imageformulations of CS8635: 20/5/12.5, 40/5/12.5, 40/10/12.5, and 40/5/25 mg (OM/AML/HCTZ) based on thepopulation PK/PD models and the demographic and baseline characteristics of the CS8635-A-U301 population.Study CS8635-A-U301: A randomized, double-blind, placebo-controlled study evaluating the efficacy and safetyof administration of high-dose triple-combination olmesartan, hydrochlorothiazide and amlodipine (40/10/25 mg)compared to the three respective high-dose dual combinations in patients with mild to severe hypertension.Fixed-dose combinations planned for marketing not part of the study: 20/5/12.5, 40/5/12.5, 40/10/12.5 and40/5/25 mg (OM/AML/HCTZ).Data and MethodsTable 1. Demographic Summary of PK SubjectsStudy Phase N M:FAge[y]Mean (SD)Weight[kg]Mean (SD)CLCR*[ml/min]Mean (SD)Model-predicted exposures from the population PK model were used in the exposure-response analysis.(Covariate-adjusted median-predicted exposures were used for subjects without PK sampling.)The analysis used seated cuff blood pressure measurements taken at the per-protocol baseline and end of theprimary efficacy analysis period (Week 8 in CS866-318 and CS8663-A-U301; Week 12 in CS8635-A-U301).Simulations of the CS-8635 blood pressure lowering effect were conducted based on the PK and exposureresponsemodels:– Primary Objective: To compare the blood pressure lowering effects of clinically un-evaluated CS-8635 dosestrengths (20/5/12.5, 40/5/12.5, 40/10/12.5, 40/5/25 mg OM/AML/HCTZ) to those of the clinically evaluateddose strength (40/10/25 mg OM/AML/HCTZ).– Patient Population used in Simulation: The CS8635-A-U301 population.Race/Ethnicity †W:B:H:A:OAll Phase I studies I 492 349:143 30.9 (7.9) 77 (13) 127 (26) 143:199:131:9:10 0CS8663-A-U301(OM+AML)CS8635-A-U301(OM+AML)– Dose combinations simulated: A full factorial design, i.e., all possible dose combinations, of OM (0/20/40 mg),AML (0/5/10 mg) and HCTZ (0/12.5/25 mg) for the placebo, mono, dual and triple combinations.Diabetes(%)III 556 283:273 54.6 (11) 94.6 (22) 99.6 (34) 351:117:69:8:11 13.7III 956 517:439 55.8 (10) 95.5 (22) 117 (43) 547:223:167:14:5 16.3Both Phase III studies III 1512 800:712 55.4 (11) 95.2 (22) 111 (41) 898:340:236:22:16 15.3All studies – 2004 1149:855 49.4 (15) 90.7 (21) 115 (38) 1041:539:367:31:26 11.6*Creatinine clearance [Cockcroft-Gault]† White/Black/Hispanic/Asian/OtherTable 2. Demographic Summary of Exposure-ResponseDataset SubjectsStudyNBaselineSBPBaselineDBPM:FAge[y]Mean (SD)Weight[kg]Mean (SD)Race/EthnicityW:B:H:A:OAll data 4873 165 (16) 102 (6.7) 2625:2248 54.8 (11) 94.9 (22) 2869:1216:654:93:41 14.1CS866-318 495 154 (13) 104 (3.1) 278:217 53.5 (11) 88.1 (18) 369:60:48:12:6 8.9CS8663-A-U301 1920 164 (17) 102 (5.6) 1043:877 54.6 (11) 95.2 (22) 1169:452:241:36:22 13.4→ non-PK subjects 1365 165 (17) 102 (5.7) 761:604 54.6 (11) 95.4 (22) 819:335:172:28:11 13.3→ PK subjects 555 163 (17) 102 (5.3) 282:273 54.6 (11) 94.6 (22) 350:117:69:8:11 13.7CS8635-A-U301 2458 169 (14) 101 (7.8) 1304:1154 55.2 (11) 96.1 (23) 1331:704:365:45:13 15.6→ non-PK subjects 1542 169 (14) 101 (7.9) 804:738 54.7 (11) 96.4 (24) 801:492:209:31:9 15.0→ PK subjects 916 168 (14) 101 (7.5) 500:416 56.1 (10) 95.5 (22) 530:212:156:14:4 16.6Diabetes(%)Population Pharmacokinetic Modeling ResultsThe concentrations of each drug were successfully described by a mammillary two-compartment model with first orderelimination and first order absorption with time lag. A covariate analysis identified demographic relationships with drugconcentrations, as follows:Figure 1. Post-predictive Check for CS8635-A-U301 for theOlmesartan Medoxomil Pharmacokinetic ModelOM concentration in plasma [ng/ml]2000150010005000Note: Concentrations dose-normalized to 40mg0 5 10 15Time from dose [hr]20 25 30Raw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th and 2.5th quantiles of the data. For sparsely sampled studies, dots areplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing the mean model prediction, and the dotted lines representing the 95%prediction interval of the simulation.Figure 2. Post-predictive Check for Phase 1 Studies (AUCss)for Olmesartan MedoxomilObserved OM AUC[ng/ml•hr]OM: Patients with lower creatinine clearances had lower clearance of the drug.AML: Older patients had lower clearance of the drug.HCTZ: Older patients, female patients, and patients with lower creatinine clearances had lower clearance of thedrug.Post-predictive checks (Figures 1-6) for each drug demonstrated that the models successfully described the patientexposure profiles.OM:AML:HCTZ:1400010000600020002000 4000 6000 8000Model-Predicted OM AUC [ng/ml•hr]10000 12000 14000Figure 3. Post-predictive Check for CS8635-A-U301 for theAmlodipine Pharmacokinetic ModelAML concentration in plasma [ng/ml]6040200Note: Concentrations dose-normalized to 10mg0 5 10 15Time from dose [hr]20 25 30Raw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th and 2.5th quantiles of the data. For sparsely sampled studies, dots areplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing the mean model prediction, and the dotted lines representing the 95%prediction interval of the simulation.Figure 4. Post-predictive Check for Phase 1 Studies (AUCss)for AmlodipineObserved AML AUC[ng/ml•hr]Figure 5. Post-predictive Check for CS8635-A-U301 for theHydrochlorothiazide Pharmacokinetic ModelHCTZ concentrationin plasma [ng/ml]8006004002006005004003002001000200 400Note: Concentrations dose-normalized to 25mg0 5 10 15 20 25 30Time from dose [hr]Raw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th and 2.5th quantiles of the data. For sparsely sampled studies, dots areplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing the mean model prediction, and the dotted lines representing the 95%prediction interval of the simulation.600Model-Predicted AML AUC [ng/ml•hr]800Figure 6. Post-predictive Check for Phase 1 Studies (AUCss)for HydrochlorothiazideObserved HCTZ AUC[ng/ml•hr]200015001000500500 10001500Model-Predicted HCTZ AUC [ng/ml•hr]2000Results from Exposure-Response Modeling of Blood PressureThe final exposure-response model for SeDBP and SeSBP related the drug effects of olmesartan medoxomil,amlodipine, and hydrochlorothiazide to their systemic exposures, AUC-OM, AUC-AML, and AUC-HCTZ, respectively.The drug effects for olmesartan medoxomil and amlodipine were described by an Emax model, whereas the drug effectfor hydrochlorothiazide was described by a linear model. The drug effect of combination therapy was greater than anyof the drug effects in monotherapy, but slightly less than their additive sum. This finding was modeled via a series ofinteraction terms. Key covariate findings in the exposure-response modeling included:For both SeDBP and SeSBP, the placebo effect varied by study and was stronger for subjects with higherbaseline BP.For OM, subjects of black race showed a weaker response than non-black subjects.For AML, subjects of lower weight showed a stronger response.Exposure-Response Analysis for Seated Diastolic Blood Pressurewhere,DBP i,j is the i th measurement within the j th subject at steady-state,η j is additive inter-subject variability in response, ande i is additive residual intra-subject variability.Exposure-Response Analysis for Seated Systolic Blood Pressurewhere,SBP i,j is the i th measurement within the j th subject at steady-state,η j is additive inter-subject variability in response, ande i is additive residual intra-subject variability.Results from Simulation ModelsTable 3. Predicted and Observed BP Lowering Effects(CS8663-A-U301, CS866-318)Table 4. Predicted and Observed BP Lowering Effects(CS8635-A-U301) Mean (SD)∆ (SeDBP)∆ (SeSBP)Treatment (mg) Observed Predicted Observed PredictedOM/AML 40/10 -17.8 (9.9) -17.8 (9.0) -31.1 (16) -31.1 (15)OM/HCTZ 40/25 -16.5 (11) -16.5 (10) -31.2 (19) -31.1 (18)AML/HCTZ 10/25 -14.8 (9.3) -14.8 (8.3) -29.0 (16) -29.0 (14)OM/AML/HCTZ 40/10/25 -21.5 (11) -21.5 (9.8) -38.1 (18) -38.1 (17)Table 5. Predicted BP Lowering Effect of CS-8635 –(Amlodipine to Benicar HCT ® ) Mean (SD)Benicar HCT ®(mg/mg)AML (mg)0 m 5 10∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP0/0 -4.0 (9.2) -4.7 (15.2) -10.7 (9.8) -17.6 (16.5) -14.1 (10.2) -22.9 (17.4)20/12.5 -12.7 (9.4) -22.4 (16.0) -16.8 (9.7) -30.4 (16.8) -18.9 (9.9) -33.7 (17.1)40/12.5 -14.1 (9.5) -24.4 (16.1) -17.9 (9.7) -32.1 (16.8) -19.8 (9.9) -35.2 (17.1)40/25 -16.6 (9.6) -29.9 (16.7) -19.8 (9.8) -35.9 (17.0) -21.4 (9.9) -38.4 (17.2)Table 6. Predicted BP Lowering Effect of CS-8635 –(HCTZ to AZOR ® ) Mean (SD)AZOR ®(mg/mg)observedmean ∆DBP[mm Hg]CS8663-A-U301observedmean ∆SBP[mm Hg]Final modelPrediction (“IPRED”) with individual-levelvariability parameterpredictedmean ∆DBP[mm Hg]HCTZ (mg)predictedmean ∆SBP[mm Hg]0 12.5 25Final modelPrediction (“PRED”) without individuallevelvariability parameterpredictedmean ∆DBP[mm Hg]predictedmean ∆SBP[mm Hg]OM Dose[mg]AML Dose[mg]HCTZ Dose[mg]0 0 0 -3.08 -2.81 -3.09 -2.88 -3.24 -3.890 10 0 -13.0 -20.2 -13.0 -20.1 -13.3 -20.00 5 0 -9.68 -15.3 -9.68 -15.3 -9.78 -15.110 0 0 -8.07 -11.9 -8.01 -11.9 -8.38 -12.810 10 0 -16.3 -25.8 -16.3 -25.8 -16.0 -25.610 5 0 -14.2 -24.7 -14.2 -24.6 -13.9 -23.520 0 0 -9.11 -13.5 -9.16 -13.7 -10.0 -15.720 10 0 -17.1 -29.1 -17.1 -29.1 -16.9 -28.620 5 0 -14.0 -23.4 -14.0 -23.5 -14.5 -24.340 0 0 -10.6 -17.1 -10.6 -17.1 -10.9 -17.140 10 0 -19.2 -30.5 -19.2 -30.6 -18.4 -31.040 5 0 -15.8 -26.2 -15.7 -26.2 -15.3 -25.3observedmean ∆DBP[mm Hg]CS866-318observedmean ∆SBP[mm Hg]Final modelPrediction (“IPRED”) with individual-levelvariability parameterpredictedmean ∆DBP[mm Hg]predictedmean ∆SBP[mm Hg]Final modelPrediction (“PRED”) without individuallevelvariability parameterpredictedmean ∆DBP[mm Hg]predictedmean ∆SBP[mm Hg]OM Dose[mg]AML Dose[mg]HCTZ Dose[mg]0 0 0 -7.51 -2.92 -7.43 -3.02 -6.59 -4.100 0 12.5 -9.28 -8.61 -9.26 -8.76 -9.03 -10.40 0 25 -12.9 -17.6 -12.8 -17.7 -12.2 -17.810 0 0 -12.7 -10.3 -12.6 -10.5 -11.6 -12.610 0 12.5 -15.3 -20.3 -15.2 -20.2 -14.5 -18.910 0 25 -18.4 -22.9 -18.3 -22.9 -17.4 -23.420 0 0 -12.4 -14.9 -12.5 -14.9 -13.4 -14.920 0 12.5 -15.8 -21.3 -15.8 -21.2 -15.9 -19.320 0 25 -18.9 -25.7 -18.9 -25.6 -19.0 -25.240 0 0 -14.4 -16.3 -14.5 -16.4 -15.0 -17.140 0 12.5 -18.4 -20.4 -18.3 -20.4 -17.7 -21.240 0 25 -21.9 -27.8 -21.7 -27.7 -20.2 -26.1∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP0/0 -4.0 (9.2) -4.7 (15.2) -6.4 (9.3) -12.5 (15.7) -8.9 (9.4) -20.3 (16.9)5/20 -15.0 (9.7) -26.5 (16.6) -16.8 (9.7) -30.4 (16.8) -18.6 (9.7) -34.4 (17.0)5/40 -16.0 (9.7) -28.2 (16.7) -17.9 (9.7) -32.1 (16.8) -19.8 (9.8) -35.9 (17.0)10/40 -18.2 (9.9) -32.0 (17.1) -19.8 (9.9) -35.2 (17.1) -21.4 (9.9) -38.4 (17.2)Figure 7. Boxplots of DBP Simulation Results for “Add toBenicar HCT ® ” ScenarioChange from Baseline [mm Hg]10 -40-10-20-30-40OMAMLHCTZPlacebo000AML 5-10.7050AML 10OM/HCTZ 20/12.5OM/HCTZ 40/12.5DBP-14.1-12.7-14.2010020012.540012.5OM/HCTZ 40/25-16.6 -16.840025OM/AML/HCTZ 20/5/12.520512.5Note: Number is mean. Dashes are 10 th , 50 th , and 90 th percentiles of population. Solid line is 25-75 th range.OM/AML/HCTZ 40/5/12.5-17.940512.5OM/AML/HCTZ 40/10/12.5-19.8 -19.8401012.5OM/AML/HCTZ 40/5/2540525OM/AML/HCTZ 40/10/25-21.4401025Figure 8. Boxplots of DBP Simulation Results for “Add toAZOR ® ” ScenarioChange from Baseline [mm Hg]Conclusions10 -40-10-20-30-40OMAMLHCTZPlacebo000HCTZ 12.5-6.50012.5HCTZ 25-8.90025AML/OM 5/20-152050AML/OM 5/40DBPOlmesartan medoxomil PK was adequately characterized by a two-compartment model with first-order absorptionand time lag; creatinine clearance was a significant predictor of the apparent oral clearance of olmesartanmedoxomil.Amlodipine PK was adequately characterized by a two-compartment model with first-order absorption and a timelag; age was a significant predictor of the apparent oral clearance of amlodipine.Hydrochlorothiazide PK was adequately characterized by a two-compartment model with first-order absorptionand a time lag; sex, age, and creatinine clearance were significant predictors of the apparent oral clearance ofhydrochlorothiazide.The blood pressure lowering effects of olmesartan medoxomil and amlodipine exposure on seated trough DBPand SBP were described by an E max model, whereas the blood pressure lowering effect of hydrochlorothiazideexposure was described by a linear model.The blood pressure lowering effects of olmesartan medoxomil, amlodipine, and hydrochlorothiazide inmonotherapy, dual combination therapy, and triple combination therapy were well characterized by a modelcomposed of the sum of the individual effects and interaction among the components.The model-predicted and simulated blood pressure lowering effects of the various tested combinations of thethree compounds were in good agreement with the observed data from the three Phase 3 studies upon which themodel was built.In the exposure-response model, black race was a covariate, decreasing the maximal possible effect on bloodpressure of olmesartan medoxomil without influencing PK parameters.In the exposure-response model, baseline seated trough DBP and SBP were covariates, with more blood pressurelowering effect associated with higher baseline blood pressure.The exposure-response model predicted the blood pressure lowering effects of triple combination therapypermutations to be superior to their respective mono and dual treatments of OM, AML, and HCTZ.The order of the blood pressure lowering effects among the different CS-8635 dose strengths was:20/5/12.5


T2416A Fixed-Dose Combination of Olmesartan Medoxomil (OM), Amlodipine (AML), and Hydrochlorothiazide (HCTZ): Use of Modeling and Simulationto Support an Understanding of the Dose Response of Intermediate Dose Combinations not Included in the Pivotal Phase 3 StudyTimothy J Carrothers, ScD 1 ; Michelle Green, PhD 1 ; Helen Moore, PhD 1 ; Shashank Rohatagi, PhD 2 ; SaeHeum Song, PhD 2 ; James Lee, PhD 2 ; Antonia Wang, PhD 2 ; Reinilde Heyrman, MD 2 ; Daniel E. Salazar, PhD 21 Pharsight Corporation, Mountain View, CA;2 Daiichi Sankyo Pharma Development, Edison, NJObjectivesDevelop population pharmacokinetic (PopPK) models of olmesartan medoxomil (OM), amlodipine (AML), andhydrochlorothiazide (HCTZ) for application to the hypertensive patient population using relevant Phase 1 andPhase 3 datasets from the clinical development programs for CS-866 (OM+HCTZ), CS-8663 (OM+AML), andCS-8635 (OM+AML+HCTZ).Characterize and quantify the effects of covariates on the oral clearances of the compounds, OM, AML, andHCTZ, including demographics (age, weight, gender, creatinine clearance, alanine aminotransferase and aspartateaminotransferase) and disease status (hypertension with and without diabetes).Develop exposure-response models that characterize the effect of the drug on seated trough diastolic bloodpressure (SeDBP) and seated trough systolic blood pressure (SeSBP) based on the data in the three relevantPhase 3 studies: CS866-318 (OM+HCTZ), CS8663-A-U301 (OM+AML), and CS8635-A-U301 (OM+AML+HCTZ).Characterize and quantify the effects of covariates (age, race, weight, sex, creatinine clearance, and baselineSeDBP or SeSBP) on the exposure-response models.Simulate the systolic and diastolic blood pressure lowering effects of clinically unevaluated market imageformulations of CS8635: 20/5/12.5, 40/5/12.5, 40/10/12.5, and 40/5/25 mg (OM/AML/HCTZ) based on thepopulation PK/PD models and the demographic and baseline characteristics of the CS8635-A-U301 population.Primary Study for AnalysisStudy CS8635-A-U301: A randomized, double-blind, placebo-controlled study evaluating the efficacy and safetyof administration of high-dose triple-combination olmesartan, hydrochlorothiazide and amlodipine (40/10/25 mg)compared to the three respective high-dose dual combinations in patients with mild to severe hypertension.Fixed-dose combinations planned for marketing not part of the study: 20/5/12.5, 40/5/12.5, 40/10/12.5 and40/5/25 mg (OM/AML/HCTZ).Data and MethodsTable 1. Demographic Summary of PK SubjectsStudy Phase N M:FAgeMean (SD)Weight[kg]Mean (SD)CLCR*[ml/min]Mean (SD)Race/Ethnicity †W:B:H:A:OAll Phase I studies I 492 349:143 30.9 (7.9) 77 (13) 127 (26) 143:199:131:9:10 0CS8663-A-U301(OM+AML)CS8635-A-U301(OM+AML)Diabetes(%)III 556 283:273 54.6 (11) 94.6 (22) 99.6 (34) 351:117:69:8:11 13.7III 956 517:439 55.8 (10) 95.5 (22) 117 (43) 547:223:167:14:5 16.3Figure 1. Post-predictive Check for CS8635-A-U301 for theOlmesartan Medoxomil Pharmacokinetic Model200015001000Note: Concentrations dose-normalized to 40mgFigure 3. Post-predictive Check for CS8635-A-U301 for theAmlodipine Pharmacokinetic Model604020Note: Concentrations dose-normalized to 10mg0 5 10 15Time from dose [hr]Table 4. Predicted and Observed BP Lowering Effects(CS8635-A-U301) Mean (SD)∆ (SeDBP) ∆ (SeSBP)Treatment (mg) Observed Predicted Observed PredictedOM/AML 40/10 -17.8 (9.9) -17.8 (9.0) -31.1 (16) -31.1 (15)OM/HCTZ 40/25 -16.5 (11) -16.5 (10) -31.2 (19) -31.1 (18)AML/HCTZ 10/25 -14.8 (9.3) -14.8 (8.3) -29.0 (16) -29.0 (14)OM/AML/HCTZ 40/10/25 -21.5 (11) -21.5 (9.8) -38.1 (18) -38.1 (17)Table 5. Predicted BP Lowering Effect of CS-8635 –(Amlodipine to Benicar HCT ® ) Mean (SD)Both Phase III studies III 1512 800:712 55.4 (11) 95.2 (22) 111 (41) 898:340:236:22:16 15.3500SBP20-4.8AML (mg)-12.5200All studies – 2004 1149:855 49.4 (15) 90.7 (21) 115 (38) 1041:539:367:31:26 11.6-20.30 m 5 10Benicar HCT ®-26.5 -28.3(mg/mg)∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP-32.1 -30.5 -32.1*Creatinine clearance [Cockcroft-Gault]-35.2 -35.9† White/Black/Hispanic/Asian/Other 0 5 10 15 20 25 300/0 -4.0 (9.2) -4.7 (15.2) -10.7 (9.8) -17.6 (16.5) -14.1 (10.2) -22.9 (17.4)-38.420/12.5 -12.7 (9.4) -22.4 (16.0) -16.8 (9.7) -30.4 (16.8) -18.9 (9.9) -33.7 (17.1)-20Time from dose [hr]40/12.5 -14.1 (9.5) -24.4 (16.1) -17.9 (9.7) -32.1 (16.8) -19.8 (9.9) -35.2 (17.1)Table 2. Demographic Summary of Exposure-Response200 40060080040/25 -16.6 (9.6) -29.9 (16.7) -19.8 (9.8) -35.9 (17.0) -21.4 (9.9) -38.4 (17.2)Raw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th and 2.5th quantiles of the data. For sparsely sampled studies, dots areModel-Predicted AML AUC [ng/ml•hr]-40Dataset Subjectsplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing the mean model prediction, and the dotted lines representing the 95%Table 6. Predicted BP Lowering Effect of CS-8635 –Age Weightprediction interval of the simulation.Baseline Baseline[kg]Race/EthnicityDiabetes(HCTZ to AZOR ® -60OMStudy N SBP DBP M:F Mean (SD) Mean (SD)W:B:H:A:O(%)) Mean (SD)AMLHCTZ12.512.5 12.5 12.5All data 4873 165 (16) 102 (6.7) 2625:2248 54.8 (11) 94.9 (22) 2869:1216:654:93:41 14.1Note: Number is mean. Dashes are 10HCTZ (mg)Figure 5. Post-predictive Check for CS8635-A-U301 for theth , 50 th , and 90 th percentiles of population. Solid line is 25-75 th range.CS866-318 495 154 (13) 104 (3.1) 278:217 53.5 (11) 88.1 (18) 369:60:48:12:6 8.9Exposure-Response Analysis for Seated Systolic Blood Pressure0 12.5 25Figure 2. Post-predictive Check for Phase 1 Studies (AUCss)Hydrochlorothiazide Pharmacokinetic ModelAZORConclusions®CS8663-A-U301 1920 164 (17) 102 (5.6) 1043:877 54.6 (11) 95.2 (22) 1169:452:241:36:22 13.4(mg/mg)∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP0/0 -4.0 (9.2) -4.7 (15.2) -6.4 (9.3) -12.5 (15.7) -8.9 (9.4) -20.3 (16.9)for Olmesartan MedoxomilNote: Concentrations dose-normalized to 25mg5/20 -15.0 (9.7) -26.5 (16.6) -16.8 (9.7) -30.4 (16.8) -18.6 (9.7) -34.4 (17.0)→ non-PK subjects 1365 165 (17) 102 (5.7) 761:604 54.6 (11) 95.4 (22) 819:335:172:28:11 13.3where,Olmesartan medoxomil PK was adequately characterized by a two-compartment model with first-order absorption5/40 -16.0 (9.7) -28.2 (16.7) -17.9 (9.7) -32.1 (16.8) -19.8 (9.8) -35.9 (17.0)600SBP i,j is the i th measurement within the j th subject at steady-state,and time lag; creatinine clearance was a significant predictor of the apparent oral clearance of olmesartan10/40 -18.2 (9.9) -32.0 (17.1) -19.8 (9.9) -35.2 (17.1) -21.4 (9.9) -38.4 (17.2)→ PK subjects 555 163 (17) 102 (5.3) 282:273 54.6 (11) 94.6 (22) 350:117:69:8:11 13.714000medoxomil.η j is additive inter-subject variability in response, andAmlodipine PK was adequately characterized by a two-compartment model with first-order absorption and a timeε500i is additive residual intra-subject variability.lag; age was a significant predictor of the apparent oral clearance of amlodipine.CS8635-A-U301 2458 169 (14) 101 (7.8) 1304:1154 55.2 (11) 96.1 (23) 1331:704:365:45:13 15.6Figure 7. Boxplots of DBP Simulation Results for “Add toHydrochlorothiazide PK was adequately characterized by a two-compartment model with first-order absorptionand a time lag; sex, age, and creatinine clearance were significant predictors of the apparent oral clearance of→ non-PK subjects 1542 169 (14) 101 (7.9) 804:738 54.7 (11) 96.4 (24) 801:492:209:31:9 15.0400Benicar HCT ® ” Scenariohydrochlorothiazide.10000The blood pressure lowering effects of olmesartan medoxomil and amlodipine exposure on seated trough DBP→ PK subjects 916 168 (14) 101 (7.5) 500:416 56.1 (10) 95.5 (22) 530:212:156:14:4 16.6DBPand SBP were described by an E 30010 max model, whereas the blood pressure lowering effect of hydrochlorothiazide-4exposure was described by a linear model.-10.7The blood pressure lowering effects of olmesartan medoxomil, amlodipine, and hydrochlorothiazide in200-14.1-12.7-14.2monotherapy, dual combination therapy, and triple combination therapy were well characterized by a model-16.6 -16.8composed of the sum of the individual effects and interaction among the components.Model-predicted exposures from the population PK model were used in the exposure-response analysis.-17.9(Covariate-adjusted median-predicted exposures were used for subjects without PK sampling.)6000-19.8 -19.8-21.4The model-predicted and simulated blood pressure lowering effects of the various tested combinations of the100-10three compounds were in good agreement with the observed data from the three Phase 3 studies upon which theThe analysis used seated cuff blood pressure measurements taken at the per-protocol baseline and end of themodel was built.primary efficacy analysis period (Week 8 in CS866-318 and CS8663-A-U301; Week 12 in CS8635-A-U301).In the exposure-response model, black race was a covariate, decreasing the maximal possible effect on bloodSimulations of the CS-8635 blood pressure lowering effect were conducted based on the PK and exposureresponsemodels:-20pressure of olmesartan medoxomil without influencing PK parameters.In the exposure-response model, baseline seated trough DBP and SBP were covariates, with more blood pressure– Primary Objective: To compare the blood pressure lowering effects of clinically un-evaluated CS-8635 dose20000 5 10 15 20 25 30-30lowering effect associated with higher baseline blood pressure.strengths (20/5/12.5, 40/5/12.5, 40/10/12.5, 40/5/25 mg OM/AML/HCTZ) to those of the clinically evaluateddose strength (40/10/25 mg OM/AML/HCTZ).Time from dose [hr]The exposure-response model predicted the blood pressure lowering effects of triple combination therapypermutations to be superior to their respective mono and dual treatments of OM, AML, and HCTZ.– Patient Population used in Simulation: The CS8635-A-U301 population.Raw data from studies plotted as grey dots.-402000 4000 6000 8000 10000 12000 14000OMThe order of the blood pressure lowering effects among the different CS-8635 dose strengths was:– Dose combinations simulated: A full factorial design, i.e., all possible dose combinations, of OM (0/20/40 mg),Blue dots represent the mean of the data, orange dots represent the 97.5th and 2.5th quantiles of the data. For sparsely sampled studies, dots areAML20/5/12.5


Timothy J Carrothers, ScD ; Mich1Pharsight Corporation, Mountain ViewObjectivesPopulation PharmacokinetDevelop population pharmacokinetic (PopPK) models of olmesartan medoxomil (OM), amlodipine (AML), andhydrochlorothiazide (HCTZ) for application to the hypertensive patient population using relevant Phase 1 andPhase 3 datasets from the clinical development programs for CS-866 (OM+HCTZ), CS-8663 (OM+AML), andCS-8635 (OM+AML+HCTZ).Characterize and quantify the effects of covariates on the oral clearances of the compounds, OM, AML, andHCTZ, including demographics (age, weight, gender, creatinine clearance, alanine aminotransferase and aspartateaminotransferase) and disease status (hypertension with and without diabetes).Develop exposure-response models that characterize the effect of the drug on seated trough diastolic bloodpressure (SeDBP) and seated trough systolic blood pressure (SeSBP) based on the data in the three relevantPhase 3 studies: CS866-318 (OM+HCTZ), CS8663-A-U301 (OM+AML), and CS8635-A-U301 (OM+AML+HCTZ).Characterize and quantify the effects of covariates (age, race, weight, sex, creatinine clearance, and baselineSeDBP or SeSBP) on the exposure-response models.Simulate the systolic and diastolic blood pressure lowering effects of clinically unevaluated market imageformulations of CS8635: 20/5/12.5, 40/5/12.5, 40/10/12.5, and 40/5/25 mg (OM/AML/HCTZ) based on thepopulation PK/PD models and the demographic and baseline characteristics of the CS8635-A-U301 population.The concentrations of each drug were successfully described byelimination and first order absorption with time lag. A covariate aconcentrations, as follows:OM: Patients with lower creatinine clearances had lower cleAML: Older patients had lower clearance of the drug.HCTZ: Older patients, female patients, and patients with lodrug.Post-predictive checks (Figures 1-6) for each drug demonstratedexposure profiles.OM:AML:Primary Study for AnalysisStudy CS8635-A-U301: A randomized, double-blind, placebo-controlled study evaluating the efficacy and safetyof administration of high-dose triple-combination olmesartan, hydrochlorothiazide and amlodipine (40/10/25 mg)compared to the three respective high-dose dual combinations in patients with mild to severe hypertension.Fixed-dose combinations planned for marketing not part of the study: 20/5/12.5, 40/5/12.5, 40/10/12.5 and40/5/25 mg (OM/AML/HCTZ).HCTZ:Data and MethodsTable 1. Demographic Summary of PK SubjectsStudy Phase N M:FAge[y]Mean (SD)Weight[kg]Mean (SD)CLCR*[ml/min]Mean (SD)Race/Ethnicity †W:B:H:A:OAll Phase I studies I 492 349:143 30.9 (7.9) 77 (13) 127 (26) 143:199:131:9:10 0CS8663-A-U301(OM+AML)CS8635-A-U301(OM+AML)Diabetes(%)III 556 283:273 54.6 (11) 94.6 (22) 99.6 (34) 351:117:69:8:11 13.7III 956 517:439 55.8 (10) 95.5 (22) 117 (43) 547:223:167:14:5 16.3Both Phase III studies III 1512 800:712 55.4 (11) 95.2 (22) 111 (41) 898:340:236:22:16 15.3All studies – 2004 1149:855 49.4 (15) 90.7 (21) 115 (38) 1041:539:367:31:26 11.6*Creatinine clearance [Cockcroft-Gault]† White/Black/Hispanic/Asian/OtherTable 2. Demographic Summary of Exposure-ResponseDataset SubjectsStudyNBaselineSBPBaselineDBPM:FAge[y]Mean (SD)Weight[kg]Mean (SD)Race/EthnicityW:B:H:A:ODiabetes(%)Figure 1. Post-predictive Check fOlmesartan MedoxomilOM concentration in plasma [ng/ml]20001500100050000 5Note: Concentrations10Time froRaw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th and 2plotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing theprediction interval of the simulation.All data 4873 165 (16) 102 (6.7) 2625:2248 54.8 (11) 94.9 (22) 2869:1216:654:93:41 14.1CS866-318 495 154 (13) 104 (3.1) 278:217 53.5 (11) 88.1 (18) 369:60:48:12:6 8.9CS8663-A-U301 1920 164 (17) 102 (5.6) 1043:877 54.6 (11) 95.2 (22) 1169:452:241:36:22 13.4→ non-PK subjects 1365 165 (17) 102 (5.7) 761:604 54.6 (11) 95.4 (22) 819:335:172:28:11 13.3→ PK subjects 555 163 (17) 102 (5.3) 282:273 54.6 (11) 94.6 (22) 350:117:69:8:11 13.7Figure 2. Post-predictive Check ffor Olmesartan Medoxo14000CS8635-A-U301 2458 169 (14) 101 (7.8) 1304:1154 55.2 (11) 96.1 (23) 1331:704:365:45:13 15.6→ non-PK subjects 1542 169 (14) 101 (7.9) 804:738 54.7 (11) 96.4 (24) 801:492:209:31:9 15.0→ PK subjects 916 168 (14) 101 (7.5) 500:416 56.1 (10) 95.5 (22) 530:212:156:14:4 16.6Model-predicted exposures from the population PK model were used in the exposure-response analysis.(Covariate-adjusted median-predicted exposures were used for subjects without PK sampling.)The analysis used seated cuff blood pressure measurements taken at the per-protocol baseline and end of theprimary efficacy analysis period (Week 8 in CS866-318 and CS8663-A-U301; Week 12 in CS8635-A-U301).Simulations of the CS-8635 blood pressure lowering effect were conducted based on the PK and exposureresponsemodels:– Primary Objective: To compare the blood pressure lowering effects of clinically un-evaluated CS-8635 dosestrengths (20/5/12.5, 40/5/12.5, 40/10/12.5, 40/5/25 mg OM/AML/HCTZ) to those of the clinically evaluateddose strength (40/10/25 mg OM/AML/HCTZ).– Patient Population used in Simulation: The CS8635-A-U301 population.– Dose combinations simulated: A full factorial design, i.e., all possible dose combinations, of OM (0/20/40 mg),AML (0/5/10 mg) and HCTZ (0/12.5/25 mg) for the placebo, mono, dual and triple combinations.Observed OM AUC[ng/ml•hr]10000600020002000 40006000Model-Predicte


Timothy J Carrothers, ScD ; Michelle Green, PhD ; Helen Moore, PhD ; Shasha1Pharsight Corporation, Mountain View, CA; 2 Daiichi Sankyo Pharma Development, Edisonesartan medoxomil (OM), amlodipine (AML), ande patient population using relevant Phase 1 andr CS-866 (OM+HCTZ), CS-8663 (OM+AML), andal clearances of the compounds, OM, AML, andne clearance, alanine aminotransferase and aspartatewithout diabetes).ect of the drug on seated trough diastolic blood(SeSBP) based on the data in the three relevantOM+AML), and CS8635-A-U301 (OM+AML+HCTZ)., weight, sex, creatinine clearance, and baselineffects of clinically unevaluated market imaged 40/5/25 mg (OM/AML/HCTZ) based on thee characteristics of the CS8635-A-U301 population.o-controlled study evaluating the efficacy and safety, hydrochlorothiazide and amlodipine (40/10/25 mg)ns in patients with mild to severe hypertension.he study: 20/5/12.5, 40/5/12.5, 40/10/12.5 andPopulation Pharmacokinetic Modeling ResultsThe concentrations of each drug were successfully described by a mammillary two-compartment model with first orderelimination and first order absorption with time lag. A covariate analysis identified demographic relationships with drugconcentrations, as follows:OM: Patients with lower creatinine clearances had lower clearance of the drug.AML: Older patients had lower clearance of the drug.HCTZ: Older patients, female patients, and patients with lower creatinine clearances had lower clearance of thedrug.Post-predictive checks (Figures 1-6) for each drug demonstrated that the models successfully described the patientexposure profiles.OM:AML:HCTZ:Figure 3. Post-predictive Check fAmlodipine PharmacokAML concentration in plasma [ng/ml]6040200Note: Concentrations d0 5 10 1Time fromRaw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th andplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing theprediction interval of the simulation.f PK Subjectseightkg]n (SD)CLCR*[ml/min]Mean (SD)ere used in the exposure-response analysis.for subjects without PK sampling.)taken at the per-protocol baseline and end of theS8663-A-U301; Week 12 in CS8635-A-U301).ere conducted based on the PK and exposuregeffects of clinically un-evaluated CS-8635 doseM/AML/HCTZ) to those of the clinically evaluated01 population.Race/Ethnicity †W:B:H:A:ODiabetes(%)(13) 127 (26) 143:199:131:9:10 06 (22) 99.6 (34) 351:117:69:8:11 13.75 (22) 117 (43) 547:223:167:14:5 16.32 (22) 111 (41) 898:340:236:22:16 15.37 (21) 115 (38) 1041:539:367:31:26 11.6f Exposure-Responsege[y]n (SD)Weight[kg]Mean (SD)Race/EthnicityW:B:H:A:ODiabetes(%)8 (11) 94.9 (22) 2869:1216:654:93:41 14.15 (11) 88.1 (18) 369:60:48:12:6 8.96 (11) 95.2 (22) 1169:452:241:36:22 13.46 (11) 95.4 (22) 819:335:172:28:11 13.36 (11) 94.6 (22) 350:117:69:8:11 13.72 (11) 96.1 (23) 1331:704:365:45:13 15.67 (11) 96.4 (24) 801:492:209:31:9 15.01 (10) 95.5 (22) 530:212:156:14:4 16.6all possible dose combinations, of OM (0/20/40 mg),o, mono, dual and triple combinations.Figure 1. Post-predictive Check for CS8635-A-U301 for theOlmesartan Medoxomil Pharmacokinetic ModelOM concentration in plasma [ng/ml]2000150010005000Note: Concentrations dose-normalized to 40mg0 5 10 15Time from dose [hr]20 25 30Raw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th and 2.5th quantiles of the data. For sparsely sampled studies, dots areplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing the mean model prediction, and the dotted lines representing the 95%prediction interval of the simulation.Figure 2. Post-predictive Check for Phase 1 Studies (AUCss)for Olmesartan MedoxomilObserved OM AUC[ng/ml•hr]1400010000600020002000 4000 6000 8000Model-Predicted OM AUC [ng/ml•hr]10000 12000 14000Figure 4. Post-predictive Check ffor AmlodipineObserved AML AUC[ng/ml•hr]Figure 5. Post-predictive Check fHydrochlorothiazide PhHCTZ concentrationin plasma [ng/ml]80060040020060050040030020010000 5200 40Model-PredicteNote: Concentration10Time frRaw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th andplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing theprediction interval of the simulation.


elle Green, PhD ; Helen Moore, PhD ; Shashank Rohatagi, PhD ; SaeHeum Song, PhD ; Jam, CA; 2 Daiichi Sankyo Pharma Development, Edison, NJc Modeling Resultsmammillary two-compartment model with first orderalysis identified demographic relationships with drugrance of the drug.er creatinine clearances had lower clearance of thehat the models successfully described the patientFigure 3. Post-predictive Check for CS8635-A-U301 for theAmlodipine Pharmacokinetic ModelAML concentration in plasma [ng/ml]6040200Note: Concentrations dose-normalized to 10mg0 5 10 15Time from dose [hr]20 25 30Figure 6. Post-predictive Check ffor HydrochlorothiazideObserved HCTZ AUC[ng/ml•hr]200015001000500500 1000Model-PredictedRaw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th and 2.5th quantiles of the data. For sparsely sampled studies, dots areplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing the mean model prediction, and the dotted lines representing the 95%prediction interval of the simulation.Results from Exposure-Responser CS8635-A-U301 for theharmacokinetic Modelose-normalized to 40mgFigure 4. Post-predictive Check for Phase 1 Studies (AUCss)for Amlodipine800The final exposure-response model for SeDBP and SeSBP relateamlodipine, and hydrochlorothiazide to their systemic exposuresThe drug effects for olmesartan medoxomil and amlodipine werefor hydrochlorothiazide was described by a linear model. The druof the drug effects in monotherapy, but slightly less than their adinteraction terms. Key covariate findings in the exposure-responFor both SeDBP and SeSBP, the placebo effect varied by sbaseline BP.For OM, subjects of black race showed a weaker responseFor AML, subjects of lower weight showed a stronger respoExposure-Response Analysis for Seated Diastolic Blood PreObserved AML AUC[ng/ml•hr]600400200where,DBP i,j is the i th measurement within the j th subject at steadyη j is additive inter-subject variability in response, andε i is additive residual intra-subject variability.5dose [hr]20 25 30th quantiles of the data. For sparsely sampled studies, dots areean model prediction, and the dotted lines representing the 95%200 400600Model-Predicted AML AUC [ng/ml•hr]800r Phase 1 Studies (AUCss)ilFigure 5. Post-predictive Check for CS8635-A-U301 for theHydrochlorothiazide Pharmacokinetic ModelHCTZ concentrationin plasma [ng/ml]600500400300200100Note: Concentrations dose-normalized to 25mgExposure-Response Analysis for Seated Systolic Blood Preswhere,SBP i,j is the i th measurement within the j th subject at steadyηj is additive inter-subject variability in response, andε i is additive residual intra-subject variability.8000OM AUC [ng/ml•hr]10000 12000 1400000 5 10 15 20 25 30Time from dose [hr]Raw data from studies plotted as grey dots.Blue dots represent the mean of the data, orange dots represent the 97.5th and 2.5th quantiles of the data. For sparsely sampled studies, dots areplotted at the median observed time for the nominal sample time.Results from model simulation drawn as red lines, with solid line representing the mean model prediction, and the dotted lines representing the 95%prediction interval of the simulation.


k Rohatagi, PhD ; SaeHeum Song, PhD ; James Lee, PhD ; Antonia Wang, PhD ; Reinilde HNJr CS8635-A-U301 for theetic Modelse-normalized to 10mgFigure 6. Post-predictive Check for Phase 1 Studies (AUCss)for Hydrochlorothiazide2000Results from Simulation MTable 3. Predicted and Observed(CS8663-A-U301, CS866-Observed HCTZ AUC[ng/ml•hr]15001000500observedmean ∆DBP[mm Hg]CS8663-A-U301observedmean ∆SBP[mm Hg]OM Dose AML Dose HCTZ Dose[mg] [mg] [mg]0 0 0 -3.08 -2.810 10 0 -13.0 -20.20 5 0 -9.68 -15.310 0 0 -8.07 -11.910 10 0 -16.3 -25.810 5 0 -14.2 -24.720 0 0 -9.11 -13.520 10 0 -17.1 -29.120 5 0 -14.0 -23.440 0 0 -10.6 -17.140 10 0 -19.2 -30.540 5 0 -15.8 -26.2Predictprme[mdose [hr]20 25 30th quantiles of the data. For sparsely sampled studies, dots areean model prediction, and the dotted lines representing the 95%r Phase 1 Studies (AUCss)500 10001500Model-Predicted HCTZ AUC [ng/ml•hr]2000Results from Exposure-Response Modeling of Blood PressureThe final exposure-response model for SeDBP and SeSBP related the drug effects of olmesartan medoxomil,amlodipine, and hydrochlorothiazide to their systemic exposures, AUC-OM, AUC-AML, and AUC-HCTZ, respectively.The drug effects for olmesartan medoxomil and amlodipine were described by an Emax model, whereas the drug effectfor hydrochlorothiazide was described by a linear model. The drug effect of combination therapy was greater than anyof the drug effects in monotherapy, but slightly less than their additive sum. This finding was modeled via a series ofinteraction terms. Key covariate findings in the exposure-response modeling included:For both SeDBP and SeSBP, the placebo effect varied by study and was stronger for subjects with higherbaseline BP.For OM, subjects of black race showed a weaker response than non-black subjects.For AML, subjects of lower weight showed a stronger response.Exposure-Response Analysis for Seated Diastolic Blood Pressurewhere,DBP i,j is the i th measurement within the j th subject at steady-state,η j is additive inter-subject variability in response, andε i is additive residual intra-subject variability.observedmean ∆DBP[mm Hg]CS866-318observedmean ∆SBP[mm Hg]OM Dose AML Dose HCTZ Dose[mg] [mg] [mg]0 0 0 -7.51 -2.920 0 12.5 -9.28 -8.610 0 25 -12.9 -17.610 0 0 -12.7 -10.310 0 12.5 -15.3 -20.310 0 25 -18.4 -22.920 0 0 -12.4 -14.920 0 12.5 -15.8 -21.320 0 25 -18.9 -25.740 0 0 -14.4 -16.340 0 12.5 -18.4 -20.440 0 25 -21.9 -27.8Table 4. Predicted and Observed(CS8635-A-U301) Mean (Treatment (mg) ObservedOM/AML 40/10 -17.8 (9.9)OM/HCTZ 40/25 -16.5 (11)AML/HCTZ 10/25 -14.8 (9.3)OM/AML/HCTZ 40/10/25 -21.5 (11)Predict∆ (SeDBP)prme[mTable 5. Predicted BP Lowering E(Amlodipine to Benicar H600AML AUC [ng/ml•hr]r CS8635-A-U301 for thermacokinetic Modeldose-normalized to 25mg80015 20 25 30m dose [hr]th quantiles of the data. For sparsely sampled studies, dots areean model prediction, and the dotted lines representing the 95%Exposure-Response Analysis for Seated Systolic Blood Pressurewhere,SBP i,j is the i th measurement within the j th subject at steady-state,η j is additive inter-subject variability in response, andε i is additive residual intra-subject variability.0 mBenicar HCT ®(mg/mg)∆SeDBP ∆SeSBP ∆SeDBP0/0 -4.0 (9.2) -4.7 (15.2) -10.7 (9.8)20/12.5 -12.7 (9.4) -22.4 (16.0) -16.8 (9.7)40/12.5 -14.1 (9.5) -24.4 (16.1) -17.9 (9.7)40/25 -16.6 (9.6) -29.9 (16.7) -19.8 (9.8)Table 6. Predicted BP Lowering E(HCTZ to AZOR ® ) Mean (S0AZOR ®(mg/mg)∆SeDBP ∆SeSBP ∆SeDBP0/0 -4.0 (9.2) -4.7 (15.2) -6.4 (9.3)5/20 -15.0 (9.7) -26.5 (16.6) -16.8 (9.7)5/40 -16.0 (9.7) -28.2 (16.7) -17.9 (9.7)10/40 -18.2 (9.9) -32.0 (17.1) -19.8 (9.9)Figure 7. Boxplots of DBP SimulaBenicar HCT ® ” ScenarioChange from Baseline [mm Hg]10 -40-10-20-30-40OMAMLHCTZPlacebo000AML 5-10.7050AML 10-14.1-12.7-14.20100OM/HCTZ 20/12.520012.5OM/HCTZ 40/12.5DBP40012.5Note: Number is mean. Dashes are 10 th , 50 th , and 90 th percentiles of population. SOM/HCTZ 40/25


s Lee, PhD ; Antonia Wang, PhD ; Reinilde Heyrman, MD ; Daniel E. Salazar, PhDr Phase 1 Studies (AUCss)1500CTZ AUC [ng/ml•hr]2000the drug effects of olmesartan medoxomil,UC-OM, AUC-AML, and AUC-HCTZ, respectively.escribed by an Emax model, whereas the drug effecteffect of combination therapy was greater than anytive sum. This finding was modeled via a series ofmodeling included:dy and was stronger for subjects with higheran non-black subjects.se.ureate,reate,odeling of Blood PressureResults from Simulation ModelsTable 3. Predicted and Observed BP Lowering Effects(CS8663-A-U301, CS866-318)observedmean ∆DBP[mm Hg]CS8663-A-U301observedmean ∆SBP[mm Hg]Final modelPrediction (“IPRED”) with individual-levelvariability parameterTable 4. Predicted and Observed BP Lowering Effects(CS8635-A-U301) Mean (SD)∆ (SeDBP)predictedmean ∆DBP[mm Hg]∆ (SeSBP)Treatment (mg) Observed Predicted Observed PredictedOM/AML 40/10 -17.8 (9.9) -17.8 (9.0) -31.1 (16) -31.1 (15)OM/HCTZ 40/25 -16.5 (11) -16.5 (10) -31.2 (19) -31.1 (18)AML/HCTZ 10/25 -14.8 (9.3) -14.8 (8.3) -29.0 (16) -29.0 (14)OM/AML/HCTZ 40/10/25 -21.5 (11) -21.5 (9.8) -38.1 (18) -38.1 (17)Table 5. Predicted BP Lowering Effect of CS-8635 –(Amlodipine to Benicar HCT ® ) Mean (SD)AML (mg)0 m 5 10Benicar HCT ®(mg/mg)∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP0/0 -4.0 (9.2) -4.7 (15.2) -10.7 (9.8) -17.6 (16.5) -14.1 (10.2) -22.9 (17.4)20/12.5 -12.7 (9.4) -22.4 (16.0) -16.8 (9.7) -30.4 (16.8) -18.9 (9.9) -33.7 (17.1)40/12.5 -14.1 (9.5) -24.4 (16.1) -17.9 (9.7) -32.1 (16.8) -19.8 (9.9) -35.2 (17.1)40/25 -16.6 (9.6) -29.9 (16.7) -19.8 (9.8) -35.9 (17.0) -21.4 (9.9) -38.4 (17.2)Table 6. Predicted BP Lowering Effect of CS-8635 –(HCTZ to AZOR ® ) Mean (SD)HCTZ (mg)predictedmean ∆SBP[mm Hg]Final modelPrediction (“PRED”) without individuallevelvariability parameterpredictedmean ∆DBP[mm Hg]predictedmean ∆SBP[mm Hg]OM Dose AML Dose HCTZ Dose[mg] [mg] [mg]0 0 0 -3.08 -2.81 -3.09 -2.88 -3.24 -3.890 10 0 -13.0 -20.2 -13.0 -20.1 -13.3 -20.00 5 0 -9.68 -15.3 -9.68 -15.3 -9.78 -15.110 0 0 -8.07 -11.9 -8.01 -11.9 -8.38 -12.810 10 0 -16.3 -25.8 -16.3 -25.8 -16.0 -25.610 5 0 -14.2 -24.7 -14.2 -24.6 -13.9 -23.520 0 0 -9.11 -13.5 -9.16 -13.7 -10.0 -15.720 10 0 -17.1 -29.1 -17.1 -29.1 -16.9 -28.620 5 0 -14.0 -23.4 -14.0 -23.5 -14.5 -24.340 0 0 -10.6 -17.1 -10.6 -17.1 -10.9 -17.140 10 0 -19.2 -30.5 -19.2 -30.6 -18.4 -31.040 5 0 -15.8 -26.2 -15.7 -26.2 -15.3 -25.3observedmean ∆DBP[mm Hg]CS866-318observedmean ∆SBP[mm Hg]Final modelPrediction (“IPRED”) with individual-levelvariability parameterpredictedmean ∆DBP[mm Hg]predictedmean ∆SBP[mm Hg]Final modelPrediction (“PRED”) without individuallevelvariability parameterpredictedmean ∆DBP[mm Hg]predictedmean ∆SBP[mm Hg]OM Dose AML Dose HCTZ Dose[mg] [mg] [mg]0 0 0 -7.51 -2.92 -7.43 -3.02 -6.59 -4.100 0 12.5 -9.28 -8.61 -9.26 -8.76 -9.03 -10.40 0 25 -12.9 -17.6 -12.8 -17.7 -12.2 -17.810 0 0 -12.7 -10.3 -12.6 -10.5 -11.6 -12.610 0 12.5 -15.3 -20.3 -15.2 -20.2 -14.5 -18.910 0 25 -18.4 -22.9 -18.3 -22.9 -17.4 -23.420 0 0 -12.4 -14.9 -12.5 -14.9 -13.4 -14.920 0 12.5 -15.8 -21.3 -15.8 -21.2 -15.9 -19.320 0 25 -18.9 -25.7 -18.9 -25.6 -19.0 -25.240 0 0 -14.4 -16.3 -14.5 -16.4 -15.0 -17.140 0 12.5 -18.4 -20.4 -18.3 -20.4 -17.7 -21.240 0 25 -21.9 -27.8 -21.7 -27.7 -20.2 -26.10 12.5 25AZOR ®(mg/mg)∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP ∆SeDBP ∆SeSBP0/0 -4.0 (9.2) -4.7 (15.2) -6.4 (9.3) -12.5 (15.7) -8.9 (9.4) -20.3 (16.9)5/20 -15.0 (9.7) -26.5 (16.6) -16.8 (9.7) -30.4 (16.8) -18.6 (9.7) -34.4 (17.0)5/40 -16.0 (9.7) -28.2 (16.7) -17.9 (9.7) -32.1 (16.8) -19.8 (9.8) -35.9 (17.0)10/40 -18.2 (9.9) -32.0 (17.1) -19.8 (9.9) -35.2 (17.1) -21.4 (9.9) -38.4 (17.2)Figure 7. Boxplots of DBP Simulation Results for “Add toBenicar HCT ® ” ScenarioChange from Baseline [mm Hg]10 -40-10-20-30-40OMAMLHCTZPlacebo000AML 5-10.7050AML 10OM/HCTZ 20/12.5OM/HCTZ 40/12.5DBP-14.1-12.7-14.2010020012.540012.5OM/HCTZ 40/25-16.6 -16.840025OM/AML/HCTZ 20/5/12.520512.5Note: Number is mean. Dashes are 10 th , 50 th , and 90 th percentiles of population. Solid line is 25-75 th range.OM/AML/HCTZ 40/5/12.5-17.940512.5OM/AML/HCTZ 40/10/12.5-19.8 -19.8401012.5OM/AML/HCTZ 40/5/2540525OM/AML/HCTZ 40/10/25-21.4401025Figure 8. Boxplots of DBP SimulaAZOR ® ” ScenarioChange from Baseline [mm Hg]Conclusions10 -40-10-20-30-40OMAMLHCTZPlacebo000HCTZ 12.5-6.50012.5HCTZ 25-8.90025DBPOlmesartan medoxomil PK was adequately characterized band time lag; creatinine clearance was a significant predictomedoxomil.Amlodipine PK was adequately characterized by a two-comlag; age was a significant predictor of the apparent oral cleaHydrochlorothiazide PK was adequately characterized by aand a time lag; sex, age, and creatinine clearance were signhydrochlorothiazide.The blood pressure lowering effects of olmesartan medoxomand SBP were described by an E max model, whereas the bloexposure was described by a linear model.The blood pressure lowering effects of olmesartan medoxommonotherapy, dual combination therapy, and triple combinacomposed of the sum of the individual effects and interactioThe model-predicted and simulated blood pressure lowerinthree compounds were in good agreement with the observemodel was built.In the exposure-response model, black race was a covariatpressure of olmesartan medoxomil without influencing PK pIn the exposure-response model, baseline seated trough DBlowering effect associated with higher baseline blood pressThe exposure-response model predicted the blood pressurepermutations to be superior to their respective mono and dThe order of the blood pressure lowering effects among the20/5/12.5


yrman, MD ; Daniel E. Salazar, PhDP Lowering Effects18)Final model(“IPRED”) with individual-levelvariability parametericted∆DBPHg]P Lowering EffectsD)∆ (SeSBP)Predicted Observed Predicted-17.8 (9.0) -31.1 (16) -31.1 (15)-16.5 (10) -31.2 (19) -31.1 (18)-14.8 (8.3) -29.0 (16) -29.0 (14)-21.5 (9.8) -38.1 (18) -38.1 (17)fect of CS-8635 –T ® ) Mean (SD)ML (mg)5 10∆SeSBP ∆SeDBP ∆SeSBP-17.6 (16.5) -14.1 (10.2) -22.9 (17.4)-30.4 (16.8) -18.9 (9.9) -33.7 (17.1)-32.1 (16.8) -19.8 (9.9) -35.2 (17.1)-35.9 (17.0) -21.4 (9.9) -38.4 (17.2)fect of CS-8635 –D)TZ (mg)predictedmean ∆SBP[mm Hg]12.5 25Final modelPrediction (“PRED”) without individuallevelvariability parameterpredictedmean ∆DBP[mm Hg]predictedmean ∆SBP[mm Hg]09 -2.88 -3.24 -3.89.0 -20.1 -13.3 -20.068 -15.3 -9.78 -15.101 -11.9 -8.38 -12.8.3 -25.8 -16.0 -25.6.2 -24.6 -13.9 -23.516 -13.7 -10.0 -15.7.1 -29.1 -16.9 -28.6.0 -23.5 -14.5 -24.3.6 -17.1 -10.9 -17.1.2 -30.6 -18.4 -31.0.7 -26.2 -15.3 -25.3Final model(“IPRED”) with individual-levelvariability parametericted∆DBPHg]delspredictedmean ∆SBP[mm Hg]Final modelPrediction (“PRED”) without individuallevelvariability parameterpredictedmean ∆DBP[mm Hg]predictedmean ∆SBP[mm Hg]43 -3.02 -6.59 -4.1026 -8.76 -9.03 -10.4.8 -17.7 -12.2 -17.8.6 -10.5 -11.6 -12.6.2 -20.2 -14.5 -18.9.3 -22.9 -17.4 -23.4.5 -14.9 -13.4 -14.9.8 -21.2 -15.9 -19.3.9 -25.6 -19.0 -25.2.5 -16.4 -15.0 -17.1.3 -20.4 -17.7 -21.2.7 -27.7 -20.2 -26.1∆SeSBP ∆SeDBP ∆SeSBP-12.5 (15.7) -8.9 (9.4) -20.3 (16.9)-30.4 (16.8) -18.6 (9.7) -34.4 (17.0)-32.1 (16.8) -19.8 (9.8) -35.9 (17.0)-35.2 (17.1) -21.4 (9.9) -38.4 (17.2)ion Results for “Add to6.6 -16.8005OM/AML/HCTZ 20/5/12.520512.5OM/AML/HCTZ 40/5/12.5id line is 25-75 th range.-17.940512.5OM/AML/HCTZ 40/10/12.5-19.8 -19.8401012.5OM/AML/HCTZ 40/5/2540525OM/AML/HCTZ 40/10/25-21.4401025Figure 8. Boxplots of DBP Simulation Results for “Add toAZOR ® ” ScenarioChange from Baseline [mm Hg]Conclusions10 -40-10-20-30-40OMAMLHCTZPlacebo000HCTZ 12.5-6.50012.5HCTZ 25-8.90025AML/OM 5/20-152050AML/OM 5/40DBPOlmesartan medoxomil PK was adequately characterized by a two-compartment model with first-order absorptionand time lag; creatinine clearance was a significant predictor of the apparent oral clearance of olmesartanmedoxomil.Amlodipine PK was adequately characterized by a two-compartment model with first-order absorption and a timelag; age was a significant predictor of the apparent oral clearance of amlodipine.Hydrochlorothiazide PK was adequately characterized by a two-compartment model with first-order absorptionand a time lag; sex, age, and creatinine clearance were significant predictors of the apparent oral clearance ofhydrochlorothiazide.The blood pressure lowering effects of olmesartan medoxomil and amlodipine exposure on seated trough DBPand SBP were described by an E max model, whereas the blood pressure lowering effect of hydrochlorothiazideexposure was described by a linear model.The blood pressure lowering effects of olmesartan medoxomil, amlodipine, and hydrochlorothiazide inmonotherapy, dual combination therapy, and triple combination therapy were well characterized by a modelcomposed of the sum of the individual effects and interaction among the components.The model-predicted and simulated blood pressure lowering effects of the various tested combinations of thethree compounds were in good agreement with the observed data from the three Phase 3 studies upon which themodel was built.In the exposure-response model, black race was a covariate, decreasing the maximal possible effect on bloodpressure of olmesartan medoxomil without influencing PK parameters.In the exposure-response model, baseline seated trough DBP and SBP were covariates, with more blood pressurelowering effect associated with higher baseline blood pressure.The exposure-response model predicted the blood pressure lowering effects of triple combination therapypermutations to be superior to their respective mono and dual treatments of OM, AML, and HCTZ.The order of the blood pressure lowering effects among the different CS-8635 dose strengths was:20/5/12.5

More magazines by this user
Similar magazines