Assessing Cardiovascular risk in different populations

Transcription

1 A-1 Assessing Cardiovascular risk in different populations Murray Stewart (GSK)

2 A-2 Studies with RSG across different risk population Short term studies low risk ADOPT newly diagnosed low risk DREAM IGT low risk RECORD medium risk BARI 2 D high risk

3 Meta Analyses of Rosiglitazone Controlled Clinical Trials Myocardial Infarction Cardiovascular Mortality GSK Integrated Clinical Trials 1.59 ( ) 1.91 ( ) FDA ( ) 1.7 ( ) Nissen & Wolski ( ) 1.64 ( ) 1 FDA Briefing Document, Statistical Review, J Mele, Table Nissen and Wolski NEJM 2007;356: A 3

4 Events of Myocardial Infarction A 4 Treatment # of Patients Rate / 100 PY HR (95% CI) with Events ICT RSG (n=8604) Comp (n=5633) ( ) RECORD RSG (n=2220) Comp (n=2227) ( ) ADOPT RSG (n=1456) MET (n=1454) ( ) GLY (n=1441) ( ) DREAM RSG (n=1325) ( ) RSG+RAM (n=1310) ( ) RAM (n=1313) Placebo (n=1321)

5 MACE Events RECORD + ADOPT + DREAM Treatment # of Rate / HR (95% CI) Rate Difference/100PY Events 100 PY (95% CI) MI/SD RSG (n=6311) Comp (n=7756) (0.82, 1.54) 0.05 ( 0.06, 0.15) Stroke RSG (n=6311) Comp (n=7756) (0.58, 1.21) 0.05 ( 0.14, 0.04) CV Mortality RSG (n=6311) Comp (n=7756) (0.57, 1.22) 0.03 ( 0.12, 0.05) MACE RSG (n=6311) Comp (n=7756) (0.83, 1.28) 0.03 ( 0.12, 0.18) Reflects Post Study Adjudication for ICT and ADOPT based on definite MI/SD and CV mortality A 5

6 MACE Events Active vs Placebo Controlled Studies (52 Studies) A-6 Events [%] RSG Control HR (95% CI) GSK FDA Pbo Cont 50 [0.7] 24 [0.5] 1.38 (0.85,2.25) Active Control 30 [0.9] 27 [1.2] 0.85 (0.50,1.43) Pbo Cont 54 [0.7] 28 [0.5] 1.53 (0.94,2.54) Active Control 16 [0.8] 14 [0.7] 1.05 (0.48,2.34) Hazard Ratio Favors RSG Favors Control

7 ICT Considerations Primarily short term trials Trials not designed to assess CV endpoints Endpoints not prospectively adjudicated Small number of events Adjudicated MACE Events ICT-42 No 101 ICT-52 No 131 FDA-52 No 109 BARI-2D Yes 455 RECORD Yes 319 Long term trial data are more reliable for assessing Longterm CV risk A-7

8 A-8 Limitations to interpretations Small numbers are inconclusive (speculation) Short term trials are not good predictors Low risk populations have small numbers! Placebo and active comparators are different Different meta-analysis of the same data do not provide independent evidence

9 Primary endpoint: CV hospitalization or CV death Favors rosiglitazone Favors control non-inferiority margin 1.20 Primary endpoint (n, events/patients) rosiglitazone 321 / 2220 control 323 / 2227 HR 0.99 (0.85,1.16) p= Hazard ratio (95% CI) Non-inferiority criterion achieved (CI excludes 1.20) Non-inferiority p-value = 0.02 A-9

10 Secondary endpoints Rosiglitazone Control HR (95% CI) p-value (n=2220) (n=2227) Death all cause ( ) 0.19 cardiovascular ( ) 0.32 Heart failure* ( ) CV death, MI or stroke ( ) 0.50 Myocardial infarction* ( ) 0.47 Stroke* ( ) 0.10 * Fatal and non-fatal MACE (major adverse cardiovascular event) Home PD, et al. Lancet 2009;373: A-10

11 RECORD: overall conclusions In a typical Type 2 diabetes population, and when compared to metformin and sulfonylureas (in combination), rosiglitazone: is non-inferior for CV hospitalisation or CV death does not increase all cause death, CV death, MI or stroke Increases heart failure A-11

12 ACCORD: NIH Presentation A-12

13 BARI-2D Subanalysis: RSG vs no TZD Rosiglitazone was used in 55% of patients randomly allocated to Insulin Sensitization group and in 3% of those allocated to Insulin Provision. Crude Incidence Rate/100 PY [Events] Rosiglitazone No TZD Relative Risk (95% CI) p-value Death 1.88 [57] 2.56 [183] 0.83 (0.58,1.18) 0.29 MACE 3.79 [105] 5.81 [350] 0.72 (0.55,0.93) 0.01 MI 2.16 [60] 3.16 [193] 0.77 (0.54,1.10) 0.15 Favors RSG RSG = rosiglitazone TZD = thiazolidinedione, i.e. rosiglitazone or pioglitazone MACE = Major Acute Clinical Event, i.e. Death, MI or stroke MI = myocardial infarction Relative Risk Favors No TZD A-13

14 A-14 Assessment of cardiovascular risk in different populations with RSG Conclusion that RSG in comparison to met/su is effective and safe treatment opition for use in type 2 diabetic patients. Label update in discussion with FDA Caveats caution with heart failure/insulin Caution in extrapolating data

15 Summary Slide What Was The CV Safety Issue Leading to Outcome Study Requirement Weight of Evidence At file fluid retention, weight gain, lipids Post marketing Surveillance heart failure and ischaemic related events RECORD showed non inferiority to met/su Presence of a Likely MOA for Risk Benefit:Risk Relationship for Drug? Is perceived risk: Generalizable to Other Drugs in Class or with Same Efficacy MOA? Lessons Learned Not established Same as other diabetes drugs Perceived risk is that this was different from other agents in the class e.g. lipids Conduct event driven, high risk study v placebo on background therapy, blinded and using probably MACE endpoint A-15