HDL en LDL cholesterol state of the art Innovatieve verlaging van LDL cholesterol: successen, gevaren, kansen Prof. dr. F.L.J. Visseren - internist Universitair Medisch Centrum Utrecht
Innovatieve verlaging van LDL-c: successen, gevaren, kansen. Frank LJ Visseren
Nieuwe geneesmiddelen: - antisense apob - CETP remming - PCSK9 remming Behandeleffect bij individuele patienten voorspellen.
Antisense ApoB
Apo B-100 as a Target Apo B Cholesterol Triglyceride VLDL Apo B-100 is a structural and functional component of lipoproteins Blocking Apo B-100 production blocks VLDL and LDL production VLDL IDL Apo A LDL1 LDL2 LDL3 Lp(a)
Antisense: Mechanism of Action DNA Transcription mrna Translation Disease-Associated Protein Traditional Drug Transcription No Translation Antisense Drug (Oligonucleotide) No Disease- Associated Proteins Produced Crooke R, et al. Adapted from: Crooke ST, ed. Antisense drug technology: principles, strategies and applications. 2nd ed. Boca Raton, FL: CRC Press, 2007:601-639.
Mean percentage change Reductions in Apo B and LDL-c Add-on in Heterozygous Familial Hypercholesterolem Reductions by dose over 6 weeks Apo B 50 100 Placebo mg/wk mg/wk 200 mg/wk 300 mg/wk (n = 8) (n = 8) (n = 8) (n = 11) (n = 9) 1% LDL-C 50 100 200 300 Placebo mg/wk mg/wk mg/wk mg/wk (n = 8) (n = 8) (n = 8) (n = 11) (n = 9) 10% 8% 13% 11% 23% P <0.05 21% P <0.05 33% P <0.01 34% P <0.01 Akdim F, et al. Am. J. Cardiol. 2010;105(10):1413-1419.
Mipomersen Reduces LDLc and Lp(a) in Homozygous FH CS5 Reduction in LDL-C over 28 weeks 3% 25% Reduction in Lp(a) over 28 weeks 8% 31% Raal FJ, et al. Lancet. 2010:375(9719):998-1006. Raal. FJ, et al. Lancet. 2010:375(9719):998-1006.
Mean percentage change in LDL-c Mean percantage change in ApoB a apob change 20 10 0 Mipomersen reduces LDLc and apob subjects Placebo Mipomersen in statin intolerant -10-20 -30-40 -50-60 0 5 9 13 17 21 25 28 32 40 50 LDLc change b 20 10 0 Timpoint (week) Placebo Mipomersen > 80% adherence to Mipomersen in statin-intolerant patients after 26 weeks of therapy! -10-20 -30-40 -50-60 0 5 9 13 17 21 25 28 32 40 50 Time point (week) Data on file, ME Visser, 2011
Summary Safety & Tolerability A: Tolerability Adverse event, preferred term Placebo (n = 17) Patients, n (%) Mipomersen (n = 34) Patients, n (%) All events 13 (76) 30 (88) Injection-site reactions, all terms 4 (24) 26 (76) Flu-like symptoms, all terms 4 (24) 10 (29) Nausea 1 (6) 6 (18) B: Safety Liver: 8-20% of Mipo patients ALT increase 3 ULN MRI: Increased liver fat fraction in Mipo patient Other: No CRP increase after 3-6 months No cardiac, renal, hematological issues
PCSK9 remming (Proprotein Convertase Subtilisin/Kexin Type 9)
Hepatic LDL-R play a key role in regulating plasma LDL-c levels
PCSK9 regulates the surface expression of hepatic LDL-Rs
Loss-of-function mutations in human PCSK9 are associated with lower LDL-c levels 1. NEJM 2006;354:1264-1272. 2. Nat Genet 2005;37:161-165 3. JACC 2010;55:2833-2842
Anti-PCSK9 monoclonal antibodies block PCSK9/LDL-R interaction and may lower LDL-C
Safety and efficacy of a monoclonal antibody to PCSK9 in patients with primary hypercholesterolemia on atorvastatin therapy JM McKenney, et al. JACC 2012
Safety and efficacy of a monoclonal antibody to PCSK9 in patients with primary hypercholesterolemia on atorvastatin therapy JM McKenney, et al. JACC 2012
CETP remming (Cholesterol Ester Transfer Protein)
Cholesterol Ester Transfer Protein (CETP)
Anacetrapib NEJM 2010;363:2406-2415
Anacetrapib (methoden) 18-80 jaar Coronaire hartziekte of >20% Framingham risico LDL-c tussen 1,3 en 2,6 mmol/l tijdens statine therapie HDL-c <1,6 mmol/l Triglyceriden <4,5 mmol/l Geen MI of vasculaire interventie <3 maanden Randomiseren tussen anacetrapib 100mg of placebo Primaire uitkomst: veranderingen lipiden en bijwerkingen NEJM 2010;363:2406-2415
Anacetrapib (baseline characteristics) NEJM 2010;363:2406-2415
Anacetrapib (resultaten) NEJM 2010;363:2406-2415
Anacetrapib (resultaten) NEJM 2010;363:2406-2415
Anacetrapib (resultaten) NEJM 2010;363:2406-2415
Evacetrapib JAMA 2011;306:2099-2109
Evacetrapib (methoden) >18 jaar HDL >1,2 (mannen) of >1,3 (vrouwen) LDL-c tussen 2,6 4,0 mmol/l; FHS risico <10% LDL-c tussen 2,6 3,4 mmol/l; FHS risico 10-20% Randomiseren tussen anacetrapib monotherapie 30mg, 100mg en 500mg of placebo Toegevoegd aan statine: randomiseren anacetrapib 100mg of placebo Primaire uitkomst: veranderingen lipiden en bijwerkingen JAMA 2011;306:2099-2109
Evacetrapib (baseline characteristics) JAMA 2011;306:2099-2109
Evacetrapib (resultaten) JAMA 2011;306:2099-2109
JAMA 2011;306:2099-2109
Evacetrapib (conclusies) Evacetrapib monotherapie: HDL-c stijgt met 54-129% LDL-c daalt met 14-36% Evacetrapib toegevoegd aan statine: HDL-c stijgt met 79-89% LDL-c daalt met 11-14% Geen noemenswaardige bijwerkingen JAMA 2011;306:2099-2109
Individualized prediction of treatment effects
Rationale Cardiovascular disease runs in my family. Do I need to use a statin? Does that mean that I will benefit too? Research shows the average patient benefits from statin.
Translating the results of trials to clinical practice Trials report a single average relative effect. Applicable to which patiënt? Average patiënt? There is a range of treatment effect! Presumably based on patiënt characteristics. One treatment fits all is an oversimplification of clinical reality.
Average patient? ------------------------------------------------------- 30% female
Average patient? 70% male ------------------------------------------------------- 30% female
Average patient? ---------------------------------------------------------------------- 30 years
Average patient? 70 years ---------------------------------------------------------------------- 30 years
Average patient? Average treatment? Current practice: providing average treatment to an average patient! Large Numbers Needed to Treat (>30 in 10 years to prevent 1 outcome event). Unable to identify individual patients that have large benefit from treatment and have little/no harm.
Not every patient benefits from treatment! Glasziou et al, BMJ. 1995 Nov 18;311(7016):1356-9.
Subgroup analyses are not the solution! Univariable analyses of patient characteristics that modify relative treatment effect Disadvantages: Multiple testing (FP) Loss of statistical power One patient characteristic at the time: Full range of treatment effect heterogeneity remains undiscovered Correlation of characteristics (FN) Does not address heterogeneity of absolute treatment effect
Then what do we need? A model for estimation of absolute treatment effect for individual patients, based on multiple characteristics together......allowing to test heterogeneity of both absolute and relative treatment effect......while minimizing the risk for FP and FN results. Applicable before start of intended therapy Also applicable in fields of medicine where risk scores are not yet available
??? Genes? Biomarkers? Imaging?? Solution: individualized prediciton of treatment effects by modelling based on available information!
Aims To show how absolute treatment effect for individual patients can be predicted based on data from randomised trials. To evaluate the net benefit of making treatment decisions based on individual predicted absolute treatment effect.
JUPITER Primary Trial Endpoint : MI, Stroke, UA/Revascularization, CV Death 0.00 0.02 0.04 0.06 0.08 Cumulative Incidence Ridker et al NEJM 2008 HR 0.56, 95% CI 0.46-0.69 P < 0.00001 0 1 2 3 4 Placebo 251 / 8901-44 % Rosuvastatin 142 / 8901 Number at Risk Rosuvastatin Placebo Follow-up (years) 8,901 8,631 8,412 6,540 3,893 1,958 1,353 983 544 157 8,901 8,621 8,353 6,508 3,872 1,963 1,333 955 534 174
Strategies for predicting treatment effect BMJ 2011;343:d5888
Model performance C-statistic 0.65 (95%CI 0.62-0.68) Concordance statistic: rcorr.cens() from Hmisc library in R C-statistic 0.66 (95%CI 0.63-0.69) Based on Somers' rank correlation for censored response variabel: Dxy / 2 + 0.5 C-statistic 0.71 (95%CI 0.68-0.74) Probability of pairs of subjects who can be ordered such that the subject with the higher predicted survival is the one who survived longer BMJ 2011;343:d5888
Distribution of individual treatment effect Framingham based Reynolds based Optimal fit model based Median 4.4% (IQR 2.6-7.0%) Mean 5.1% Median 4.2% (IQR 2.5-7.1%) Mean 5.4% Median 3.9% (IQR 2.5-6.1%) 4.5% BMJ 2011;343:d5888
Individualized treatment effect prediction Age Gender Current smoking BP lowering meds Family Hx CHD LDLc hscrp Diabetes 60 years male no yes positive 2.8 mmol/l 4.3 mg/l No
Individualized treatment effect prediction Age Gender Current smoking BP lowering meds Family Hx CHD LDLc hscrp Diabetes 55 years female yes no negative 2.8 mmol/l 4.3 mg/l No
Net benefit of individualized treatment effect prediction (based on JUPITER) BMJ 2011;343:d5888
A prediction rule for the effect of high-dose statin therapy in patients with stable coronary artery disease based on data from the TNT and IDEAL trials [to be submitted soon] J.A. Dorresteijn, S.M. Boekholdt, Y. van der Graaf, J.J. Kastelein, J.C. LaRosa, T.R. Pedersen, D.A. DeMicco, P.M Ridker, N.R. Cook, F.L. Visseren
Model presentation
Dorresteijn et al. Eur Heart J 2011;32:2962-2969
Chol: RR: Stolling: Verhoogd risico Vaatziekten Diabetes
Conclusies Nieuwe therapie voor LDL-c reductie in aantocht!! Antisense ApoB, PCSK9 remming en CETP-remming. Mogelijkheden voor schatten van individueel behandeleffect in ontwikkeling.