1. www.ascotstudy.org Cardiovascular and All-Cause Mortality Outcomes Among Hypertensive Patients With Moderate Renal Dysfunction in the ASCOT- LLA, and its Extended Follow-up Dr Ajay K Gupta International Centre for Circulatory Health, ICTU, Imperial College London, UK Presented on Behalf of the ASCOT Investigators

2. www.ascotstudy.org Background: CKD and Cardiovascular morbidity and mortality The Lancet, Vol 375, Pages 2073 - 2081, 12 June 2010 An eGFR less than 60 mL/min/1 ・ 73 m2 is an independent predictor of all-cause mortality and cardiovascular mortality in the general population. Stag e Description GFR (mL/min/1.73 m 2 ) Clinical term 1Kidney damage with normal or  GFR >90 2 Mild  GFR 60-89 3 Moderate  GFR 30-59 CKD 4 Severe  GFR 15-29Advanced CKD 5Kidney failure<15 or dialysis ESRD

3. www.ascotstudy.org Trial evidence: statin use among those with CKD In clinical trials on chronic kidney disease (CKD) patients on haemodialysis, the use of statins was not associated with any cardiovascular (CV) benefits (AURORA trial & 4D study). Among non-dialysis CKD patients, the use of statins also failed to show any extra CV benefits (PREVEND-IT & ALERT). Among hypertensive patients with renal dysfunction in the lipid-lowering arm of the ALLHAT trial, the use of pravastatin had no added CV benefits, beyond that of usual care.

4. www.ascotstudy.org Statin among those with CKD In contrast, in post-hoc secondary analyses of the JUPITER trial, CARDS & TNT, allocation to the statin therapy was found to have protective effects on some, but not all, CV outcomes among patients with renal dysfunction. In the SHARP trial, among CKD patients, the use of ezetimibe/simvastatin combination was associated with: –17% and 19% (significant) reduction in major vascular events and stroke, respectively. –No significant impact on total coronary events or all- cause mortality.

5. www.ascotstudy.org 0.50.7511.52 Trial Events (% pa) Allocated LDL-C reduction Allocated control Risk ratio (RR) per mmol/L LDL-C reduction p LDL-C reduction better Control better 99% or95% CI Comparison of the impact of statin on non-fatal MI in the Renal and non-Renal trials 4D 33 (1.91)35 (2.02) AURORA 91 (1.97)107 (2.33) ALERT54 (1.03)65 (1.24) SHARP 134 (0.71)159 (0.85)  3 2 =0.3 (p = 0.96) Subtotal: 4 renal trials 312 (1.02)366 (1.21)0.83 (0.70 - 0.98) 0.03 23 other trials3307 (0.97)4386 (1.29)0.73 (0.70 - 0.76) <0.0001 All trials 3619 (0.97)4752 (1.29)0.74 (0.70 - 0.77) <0.0001 Difference between renal and non- renal trials:  1 2 = 2.2 (p = 0.14) *modified from the SHARP presentation

6. www.ascotstudy.org 0.50.7511.52 Trial Events (% pa) Allocated LDL-C reduction Allocated control Risk ratio (RR) per mmol/L LDL-C reduction p LDL-C reduction better Control better 99% or95% CI 4D151 (8.52)167 (9.36) AURORA 324 (6.87)324 (6.86) ALERT 66 (1.23)73 (1.36) SHARP 361 (1.82)388 (1.97)  3 2 =0.9 (p = 0.82) Subtotal: 4 renal trials902 (2.85)952 (3.01)0.94 (0.85 - 1.04) 0.27 23 other trials 3679 (1.05)4230 (1.21)0.85 (0.81 - 0.89) <0.0001 All trials 4581 (1.20)5182 (1.36)0.86 (0.83 - 0.90) <0.0001 Difference between renal and non- renal trials:  1 2 =3.8 (p = 0.05) Comparison of impact of statin on Vascular death in the Renal vs Non-Renal trials *modified from the SHARP presentation

7. www.ascotstudy.org ASCOT-LLA-Extension Early closure of ASCOT-LLA  Median follow-up 3.3 years  Atorvastatin versus placebo: o 36% reduction in the primary endpoint o 27% reduction in stroke ASCOT-LLA-extended  Offering atorvastatin 10 mg daily to all patients in LLA  Continued for a further 2.2 years until the closure of ASCOT-BPLA Statin usage at the end of LLA-extended Atorvastatin (n=4978)Placebo (n=4916) AtorvastatinOther statinsAtorvastatinOther satins End of ASCOT-LLA4113 (82.6) 54 (1.1) 415 (8.4)220 (4.5) End of ASCOT-BPLA*3122 (62.7)200 (4.0) 2752 (56.0)337 (6.9) Values are n (%) *Also the end of LLA-extension

8. www.ascotstudy.org Methods Based on mean estimated glomerular filtration rate (eGFR) (MDRD- method), patients at baseline were allocated to have a moderate renal damage (CKD) (eGFR 30-60 mL/min/1.73 m 2 ) or not. Outcomes: –Non-fatal myocardial infarction (MI) plus fatal CHD plus revascularisation. –Secondary outcomes: total coronary events, fatal and non-fatal stroke, CV mortality & all-cause mortality.

9. www.ascotstudy.org Statistical Methods For each of these outcomes, a separate Cox model was developed to assess the effect of statin therapy, after adjusting for a priori confounders: age, sex, race, socio-economic status (age at leaving education), systolic BP and allocated BP treatment, among those with and without CKD at baseline. Heterogeneity in the treatment effect between those with or without CKD was assessed using an appropriate interaction test These analyses were repeated using time to first event during the extended follow-up in the ASCOT-LLA-extended.

10. www.ascotstudy.org Trial Profile: patients with CKD in the ASCOT- lipid lowering arm 10,305 randomised 10,235 evaluable 65 excluded because of protocol violations 5,132 in statin arm5,103 in placebo arm 4,083(80.0%) without CKD 1,020(20.0 %) with CKD 4,130 (80.5%) without CKD 1,002(19.5%) with CKD 5 excluded as serum creatinine at baseline > 200 μ mol/L

11. www.ascotstudy.org Baseline characteristics Non CKDCKD Mean (SD)/% Age62.0 (8.4)67.9 (7.4) Statin treatment allocation (%)50.349.6 Amlodipine-based treatment allocation (%)50.646.9 Male sex (%)85.464.5 Race (%) White Caucasian94.395.9 Age at leaving full-time education (%) ≥19 years20.816.5 BMI (kg/m 2 )28.7 (4.6)28.6 (5.0) Current (or ex <1 year) smoker69.162.3 Alcohol (units/week)8.7 (12.1)5.4 (8.9) Diabetes (%)26.626.9 Presence of LVH (%)22.725.1 Previous antihypertensive treatment (%)79.983.7 Systolic BP (mm Hg)163.6 (17.6)166.8 (18.6) Glucose (mmol/L)*6.2 (2.1)6.2 (2.0) Total cholesterol (mmol/L)5.5 (0.8) HDL-cholesterol (mmol/L)1.3 (0.4) LDL-cholesterol (mmol/L)*3.4 (0.7)3.5 (0.7) Serum creatinine (µmol/L)*94.2 (11.1)116.5 (19.3) eGFR (ml/min/1.73 m 2 )*73.6 (9.1)53.2 (6.0) Baseline characteristics, stratified by CKD status

12. www.ascotstudy.org Lipids Levels During ASCOT-LLA and LLA-extended *Lipid closeout visit (end of ASCOT-LLA) Atorvastatin Placebo

13. www.ascotstudy.org Mean LDL-cholesterol among those with CKD, stratified by treatment dif=mean difference in change in LDL-C from baseline to year 3 between atorvastatin (-1.2) and placebo (-0.3) LDL-C: low-density lipoprotein cholesterol; CKD: chronic kidney disease; Ator: atorvastatin CKD & Atorvastatin dif=0.9 P<0.0001 CKD & Placebo

14. www.ascotstudy.org Non-CKD (8,061)CKD (2,022) Events (n) Rates (95% CI) Events (n) Rates (95% CI) NF MI + fatal CHD1997.7 (6.7 to 8.8)507.8 (5.9 to 10.3) All-cause mortality28110.7 (9.6 to 12.1)10516.2 (13.4 to 19.6) Fatal + NF strokes1535.9 (5.0 to 6.9)528.1 (6.2 to 10.6) CV death + MI + stroke40515.8 (14.3 to 17.4)12119.1 (16.0 to 22.8) NF MI + fatal CHD + revascularisation procedures 26410.2 (9.1 to 11.5)6910.8 (8.5 to 13.6) CV death1124.3 (3.6 to 5.2)396.0 (4.4 to 8.2) Total coronary events32812.8 (11.5 to 14.2)9014.2 (11.5 to 17.4) Total coronary events + revascularisation procedures 64625.6 (23.7 to 27.7)21034.1 (29.8 to 39.0) CI: confidence interval; NF: non-fatal; (those with missing value of serum creatinine at baseline were excluded from these analyses) Event rates of cardiovascular outcomes and death, among those with and without presence of CKD

15. www.ascotstudy.org Interaction Test Placebo p=0.26 p=0.10 p=0.70 p=0.35 p=0.89 NF MI+F CHD+ Revas Non-CKD CKD NF MI+NF Stroke +CV Death Non-CKD CKD CV Death Non-CKD CKD Total Stroke Non-CKD CKD All Cause Mortality Non-CKD CKD Total Coronary Events Non-CKD CKD p=0.97 161 37 240 62 59 22 90 27 152 56 190 52 Atorvastatin 103 32 165 59 53 17 63 25 129 49 138 38 0.50.60.70.80.911.21.5 Total events 264 69 405 121 112 39 153 52 281 105 328 90 Favors atorvastatin Favors placebo Hazard ratio and 95% CI (log scale) Adjusted for age, sex, baseline SBP, ethnicity, socio-economic status, randomised BP treatment Cardiovascular events and deaths associated with statin therapy among those with and without CKD (ASCOT-LLA)

16. www.ascotstudy.org NF MI+F CHD+ Revas Non-CKD CKD NF MI+NF Stroke+CV Death Non-CKD CKD CV Death Non-CKD CKD Total Stroke Non-CKD CKD All Cause Mortality Non-CKD CKD Total Coronary Events Non-CKD CKD Atorvastatin 170 54 250 98 93 40 110 53 269 107 240 70 Interaction Test p=0.09 p=0.06 p=0.59 p=0.18 p=0.58 p=0.49 Total events 437 114 615 205 210 84 263 108 590 222 569 156 Placebo 267 60 365 107 117 44 153 55 321 115 329 86 0.50.60.70.80.911.21.5 Favors atorvastatin Favors placebo Hazard ratio and 95% CI (log scale) Adjusted for age, sex, baseline SBP, ethnicity, socio-economic status, randomised BP treatment Cardiovascular events and deaths associated with statin therapy, after 2 years of further follow- up (LLA-extended)

17. www.ascotstudy.org Conclusions Hypertensive patients with CKD are at a higher risk of cardiovascular events and death. In the ASCOT LLA –Among those without CKD, allocation to atorvastatin therapy was associated with a significant reduction in coronary and stroke outcomes, but not for death. –There was no difference (statistically) in the effect of statin therapy among those with and without renal damage. In ASCOT-LLA extended: –Statin therapy was associated with significant reduction in CV outcomes and death among those without CKD, and similar trends were noted among those with CKD. –The benefits of statin therapy remained unchanged, 2 years after the ASCOT-LLA closure. This was despite similar lipid levels at the end of LLA-extended..

18. www.ascotstudy.org Acknowledgements All ASCOT patients and investigators Professor Peter Sever, Professor Neil R Poulter, Bjorn Dahlof, C L Chang and staff at ICCH, Imperial College London, UK

19. www.ascotstudy.org Thank you for your attention

20. www.ascotstudy.org HR: adjusted for age, sex, baseline SBP, ethnicity, socio-economic status (education completion age) and randomised BP treatment Kaplan-Meier curve for total coronary events by CKD status

21. www.ascotstudy.org HR: adjusted for age, sex, baseline SBP, ethnicity, socio-economic status (education completion age) and randomised BP treatment Kaplan-Meier curve for fatal and non-fatal stroke by CKD status

22. www.ascotstudy.org HR: adjusted for age, sex, baseline SBP, ethnicity, socio-economic status (education completion age) and randomised BP treatment Kaplan-Meier curve for all-cause mortality by CKD status

23. www.ascotstudy.org HR: adjusted for age, sex, baseline SBP, ethnicity, socio-economic status (education completion age) and randomised BP treatment Kaplan-Meier curve for non-fatal MI and fatal CHD and revascularisation by chronic kidney disease status

24. www.ascotstudy.org HR: adjusted for age, sex, baseline SBP, ethnicity, socio-economic status (education completion age) and randomised BP treatment Kaplan-Meier curve for non-fatal MI, non-fatal stroke and CV deaths by chronic kidney disease status

25. www.ascotstudy.org HR: adjusted for age, sex, baseline SBP, ethnicity, socio-economic status (education completion age) and randomised BP treatment Kaplan-Meier curve for cardiovascular deaths by chronic kidney disease status

26. www.ascotstudy.org Death from Any Cause Total Events = 51,424 Relationship Between Estimated GFR (eGFR) and Clinical Outcomes eGFR (mL/min/1.73 m 2 ) Age-Standardized Event Rate (per 100 Person-Yr) Go AS, et al. N Engl J Med. 2004;351:1296-305. Cardiovascular Events Total Events = 139,011 Cardiovascular Events Total Events = 139,011 Any Hospitalization Total Events = 554,651 Any Hospitalization Total Events = 554,651  60 45– 59 30– 44 15– 29 <15  60 45– 59 30– 44 15– 29 <15  60 45– 59 30– 44 15– 29 <15

27. www.ascotstudy.org Background In the community, patients with moderate renal damage, as compared with those without, are at a higher risk of cardiovascular (CV) morbidity and mortality. Statin therapy is a well recognised strategy to reduce CV morbidity and mortality among high CV risk individuals. However, it is controversial whether the routine use of statins among patients with moderate renal dysfunction provides similar CV benefits, as seen among those with high CV risk but no renal damage.

28. www.ascotstudy.org ASCOT-LLA and LLA-extended: overview ASCOT-LLA : 10,305 hypertensive patients, with no pre- existing coronary heart disease (CHD), were randomised, using a 2  2 factorial design, to receive placebo or atorvastatin (10 mg/d). Median follow-up: 3.3 years ASCOT-LLA (extended): Patients were further followed up for an average of 2 years after closure of the LLA (as a part of the blood pressure lowering arm of the ASCOT trial [ASCOT- BPLA]). Median follow-up: 5.5 years

29. www.ascotstudy.org Changes in LDL cholesterol until the end of LLA extended Atorvastatin Placebo mmol/L mg/Dl

30. www.ascotstudy.org ASCOT Trial Design placebo atorvastatin 10 mg Double-blind ASCOT-LLA 10,305 patients TC ≤6.5 mmol/L (250 mg/dL) atenolol ± bendroflumethiazide amlodipine ± perindopril 19,257 hypertensive patients PROBE design

31. www.ascotstudy.org 1/19 268/683 1474/8448 19/431 44/160 9741 246/636 636 66/1052 1052 10/73 73 11 502 0/17 182/646 1174/8376 15/433 20/150 9622 205/619 619 46/1050 1050 6/70 70 11 361 Effect of Statins Compared With Placebo or No Treatment on CV Events in Pre-Dialysis, Dialysis, and Transplant Patients Giovanni FM. BMJ 2008;336;645-651. Pre-dialysis patients Fried 2001 HPS 2003 PPP 2004 PEVEND IT 2004 Lemos 2005 Subtotal (95% CI) Total events: 1391 (statin), 1806 (placebo) Test for heterogeneity: X 2 =5.78, df=4, P=0.22, I 2 =30.7% Test for overall effect: z=4.58, P<0.001 Dialysis patients 4D trial 2005 Subtotal (95% CI) Total events: 205 (statin), 246 (placebo) Test for heterogeneity: not applicable Test for overall effect: z=2.05, P=0.04 Transplant patients Holdaas 2003 Subtotal (95% CI) Total events: 46 (statin), 66 (placebo) Test for heterogeneity: not applicable Test for overall effect: z=1.92, P=0.05 Mixed population ( pre-dialysis and dialysis patients) Stegmayr 2005 Sub total (95% CI) Total events: 6 (statin), 10 (placebo) Test for heterogeneity: not applicable Test for overall effect: z=0.96, P=0.34 Total (95% CI) Total events: 1648 (statin), 2128 (placebo) Test for heterogeneity: X 2 =7.68, df=7, P=0.36, I 2 =8.9% Test for overall effect: z=6.72, P<0.001 0.01 0.37 (0.02 to 8.53) 0.72 (0.62 to 0.84) 0.80 (0.75 to 0.86) 0.79 (0.40 to 1.53) 0.48 (0.30 to 0.78) 0.75 (0.66 to 0.85) 0.86 (0.74 to 0.99) 0.70 (0.48 to 1.01) 0.63 (0.24 to 1.63) 0.78 (0.73 to 0.84) 2001 2003 2004 2005 2003 2005 0.05 18.12 54.92 1.15 2.18 76.43 19.33 3.69 0.56 100.00 0.1110100 Favors placebo Favors statin Study or subcategoryStatin n/N Placebo n/N Relative risk (random) (95% CI) Weight (%) Relative risk (random) (95% CI) Year Only studies with at least 1 event are included in the plot

32. www.ascotstudy.org