Supplementary appendix

This appendix formed part of the original submission and has been peer reviewed.
We post it as supplied by the authors.
Supplement to: Cholesterol Treatment Trialists (CTT) Collaborators. The effects of
lowering LDL cholesterol with statin therapy in people at low risk of vascular disease:
meta-analysis of individual data from 27 randomised trials. Lancet 2012; published
online May 17. DOI:10.1016/S0140-6736(12)60367-5.

Online webappendix
The effects of lowering LDL cholesterol with statin therapy in people at low-risk of vascular
disease: meta-analysis of individual data from 27 randomised trials
Webtables
Cox proportional hazard models predicting the risk of a first major vascular event in participants allocated
to control (model 1) or less statin (model 2)

Comparison of the observed (95% CI) and predicted rates of major vascular events in participating trials

Mean difference in plasma lipid concentrations at 1 year in participants at different levels of risk

Eligibility of participants without prior vascular disease in each CTT risk category for statin therapy
under current major guidelines

Webfigures
Effects on major vascular events per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk,
by baseline age and gender

Effects on major vascular events per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk,
by baseline LDL cholesterol (mmol/L)

Effects on nonfatal myocardial infarctions, CHD deaths and major coronary events per 1.0 mmol/L
reduction in LDL cholesterol at different levels of risk

Effects on stroke subtypes per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk

Effects on major coronary events, strokes, coronary revascularisation procedures and major vascular events
per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk in the 22 statin vs control trials

10

Effects on major coronary events, strokes, coronary revascularisation procedures and major vascular events
per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk in the 5 trials of more vs less statin

11

Effects on major coronary events per 1.0 mmol/L reduction in LDL cholesterol at different levels of
risk, by history of vascular disease

12

Effects on causespecific mortality per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk

13

Effects on any deaths per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk, by history of vascular
disease and overall
14

Statistical Appendix

15

Webtable 1: Cox proportional hazard models predicting the risk of a first major vascular event in
participants allocated to control (model 1) or less statin (model 2)

Parameter

Model 1
Statin vs. Control
Hazard ratio (95% CI)

A. Baseline characteristics
Male gender
Current smoker
Age (per 10 years)
Natural logarithm of HDL (per 1
lnmmol/L)
LDL (per 1 mmol/L)
Treatment for hypertension
Systolic BP (per 20mmHg)
Diastolic BP (per 10mmHg)
Creatinine (per 50mol/L) *
History of MI
History of other CHD, but no MI
History of stroke
History of PAD
Other/nonspecific vascular disease
history
History of diabetes mellitus

Model 2
More vs. less statin
Hazard ratio (95% CI)

1.56 (1.45 - 1.69)

1.43 (1.28 - 1.60)
1.41 (1.35 - 1.48)

1.07 (0.98 - 1.16)

1.27 (1.18 - 1.38)
1.12 (1.08 - 1.16)

0.69
1.14
1.23
1.18
0.95
1.18
2.50
1.83
1.35
1.24

0.76 (0.68 - 0.86)

1.19 (1.13 - 1.25)
1.15 (1.08 - 1.22)

(0.63 - 0.75)
(1.11 - 1.18)
(1.17 - 1.29)
(1.13 - 1.23)
(0.93 - 0.97)
(1.12 - 1.23)
(2.23 - 2.81)
(1.69 - 1.97)
(1.25 - 1.46)
(1.16 - 1.32)

1.22 (1.06 - 1.42)

1.53 (1.45 - 1.62)

B. Interaction terms #
Age and history of MI (per 10 years)
Age and history of other CHD (per
10 years)
Systolic BP and history of MI
Systolic BP and other CHD
Current smoker and male gender
Male gender and history of MI

0.77
0.91
0.89
0.80
0.78

(0.71 - 0.83)
(0.86 - 0.95)
(0.84 - 0.94)
(0.71 - 0.9)
(0.70 - 0.87)

C. Trial-specific terms (to model

average risk)
SSSS
WOSCOPS
CARE
Post-CABG
AFCAPS/TexCAPS
LIPID
GISSI-P
ASCOT-LLA
PROSPER
CARDS
ALERT
ALLHAT-LLT
LIPS
ALLIANCE
ASPEN
4D
MEGA
JUPITER
GISSI-HF
AURORA

1.95
0.91
1.25
0.88
0.55
1.06
0.74
0.65
2.00
0.67
0.87
1.19
1.11
1.51
0.81
1.93
0.32
0.48
0.47
2.78

(1.78 - 2.13)
(0.79 - 1.04)
(1.13 - 1.39)
(0.72 - 1.07)
(0.47 - 0.65)
(0.98 - 1.15)
(0.59 - 0.92)
(0.56 - 0.75)
(1.80 - 2.22)
(0.55 - 0.81)
(0.73 - 1.04)
(1.07 - 1.32)
(0.87 - 1.42)
(1.33 - 1.73)
(0.67 - 0.97)
(1.63 - 2.28)
(0.26 - 0.39)
(0.41 - 0.57)
(0.37 - 0.60)
(2.46 - 3.14)

1.19 (1.10 - 1.28)

1.19 (1.07 - 1.32)
1.48 (1.33 - 1.64)
1.38 (1.24 - 1.54)

1.40 (1.29 - 1.51)

0.74 (0.70 - 0.79)

CORONA
A to Z
PROVE-IT
TNT
IDEAL

0.95 (0.79 - 1.16)

0.59
1.76
1.37
1.21

(0.40 - 0.86)
(1.39 - 2.23)
(1.24 - 1.51)
(1.09 - 1.35)

D. Trial and period-specific terms

GISSI-P; months 0 to 6
4.27 (3.21 - 5.67)
LIPS; months 0 to 6
4.00 (2.82 - 5.66)
LIPS; months 7 to 12
2.96 (1.99 - 4.40)
GISSI-HF; months 0 to 6
1.75 (1.20 - 2.56)
A to Z; months 0 to 6
7.77 (5.03 - 12.02)
A to Z; months 7 to 12
2.44 (1.51 - 3.95)
PROVE-IT; months 0 to 6
3.78 (2.77 - 5.16)
PROVE-IT; months 7 to 12
2.02 (1.46 - 2.79)
TNT; months 0 to 6
1.41 (1.11 - 1.80)
IDEAL; months 0 to 6
3.12 (2.48 - 3.93)
IDEAL; months 7 to 12
1.39 (1.10 - 1.75)
HDL= high-density lipoprotein cholesterol. LDL= low-density lipoprotein cholesterol. BP=blood pressure.
MI=myocardial infarction. CHD=coronary heart disease. PAD=peripheral arterial disease
* missing creatinine values at randomisation in ASCOT were replaced with creatinine measured at screening; in
AFCAPS/TexCAPS, AURORA and 4D creatinine data were either not available (AFCAPS/TexCAPS) or not relevant (dialysis
patients in AURORA/4D) and so centred values of 0 were used
centred around study mean
defined as history of myocardial infarction or stroke for ALLHAT; history of heart failure in GISSI-HF and CORONA; carotid
artery disease, carotid stenosis 50%, carotid endarterectomy and abdominal aortic aneurysm in AURORA
# interpretation of the effects of the separate characteristics in these interactions should be based both on relevant main effects
(part A) and the interaction effects (part B)

Webtable 2: Comparison of the observed (95% CI) and predicted rates of major vascular events in
participating trials
Study

Duration
(years)*

Observed MVEs (%) (95% CI)

Average predicted
MVEs (%)

Statin vs. Control

SSSS
WOSCOPS
CARE
Post-CABG
AFCAPS/TexCAPS
LIPID
GISSI-P
HPS
ASCOT-LLA
PROSPER
CARDS
ALERT
ALLHAT-LLT
LIPS
ALLIANCE
ASPEN
4D
MEGA
JUPITER
GISSI-HF
AURORA
CORONA

5
5
5
5
5
5
2
5
4
4
5
5
5
4
5
4
5
5
5
5
5
3

33.8%
10.0%
27.3%
22.9%
5.4%
22.4%
11.1%
19.5%
6.9%
19.3%
10.9%
12.5%
16.1%
27.3%
28.0%
12.3%
39.9%
3.2%
5.3%
9.4%
33.9%
15.1%

(31.8% - 35.8%)
( 8.9% - 11.1%)
(25.3% - 29.3%)
(15.5% - 30.4%)
( 4.6% - 6.2%)
(21.2% - 23.7%)
( 9.7% - 12.4%)
(18.7% - 20.3%)
( 6.1% - 7.8%)
(17.6% - 21.0%)
( 8.9% - 12.9%)
(10.4% - 14.6%)
(15.0% - 17.2%)
(23.9% - 30.7%)
(25.1% - 30.9%)
(10.4% - 14.3%)
(33.6% - 46.2%)
( 2.7% - 3.8%)
( 4.2% - 6.5%)
( 7.9% - 10.8%)
(30.7% - 37.1%)
(13.5% - 16.7%)

33.7%
10.1%
27.2%
17.4%
5.7%
22.9%
10.9%
19.5%
7.3%
20.2%
11.4%
12.8%
15.8%
26.8%
27.6%
12.5%
39.4%
3.2%
4.9%
10.6%
34.7%
15.3%

More vs. Less statin

A to Z
PROVE-IT
TNT
IDEAL
SEARCH

2
2
5
5
5

13.3%
22.6%
23.4%
25.8%
17.1%

(11.8% - 14.8%)
(20.7% - 24.5%)
(22.2% - 24.6%)
(24.4% - 27.1%)
(16.2% - 18.1%)

13.3%
22.7%
23.4%
25.7%
17.2%

Risk categories
<5%
5
2.8% ( 2.4% - 3.2%)
3.4%
5%, <10%
5
7.4% ( 6.9% - 7.9%)
7.3%
10%, <20%
5
15.9% (15.5% - 16.4%)
15.4%
20%, <30%
5
24.7% (24.0% - 25.3%)
24.3%
30%
5
38.1% (37.0% - 39.2%)
38.1%
MVE= major vascular event.
*Duration over which rates of major vascular events compared: 5 years or the latest year with available Kaplan-Meier estimate of
MVE within 50 days from end of that year.
Estimated using Kaplan-Meier survival methods among participants allocated to the control or less statin arm, respectively.

Webtable 3: Mean difference in plasma lipid concentrations at 1 year in participants at different

levels of risk
Estimated 5-year risk of

Total cholesterol

LDL cholesterol

HDL cholesterol

Triglycerides

major vascular event

(mmol/L*)

(mmol/L*)

(mmol/L*)

(mmol/L*)

<5%

-0.94

-0.88

0.034

-0.19

5%, <10%

-1.08

-0.96

0.031

-0.25

10%, <20%

-1.14

-0.99

0.045

-0.27

20%, <30%

-1.26

-1.10

0.032

-0.24

30%

-1.31

-1.21

0.034

-0.23

-1.22

-1.08

0.038

-0.26

10%, <20%

-0.52

-0.44

0.006

-0.19

20%, <30%

-0.65

-0.53

-0.011

-0.24

30%

-0.70

-0.58

-0.013

-0.30

-0.61

-0.51

-0.005

-0.23

Statin vs. Control

Subtotal (22 trials)

More vs. Less statin

Subtotal (5 trials)

LDL= low-density lipoprotein cholesterol. HDL= high-density lipoprotein cholesterol.

* To convert values from mmol/L to mg/dL, divide triglycerides by 0.01129 and other lipids by 0.02586.
In trials where the LDL-C at 1 year was missing (or >10mmol/L) the baseline value was assigned. In some studies, only a subsample of participants were selected for 1 year blood samples: HPS (~4% of trial participants); SEARCH (~3% of participants). In
other studies, no samples at 1 year were taken and so other blood samples were used: In A to Z and PROVE-IT, blood samples
taken at 8 months were used; in ALLHAT samples taken at 2 years (from a random sample of 10% of participants randomised to
pravastatin and 5% of participants randomised to usual care) were used. In ALLIANCE, lipid differences at 1 year in the usual
care group were interpolated from those at baseline and final follow-up because the 1 year bloods were assayed in different
laboratories depending on treatment allocation.
Average values weighted by the trial and risk category-specific variances of the 'logrank' O minus E statistics for major vascular
events.
Includes 141 participants (48 [4 MVEs] from A to Z and 93 [11 MVEs] from SEARCH) with an estimated 5-year risk of MVE
between 5% and 10%

Webtable 4: Eligibility of participants without prior vascular disease in each CTT risk category for
statin therapy under current major guidelines
CTT risk category
(among participants without prior vascular disease)
Estimated 5Observed MCE
Observed vascular
year MVE risk
event rate
death rate
(% per annum)*
(% per annum)*

Broad eligibility under current

guidelines
ATP-III1
ESC task
NICE3
2
force

<5%
0.2
0.1

5%, <10%
0.8
0.3

10%, <20%
1.6
1.0

20%, <30%
3.2
2.3

30%
5.6
5.8

CTT= Cholesterol Treatment Collaboration. MCE= major coronary event, defined as non-fatal myocardial
infarction or coronary death. * among placebo-allocated participants
1.

The Adult Treatment Panel III (ATP III) of the National Cholesterol Education program in the US recommends
considering drug therapy for cholesterol modification for individuals without history of vascular disease but with
multiple risk factors that confer a 10-year risk for major coronary events (i.e. myocardial infarction and coronary death)
>10% which is considered moderately high or high risk. Observed annual major coronary event rate was less than 1% in
those allocated to control intervention in the two lowest risk categories in CTT and therefore these participants are
ineligible for statin intervention under ATP III.
Expert Panel on Detection Evaluation and Treatment of High Blood Cholesterol in Adults. Executive Summary of the Third Report of
the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood
Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001;285(19):2486-97.
Grundy SM, Cleeman JI, Merz CNB et al. Implications of Recent Clinical Trials for the National Cholesterol Education Program
Adult Treatment Panel III Guidelines. Circulation 2004;110(2):227-39.

2.

The Fourth Joint Task Force of the European Society of Cardiology (ESC) and Other Societies on Cardiovascular
Disease Prevention in Clinical Practice and the ESC/EAS Guidelines for the management of dyslipidaemias recommend
consideration of statin treatment typically for individuals without known vascular disease but at estimated 10-year risk of
a fatal atherosclerotic event (including heart attack, stroke, aneurysm of the aorta, or other) of at least 5%. Observed
annual vascular death rate was less than 0.5% in those participants without vascular disease allocated to control
intervention in the two lowest risk categories in CTT and therefore these participants are ineligible for statin intervention
under The Fourth Joint Task Force of the European Society of Cardiology and Other Societies on Cardiovascular Disease
Prevention in Clinical Practice.
Graham I, Atar D, Borch-Johnsen K et al. European guidelines on cardiovascular disease prevention in clinical practice: full text.
Fourth Joint Task Force of the European Society of Cardiology and other societies on cardiovascular disease prevention in clinical
practice (constituted by representatives of nine societies and by invited experts). European Journal of Cardiovascular Prevention &
Rehabilitation 2007;14:Suppl-113.
Reiner Z, Catapano AL, De Backer G, Graham I, Taskinen MR, Wiklund O et al. ESC/EAS Guidelines for the management of
dyslipidaemias: the Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European
Atherosclerosis Society (EAS). Eur Heart J. 2011;32(14):1769-818.

3.

The National Institute for Health and Clinical Excellence (NICE) in the National Health Service (NHS) in England and
Wales recommends lipid modification drug therapy for individuals without known cardiovascular disease (CVD) who
have 20% or greater 10-year risk of developing CVD (defined as myocardial infarction, CHD death, angina, stroke or
transient ischemia). As no total CVD endpoint was available in CTT data, an approximate estimation of risk was carried
out to allow interpretation. Total CVD was evaluated by multiplying observed vascular death rates within the risk
categories by a factor of 3 to 4 (as indicated by the SCORE data). The total estimated 10-year CVD risk among those
allocated to control intervention in the two lowest risk categories in CTT was less than 20% and therefore these
participants are ineligible for statin intervention under NICE guidelines.
Cooper A, Nherera L, Calvert N et al. Clinical Guidelines and Evidence Review for Lipid Modification: cardiovascular risk
assessment and the primary and secondary prevention of cardiovascular disease. London:National Collaborating Centre for Primary
Care and Royal College of General Practitioners; 2008.
Reiner Z, Catapano AL, De Backer G, Graham I, Taskinen MR, Wiklund O et al. ESC/EAS Guidelines for the management of
dyslipidaemias: the Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European
Atherosclerosis Society (EAS). Eur Heart J. 2011;32(14):1769-818.

Webfigure 1: Effects on major vascular events per 1.0 mmol/L reduction in LDL cholesterol at different
levels of risk, by baseline age and gender
5year MVE risk
at baseline

Events (% per annum)

Statin/more

Control/less

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

30%

110 (0.40)
295 (1.03)
1336 (2.81)
1410 (4.86)
780 (7.74)

145 (0.51)
433 (1.53)
1615 (3.55)
1691 (5.88)
963 (9.93)

0.69 (0.49 0.98)

0.65 (0.54 0.79)
0.74 (0.66 0.83)
0.81 (0.74 0.90)
0.81 (0.73 0.89)

Subtotal

3931 (2.75)

4847 (3.45)

0.77 (0.74 0.81)

p<0.0001

30%

52 (0.35)
212 (1.08)
1380 (2.84)
1637 (4.58)
1116 (7.20)

92 (0.64)
295 (1.52)
1622 (3.38)
1993 (5.70)
1439 (9.65)

0.56 (0.35 0.89)

0.70 (0.55 0.90)
0.78 (0.70 0.87)
0.81 (0.74 0.89)
0.78 (0.71 0.85)

Subtotal

4397 (3.28)

5441 (4.13)

0.78 (0.75 0.81)

p<0.0001

30%

5 (0.25)
97 (1.43)
898 (3.48)
1061 (4.83)
891 (8.19)

17 (0.81)
119 (1.82)
958 (3.66)
1235 (5.87)
1056 (9.96)

0.37 (0.13 1.08)

0.79 (0.56 1.10)
0.90 (0.79 1.04)
0.81 (0.72 0.91)
0.81 (0.71 0.91)

Subtotal

2952 (4.37)

3385 (5.09)

0.83 (0.78 0.87)

p<0.0001

Trend test

Age at baseline
60 years
< 5%
5%,<10%
10%,<20%
20%,<30%

21=7.37
(p=0.007)

>60, 70 years
< 5%
5%,<10%
10%,<20%
20%,<30%

21=0.96
(p=0.3)

>70 years
< 5%
5%,<10%
10%,<20%
20%,<30%

21=0.42
(p=0.5)

Test for trend in overall effects across age groups: 21= 3.29 (p=0.1)
Gender
Female
< 5%
5%,<10%
10%,<20%
20%,<30%
30%

78 (0.31)
196 (1.25)
956 (3.04)
680 (4.94)
429 (8.33)

119 (0.48)
232 (1.48)
1071 (3.36)
750 (5.68)
522 (10.41)

0.57 (0.38 0.86)

0.84 (0.64 1.10)
0.88 (0.77 1.00)
0.88 (0.76 1.02)
0.79 (0.67 0.94)

Subtotal

2339 (2.57)

2694 (2.98)

0.84 (0.79 0.89)

p<0.0001

30%

89 (0.46)
408 (1.04)
2658 (2.94)
3428 (4.70)
2358 (7.53)

135 (0.67)
615 (1.60)
3124 (3.55)
4169 (5.83)
2936 (9.72)

0.66 (0.46 0.95)

0.64 (0.55 0.75)
0.76 (0.70 0.83)
0.80 (0.75 0.85)
0.79 (0.74 0.84)

Subtotal

8941 (3.53)

10979 (4.42)

0.78 (0.76 0.80)

p<0.0001

Male
< 5%
5%,<10%
10%,<20%
20%,<30%

21=0.23
(p=0.6)

21=6.73
(p=0.009)

Difference in overall effect between men and women: 21= 5.23 (p=0.02)

0.5
99% or

95% limits

0.75

Statin/more
better

1.25

1.5

Control/less
better

Webfigure 2: Effects on major vascular events per 1.0 mmol/L reduction in LDL cholesterol at different
levels of risk, by baseline LDL cholesterol (mmol/L)
5year MVE risk
at baseline

Events (% per annum)

Statin/more

Control/less

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

30%

75 (0.37)
245 (1.00)
2356 (3.01)
2438 (4.86)
1462 (8.54)

132 (0.65)
369 (1.54)
2686 (3.50)
2874 (5.87)
1671 (10.03)

0.55 (0.39 0.79)

0.62 (0.50 0.76)
0.76 (0.69 0.85)
0.78 (0.71 0.86)
0.83 (0.73 0.93)

Subtotal

6576 (3.46)

7732 (4.14)

0.77 (0.73 0.80)

p<0.0001

30%

92 (0.38)
359 (1.18)
1258 (2.88)
1670 (4.57)
1325 (6.85)

122 (0.50)
478 (1.58)
1509 (3.50)
2045 (5.72)
1787 (9.64)

0.74 (0.48 1.13)

0.75 (0.63 0.90)
0.82 (0.75 0.90)
0.84 (0.78 0.90)
0.78 (0.73 0.83)

Subtotal

4704 (3.05)

5941 (3.90)

0.81 (0.78 0.83)

p<0.0001

<3.5
< 5%
5%,<10%
10%,<20%
20%,<30%

3.5
< 5%
5%,<10%
10%,<20%
20%,<30%

Trend test

21=11.52
(p=0.0007)

21=0.21
(p=0.6)

Heterogeneity between participants at different levels of LDL cholesterol:

21= 3.26 (p=0.07)
0.5
99% or

95% limits

0.75

Statin/more
better

1.25

1.5

Control/less
better

Trial and LDL subgroupspecific LDL weights used in these subgroup analyses.

Webfigure 3: Effects on nonfatal myocardial infarctions, CHD deaths and major coronary events per 1.0 mmol/L
reduction in LDL cholesterol at different levels of risk
Events (% per annum)

5year MVE risk

at baseline

Statin/more

Control/less

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

Nonfatal myocardial infarction

38 (0.08)
< 5%
228 (0.41)
5%,<10%
1114 (0.88)
10%,<20%
1241 (1.34)
20%,<30%
980 (2.48)
30%

79 (0.17)
366 (0.67)
1404 (1.12)
1618 (1.77)
1271 (3.27)

0.46 (0.28 0.76)

0.60 (0.49 0.74)
0.73 (0.65 0.83)
0.74 (0.67 0.82)
0.77 (0.71 0.85)

3601 (1.00)

4738 (1.33)

0.74 (0.71 0.77)

p<0.0001

30%

14 (0.03)
63 (0.11)
582 (0.45)
633 (0.66)
610 (1.47)

9 (0.02)
85 (0.15)
635 (0.49)
776 (0.81)
785 (1.89)

1.60 (0.57 4.47)

0.73 (0.47 1.11)
0.84 (0.70 1.00)
0.82 (0.71 0.95)
0.77 (0.69 0.87)

Overall

1902 (0.52)

2290 (0.62)

0.80 (0.76 0.85)

p<0.0001

30%

50 (0.11)
276 (0.50)
1644 (1.29)
1789 (1.93)
1471 (3.73)

88 (0.19)
435 (0.79)
1973 (1.57)
2282 (2.49)
1887 (4.86)

0.57 (0.36 0.89)

0.61 (0.50 0.74)
0.77 (0.69 0.85)
0.77 (0.71 0.83)
0.78 (0.72 0.84)

Overall

5230 (1.45)

6665 (1.87)

0.76 (0.73 0.79)

p<0.0001

Overall

CHD death
< 5%
5%,<10%
10%,<20%
20%,<30%

Trend test

21=9.04
(p=0.003)

21=1.13
(p=0.3)

Major coronary event

< 5%
5%,<10%
10%,<20%
20%,<30%

0.5
99% or

95% limits

0.75

Statin/more
better

1.25

21=5.66
(p=0.02)

1.5

Control/less
better

Webfigure 4: Effects on stroke subtypes per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk

Events (% per annum)

5year MVE risk

at baseline

Statin/more

Control/less

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

30%

45 (0.10)
101 (0.23)
479 (0.41)
498 (0.56)
373 (1.00)

60 (0.14)
151 (0.34)
549 (0.47)
608 (0.69)
466 (1.24)

0.69 (0.39 1.23)

0.66 (0.49 0.90)
0.84 (0.71 1.00)
0.80 (0.68 0.94)
0.81 (0.68 0.96)

Overall

1496 (0.45)

1834 (0.55)

0.79 (0.74 0.85)

p<0.0001

30%

20 (0.05)
25 (0.06)
82 (0.07)
91 (0.10)
57 (0.15)

18 (0.04)
23 (0.05)
74 (0.07)
71 (0.08)
44 (0.12)

1.19 (0.42 3.36)

1.01 (0.49 2.08)
1.04 (0.68 1.59)
1.26 (0.84 1.88)
1.26 (0.75 2.12)

Overall

275 (0.08)

230 (0.07)

1.15 (0.97 1.38)

p= 0.11

30%

6 (0.01)
64 (0.11)
236 (0.18)
192 (0.21)
141 (0.36)

12 (0.03)
66 (0.12)
284 (0.22)
221 (0.24)
151 (0.38)

0.53 (0.15 1.85)

0.99 (0.60 1.61)
0.84 (0.65 1.07)
0.87 (0.68 1.12)
0.92 (0.70 1.20)

Overall

639 (0.18)

734 (0.20)

0.87 (0.79 0.97)

p= 0.01

Ischaemic stroke
< 5%
5%,<10%
10%,<20%
20%,<30%

Trend test

21=0.86
(p=0.4)

Haemorrhagic stroke
< 5%
5%,<10%
10%,<20%
20%,<30%

Unknown stroke
< 5%
5%,<10%
10%,<20%
20%,<30%

0.5
99% or

95% limits

0.75

Statin/more
better

1.25

21=0.60
(p=0.4)

21=0.25
(p=0.6)

1.5

Control/less
better

Webfigure 5: Effects on major coronary events, strokes, coronary revascularisation procedures and major vascular
events per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk in the 22 statin vs control trials
Events (% per annum)

5year MVE risk

at baseline

Statin

Control

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

30%

50 (0.11)
271 (0.49)
949 (1.18)
1155 (1.93)
1080 (3.75)

88 (0.19)
432 (0.79)
1219 (1.54)
1526 (2.57)
1427 (5.12)

0.57 (0.36 0.89)

0.61 (0.50 0.74)
0.76 (0.69 0.85)
0.78 (0.71 0.85)
0.78 (0.72 0.84)

Overall

3505 (1.30)

4692 (1.76)

0.76 (0.73 0.79)

p<0.0001

30%

71 (0.16)
189 (0.34)
578 (0.72)
564 (0.95)
436 (1.54)

90 (0.20)
238 (0.43)
677 (0.85)
646 (1.08)
484 (1.73)

0.74 (0.46 1.19)

0.77 (0.60 0.98)
0.85 (0.74 0.98)
0.87 (0.76 1.00)
0.88 (0.76 1.01)

Overall

1838 (0.68)

2135 (0.80)

0.85 (0.81 0.90)

p<0.0001

Coronary revascularisation
73 (0.16)
< 5%
221 (0.40)
5%,<10%
879 (1.10)
10%,<20%
1289 (2.21)
20%,<30%
757 (2.67)
30%

135 (0.30)
340 (0.62)
1090 (1.39)
1602 (2.77)
1002 (3.65)

0.52 (0.35 0.75)

0.63 (0.51 0.79)
0.77 (0.68 0.86)
0.82 (0.75 0.89)
0.77 (0.70 0.85)

3219 (1.21)

4169 (1.58)

0.77 (0.74 0.80)

p<0.0001

30%

167 (0.38)
596 (1.09)
2133 (2.74)
2607 (4.63)
1940 (7.23)

254 (0.56)
840 (1.56)
2566 (3.37)
3175 (5.73)
2422 (9.48)

0.62 (0.47 0.81)

0.69 (0.60 0.79)
0.80 (0.74 0.86)
0.83 (0.78 0.88)
0.80 (0.75 0.85)

Overall

7443 (2.86)

9257 (3.62)

0.80 (0.78 0.82)

p<0.0001

Trend test

Major coronary event

< 5%
5%,<10%
10%,<20%
20%,<30%

21=5.86
(p=0.02)

Any stroke
< 5%
5%,<10%
10%,<20%
20%,<30%

Overall

21=1.63
(p=0.2)

21=6.02
(p=0.01)

Major vascular event

< 5%
5%,<10%
10%,<20%
20%,<30%

0.5
99% or

95% limits

0.75
Statin
better

1.25

21=5.29
(p=0.02)

1.5

Control
better

10

Webfigure 6: Effects on major coronary events, strokes, coronary revascularisation procedures and major vascular
events per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk in the 5 trials of more vs less statin
Events (% per annum)

5year MVE risk,

less statin

More statin

Less statin

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

30%

700 (1.49)
634 (1.92)
391 (3.67)

757 (1.61)
756 (2.35)
460 (4.19)

0.79 (0.57 1.10)

0.68 (0.52 0.89)
0.80 (0.59 1.09)

Overall

1725 (1.90)

1973 (2.19)

0.74 (0.65 0.85)

p<0.0001

30%

220 (0.46)
217 (0.65)
135 (1.23)

232 (0.49)
254 (0.77)
177 (1.56)

0.90 (0.51 1.59)

0.69 (0.44 1.09)
0.70 (0.42 1.18)

Overall

572 (0.62)

663 (0.72)

0.74 (0.59 0.92)

p= 0.007

Coronary revascularisation
830 (1.82)
10%,<20%*
917 (2.92)
20%,<30%
503 (4.98)
30%

973 (2.15)
1115 (3.68)
653 (6.40)

0.66 (0.51 0.86)

0.66 (0.54 0.81)
0.65 (0.51 0.84)

2250 (2.58)

2741 (3.20)

0.66 (0.60 0.73)

p<0.0001

30%

1489 (3.35)
1501 (4.93)
847 (8.79)

1636 (3.71)
1744 (5.95)
1036 (10.74)

0.75 (0.61 0.92)

0.70 (0.59 0.83)
0.72 (0.59 0.88)

Overall

3837 (4.54)

4416 (5.32)

0.72 (0.66 0.78)

p<0.0001

Major coronary event

10%,<20%*
20%,<30%

Any stroke
10%,<20%*
20%,<30%

Overall

Major vascular event

10%,<20%*
20%,<30%

0.5
99% or

0.75

1.25

Trend test

21=0.01
(p=0.9)

21=0.63
(p=0.4)

21=0.01
(p=0.9)

21=0.12
(p=0.7)

1.5

95% limits

More statin
Less statin
better
better
*Includes 141 participants (48 from A to Z and 93 from SEARCH) with an estimated 5year risk of MVE less than 10%.

11

Webfigure 7: Effects on major coronary events per 1.0 mmol/L reduction in LDL cholesterol at different levels of
risk, by history of vascular disease
5year MVE risk
at baseline

Events (% per annum)

Control/less

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

Participants without vascular disease

47 (0.11)
80 (0.19)
< 5%
238 (0.50)
394 (0.83)
5%,<10%
447 (1.29)
557 (1.61)
10%,<20%
173 (2.51)
217 (3.15)
20%,<30%
79 (4.52)
91 (5.64)
30%

0.59 (0.37 0.96)

0.58 (0.48 0.72)
0.78 (0.65 0.93)
0.80 (0.60 1.06)
0.76 (0.50 1.17)

Statin/more

984 (0.73)

1339 (1.00)

0.71 (0.65 0.77)

p<0.0001

Participants with vascular disease

3 (0.14)
< 5%
38 (0.50)
5%,<10%
1197 (1.30)
10%,<20%
1616 (1.88)
20%,<30%
1392 (3.69)
30%

8 (0.37)
41 (0.55)
1416 (1.56)
2065 (2.44)
1796 (4.82)

0.41 (0.09 1.76)

0.86 (0.49 1.50)
0.76 (0.67 0.86)
0.77 (0.70 0.83)
0.78 (0.72 0.84)

4246 (1.88)

5326 (2.39)

0.77 (0.74 0.80)

p<0.0001

30%

50 (0.11)
276 (0.50)
1644 (1.29)
1789 (1.93)
1471 (3.73)

88 (0.19)
435 (0.79)
1973 (1.57)
2282 (2.49)
1887 (4.86)

0.57 (0.36 0.89)

0.61 (0.50 0.74)
0.77 (0.69 0.85)
0.77 (0.71 0.83)
0.78 (0.72 0.84)

Overall

5230 (1.45)

6665 (1.87)

0.76 (0.73 0.79)

p<0.0001

Subtotal

Subtotal

Trend test

21=6.08
(p=0.01)

21=0.23
(p=0.6)

All participants
< 5%
5%,<10%
10%,<20%
20%,<30%

21=5.66
(p=0.02)

Heterogeneity between participants without and with vascular disease:

21= 2.80 (p=0.09)
0.5
99% or

95% limits

0.75

Statin/more
better

1.25

1.5

Control/less
better

12

Webfigure 8: Effects on causespecific mortality per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk

Events (% per annum)

5year MVE risk

at baseline

Statin/more

Control/less

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

30%

14 (0.03)
63 (0.11)
582 (0.45)
633 (0.66)
610 (1.47)

9 (0.02)
85 (0.15)
635 (0.49)
776 (0.81)
785 (1.89)

1.60 (0.57 4.47)

0.73 (0.47 1.11)
0.84 (0.70 1.00)
0.82 (0.71 0.95)
0.77 (0.69 0.87)

Overall

1902 (0.52)

2290 (0.62)

0.80 (0.76 0.85)

p<0.0001

30%

51 (0.11)
172 (0.31)
578 (0.45)
688 (0.72)
466 (1.12)

64 (0.14)
185 (0.33)
623 (0.48)
763 (0.80)
486 (1.17)

0.79 (0.47 1.32)

0.90 (0.68 1.20)
0.91 (0.78 1.05)
0.91 (0.80 1.03)
0.98 (0.83 1.14)

Overall

1955 (0.53)

2121 (0.58)

0.92 (0.87 0.98)

p= 0.006

30%

10 (0.02)
42 (0.07)
170 (0.13)
152 (0.16)
147 (0.35)

13 (0.03)
37 (0.07)
180 (0.14)
165 (0.17)
138 (0.33)

0.77 (0.25 2.35)

1.12 (0.62 2.03)
0.90 (0.67 1.22)
0.94 (0.71 1.25)
1.08 (0.81 1.45)

Overall

521 (0.14)

533 (0.15)

0.98 (0.86 1.10)

p= 0.69

30%

4 (0.01)
33 (0.06)
143 (0.11)
123 (0.13)
117 (0.28)

6 (0.01)
23 (0.04)
153 (0.12)
129 (0.14)
124 (0.30)

0.64 (0.12 3.45)

1.48 (0.73 3.00)
0.88 (0.63 1.22)
0.94 (0.68 1.30)
0.96 (0.69 1.32)

Overall

420 (0.11)

435 (0.12)

0.95 (0.83 1.09)

p= 0.48

30%

79 (0.18)
310 (0.55)
1473 (1.14)
1596 (1.67)
1340 (3.23)

92 (0.20)
330 (0.59)
1591 (1.23)
1833 (1.92)
1533 (3.69)

0.87 (0.58 1.31)

0.92 (0.74 1.13)
0.88 (0.79 0.97)
0.88 (0.81 0.96)
0.87 (0.80 0.95)

Overall

4798 (1.30)

5379 (1.47)

0.88 (0.84 0.91)

p<0.0001

30%

116 (0.26)
270 (0.48)
1054 (0.81)
963 (1.01)
681 (1.64)

101 (0.22)
309 (0.55)
1104 (0.86)
941 (0.99)
705 (1.70)

1.16 (0.80 1.68)

0.88 (0.71 1.09)
0.94 (0.83 1.07)
1.00 (0.89 1.13)
0.96 (0.83 1.10)

Overall

3084 (0.84)

3160 (0.86)

0.96 (0.92 1.01)

p= 0.16

30%

232 (0.52)
639 (1.14)
2651 (2.04)
2683 (2.80)
2165 (5.22)

244 (0.54)
710 (1.27)
2827 (2.19)
2903 (3.04)
2403 (5.78)

0.97 (0.76 1.24)

0.89 (0.77 1.03)
0.91 (0.84 0.98)
0.92 (0.86 0.99)
0.89 (0.83 0.96)

Overall

8370 (2.27)

9087 (2.47)

0.91 (0.88 0.93)

p<0.0001

CHD death
< 5%
5%,<10%
10%,<20%
20%,<30%

Other cardiac death

< 5%
5%,<10%
10%,<20%
20%,<30%

Stroke death
< 5%
5%,<10%
10%,<20%
20%,<30%

Other vascular death

< 5%
5%,<10%
10%,<20%
20%,<30%

Trend test

21=1.13
(p=0.3)

21=1.25
(p=0.3)

21=0.54
(p=0.5)

21=0.08
(p=0.8)

Any vascular death

< 5%
5%,<10%
10%,<20%
20%,<30%

Nonvascular death
< 5%
5%,<10%
10%,<20%
20%,<30%

21=0.18
(p=0.7)

21=0.02
(p=0.9)

Any death*
< 5%
5%,<10%
10%,<20%
20%,<30%

0.5
99% or

95% limits

0.75

1.25

21=0.22
(p=0.6)

1.5

Statin/more
Control/less
better
better
*Includes 488 (statin/more statin) vs 548 (control/less statin) deaths of unknown cause

13

Webfigure 9: Effects on any deaths per 1.0 mmol/L reduction in LDL cholesterol at different levels of risk,
by history of vascular disease and overall
5year MVE risk
at baseline

Deaths (% per annum)

Control/less

RR (CI) per 1.0 mmol/L reduction in LDL cholesterol

Participants without vascular disease

164 (0.38)
177 (0.41)
< 5%
372 (0.77)
446 (0.93)
5%,<10%
703 (1.99)
778 (2.19)
10%,<20%
363 (5.13)
339 (4.73)
20%,<30%
192 (10.76)
192 (11.44)
30%

0.94 (0.71 1.26)

0.83 (0.69 0.99)
0.88 (0.76 1.02)
1.06 (0.86 1.32)
0.94 (0.70 1.25)

Statin/more

1794 (1.33)

1932 (1.42)

0.91 (0.85 0.97)

p= 0.007

Participants with vascular disease

68 (3.06)
< 5%
267 (3.48)
5%,<10%
1948 (2.07)
10%,<20%
2320 (2.62)
20%,<30%
1973 (4.97)
30%

67 (3.10)
264 (3.50)
2049 (2.19)
2564 (2.91)
2211 (5.54)

1.04 (0.65 1.68)

1.00 (0.80 1.26)
0.92 (0.84 1.00)
0.90 (0.84 0.97)
0.89 (0.83 0.96)

6576 (2.83)

7155 (3.09)

0.90 (0.87 0.93)

p<0.0001

30%

232 (0.52)
639 (1.14)
2651 (2.04)
2683 (2.80)
2165 (5.22)

244 (0.54)
710 (1.27)
2827 (2.19)
2903 (3.04)
2403 (5.78)

0.97 (0.76 1.24)

0.89 (0.77 1.03)
0.91 (0.84 0.98)
0.92 (0.86 0.99)
0.89 (0.83 0.96)

Overall

8370 (2.27)

9087 (2.47)

0.91 (0.88 0.93)

p<0.0001

Subtotal

Subtotal

Trend test

21=1.57
(p=0.2)

21=1.82
(p=0.2)

All participants
< 5%
5%,<10%
10%,<20%
20%,<30%

21=0.22
(p=0.6)

Heterogeneity between participants without and with vascular disease:

21= 0.03 (p=0.9)
0.5
99% or

95% limits

0.75

Statin/more
better

1.25

1.5

Control/less
better

179 (statin/more statin) vs 210 (control/less statin) deaths of unknown cause are included among participants without vascular disease.
309 (statin/more statin) vs 338 (control/less statin) deaths of unknown cause are included among participants with vascular disease.

14


Statisticalappendix
Estimatingthefiveyearriskofmajorvasculareventamongthe174,149participantsin27randomised
trialsofstatintherapy

The5yearriskofamajorvascularevent(firstnonfatalmyocardialinfarction,coronarydeath,strokeor
coronaryrevascularisationprocedure)wasestimatedusingseparateCoxproportionalhazardsmodelsfor
the67,000patientsallocatedthecontrolregimeninthe22trialsofstatinversuscontrol(model1)andthe
20,000patientsallocatedthelessintensivestatinregimeninthe5trialsofmoreversuslessstatin(model
2).Theresultsfromthesetworegressionmodelswerethenappliedtoallpatients(includingthoseinthe
activetreatmentarms),asdescribedbelow.

Forpatientiinstudyjwithallocatedtreatmentk(wherek=0correspondstothecontrol/lessstatin
treatmentandk=1correspondstothestatin/morestatintreatment),thehazardfunctioninthe
control/lessstatingroupwasmodelledbytheregressionequation:





whereh0(t)isthebaselinehazardfunction,isanoverallinterceptterm,jrepresentstheeffectofstudyj
relativetotheHeartProtectionStudyformodel1ortheStudyoftheEffectivenessofAdditional
ReductionsinCholesterolandHomocysteineformodel2(seeWebtable1,termsC),representsavector
ofloghazardratioscorrespondingtothepatientssetofbaselinecharacteristicsxij0(centredaroundstudy
whereappropriate:seeWebtable1,termsA),representsavectorofloghazardratios
means
correspondingtointeractionswij0betweenvariousbaselinecharacteristics(seeWebtable1,termsB),and
representsavectorofloghazardratioscorrespondingtotrialspecifictimedependenteffectszj(t)
(definedforinitialsixmonthlytimeperiods:seeWebtable1,termsD).

Foreachofthetworegressionmodels,thebaselinecharacteristicsxijandinteractionswijwereselected
usingbackwardelimination,withfactorsremaininginthemodeliftheywerestatisticallysignificantatthe
1%level(ageandsexweretobeincludedinbothmodelsirrespectiveofstatisticalsignificance).The
baselinecharacteristicsincludedinthefinalmodelsareshowninWebtable1.Thetrialspecifictime
dependenteffectszj(t)weredefinedforinitialsixmonthlytimeperiodsandabackwardselimination
strategywithstatisticalsignificanceat1%wasemployedtoselecttheeffectsremaininginthemodels.

TheCoxmodelsprovideestimatesofloghazardratios,butprovidenodirectestimateofthebaseline
hazard .However,anestimateofthecumulativehazardfunction canberecoveredby
estimationofbaselinehazardcontributionsatfailuretimesusingtheKalbfleischandPrenticemethodand,
fromthat,anestimateofthebaselinecumulativesurvivalS0(t)=exp(H0(t))canbemade.




15

Separatingstudyparticipantsaccordingtobaseline5yearmajorvasculareventrisk

Thepredicted5yearriskofamajorvasculareventforallpatientswasestimatedby:

 att=5years

Patientswithmissingbaselinecharacteristicsemployedintheriskmodelswereexcludedfromthe
estimationofmodels1and2,buttheirvalueswereimputedforthepurposeofpredicting5yearriskofa
majorvascularevent.Occasionalmissingage,gender,treatmentforhypertensionwereimputedusing
studyspecificmean(age)ormedian(gender,treatmentforhypertension).MissingdataforLDLC(1.7%),
HDLC(0.7%),bloodpressure(0.4%)andcreatinine(1.4%)wereimputedusingstudyspecificmeanvalues
byage,genderandtreatmentforhypertension.

Trialparticipantswerecategorisedintooneoffivebaselinecategoriesof5yearrisk:<5%;5to<10%;10to
<20%;20to<30%;and30%orlarger.Theproportionateandabsoluteeffectsofallocationtostatinor
morestatininterventiononspecificendpointswasthenestimatedseparatelywithineachofthese
subgroups(asdescribedinthemainstatisticalmethodssection).



16