Doi:10.1016/s0140-6736(05)67394-

Articles
Efficacy and safety of cholesterol-lowering treatment:
prospective meta-analysis of data from 90 056 participants
in 14 randomised trials of statins
Cholesterol Treatment Trialists’ (CTT) Collaborators* Summary
Background Results of previous randomised trials have shown that interventions that lower LDL cholesterol
Lancet 2005; 366: 1267–78
concentrations can significantly reduce the incidence of coronary heart disease (CHD) and other major vascular Published online
events in a wide range of individuals. But each separate trial has limited power to assess particular outcomes or
particular categories of participant.
Methods A prospective meta-analysis of data from 90 056 individuals in 14 randomised trials of statins was done.
Weighted estimates were obtained of effects on different clinical outcomes per 1·0 mmol/L reduction in LDL Epidemiological Studies Unit
cholesterol.
Findings During a mean of 5 years, there were 8186 deaths, 14 348 individuals had major vascular events, and 5103 ctt@ctsu.ox.ac.uk
developed cancer. Mean LDL cholesterol differences at 1 year ranged from 0·35 mmol/L to 1·77 mmol/L (mean
1·09) in these trials. There was a 12% proportional reduction in all-cause mortality per mmol/L reduction in LDL
cholesterol (rate ratio [RR] 0·88, 95% CI 0·84–0·91; pϽ0·0001). This reflected a 19% reduction in coronary mortality Research Council Clinical Trials
(0·81, 0·76–0·85; pϽ0·0001), and non-significant reductions in non-coronary vascular mortality (0·93, 0·83–1·03; Centre (CTC), Mallett Street
p=0·2) and non-vascular mortality (0·95, 0·90–1·01; p=0·1). There were corresponding reductions in myocardial
infarction or coronary death (0·77, 0·74–0·80; pϽ0·0001), in the need for coronary revascularisation (0·76, Australia
0·73–0·80; pϽ0·0001), in fatal or non-fatal stroke (0·83, 0·78–0·88; pϽ0·0001), and, combining these, of 21% in ctt@ctc.usyd.edu.au
any such major vascular event (0·79, 0·77–0·81; pϽ0·0001). The proportional reduction in major vascular events *Collaborators listed at end of
differed significantly (pϽ0·0001) according to the absolute reduction in LDL cholesterol achieved, but not otherwise. paper
These benefits were significant within the first year, but were greater in subsequent years. Taking all years together,
the overall reduction of about one fifth per mmol/L LDL cholesterol reduction translated into 48 (95% CI 39–57)
fewer participants having major vascular events per 1000 among those with pre-existing CHD at baseline, compared
with 25 (19–31) per 1000 among participants with no such history. There was no evidence that statins increased the
incidence of cancer overall (1·00, 0·95–1·06; p=0·9) or at any particular site.
Interpretation Statin therapy can safely reduce the 5-year incidence of major coronary events, coronary
revascularisation, and stroke by about one fifth per mmol/L reduction in LDL cholesterol, largely irrespective of the
initial lipid profile or other presenting characteristics. The absolute benefit relates chiefly to an individual’s absolute
risk of such events and to the absolute reduction in LDL cholesterol achieved. These findings reinforce the need to
consider prolonged statin treatment with substantial LDL cholesterol reductions in all patients at high risk of any
type of major vascular event.
Introduction
inhibitors (statins).10 But discussion among the principal Results of observational studies in different populations investigators of ongoing large-scale randomised trials of indicate a continuous positive relationship between these treatments suggested that some uncertainties coronary heart disease (CHD) risk and blood cholesterol about their effects were likely to persist unless there was concentrations that extends well below the range seen in a systematic meta-analysis of the findings; although the many developed populations, without any definite individual trials might be large enough to show effects threshold below which a lower cholesterol concentration on the aggregate of all coronary events, they might well is not associated with lower risk.1,2 Despite this evidence, over estimate or under estimate any effects on coronary there has been substantial uncertainty about the effects death or on other specific vascular or non-vascular on mortality and major morbidity of lowering blood outcomes, especially when particular subgroups of participants were considered. Hence, in 1994, the Definitive assessment of whether a substantial decision was made to undertake periodic meta-analyses reduction in LDL cholesterol concentrations would be of individual participant data on mortality and morbidity beneficial was facilitated by the development of potent from all relevant large-scale randomised trials of lipid- cholesterol-lowering drugs, such as the 3-hydroxy-3- modifying treatments whose first results would be methylglutaryl coenzyme A (HMG-CoA) reductase reported subsequently. This report is of the results from www.thelancet.com Vol 366 October 8, 2005
Articles
Treatment
Age range
Diabetes
Baseline history of vascular disease (%)
recruitment
publication duration
comparison
of primary
of follow- (mg/day)†
patients
up (years)*
21 575 (24%) 18 686 (21%) 28 725 (32%) 13 406 (15%) 4S=Scandinavian Simvastatin Survival Study.13 WOSCOPS=West of Scotland Coronary Prevention Study.14 CARE=Cholesterol And Recurrent Events.15 Post-CABG=Post-Coronary Artery Bypass Graft.16 AFCAPS/TexCAPS=AirForce/Texas Coronary Atherosclerosis Prevention Study.17 LIPID=Long-term Intervention with Pravastatin in Ischaemic Disease.18 GISSI Prevention=Gruppo Italiano per lo Studio della Sopravvivenza nell’Infarto Miocardico.19LIPS=Lescol Intervention Prevention Study.20 HPS=Heart Protection Study.21 PROSPER=PROspective Study of Pravastatin in the Elderly at Risk.22 ALLHAT-LLT=Antihypertensive and Lipid-Lowering Treatment to Prevent HeartAttack Trial.23 ASCOT-LLA=Anglo-Scandinavian Cardiac Outcomes Trial-Lipid Lowering Arm.24 ALERT=Assessment of Lescol in Renal Transplantation.25 CARDS=Collaborative Atorvastatin Diabetes Study.26 S=simvastatin.
L=lovastatin. P=pravastatin. F=fluvastatin. A=atorvastatin. *Mean duration of follow-up based on survival times within each trial. Overall mean is weighted by trial-specific variances of logrank (o–e) for major vascular events.
†All trials included dietary intervention: WOSCOPS, GISSI Prevention, LIPID, 4S, LIPS, and HPS provided dietary advice; AFCAPS/TexCAPS, Post-CABG, CARE, ALLHAT, ALERT, CARDS, and PROSPER used American HeartAssociation (AHA) Step 1 diet and CARE intensified to Step 2 diet if LDL cholesterol у4·5 mmol/L (175 mg/dL). ‡Other CHD includes patients with a history of other symptomatic CHD but excludes those with a history of MI (asalready counted in MI column). §Other vascular includes history of intracerebral bleed, transient ischaemic attack, ischaemic stroke, unknown stroke, and peripheral artery disease. ¶None includes individuals without a history ofMI, symptomatic CHD, intracerebral bleed, transient ischaemic attack, ischaemic stroke, unknown stroke, or peripheral artery disease.
Table: Baseline characteristics and eligibility criteria of participating trials
the first cycle of such meta-analyses, and involves only (defined as non-fatal myocardial infarction [MI] or CHD death) in particular prespecified subgroups, and ofeffects on stroke, cancer, and vascular procedures. In addition, we have analysed the effects on major vascular Study eligibility
events (defined as the combined outcome of major A protocol for the Cholesterol Treatment Trialists’ coronary event, non-fatal or fatal stroke, or coronary (CTT) Collaboration was agreed in November, 1994, revascularisation) in different circumstances.
before the results of any of the relevant trials becameavailable, and was published the next year.11 Properly Statistical analysis
randomised trials were eligible for inclusion if: (i) the For every trial, the logrank Observed-minus-Expected main effect of at least one of the trial interventions was statistic (o–e) and its variance (v) were calculated from to modify lipid levels; (ii) the trial was unconfounded the results during every year of follow-up.12 For an with respect to this intervention (ie, no other differences unweighted meta-analysis, these (o–e) values, one from in risk factor modification between the relevant every trial, would be summed to produce a grand treatment groups were intended); and (iii) the trial total (G), with variance (V) equal to the sum of their aimed to recruit at least 1000 participants with separate variances. The value exp (G/V) would then be the overall event rate ratio (with ␹2 for heterogeneity between the effects in different trials equal to S–G2/V, Prespecified analyses of major outcomes
where S is the sum of [o–e]2/v for each trial). For the The principal planned analyses are described in the main LDL-weighted meta-analyses, let the mean published protocol.11 Briefly, the primary meta-analyses absolute difference in LDL cholesterol (mmol/L) after were to be of the effects on clinical outcome in each trial 1 year between those allocated active treatment and weighted by the absolute LDL cholesterol difference in those allocated control in a particular trial be w. The that trial at the end of the first year of follow-up, and are logrank (o–e) for that trial is then multiplied by the reported as the effects per 1·0 mmol/L (39 mg/dL)
weight w, and its variance by w2, and these weighted reduction in LDL cholesterol. The main prespecified values for every trial are then summed to produce a outcomes were all-cause mortality, CHD mortality, and weighted grand total (G ) and its variance (V ). The non-CHD mortality. Secondary analyses were to be of value exp(G /V ) is then the one-step weighted estimate effects on CHD death and on major coronary events of the event rate ratio (RR) per 1·0 mmol/L reduction in
www.thelancet.com Vol 366 October 8, 2005
Articles
LDL cholesterol (with ␹2 for heterogeneity between the Cause of death
Events (%)
effects per mmol in different trials equal to S–G 2/V ).
Treatment
In the figures and in the text, summary rate ratios are presented with 95% CI, whilst those derived from Vascular causes:
secondary or subgroup analyses are 99% CI. For subgroup analyses, the weighted results were 1548 (3·4%) 1960 (4·4%)
0·81 (0·76 –0·85)
calculated separately in every subgroup and were thencompared with standard ␹2 tests for heterogeneity or, where appropriate, for trend. Where many subgroup analyses were to be done (eg, of sex, age, initial blood Any non-CHD vascular
554 (1·2%)
593 (1·3%)
0·93 (0·83 –1·03)
pressure, etc), the separate ␹2 statistics for each weresummed (as were their degrees of freedom) to yield aglobal test for heterogeneity that can help make Any vascular
2102 (4·7%) 2553 (5·7%)
0·83 (0·79 –0·87)
allowance for the multiplicity of comparisons.12 Non-vascular causes:
Role of the funding sources
The funding sources had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The writing committee had full access to all the data in the study and had final responsibility forthe decision to submit for publication.
Any non-vascular
1730 (3·8%) 1801 (4·0%)
0·95 (0·90 –1·01)
Any death
3832 (8·5%) 4354 (9·7%)
0·88 (0·84 –0·91)
For the first cycle of analyses, individual participant datawere available from 14 trials of statin therapy (table),13–26but not from one other eligible trial.27 Data were obtained on 90 056 participants, of whom 42 131 (47%) had pre-existing CHD, 21 575 (24%) were women, 18 686 (21%) had a history of diabetes, and 49 689 (55%) had a history of hypertension. The mean pre-treatment Figure 1: Proportional effects on cause-specific mortality per mmol/L LDL cholesterol reduction
LDL cholesterol was 3·79 mmol/L, and ranged from Diamonds=totals and subtotals (95%CI). Squares=individual categories (horizontal lines are 99% CIs). Area ofsquare proportional to amount of statistical information in that category. RRs are weighted to represent reduction 3·03 mmol/L in CARDS26 to 4·96 mmol/L in in rate per 1 mmol/L LDL cholesterol reduction achieved by treatment at 1 year after randomisation. 26 active WOSCOPS14 (webtable 128). Overall in these trials, the versus 31 control deaths in the Post-CABG trial could not be subclassified into vascular and non-vascular causes, weighted average difference in LDL cholesterol at 1 year but were known not to be due to CHD and were assigned to other non-vascular deaths. was 1·09 mmol/L. The weighted mean duration of
follow-up among survivors was 4·7 years, and ranged
0·81, See Lancet Online
from 2 years in the GISSI Prevention trial19 to 6 years in 95% CI 0·76–0·85; pϽ0·0001) per mmol/L reduction for webtables 1 and 2 the LIPID trial.18 Details of the design of individual trials in LDL cholesterol. There were also non-significant are shown in the table and in webtable 1.
reductions in deaths from stroke (RR 0·91,99% CI 0·74–1·11; p=0·2), from other vascular causes Cause-specific mortality
(RR 0·95, 99% CI 0·78–1·16; p=0·5), and from non- There were a total of 8186 deaths, including 4655 (57%) vascular causes (RR 0·95, 95% CI 0·90–1·01; p=0·1).
from vascular causes and 3531 (43%) from non-vascular Among the non-vascular causes of death, there was no causes (webtable 2). During the scheduled treatment evidence that lowering LDL cholesterol with a statin period, there were 3832 (8·5%) deaths among the adversely affected the risk of death from cancer, 45 054 participants allocated a statin compared with respiratory disease, trauma, or other or unknown 4354 (9·7%) among the 45 002 controls. This difference represents a 12% proportional reduction in all-cause The 19% proportional reduction in CHD death per mortality per mmol/L LDL cholesterol reduction mmol/L LDL cholesterol reduction translated into (RR 0·88, 95% CI 0·84–0·91; pϽ0·0001; figure 1). In 14 (95% CI 9–19) fewer deaths per 1000 among an unweighted analysis, a slightly larger mortality participants with pre-existing CHD, compared with reduction of 13% (RR 0·87, 0·84–0·91; pϽ0·0001) was 4 (1–7) fewer per 1000 among participants who did not See Lancet Online
found, chiefly because the mean LDL cholesterol have pre-existing CHD (see webfigure 1i). The for webfigures 1 and 2 reduction at 1 year in these trials was 1·09 mmol/L.
proportional reduction in the risk of CHD death per The weighted average relative reduction of 12% in all- mmol/L lower LDL cholesterol was similar in all of the cause mortality was attributable mainly to the 19% prespecified subgroups examined (global heterogeneity proportional reduction in CHD deaths (1548 [3·4%] www.thelancet.com Vol 366 October 8, 2005
Articles
all years together, the overall incidence of major Endpoint
Events (%)
coronary events was reduced by about one quarter per Treatment
mmol/L reduction in LDL cholesterol amongparticipants with a previous history of MI or other CHD, as well as among those without any pre-existing CHD (figure 5). But since the absolute risk of eventswas higher among participants with pre-existing CHD, Any major coronary event
3337 (7·4%)
4420 (9·8%)
0·77 (0·74–0·80)
this reduction of about a quarter per mmol/L LDL cholesterol reduction translated into 30 (95% CI 24–37) fewer such participants having major coronary events per 1000 during an average of 5 years, compared with18 (14–23) fewer among participants who did not have Any coronary revascularisation
2620 (5·8%)
3434 (7·6%)
0·76 (0·73–0·80)
pre-existing CHD (figure 6 and webfigure 1ii). The proportional reduction in the incidence of major coronary events per mmol/L LDL cholesterol reduction Any stroke
1340 (3·0%)
1617 (3·7%)
0·83 (0·78–0·88)
was also about one quarter in all of the otherprespecified subgroups (global test for heterogeneity Any major vascular event
6354 (14·1%) 7994 (17·8%)
0·79 (0·77–0·81)
p=0·4; figure 5). Indeed, separately significantreductions in major coronary events were noted withinall of these subgroups of baseline characteristics, including: those aged older than 65 years; women; those treated for hypertension; those with a diastolic blood pressure above 90 mm Hg; and those with a Figure 2: Proportional effects on major vascular events per mmol/L LDL cholesterol reduction
history of diabetes (all pϽ0·0001). In addition to the Symbols and conventions as in figure 1. Broken vertical line indicates overall RR for any type of major vascular prespecified subgroups in figure 5, there were event. CABG=coronary artery bypass graft. PTCA=percutaneous transluminal coronary angioplasty. LIPS only significant reductions in major coronary events in other provided data on fatal strokes20 and so does not contribute to the stroke analyses.
with pretreatment LDL cholesterol of 2·6 mmol/L or Major coronary events
less (200 [6·0%] statin vs 247 [7·4%] control; RR 0·75, Data were available on 7757 first major coronary events 99% CI 0·56–1·01; p=0·01); diabetic individuals after randomisation, with 4770 participants having a without pre-existing vascular disease (368 [5·4%] vs 475 non-fatal MI and 2987 dying from CHD without having [7·1%]; RR 0·74, 99% CI 0·62–0·88; pϽ0·0001); and a non-fatal MI (see webtable 2). Overall, there was a people aged 75 years or older when randomised (385 highly significant 23% proportional reduction in the [10·6%] vs 470 [12·8%]; RR 0·82, 99% CI 0·70–0·96; incidence of first major coronary events per mmol/L LDL cholesterol reduction (3337 [7·4%] statin vs 4420[9·8%] control: RR 0·77, 95% CI 0·74–0·80; Coronary revascularisation
pϽ0·0001), which included a 26% reduction in non- Data were available on 6054 first coronary fatal MI (RR 0·74, 99% CI 0·70–0·79; pϽ0·0001; revascularisation procedures after randomisation figure 2). There was a significant trend (␹2 =10·5, (webtable 2). Overall, there was a significant 24% p=0·001) towards greater proportional reductions in proportional reduction in the incidence of first coronary major coronary events being associated with greater revascularisation (RR 0·76, 95% CI 0·73–0·80; mean absolute LDL cholesterol reductions in the pϽ0·0001) per mmol/L LDL cholesterol reduction, See Lancet Online
different trials (figure 3 and webfigure 3i), but no with similar proportional reductions in coronary artery significant heterogeneity between the relative effects grafting and angioplasty (figure 2). There was a after weighting for the absolute LDL cholesterol significant trend (␹2 =13·7, p=0·0002) towards greater reduction (␹2 =7·3, p=0·9; webfigure 3i). proportional reductions in coronary revascularisation The large number of major coronary events allowed being associated with greater mean absolute LDL the effects of lowering LDL cholesterol with a statin in cholesterol reductions in the different trials (webfigures different circumstances to be assessed reasonably 3ii and 4i), but no significant heterogeneity between the relative effects after weighting for the absolute LDL randomisation: even during the first year there was a highly significant 14% proportional reduction in major The effect on coronary revascularisations of lowering coronary events (RR 0·86, 99% CI 0·77–0·95; LDL cholesterol with a statin did not reach significance pϽ0·0001) per mmol/L, and there were highly during the first year after randomisation (RR 0·95, significant reductions of about 20–30% in every 99% CI 0·84–1·08; p=0·2), but there were clearly separate year thereafter (all pϽ0·0001; figure 4). Taking significant yearly reductions of between about 25% and www.thelancet.com Vol 366 October 8, 2005
Articles
Major coronary events
Endpoint
Events (%)
Treatment
Year 0–1:
Major vascular event
1747 (3·9%) 1951 (4·3%)
0·90 (0·85–0·96)
Year 1–2:
Major coronary event
Major vascular event
1231 (2·9%) 1603 (3·8%)
0·78 (0·73–0·83)
Year 2–3:
Major vascular event
1151 (2·8%) 1543 (3·9%)
0·74 (0·69–0·79)
Major vascular events
Year 3–4:
Major coronary event
Major vascular event
946 (2·6%) 1306 (3·8%)
0·72 (0·67–0·78)
Year 4–5:
Major vascular event
811 (2·9%)
993 (3·7%)
0·79 (0·74–0·86)
Major vascular event
468 (2·8%)
598 (3·8%)
0·74 (0·67–0·82)
Figure 3: Relation between proportional reduction in incidence of major
coronary events and major vascular events and mean absolute LDL
Trend test for major vascular events: ␹2 cholesterol reduction at 1 year
Square represents a single trial plotted against mean absolute LDL cholesterol reduction at 1 year, with vertical lines above and below corresponding to one SEof unweighted event rate reduction. Trials are plotted in order of magnitude of Figure 4: Proportional effects on major vascular events per mmol/L LDL cholesterol reduction by year
difference in LDL cholesterol difference at 1 year (webtable 1). For each Symbols and conventions as in figure 1. For each time period, RRs weighted by trial-specific LDL cholesterol outcome, regression line (which is forced to pass through the origin) represents reductions at 1 year relate to participants at risk of a first events (as do percentages).
weighted event rate reduction per mmol/L LDL cholesterol reduction. 30% during each of the subsequent 5 years (all Data were available on a total of 2957 first strokes pϽ0·001; figure 4). During an average of 5 years of after randomisation (webtable 2). There were 2282 treatment, the reduction in the overall incidence of coronary revascularisation of about one quarter per trials13,15,17–19,21,22,24,26 that sought information on stroke mmol/L LDL cholesterol reduction translated into type, of which 204 (9%) were attributed definitely to 27 (95% CI 20–34) fewer participants having such haemorrhage, 1565 (69%) were confirmed to be procedures per 1000 among those with pre-existing ischaemic, and 513 (22%) were of unknown type.
CHD at baseline, compared with 12 (9–16) fewer Overall, there was a significant 17% proportional among participants with no such history (figure 6 and reduction in the incidence of first stroke of any type webfigure 1iii). The proportional reduction in the (1340 [3·0%] statin vs 1617 [3·7%] control; RR 0·83, incidence of coronary revascularisation procedures per 95% CI 0·78–0·88; pϽ0·0001) per mmol/L lower LDL mmol/L LDL cholesterol reduction was about one cholesterol (figure 2). As was the case for major quarter in all of the prespecified subgroups (global coronary events and revascularisations, there was a heterogeneity p=0·9; webfigure 2ii).
significant trend (␹2 =6·8, p=0·009) towards greater www.thelancet.com Vol 366 October 8, 2005
Articles
Events (%)
Heterogeneity/
Treatment
trend test
Previous disease:
Age (years):
р65
Treated hypertension:
Yes
History of diabetes:
Yes
Diastolic blood pressure (mm Hg):
р90
Total cholesterol (mmol/L):
р5·2
LDL cholesterol (mmol/L):
р3·5
HDL cholesterol (mmol/L):
р0·9
Triglycerides (mmol/L):
р1·4
3337 (7·4%)
4420 (9·8%)
0·77 (0·74–0·80)
Global test for heterogeneity: ␹2 = 15·1; p = 0·4 Figure 5: Proportional effects on major coronary events per mmol/L LDL cholesterol reduction subdivided by baseline prognostic factors
Symbols and conventions as in figure 1.
proportional reductions in stroke being associated with haemorrhagic stroke (RR 1·05, 99% CI 0·78–1·41; greater mean absolute LDL cholesterol reductions in p=0·7; figure 2). The overall reduction in presumed the different trials (webfigures 3iii and 4ii). ischaemic stroke reflected a highly significant 22% This overall reduction in stroke reflected a highly proportional reduction in confirmed ischaemic significant 19% proportional reduction (RR 0·81, stroke (RR 0·78, 99% CI 0·70–0·87; pϽ0·0001) per 99% CI 0·74–0·89; pϽ0·0001) in strokes not attributed mmol/L LDL cholesterol reduction and a 12% to haemorrhage (ie, presumed ischaemic) per mmol/L proportional reduction in stroke of unknown type LDL cholesterol reduction, and no apparent difference in (RR 0·88, 99% CI 0·75–1·02; p=0·03). There was no www.thelancet.com Vol 366 October 8, 2005
Articles
significant reduction in stroke during the first year Participants with previous MI or CHD
after randomisation (RR 0·96, 99% CI 0·79–1·17; p=0·6), but there were significant reductions of about 20–25% during each of the subsequent 3 years and favourable trends thereafter (figure 4). During an average of 5 years of treatment, the reduction in theoverall incidence of stroke of about one sixth per mmol/L LDL cholesterol reduction translated into eight (95% CI 4–12) fewer participants having any stroke per1000 among those with pre-existing CHD at baseline, compared with five (1–8) fewer per 1000 among participants with no such history (figure 6 andwebfigure 1iv). Major vascular events
Data were available on 14 348 first major vascular
Participants without previous MI or CHD
events after randomisation, with 7757 participants having had a major coronary event, 6054 having had a coronary revascularisation procedure, and 2957having had a stroke (and some having more than one such event; webtable 2). Overall, there was a highly significant 21% proportional reduction in theincidence of major vascular events (RR 0·79, 95% CI 0·77–0·81; pϽ0·0001) per mmol/L LDL cholesterol reductions in major coronary events, coronary revascularisation procedures, and strokes (figure 2).
There was a significant trend (␹2 =26·4; pϽ0·0001) towards greater proportional reductions in majorvascular events being associated with greater LDL Figure 6: 5-year absolute benefits on particular vascular outcomes per mmol/L LDL cholesterol reduction in
cholesterol reductions in the different trials (figure 3 participants with and without previous MI or CHD
and webfigure 3iv), but no significant heterogeneity Many participants had more than one type of outcome, so sum of absolute differences for separate outcomesexceeds total number of participants avoiding at least one major vascular event.
between the relative effects after weighting for theabsolute LDL cholesterol reduction (␹2 =10·1; p=0·7; all of these subgroups considered separately (all Given the large number of major vascular events, the pϽ0·0001). There were also significant reductions in effects of lowering LDL cholesterol with a statin could risk per mmol/L LDL cholesterol reduction in a be examined particularly reliably in various different number of other subgroups, including: individuals circumstances (although analysis of this outcome had with pretreatment LDL cholesterol of 2·6 mmol/L or not been prespecified). There was a significant 10% less (383 [11·5%] vs 476 [14·3%]; RR 0·73, 99% CI proportional reduction (RR 0·90, 95% CI 0·85–0·96; 0·58–0·90; p=0·0001) or even of 2·0 mmol/L or less p=0·0006) in major vascular events during the first (75 [10·2%] vs 91 [12·9%]; RR 0·66, 99% CI 0·38–1·14; year after randomisation, and this was followed by p=0·05); diabetic individuals without previously known highly significant yearly reductions of around one vascular disease (713 [10·4%] vs 884 [13·1%]; RR 0·75, quarter during every subsequent year (all pϽ0·0001; 99% CI 0·66–0·86; pϽ0·0001); and those individuals figure 4). Taking all years together, the overall aged 75 years or older when randomised (612 [16·8%] reduction of about one fifth per mmol/L LDL vs 721 [19·7%]; RR 0·82, 99% CI 0·72–0·93; cholesterol reduction translated into 48 (95% CI 39–57) fewer participants having major vascular events per1000 among those with pre-existing CHD at baseline, compared with 25 (19–31) fewer per 1000 among The present analyses are of the 5103 first incident participants with no such history (figure 6 and cancers recorded after randomisation, excluding non- fatal recurrences of previously diagnosed cancers, but The incidence of major vascular events was reduced including any deaths from such recurrences. Non- by about one fifth per mmol/L LDL cholesterol melanoma skin cancers were not recorded routinely in reduction in every prespecified subgroup (global these trials, and so are not included in the analyses.
heterogeneity p=0·5; figure 7), and was significant in Overall, there was no evidence that lowering LDL www.thelancet.com Vol 366 October 8, 2005
Articles
Events (%)
Heterogeneity/
Treatment
trend test
Previous disease:
Age (years):
р65
Treated hypertension:
Yes
History of diabetes:
Yes
Diastolic blood pressure (mm Hg):
р90
Total cholesterol (mmol/L):
р5·2
LDL cholesterol (mmol/L):
р3·5
HDL cholesterol (mmol/L):
р0·9
Triglycerides (mmol/L):
р1·4
6354 (14·1%)
7994 (17·8%)
0·79 (0·77–081)
Figure 7: Proportional effects on major vascular events per mmol/L LDL cholesterol reduction subdivided by baseline prognostic factors
Symbols and conventions as in figure 1.
cholesterol by 1·0 mmol/L with statin therapy increased no apparent excesses among any particular site-specific 95% CI 0·95–1·06; p=0·9; figure 8), and the results ofunweighted analyses were similar (webfigure 3v).
Rhabdomyolysis
Furthermore, there was no evidence of an excess Information on rhabdomyolysis was available from all incidence of cancer emerging with increasing duration but one23 of the 14 trials (9 [0·023%] of 39 884 patients of treatment (␹2 for trend=0·6; p=0·4; figure 9).
vs 6 [0·015%] of 39 817 allocated Moreover, when cancer was analysed by site, there were control), and the 5-year excess risk with statin was small www.thelancet.com Vol 366 October 8, 2005
Articles
and not significant (absolute excess 0·01% [SE 0·01]; Events (%)
first cancer
Treatment
Discussion
The main objective of this collaboration is to provide reliable assessments of the major benefits and risks oflipid-modifying treatments, and this first cycle of meta- analyses has specifically addressed the effects of lowering LDL cholesterol with statins. We aimed to minimise both systematic and random errors by bringing together individual participant data from all 2567 (6·4%)
2536 (6·4%)
1·00 (0·95–1·06)
eligible large randomised trials comparing statintherapy versus control, and by prespecifying the main analyses.11 Furthermore, by weighting the results in individual trials and subgroups by the size of the achieved LDL cholesterol reductions, we were able to adjust for the potential confounding effects of such Figure 8: Proportional effects on cancer incidence per mmol/L LDL cholesterol reduction by site
differences. The results of this meta-analysis help to Symbols and conventions as in figure 1. For every type of cancer, analyses are of number of participants whose first clarify the relationship between the reductions in LDL recorded cancer after randomisation was of that type. ASCOTT-LLA only provided data on fatal cancers24 and sodoes not contribute to these analyses.
cholesterol and the effects on the incidence of differentvascular outcomes, the magnitude of the benefits indifferent circumstances, the time course over which Events (%)
Treatment
such benefits emerge, and the safety of the statin Benefit versus LDL cholesterol reduction
The results of this meta-analysis are consistent with there being an approximately linear relationship between the absolute reductions in LDL cholesterol achieved in these trials and the proportional reductions in the incidence of coronary and other major All times
2567 (6·4%)
2536 (6·4%)
1·00 (0·95–1·06)
vascular events. This finding is reinforced by those ofsome direct randomised comparisons of different statin regimens,29–31 which also indicate that larger LDL cholesterol reductions produce larger reductions in vascular disease risk. (Further evidence will beprovided by other such trials that are still ongoing.32,33) Figure 9: Proportional effects on cancer incidence per mmol/L LDL cholesterol reduction by year
Overall among the trials included in the present meta- Symbols and conventions as in figures 1 and 4. ASCOTT-LLA only provided data on fatal cancers24 and so does not analysis, the difference in LDL cholesterol at 5 years was about 0·8 mmol/L (chiefly reflecting non-compliance with the allocated treatments). The ratio of Benefits in different subgroups
the average LDL cholesterol difference for the whole A wide range of different types of participants was study period to the average difference measured at included in the 14 trials that contributed to this meta- 1 year was therefore about 0·9. Consequently, a analysis, so it was possible to explore the effects of reduction in LDL cholesterol of 1 mmol/L that is lowering LDL cholesterol with statin therapy in many sustained for 5 years may well produce a proportional reduction in major vascular events of about 23% (rather according to the subgroup-specific LDL cholesterol than the 21% reduction observed in the weighted differences between the treatment groups made it analysis). In many circumstances, full compliance with possible to allow for any differences between the LDL available regimens can reduce LDL cholesterol by cholesterol reductions in different subgroups. For substantially more than 1 mmol/L,10 and the present example, the absolute reduction in LDL cholesterol with results suggest that such reductions would produce a particular dose of a statin tends to be smaller among greater effects on vascular outcomes. For example, a those presenting with lower LDL cholesterol levels than reduction of 1·5 mmol/L in LDL cholesterol with among those with higher levels,21,34 but the proportional sustained statin therapy might well be expected to reduction in the event rate per mmol/L reduction in LDL reduce the incidence of major vascular events by about cholesterol was largely independent of the presenting level. That is, the results of the present analyses indicate www.thelancet.com Vol 366 October 8, 2005
Articles
that while lowering LDL cholesterol from 4 mmol/L to cancer with statin therapy among people aged older than 3 mmol/L reduces the risk of vascular events by about 70 years in another contributing trial22 was not 23%, lowering LDL cholesterol from 3 mmol/L to confirmed by the findings in the other trials (RR 1·03, 2 mmol/L also reduces (residual) risk by about 23%. So, 99% CI 0·91–1·16). Conversely, based on the results of an LDL cholesterol reduction of 2 mmol/L might be non-randomised observational studies, it has been expected to reduce risk by as much as 40% (ie, RRs of suggested that statin therapy might reduce the incidence 0·77✕0·77 yielding a combined RR of 0·59). The of various cancers (including colorectal41 and prostate proportional reductions in major vascular event rates per cancer42), but the results of the present meta-analysis of mmol/L LDL cholesterol reduction were very similar in all randomised trials do not support such claims. Although of the subgroups examined, including not just individuals the findings of this meta-analysis provide reassurance presenting with LDL cholesterol below 2·6 mmol/L that lowering LDL cholesterol with statin therapy does (100 mg/dL), but other groups for whom there had not increase the risk of non-vascular mortality and previously been uncertainty (such as diabetic individuals cancer during an average of 5 years, extended follow-up without pre-existing vascular disease, and people aged beyond the study treatment periods (perhaps through national registries) is warranted to identify whether anyadverse effects might emerge in the longer term. Evolution of benefits over time
Further evidence of the safety of the statin regimens There have been conflicting reports about how rapidly studied is also provided by the extremely low incidence benefits emerge after statin therapy is commenced, with of rhabdomyolysis (5 year excess: 0·01%, SE 0·01).
some trials reporting little or no reduction in vascular However, none of the trials in the meta-analysis involved events within the first year of treatment,13,15 and one trial22 a high-dose statin regimen and, since the risk of reporting no reduction in stroke with 3 years of treatment, myopathy is dose-dependent,43 the possibility that higher whereas other trials have reported more rapid benefits.29,35 doses would result in clinically relevant adverse effects In the present meta-analysis, there was a highly cannot be excluded. Information on episodes of raised significant 10% proportional reduction in major vascular liver enzymes was not sought for the meta-analysis, but events per mmol/L LDL cholesterol reduction during the results of other studies have shown that statins rarely first year (chiefly reflecting the observed 14% proportional induce hepatitis.44 In summary, therefore, the potential reduction in major coronary events) and larger reductions hazards of lowering LDL cholesterol with these statin of about 20–30% per mmol/L during every successive regimens seemed to be extremely small in relation to the year of treatment. There was limited power, however, to clear benefits in many circumstances.
assess how early the separate effects on major coronaryevents, coronary revascularisations, and strokes emerged.
Effects on total mortality
The survival analyses showed that the beneficial effects of Overall among the participants included in this meta- lowering LDL cholesterol with a statin accumulated analysis, statin therapy produced a clear reduction in all- during an average of 5 years of treatment, so the absolute cause mortality. Even so, the effects on vascular and non- benefits increased with continuing treatment. Since this vascular mortality considered separately may be more meta-analysis includes only the effects on first events, it widely generalisable to different populations in which the underestimates the absolute benefits of continued statin proportions of deaths from such causes differ. Similar therapy because the incidence of subsequent vascular proportional reductions in mortality attributed to events has also been shown to be reduced.36 coronary heart disease and in the incidence of majorvascular events were found among a wide range of Safety of lowering cholesterol
individuals, while no adverse effect was observed on non- Previously, the results of some observational studies7,37,38 vascular mortality or morbidity in any of the different and early randomised trials4,5,39,40 had raised concerns that circumstances studied. So, in populations where the lowering blood cholesterol concentrations might proportion of deaths from occlusive vascular disease is increase the risks of various non-vascular causes of lower than in the meta-analysis, a given proportional death and of particular cancers (eg, gastrointestinal, reduction in vascular mortality would be expected to respiratory, and haematological). In the present meta- translate into a smaller proportional reduction in all- analysis, however, there was no evidence that lowering cause mortality. By contrast, in populations at high risk LDL cholesterol by 1 mmol/L with 5 years of statin of vascular death (such as individuals with pre-existing therapy increased the risks of any specific non-vascular occlusive vascular disease), both the proportional and cause of death or of any specific type of cancer. One of absolute reductions in all-cause mortality would be the trials15 included in the present meta-analysis reported a possible excess risk of breast cancer inwomen with statin therapy, but this finding was not Implications
confirmed in the other contributing trials (RR 1·01, The results of the present meta-analysis indicate that the 99% CI 0·73–1·40). Similarly, an apparent excess risk of proportional reductions in the incidence of major www.thelancet.com Vol 366 October 8, 2005
Articles
coronary events, coronary revascularisations, and strokes ALERT (Assessment of Lescol in Transplantation): H Holdaas; ALLHAT were approximately related to the absolute reductions in (Antihypertensive Lipid Lowering Heart Attack Trial): D Gordon, LDL cholesterol achieved with the statin regimens B Davis; ASCOT (Anglo-Scandinavian Cardiac Outcomes Trial): B Dahlof, N Poulter, P Sever; BIP (Bezafibrate Infarction Prevention studied, and that the proportional reductions in such Study): U Goldbourt, E Kaplinsky; CARDS (Collaborative Atorvastatin major vascular events per mmol/L LDL cholesterol Diabetes Study): HM Colhoun, DJ Betteridge, PN Durrington, reduction were similar irrespective of the pretreatment GA Hitman, J Fuller, A Neil; 4D (Der Deutsche Diabetes Dialyse study): cholesterol concentrations or other characteristics (eg, C Wanner, V Krane; CARE (Cholesterol And Recurrent Events Study): F Sacks, L Moyé, M Pfeffer; CM Hawkins, E Braunwald; FIELD age, sex, or pre-existing disease) of the study participants.
(Fenofibrate Intervention and Event Lowering in Diabetes): P Barter, Current treatment guidelines are based on lowering LDL A Keech, J Simes; GISSI Prevention (Gruppo Italiano per lo Studio della cholesterol to particular target levels, with somewhat Sopravvivenza nell’Infarto miocardico): R Marchioli, G Tognoni, lower targets for people at higher risk of coronary MG Franzosi, A Maggioni; HIT (Veteran Administration Low HDLIntervention Trial): H Rubins, S Robins; HPS (Heart Protection Study): events.45,46 The results of this meta-analysis suggest, R Collins, J Armitage, A Keech, S Parish, R Peto, P Sleight; LIPS (Lescol however, that this strategy may not realise the full Intervention Prevention Study): PW Serruys; LDS (Lipids in Diabetes potential of such treatment. First, assessment of baseline Study): R Holman; LEADER (Lower Extremity Arterial Disease Event risk should be based on any type of occlusive vascular Reduction trial): T Meade; LIPID (Long-term Intervention withPravastatin in Ischaemic Disease): J Simes, A Keech, S MacMahon, event (rather than on coronary events alone), since I Marschner, A Tonkin, J Shaw; Post CABG (Post-Coronary Artery lowering LDL cholesterol with a statin lowers the risks Bypass Graft study): G Knatterud; PPP (Pravastatin Pooling Project): not just of coronary events but also of revascularisation C Furberg, R Byington; PROSPER (Prospective Study of Pravastatin inthe Elderly at Risk): P Macfarlane, S Cobbe, I Ford, M Murphy, procedures and of ischaemic strokes. Secondly, treat- G J Blauw, C Packard, J Shepherd; 4S (Scandinavian Simvastatin ment goals for statin treatment should aim chiefly to Survival Study): J Kjekshus, T Pedersen, L Wilhelmsen; SEARCH (Study achieve substantial absolute reductions in LDL choles- of Effectiveness of Additional Reductions in Cholesterol and terol (rather than to achieve particular target levels of LDL Homocysteine): R Collins, J Armitage, S Parish, R Peto, P Sleight;SHARP (Study of Heart and Renal Protection): C Baigent, A Baxter, cholesterol), since the risk reductions are proportional to R Collins, M Landray; TNT (Testing New Targets): J La Rosa; WHI the absolute LDL cholesterol reductions. Full compliance (Women’s Health Initiative): J Roussouw, J Probstfield; WOSCOPS with available statin regimens can reduce LDL choles- (West of Scotland Coronary Prevention Study): J Shepherd, S Cobbe, terol by at least 1·5 mmol/L in many circumstances, and P Macfarlane, I Ford.
Other members—M Flather, J Tobert, J Varigos, S Yusuf.
hence might be expected to reduce the incidence of major CTT secretariat—A Baxter, C Baigent, L Blackwell, G Buck, R Collins, vascular events by about one third. Ensuring that patients J Emberson, PM Kearney, A Keech, A Kirby, I Marschner, C Pollicino, at high 5-year risk of any type of occlusive major vascular event achieve and maintain a substantial reduction in Observers—Bristol-Myers Squibb: M Mellies, M McGovern, J Barclay, R Belder; Merck: Y Mitchell, T Musliner; Laboratoires Fournier: LDL cholesterol would result in major clinical and public- J-C Ansquer; Bayer: M Llewellyn; Novartis Pharma: M Bortolini; AstraZeneca: B Bryzinski; G Olsson J Pears; Pfizer: C Shear, C Newman. Contributors
All members of the writing committee contributed to the collection and
References
analysis of the data, and to the preparation of the manuscript. All Stamler J, Vaccaro O, Neaton JD, Wentworth D, for the Multiple Risk collaborators had an opportunity to contribute to the interpretation of Factor Intervention Trial Research Group. Diabetes, other risk factors, the results and to the drafting of the manuscript. The writing committee and 12-year cardiovascular mortality for men screened in the Multiple accepts full responsibility for the content of this paper. Risk Factor Intervention Trial. Diabetes Care 1993; 16: 434–44.
Chen Z, Peto R, Collins R, et al. Serum cholesterol concentration Conflict of interest statement
and coronary heart disease in population with low cholesterol Most of the trials in this report were supported by research grants from concentrations. BMJ 1991; 303: 276–82.
the pharmaceutical industry. Some members of the writing committee Rossouw JE, Lewis B, Rifkind BM. The value of lowering cholesterol (CB, AK, RP, RC, and JS) have had the costs of participating in scientific after myocardial infarction. N Engl J Med 1990; 323: 1112–19.
meetings reimbursed by the pharmaceutical industry. All other Oliver MF. Might treatment of hypercholesterolaemia increase non- members of the writing committee declare that they have no conflict of cardiac mortality? Lancet 1991; 337: 1529–31.
Davey Smith G, Pekkanen J. Should there be a moratorium on theuse of cholesterol lowering drugs? Acknowledgments
BMJ 1992; 304: 431–34.
The Clinical Trial Service Unit and Epidemiological Studies Unit Law MR, Thompson SG, Wald NJ. Assessing possible hazards of (CTSU) in the UK and National Health and Medical Research Council reducing serum cholesterol. BMJ 1994; 308: 373–79.
Clinical Trials Centre (CTC) in Australia coordinate this collaboration Jacobs D, Blackburn H, Higgins M, et al. Report of the Conferenceon low blood cholesterol: mortality associations. jointly. The CTSU is supported by the UK Medical Research Council, 86: 1046–60.
British Heart Foundation, Cancer Research UK, Wellcome Trust Law MR, Wald NJ, Thompson SG. By how much and how quickly Cardiovascular Research Initiative, and, previously, the European does reduction in serum cholesterol concentration lower risk of Community Biomed Programme. The CTC is supported by a ischaemic heart disease? BMJ 1994; 308: 367–72.
programme grant from the Australian National Health and Medical Muldoon MF, Manuck SB, Matthews KA. Lowering cholesterol Research Council, and a grant from the National Heart Foundation, concentrations and mortality: a quantitative review of primary prevention trials. BMJ 1990; 301: 309–14.
Current membership of the CTT Collaboration
Tobert JA. Lovastatin and beyond: the history of the HMG-CoA Writing committee—C Baigent, A Keech, PM Kearney, L Blackwell, reductase inhibitors. Nat Rev Drug Discov 2003; 2: 517–26.
G Buck, C Pollicino, A Kirby, T Sourjina, R Peto, R Collins, J Simes.
Cholesterol Treatment Trialists’ (CTT) Collaboration. Protocol for a Collaborating trialists—AFCAPS/TEXCAPS (AirForce/Texas Coronary prospective collaborative overview of all current and planned Atherosclerosis Prevention Study): JR Downs, A Gotto, M Clearfield; randomized trials of cholesterol treatment regimens.
Am J Cardiol 1995; 75: 1130–34.
www.thelancet.com Vol 366 October 8, 2005
Articles
Early Breast Cancer Trialists’ Collaborative Group. Treatment of Friedewald WT, Levy RI, Fredrickson DS. Estimation of the early breast cancer: worldwide evidence 1985–1990. Oxford: concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972; 18:
Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart Cannon CP, Braunwald E, McCabe CH, et al. Comparison of disease: the Scandinavian Simvastatin Survival Study (4S). Lancet intensive and moderate lipid lowering with statins after acute 1994; 344: 1383–89.
coronary syndromes. N Engl J Med 2004; 350: 1495–504.
Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart Nissen SE, Tuzcu EM, Schoenhagen P, et al. Effect of intensive disease with pravastatin in men with hypercholesterolemia. compared with moderate lipid-lowering therapy on progression of N Engl J Med 1995; 333: 1301–07.
coronary atherosclerosis: a randomized controlled trial. JAMA Sacks FM, Pfeffer MA, Moye LA, et al. The effect of pravastatin on 2004; 291: 1071–80.
coronary events after myocardial infarction in patients with average LaRosa JC, Grundy SM, Waters DD, et al. Intensive lipid lowering cholesterol levels. N Engl J Med 1996; 335: 1001–09.
with atorvastatin in patients with stable coronary disease. The Post Coronary Artery Bypass Graft Trial Investigators. The N Engl J Med 2005; 352: 1425–35.
effect of aggressive lowering of low-density lipoprotein cholesterol SEARCH investigators. Study of the Effectiveness of Additional levels and low-dose anticoagulation on obstructive changes in Reductions in Cholesterol and Homocysteine.
saphenous-vein coronary-artery bypass grafts. N Engl J Med 1997; http://www.ctsu.ox.ac.uk/projects/search.shtml (accessed Sept 18, 336: 153–62.
Downs JR, Clearfield M, Weis S, et al. Primary prevention of acute Pedersen TR, Faergeman O, Holme I, Olsson AG, Tikkanen MJ.
coronary events with lovastatin in men and women with average Effect of greater LDL-C reductions on prognosis: the Incremental cholesterol levels: results of AFCAPS/TexCAPS. JAMA 1998; 279:
Decrease in Endpoints through Aggressive Lipid Lowering (IDEAL) trial. Atherosclerosis 1999; 144: 38.
The Long-Term Intervention with Pravastatin in Ischaemic Disease Cullen P, Assmann G. Treatment goals for low-density lipoprotein (LIPID) Study Group. Prevention of cardiovascular events and cholesterol in the secondary prevention of coronary heart disease: death with pravastatin in patients with coronary heart disease and a absolute levels or extent of lowering. Am J Cardiol 1997; 80:
broad range of initial cholesterol levels. N Engl J Med 1998; 339:
Schwartz GG, Olsson AG, Ezekowitz MD, et al. Effects of GISSI Prevenzione Investigators (Gruppo Italiano per lo Studio atorvastatin on early recurrent ischemic events in acute coronary della Sopravvivenza nell’Infarto Miocardico). Results of the low- syndromes: the MIRACL study—a randomized controlled trial.
dose (20 mg) pravastatin GISSI Prevenzione trial in 4271 patients JAMA 2001; 285: 1711–18.
with recent myocardial infarction: do stopped trials contribute to Heart Protection Study Collaborative Group. MRC/BHF Heart overall knowledge? Ital Heart J 2000; 1: 810–20.
Protection Study of cholesterol-lowering with simvastatin in 5963 Serruys PW, de Feyter P, Macaya C, et al. Fluvastatin for people with diabetes: a randomised placebo-controlled trial. Lancet prevention of cardiac events following successful first percutaneous 2003; 361: 2005–16.
coronary intervention: a randomized controlled trial. JAMA 2002; Cummings P, Psaty BM. The association between cholesterol and 287: 3215–22.
death from injury. Ann Intern Med 1994; 120: 848–55.
Heart Protection Study Collaborative Group. MRC/BHF Heart Law MR, Thompson SG. Low serum cholesterol and the risk of Protection Study of cholesterol lowering with simvastatin in 20 536 cancer: an analysis of the published prospective studies. Cancer high-risk individuals: a randomised placebo-controlled trial. Lancet Causes Control 1991; 2: 253–61.
2002; 360: 7–22.
The Lipid Research Clinics Coronary Primary Prevention Trial Shepherd J, Blauw GJ, Murphy MB, et al. Pravastatin in elderly results. I: reduction in incidence of coronary heart disease. JAMA individuals at risk of vascular disease (PROSPER): a randomised 1984; 251: 351–64.
controlled trial. Lancet 2002; 360: 1623–30.
Frick MH, Elo O, Haapa K, et al. Helsinki Heart Study: primary- The ALLHAT Officers and Coordinators for the ALLHAT prevention trial with gemfibrozil in middle-aged men with Collaborative Research Group. Major outcomes in moderately dyslipidemia: safety of treatment, changes in risk factors, and hypercholesterolemic, hypertensive patients randomized to incidence of coronary heart disease. N Engl J Med 1987; 317:
pravastatin vs usual care. JAMA 2002; 288: 2998–3007.
Sever PS, Dahlof B, Poulter NR, et al. Prevention of coronary and Poynter JN, Gruber SB, Higgins PDR, et al. Statins and the risk of stroke events with atorvastatin in hypertensive patients who have colorectal cancer. N Engl J Med 2005; 352: 2184–92.
average or lower-than-average cholesterol concentrations, in the Shannon J, Tewoderos S, Garzotto M, et al. Statins and prostate Anglo-Scandinavian Cardiac Outcomes Trial-Lipid Lowering Arm cancer risk: a case-control study. Am J Epidemiol 2005; 162: 318–25.
(ASCOT-LLA): a multicentre randomised controlled trial. Lancet Thompson PD, Clarkson P, Karas RH. Statin-associated myopathy.
2003; 361: 1149–58.
JAMA 2003; 289: 1681–90.
Holdaas H, Fellstrom B, Jardine AG, et al. Effect of fluvastatin on cardiac outcomes in renal transplant recipients: a multicentre, Am J Cardiol 2002; 89:
ranomised, placebo-controlled trial. Lancet 2003; 361: 2024–31.
Grundy SM, Cleeman JI, Merz CN, et al. Implications of recent Colhoun HM, Betteridge DJ, Durrington PN, et al. Primary clinical trials for the national cholesterol education program adult prevention of cardiovascular disease with atorvastatin in type 2 diabetes in the Collaborative Atorvastatin Diabetes Study (CARDS): Circulation 2004; 110: 227–39.
multicentre randomised placebo-controlled trial. Lancet 2004; 364:
De Backer G, Ambrosioni E, Borch-Johnsen K, et al. European guidelines on cardiovascular disease prevention in clinical practice:Third Joint Task Force of European and other Societies on Athyros VG, Papageorgiou AA, Mercouris BR, et al. Treatment Cardiovascular Disease Prevention in Clinical Practice.
with atorvastatin to the National Cholesterol Educational Programgoal versus ‘usual’ care in secondary coronary heart disease Atherosclerosis 2004; 171: 145-55.
prevention. Curr Med Res Opin 2002; 18: 220–28.
www.thelancet.com Vol 366 October 8, 2005

Source: http://www.saofranciscoodontologia.com.br/images/online/Cholesterol_Safety_and_Efficacy_Metanalise_Lancet_downloads.pdf