Am J Cardiovasc Drugs

DOI 10.1007/s40256-015-0139-y

CURRENT OPINION

Meta-Analyses of Statin Therapy for Primary Prevention Do Not

Answer Key Questions: An Empirical Appraisal of 5 Years
of Statin Meta-Analyses
Chetan Huded1 • Vinay Prasad2

Ó Springer International Publishing Switzerland (outside the USA) 2015

Abstract Although meta-analyses of statins in primary conclude that most recent meta-analyses of statins for
prevention are designed to provide doctors and patients primary prevention do not adequately address the question
with better evidence about the risks and potential benefits they seek to answer.
of treatment, they may ignore important patient-centered
outcomes and concerns. We examined all meta-analyses of
statins for primary prevention over the last 5 years. We
Key Points
assessed whether each meta-analysis addressed five key
points: whether authors examined endpoints based on the
Whether or not healthy people should take statin
use of statin therapy, and not stratified by low-density
medications as primary prevention has become a
lipoprotein reduction; whether authors included only
controversial topic, and is frequently debated.
studies of statin versus placebo, and not varying doses or
brands of statin; whether authors considered commonly Meta-analyses often claim to adjudicate the debate
cited harms; whether secondary prevention patients were by providing a summary estimate of statins’ benefits
excluded; and, whether overall mortality was examined. and harms.
We examined 189 articles to identify 24 meta-analyses of In an empirical analysis, we found that few meta-
statins that made claims regarding primary prevention. Six analyses properly frame the primary prevention
studies (25 %) reported outcomes as a function of reduc- question. Specifically, they do not always report
tion in serum lipid levels rather than treatment received. outcomes in terms of taking a statin or not, exclude
Seven studies (29 %) included trials of high-dose versus randomized controlled trials of a high- versus low-
low-dose statin in their analysis. Five studies (21 %) did dose or two different statins, exclude patients with
not examine all-cause mortality. The majority of studies established cardiovascular disease, report all-cause
(n = 21, 88 %) failed to exclude patients with known mortality, and assess at least two of three safety
cardiovascular disease, and 22 (92 %) studies failed to concerns.
assess two of three common safety concerns. Nevertheless,
most (n = 20, 83 %) meta-analyses supported the use of
statins in primary prevention. Based on our findings, we

1 Introduction
& Vinay Prasad
[email protected] The benefits of statin therapy in secondary prevention are
1 well known. Evidence supports their use in patients after
Department of Medicine, Northwestern University, Chicago,
IL, USA myocardial infarction (MI) [1], post stroke [2], in patients
2 with angina and elevated cholesterol [3], and even in
Medical Oncology Branch, National Cancer Institute,
National Institutes of Health, 10 Center Dr. 10/12N226, patients with peripheral arterial disease (PAD) [4]. For
Bethesda, MD 20892, USA patients with known coronary heart disease (CHD),
 C. Huded, V. Prasad

improvements have been shown not only for future car- to assess all meta-analyses of statins in primary prevention
diovascular events, but also for all-cause mortality [1, 5]. over the last 5 years.
However, there are patient populations known to not ben-
efit from statins. Statins have failed to improve outcomes in
patients with heart failure [6, 7], and for patients on 2 Meta-Analysis Selection Strategy
hemodialysis [8, 9]. Debate continues regarding their use in
primary prevention [10]. We performed a systematic review of statin meta-analyses.
Whether or not statins should be recommended to Meta-analyses of statin therapy in primary prevention were
healthy patients, including those with risk factors but no obtained via a search of the MEDLINE database using the
history of cardiovascular disease (CVD), has emerged as a following terms: statin(s) primary prevention, meta-analysis
contentious topic in medicine [11–13]. Pooled analysis of statin(s), pooled analysis of statin(s), statin(s) benefits,
randomized controlled trials (RCTs) of statins for this statin(s) cardiovascular prevention, and HMG CoA reduc-
purpose have yielded mixed results [10]. While some RCTs tase inhibitor(s) meta-analysis. The search was filtered to
have shown benefit in reducing cardiovascular events, the include only meta-analyses published after 1 December
absolute risk reduction has been small [14–16]. Harms 2007 and before 18 September 2012, the date of our search.
related to statin therapy have been inconsistently docu- A total of 189 articles were retrieved and read in full
mented, and improvements in all-cause mortality have not (Fig. 1). Articles were excluded if the authors did not
been well established. attempt to analyze whether statins were beneficial in pri-
Meta-analyses of statins in primary prevention attempt mary prevention. Articles regarding novel therapeutic uses
to resolve the dispute. In 2012, the Cholesterol Trialists’ of statins, such as the prevention of aortic stenosis, atrial
Group (CTT) published their findings in The Lancet [17]. fibrillation, infection, renal impairment, and malignancy,
The group analyzed pooled data from 27 RCTs and were excluded, as were those that studied whether statins
134,537 participants. They found that, for every 18 mg/dl confer peri-procedural benefits. Articles assessing statins’
that statins lowered cholesterol, patients benefitted from a impact on surrogate endpoints only, including vascular
21 % reduction in vascular events. However, the CTT’s imaging (computed tomography [CT] angiogram, intravas-
analysis has several limitations. First, the authors weighed cular ultrasound [IVUS], carotid intima-media thickness
trials based on change in lipid levels, meaning that end- [IMT]) or serum lipid measurements, were excluded.
points were given more importance in trials where levels of Studies limited to secondary prevention, or those pertaining
low-density lipoprotein (LDL) fell the most. However, this to specific subpopulations (i.e. end-stage renal disease and
analysis does not reflect the real-world decision to take a familial hypercholesterolemia) were also excluded. All
statin or not, realizing that a priori we cannot know how other articles were included. Our study was conducted
much LDL reduction a patient may experience. Addition- between 18 September 2012 and 1 February 2013.
ally, weighing outcomes based on response is a flawed
analytic technique that places undue weight on patients 2.1 Design Choices that Statin Meta-Analyses
who respond to a therapy [18, 19]. Second, the analysis did Should Embrace to Address the Primary
not exclude trials comparing two different statins (or dif- Prevention Question
ferent doses). Such trials may confound the results, as both
groups may experience most of the harms of the medica- The following five points were examined in each article:
tion, but only one may experience disproportionate bene- (1) outcomes should be reported in terms of taking a statin
fits. Third, the article did not consider potential harms, or not, and not based on changes in serum lipid concen-
including common patient concerns such as incident dia- trations; (2) analyses should exclude RCTs of statin versus
betes, cancer, or cognitive dysfunction related to long-term statin (high vs. low dose or two different statins); (3)
statin therapy. Two other common points of contention in patients with established CVD, for whom statin therapy is
most meta-analyses on this topic are a failure to examine indicated for secondary prevention, should be excluded; (4)
overall mortality and failure to exclude secondary pre- all-cause mortality should be reported rather than only
vention patients. However, the CTT did meet these metrics. disease-specific mortality; (5) at least two of the following
An updated analysis in 2015 by the same group continues safety endpoints should be reported: cognitive dysfunction,
to be subject to the same criticism [20]. new-onset diabetes or worsening glycemic control in prior
These five criteria are key preconditions for a meta- diabetics, new-onset malignancy or malignancy-related
analysis to adequately address whether a patient should death. These five criteria were pre-specified and chosen
take a statin for primary prevention. Using them, we set out because they more precisely identify the question in
5 Years of Statins’ Meta-analyses

189 arcles idenfied through database search and

reviewed by hand

165 arcles were excluded

82 tested stan therapy for indicaons not related to CV outcomes
34 did not pertain to risks/benefits of stan therapy
13 were commentary or review arcles
9 tested secondary prevenon only
7 compared stans to non-stan lipid lowering therapy rather than placebo
5 studied familial hypercholesterolemia
5 reported effect of stans on lab surrogates (CV risk markers or lipid levels)
4 compared stan regimens (either type or dose of stan)
3 studied ESRD or renal transplant paents
3 reported effect of stans on vascular imaging surrogates (CT, IVUS, CIMT)

24 arcles were included in the analysis

Fig. 1 A search of the MEDLINE database resulted in 189 articles. excluded. CIMT carotid intima-media thickness, CT computed
The articles were reviewed manually, and only meta-analyses relating tomography, CV cardiovascular, ESRD end-stage renal disease, IVUS
to the risks and benefits of statin therapy for the primary prevention of intravascular ultrasound
cardiovascular endpoints were included. In total, 165 articles were

primary prevention: does taking a statin (not knowing how low-dose statins, or two different statin medications, in
much LDL will fall) improve survival in patients who do their analysis. The majority of studies (88 %, n = 21)
not otherwise have an indication for statins, and if it does, included at least some patients with known CVD, defined
does it do so without significant toxicity or harms [10]. as established CHD, peripheral arterial disease, or cere-
Primary RCTs included in each meta-analysis were brovascular disease, including prior transient ischemic
reviewed when questions could not be answered by review attack (TIA) or stroke (Table 2).
of the meta-analysis alone. All meta-analyses were A recent study by the CTT [17] included both primary
reviewed by one reviewer (CH). prevention and secondary prevention populations, although
all-cause mortality was reported separately in the popula-
2.2 Meta-Analysts’ Opinion of Statins tion without established vascular disease. A study by Ray
et al. [28] included RCTs that enrolled secondary preven-
The authors’ commentary on the use of statin therapy in tion patients, but the authors obtained the primary datasets
primary prevention was classified into one of three groups: for those RCTs and meticulously excluded patients with
(1) positive—the authors support the use of statins in pri- known CVD. A study assessing primary prevention in
mary prevention; (2) equivocal—the authors make no women by Mora et al. [31] relied exclusively on three RCTs
definitive claims regarding the use or avoidance of statins that excluded all patients with established CVD (AFCAPS/
in primary prevention; or (3) negative—the authors suggest TexCAPS [Airforce/Texas Coronary Atherosclerosis
caution regarding the use of statins in primary prevention. Prevention Study] [15], MEGA [primary prevention of
cardiovascular disease with pravastatin in Japan] [44], and
JUPITER [Rosuvastatin to prevent vascular events in men
3 The Results of Our Analysis and women with elevated C-Reactive Protein] [16]).
Among these three trials that rigorously excluded patients
Our search strategy yielded 189 articles, which were with known CVD, only the CTT study reported a benefit on
reviewed in full; 24 assessed the effect of statins in primary all-cause mortality [per 1.0 mmol/l reduction in LDL, rel-
prevention and were included in this study [17, 21–43] ative risk 0.91; 95 % confidence interval (CI) 0.85–0.97;
(Table 1). Six studies (25 %) reported outcomes as a p = 0.007). However, as noted, that study was one of six
function of reduction in serum lipid levels rather than based that reported endpoints based on reduction in LDL levels
simply on whether patients were assigned to statin therapy. rather than by treatment arm. All-cause mortality was
Seven studies (29 %) included trials of high-dose versus reported in all but five studies (21 %) (Table 2).
 C. Huded, V. Prasad

Table 1 Each meta-analysis was assessed for achievement of the five benchmarks described above
Article Reports outcomes Excludes trials comparing Excludes patients Reports Reports at least Study
based on taking a two different statins or taking statin all-cause two of three conclusions
statin or not a statin at two different doses for secondary mortality common safety
prevention concernsa

Mihaylova et al. [17] Y Y Y ?

Kostis et al. [21] Y Y ?
Chen et al. [22] Y Y Y ?
Tonelli et al. [23] Y Y Y Y ?
Chan et al. [24] Y ?
Mills et al. [25] Y Y Y Y ?
Taylor et al. [26] Y Y Y ?
Baigent et al. [27] Y ?
Ray et al. [28] Y Y Y Y -
Kizer et al. [29] Y ?
Bukkapatnam et al. [30] Y Y Y ?/-
Mora et al. [31] Y Y Y Y ?
Petretta et al. [32] Y Y Y -
Sheng et al. [33] Y ?
Brugts et al. [34] Y Y Y ?
Amarenco and Labreuche Y ?
[35]
Navaneethan et al. [36] Y Y ?
Delahoy et al. [37] ?
Mills et al. [38] Y Y Y ?
Strippoli et al. [39] Y Y Y ?/-
Messerli et al. [40] Y Y Y ?
Kearney et al. [41] Y Y ?
O’Regan et al. [42] Y Y Y ?
Henyan et al. [43] Y Y ?
a
Three common safety concerns include incident diabetes or glycemic dysfunction, cognitive dysfunction, and incident or fatal malignancy
Y indicates the designated benchmark was achieved, ? indicates the authors support the use of statins in primary prevention, - indicates the
authors suggest caution regarding the use of statins in primary prevention, ?/- indicates the authors make no definitive claims regarding the use
or avoidance of statins in primary prevention

Table 2 The percentage of meta-analyses not achieving each of the proposed benchmarks is shown
Benchmark Not achieveda

The article reports outcomes based on taking a statin or not taking a statin 6 (25)
The article excludes trials comparing two different statins or a statin at two different doses 7 (29)
The article excludes patients who are using a statin for secondary prevention 21 (88)
The article reports all-cause mortality 5 (21)
The article reports at least two out of three common safety concerns (diabetes, cognitive dysfunction, malignancy) 22 (92)
a
Data are presented as n (%)

Commonly purported adverse effects of statin therapy reported the effects of statin therapy on cognitive function,
were generally under-reported by these 24 articles. Of the and only two studies (8 %) reported incident diabetes. Nine
24 studies, 22 (92 %) did not meet our prespecified criteria studies (38 %) reported effects of statins on malignancy-
of reporting outcomes on at least two of the three safety related death or incident malignancy (Table 3).
endpoints defined above (diabetes, cognitive dysfunction, None of the meta-analyses achieved all of the five
malignancy-related events). None of the 24 studies benchmarks. Four studies (17 %) achieved four benchmarks,
5 Years of Statins’ Meta-analyses

Table 3 Each meta-analysis was assessed for the monitoring and reporting of three commonly reported risks of statin therapy
Article Reports on incident diabetes Reports on cognitive Reports on incident
or glycemic dysfunction dysfunction or fatal malignancy

Mihaylova et al. [17] Y

Kostis et al. [21]
Chen et al. [22]
Tonelli et al. [23] Y Y
Chan et al. [24] Y
Mills et al. [25] Y Y
Taylor et al. [26] Y
Baigent et al. [27] Y
Ray et al. [28]
Kizer et al. [29]
Bukkapatnam et al. [30] Y
Mora et al. [31]
Petretta et al. [32]
Sheng et al. [33]
Brugts et al. [34] Y
Amarenco and Labreuche [35]
Navaneethan et al. [36]
Delahoy et al. [37]
Mills et al. [38] Y
Strippoli et al. [39]
Messerli et al. [40]
Kearney et al. [41]
O’Regan et al. [42]
Henyan et al. [43]
Y indicates the designated harm was assessed and reported in the manuscript

ten (42 %) achieved three benchmarks, four (17 %) achieved useful to clarify these discrepancies. Our findings highlight
two benchmarks, five (21 %) achieved one benchmark, and limitations in our most recent pooled datasets intended to
one study (4 %) achieved none of the benchmarks. Despite address the topic of statin therapy in primary prevention.
these findings, the overwhelming majority (n = 20, 83 %) Remarkably, 88 % of the meta-analyses on this topic in
of trials supported the use of statins in primary prevention. Of the past 5 years have included patients with established
the four other trials, two (8 %) reported negative findings and CVD. Three primary trials—the HPS (Heart Protection
two (8 %) reached equivocal findings. Study) [46, 47], the ASCOT-LLA (Anglo-Scandinavian
Cardiac Outcomes Trial) [47], the ALLHAT-LLT (Anti-
hypertensive and Lipid-Lowering to Prevent Heart Attack
4 Making Sense of the Statins Meta-Analysis Trial) [48]—were commonly included in primary preven-
Literature tion meta-analyses although these studies included patients
with established vascular disease. The ALLHAT cohorts
Our results provide evidence for why the use of statin included up to 17 % of patients with established CHD,
therapy in patients without established CVD is a con- while the ASCOT population included 10 % with known
tentious issue in preventive medicine. Patients are symptomatic cerebrovascular disease, 5 % with PAD, and
increasingly exposed to the idea that taking statins to treat over 3 % with ‘‘other’’ CVD at baseline. The HPS trial was
CVD before the manifestation of overt symptoms offers a mixed primary and secondary prevention population,
significant benefits [45]. However, clinical trials on this with known CHD in over 20 % of the study population.
topic have yielded mixed results—many trials have shown Inclusion of patients with established CVD introduces
improvements in cardiovascular event rates, but the effect considerable bias in favor of statin therapy given that the
on all-cause mortality has not been reliably replicated [10, benefits of statin therapy in such patients are well docu-
14, 16, 44]. Pooled data in meta-analyses should prove mented. The inclusion of secondary prevention patients is
 C. Huded, V. Prasad

not a necessary consequence of pooled analyses, as some cancer have been voiced since at least the 1990s [56, 57].
authors did perform a review of the individual patient-level Links to incident diabetes gained prominence in 2008,
data for trials with mixed populations and meticulously largely with the dissemination of results from the JUPITER
excluded patients with known CVD in their analysis [28]. study (only five meta-analyses we included were published
Although the updated meta-analysis by the Cochrane col- prior) [16], and concern that statin use could precipitate
laboration on this topic fell outside our study dates, it also cognitive dysfunction have been present for over a decade
includes secondary prevention patients, as long as these [58]. Of course, we agree with others that simply being
patients accounted for \10 % of the total in a given trial linked to diabetes mellitus should not invalidate statins for
[49]. We acknowledge that including these patients is all their uses [59]; instead, we merely asked whether meta-
neither inherently wrong nor incorrect but, rather, shifts the analysts quantified potential benefits and harms.
question away from the primary prevention question.
Over one-fifth (21 %) of the meta-analyses analyzed did 4.1 Important Caveats
not report the effect of statin therapy in primary prevention
on all-cause mortality. All of these five studies ultimately Our study has several limitations. We did not extract
concluded that statin therapy is beneficial in this setting information regarding reported conflicts of interest among
despite the omission of overall mortality data. authors of meta-analyses or sources of funding, and whether
Instituting statin therapy in primary prevention should these are associated with the conclusions of meta-analysts.
include a process of shared decision making with patients Other groups have shown that industry-supported meta-
based on the risks and benefits. Our study illustrates that analyses are more likely to reach favorable conclusions, are
the reporting of risks has been limited. Although serum less transparent, and appreciate fewer methodological
markers of liver and muscle dysfunction have been closely reservations than corresponding Cochrane reviews [60].
monitored in the primary trial literature, there is little However, when it comes to the use of statins, nearly all
evidence on other safety concerns. The JUPITER [16] trial, meta-analyses support the use of these drugs. In our study,
published in 2008, demonstrated a 0.6 % absolute risk only two meta-analyses reached negative conclusions, and
increase in the development of diabetes (p = 0.01) in the two were equivocal. The vast majority 20 (83 %) reached
rosuvastatin group compared with placebo. This effect favorable conclusions. Thus, studying any association
correlates with a new diagnosis of diabetes in one out of between funding source and conclusion would be under-
every 167 patients treated with rosuvastatin [50]. In our powered to draw firm conclusions. Other groups have
investigation, only two studies (8 %) reported on the risk of examined the conclusions of articles assessing the cost
incident diabetes. However, in 2012, the US FDA expan- effectiveness of statins [61]. Indeed, they found that
ded the safety label on statins to include the risk of new- industry sponsorship was strongly associated with favorable
onset diabetes as well as the risk of cognitive dysfunction cost-effectiveness conclusions for statins in primary pre-
[51]. Although no durable association between statin vention. Specifically, the percent of studies with negative or
therapy and cognitive dysfunction has been demonstrated neutral conclusions was 0 % for industry-sponsored inves-
in limited studies [52], the monitoring and reporting of tigations and 57.9 % for non-conflicted studies (p \ 0.001).
cognitive outcomes should be performed meticulously in Additionally, we did not examine outcomes by sex.
studies of statin therapy. The risk of incident malignancy in Prior meta-analyses have reached divergent conclusions
patients treated with statins has also not been reliably about whether the risk benefit profile for statins in men is
demonstrated [53, 54]. However, despite commonly cited comparable to that profile in women [5, 21]. Women have
concerns regarding neoplastic effects of statins, malig- historically been under-represented in statin trials, and their
nancy-related outcomes have been reported by less than outcomes may be different than those for men. We believe
half of the studies analyzed here (38 %, n = 9). this topic is of immense importance, but did not pre-specify
Some may contend that we are holding meta-analysts to it prior to our data collection. Future researchers may
an unfair standard. The FDA did not mandate warnings that consider specifically examining this question, using the
statins may be linked to incident diabetes or cognitive framework we outline.
dysfunction until February 2012 [55], therefore it is
unreasonable to ask whether older analyses examined these
concerns. Yet, we granted credit to studies that addressed 5 Conclusions
two of three of these concerns, and two studies did, and the
harms we queried (cancer/incident diabetes/cognitive dys- Whether or not patients without established CVD should
function) have been widely known and discussed for years. use statins for primary prevention remains a topic of
Concerns that statins may be linked to higher rates of intense debate. Many authors have advocated widespread
5 Years of Statins’ Meta-analyses

adoption of statin therapy for prevention purposes, while 12. Blaha MJ, Nasir K, Blumenthal RS. Statin therapy for healthy
others endorse a more cautious approach. We analyzed five men identified as ‘‘increased risk’’. Jama. 2012;307:1489–90.
13. Steinberg D, Grundy SM. The case for treating hypercholes-
design choices that a meta-analysis should embrace in terolemia at an earlier age: moving toward consensus. J Am Coll
order to provide sufficient evidence for the use of statins in Cardiol. 2012;60:2640–2.
primary prevention—choices that more accurately define 14. Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart
the primary prevention question. Our investigation is the disease with pravastatin in men with hypercholesterolemia. West
of Scotland Coronary Prevention Study Group. N Engl J Med.
first study to systematically appraise meta-analyses for 1995;333:1301–7.
statins in primary prevention based on key questions that 15. Downs JR, Clearfield M, Weis S, et al. Primary prevention of
face providers and patients. Future meta-analyses, partic- acute coronary events with lovastatin in men and women with
ularly those that promote the use of statins in primary average cholesterol levels: results of AFCAPS/TexCAPS. Air
Force/Texas Coronary Atherosclerosis Prevention Study. Jama.
prevention, should aim to address the concerns raised here. 1998;279:1615–22.
16. Ridker PM, Danielson E, Fonseca FA, et al. Rosuvastatin to
Compliance with Ethical Standards This article was fully com- prevent vascular events in men and women with elevated C-re-
pliant with all ethical standards. active protein. N Engl J Med. 2008;359:2195–207.
17. Mihaylova B, Emberson J, Blackwell L, et al. The effects of
Conflicts of interest The views and opinions of Dr. Prasad do not lowering LDL cholesterol with statin therapy in people at low
necessarily reflect those of the National Cancer Institute or the risk of vascular disease: meta-analysis of individual data from 27
National Institutes of Health. Chetan Huded and Vinay Prasad report randomised trials. Lancet. 2012;380:581–90.
no disclosures or conflicts of interest in preparing this manuscript. 18. Newman DH, Saini V, Brody H, et al. Statins for people at low
Neither Chetan Huded or Vinay Prasad received any funding for the risk of cardiovascular disease. Lancet 2012;380:1814 (author
present manuscript. reply 7–8).
19. Anderson JR, Cain KC, Gelber RD. Analysis of survival by tumor
Author contributions Chetan Huded and Vinay Prasad contributed response. J Clin Oncol. 1983;1:710–9.
to the data analysis and writing of the manuscript. 20. Mosca L. Sex, statins, and statistics. Lancet. 2015;385:1368–9.
21. Kostis WJ, Cheng JQ, Dobrzynski JM, Cabrera J, Kostis JB.
Meta-analysis of statin effects in women versus men. J Am Coll
Cardiol. 2012;59:572–82.
22. Chen YH, Feng B, Chen ZW. Statins for primary prevention of
References cardiovascular and cerebrovascular events in diabetic patients
without established cardiovascular diseases: a meta-analysis. Exp
1. Sacks FM, Pfeffer MA, Moye LA, et al. The effect of pravastatin Clin Endocrinol Diabetes. 2012;120:116–20.
on coronary events after myocardial infarction in patients with 23. Tonelli M, Lloyd A, Clement F, et al. Efficacy of statins for
average cholesterol levels. Cholesterol and Recurrent Events primary prevention in people at low cardiovascular risk: a meta-
Trial investigators. N Engl J Med. 1996;335:1001–9. analysis. CMAJ. 2011;183:11.
2. Amarenco P, Bogousslavsky J, Callahan A 3rd, et al. High-dose 24. Chan DK, O’Rourke F, Shen Q, Mak JC, Hung WT. Meta-
atorvastatin after stroke or transient ischemic attack. N Engl J analysis of the cardiovascular benefits of intensive lipid lowering
Med. 2006;355:549–59. with statins. Acta Neurol Scand. 2011;124:188–95.
3. Randomised trial of cholesterol lowering in. 4444 patients with 25. Mills EJ, Wu P, Chong G, et al. Efficacy and safety of statin
coronary heart disease: the Scandinavian Simvastatin Survival treatment for cardiovascular disease: a network meta-analysis of
Study (4S). Lancet. 1994;344:1383–9. 170,255 patients from 76 randomized trials. QJM.
4. Schillinger M, Exner M, Mlekusch W, et al. Statin therapy 2011;104:109–24.
improves cardiovascular outcome of patients with peripheral 26. Taylor F, Ward K, Moore TH, et al. Statins for the primary
artery disease. Eur Heart J. 2004;25:742–8. prevention of cardiovascular disease. Cochrane Database Syst
5. Gutierrez J, Ramirez G, Rundek T, Sacco RL. Statin therapy in Rev. 2011. doi:10.1002/14651858.CD004816.pub4.
the prevention of recurrent cardiovascular events: a sex-based 27. Baigent C, Blackwell L, Emberson J, et al. Efficacy and safety of
meta-analysis. Arch Intern Med. 2012;172:909–19. more intensive lowering of LDL cholesterol: a meta-analysis of
6. Kjekshus J, Apetrei E, Barrios V, et al. Rosuvastatin in older data from 170,000 participants in 26 randomised trials. Lancet.
patients with systolic heart failure. N Engl J Med. 2010;376:1670–81.
2007;357:2248–61. 28. Ray KK, Seshasai SR, Erqou S, et al. Statins and all-cause
7. Tavazzi L, Maggioni AP, Marchioli R, et al. Effect of rosuvas- mortality in high-risk primary prevention: a meta-analysis of 11
tatin in patients with chronic heart failure (the GISSI-HF trial): a randomized controlled trials involving 65,229 participants. Arch
randomised, double-blind, placebo-controlled trial. Lancet. Intern Med. 2010;170:1024–31.
2008;372:1231–9. 29. Kizer JR, Madias C, Wilner B, et al. Relation of different mea-
8. Fellstrom BC, Jardine AG, Schmieder RE, et al. Rosuvastatin and sures of low-density lipoprotein cholesterol to risk of coronary
cardiovascular events in patients undergoing hemodialysis. artery disease and death in a meta-regression analysis of large-
N Engl J Med. 2009;360:1395–407. scale trials of statin therapy. Am J Cardiol. 2010;105:1289–96.
9. Wanner C, Krane V, Marz W, et al. Atorvastatin in patients with 30. Bukkapatnam RN, Gabler NB, Lewis WR. Statins for primary
type 2 diabetes mellitus undergoing hemodialysis. N Engl J Med. prevention of cardiovascular mortality in women: a systematic
2005;353:238–48. review and meta-analysis. Prev Cardiol. 2010;13:84–90.
10. Prasad V. Statins primary prevention overall mortality. Ann 31. Mora S, Glynn RJ, Hsia J, MacFadyen JG, Genest J, Ridker PM.
Intern Med. 2014;160:867–9. doi:10.7326/M13-2974. Statins for the primary prevention of cardiovascular events in
11. Redberg RF, Katz MH. Reassessing benefits and risks of statins. women with elevated high-sensitivity C-reactive protein or dys-
N Engl J Med. 2012;367:776 (author reply). lipidemia: results from the Justification for the Use of Statins in
 C. Huded, V. Prasad

Prevention: An Intervention Trial Evaluating Rosuvastatin 46. Collins R, Armitage J, Parish S, Sleigh P, Peto R. MRC/BHF
(JUPITER) and meta-analysis of women from primary prevention Heart Protection Study of cholesterol-lowering with simvastatin
trials. Circulation. 2010;121:1069–77. in 5963 people with diabetes: a randomised placebo-controlled
32. Petretta M, Costanzo P, Perrone-Filardi P, Chiariello M. Impact trial. Lancet. 2003;361:2005–16.
of gender in primary prevention of coronary heart disease with 47. Sever PS, Dahlof B, Poulter NR, et al. Prevention of coronary and
statin therapy: a meta-analysis. Int J Cardiol. 2010;138:25–31. stroke events with atorvastatin in hypertensive patients who have
33. Sheng X, Wei L, Murphy MJ, MacDonald TM. Statins and total average or lower-than-average cholesterol concentrations, in the
(not LDL) cholesterol concentration and outcome of myocardial Anglo-Scandinavian Cardiac Outcomes Trial-Lipid Lowering
infarction: results from a meta-analysis and an observational Arm (ASCOT-LLA): a multicentre randomised controlled trial.
study. Eur J Clin Pharmacol. 2009;65:1071–80. Lancet. 2003;361:1149–58.
34. Brugts JJ, Yetgin T, Hoeks SE, et al. The benefits of statins in 48. Margolis KL, Dunn K, Simpson LM, et al. Coronary heart disease
people without established cardiovascular disease but with car- in moderately hypercholesterolemic, hypertensive black and non-
diovascular risk factors: meta-analysis of randomised controlled black patients randomized to pravastatin versus usual care: the
trials. Bmj. 2009. doi:10.1136/bmj.b2376. antihypertensive and lipid lowering to prevent heart attack trial
35. Amarenco P, Labreuche J. Lipid management in the prevention (ALLHAT-LLT). Am Heart J. 2009;158:948–55.
of stroke: review and updated meta-analysis of statins for stroke 49. Taylor F, Huffman MD, Macedo AF, et al. Statins for the primary
prevention. Lancet Neurol. 2009;8:453–63. prevention of cardiovascular disease. Cochrane Database Syst
36. Navaneethan SD, Pansini F, Perkovic V, et al. HMG CoA Rev. 2013;1:CD004816.
reductase inhibitors (statins) for people with chronic kidney 50. Daniels SR, Greer FR. Committee on N. Lipid screening and
disease not requiring dialysis. Cochrane Database Syst Rev 2009. cardiovascular health in childhood. Pediatrics. 2008;122:198–208.
doi:10.1002/14651858.CD007784. 51. FDA Drug Safety Communication: Important safety label chan-
37. Delahoy PJ, Magliano DJ, Webb K, Grobler M, Liew D. The ges to cholesterol-lowering statin drugs. http://www.fda.gov/
relationship between reduction in low-density lipoprotein Drugs/DrugSafety/ucm293101.htm. Accessed 3 June 2015.
cholesterol by statins and reduction in risk of cardiovascular 52. Rojas-Fernandez CH, Cameron JC. Is statin-associated cognitive
outcomes: an updated meta-analysis. Clin Ther. 2009;31:236–44. impairment clinically relevant? A narrative review and clinical
38. Mills EJ, Rachlis B, Wu P, Devereaux PJ, Arora P, Perri D. recommendations. Ann Pharmacother. 2012;46:549–57.
Primary prevention of cardiovascular mortality and events with 53. Bonovas S, Sitaras NM. Does pravastatin promote cancer in
statin treatments: a network meta-analysis involving more than elderly patients? A meta-analysis. CMAJ. 2007;176:649–54.
65,000 patients. J Am Coll Cardiol. 2008;52:1769–81. 54. Alsheikh-Ali AA, Trikalinos TA, Kent DM, Karas RH. Statins,
39. Strippoli GF, Navaneethan SD, Johnson DW, et al. Effects of low-density lipoprotein cholesterol, and risk of cancer. J Am Coll
statins in patients with chronic kidney disease: meta-analysis and Cardiol. 2008;52:1141–7.
meta-regression of randomised controlled trials. BMJ. 2008;336: 55. FDA adds diabetes warning to statin label. http://www.
645–51. medpagetoday.com/Cardiology/Dyslipidemia/31408. Accessed 3
40. Messerli FH, Pinto L, Tang SS, et al. Impact of systemic June 2015.
hypertension on the cardiovascular benefits of statin therapy—a 56. Newman TB, Hulley SB. Carcinogenicity of lipid-lowering
meta-analysis. Am J Cardiol. 2008;101:319–25. drugs. JAMA. 1996;275:55–60.
41. Kearney PM, Blackwell L, Collins R, et al. Efficacy of choles- 57. Bjerre LM, LeLorier J. Do statins cause cancer? A meta-analysis
terol-lowering therapy in 18,686 people with diabetes in 14 of large randomized clinical trials. Am J Med. 2001;110:716–23.
randomised trials of statins: a meta-analysis. Lancet. 2008;371: 58. Wagstaff LR, Mitton MW, Arvik BM, Doraiswamy PM. Statin-
117–25. associated memory loss: analysis of 60 case reports and review of
42. O’Regan C, Wu P, Arora P, Perri D, Mills EJ. Statin therapy in the literature. Pharmacotherapy. 2003;23:871–80.
stroke prevention: a meta-analysis involving 121,000 patients. 59. Goldfine AB. Statins: is it really time to reassess benefits and
Am J Med. 2008;121:24–33. risks? N Engl J Med. 2012;366:1752–5.
43. Henyan NN, Riche DM, East HE, Gann PN. Impact of statins on risk 60. Jorgensen AW, Hilden J, Gotzsche PC. Cochrane reviews com-
of stroke: a meta-analysis. Ann Pharmacother. 2007;41:1937–45. pared with industry supported meta-analyses and other meta-anal-
44. Nakamura H, Arakawa K, Itakura H, et al. Primary prevention of yses of the same drugs: systematic review. BMJ. 2006;333:782.
cardiovascular disease with pravastatin in Japan (MEGA Study): a 61. Catalá-López F, Sanfélix-Gimeno G, Ridao M, Peiró S. When are
prospective randomised controlled trial. Lancet. 2006;368:1155–63. statins cost-effective in cardiovascular prevention? A systematic
45. Ridker PM, Wilson PW. A trial-based approach to statin guide- review of sponsorship bias and conclusions in economic evalu-
lines. JAMA. 2013;310(11):1123–4. ations of statins. PloS one. 2013;8:e69462.