Open access Research

Statins for the primary prevention of

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
cardiovascular disease: an overview of
systematic reviews
Paula Byrne,1 John Cullinan,1 Amelia Smith,2 Susan M Smith3

To cite: Byrne P, Cullinan J, Abstract

Smith A, et al. Statins for Objective  To synthesise evidence from exclusively primary
Strengths and limitations of this study
the primary prevention of prevention data on the effectiveness of statins for prevention
cardiovascular disease: ►► Overview of systematic reviews that reviewed ex-
of cardiovascular disease (CVD), including stroke, and
an overview of systematic clusively primary prevention data on statins.
outcomes stratified by baseline risk and gender.
reviews. BMJ Open ►► Transparent search strategy, published protocol and
2019;9:e023085. doi:10.1136/ Design  Overview of systematic reviews (SRs) using
validated instruments to assess the methodological
bmjopen-2018-023085 Revised-AMSTAR approach to assess quality.
quality of included reviews.
Data sources  Cochrane Database of Systematic Reviews,
►► Prepublication history and ►► Synthesised evidence from systematic reviews with
MEDLINE, Embase, PubMed, Scopus and PROSPERO to
additional material for this a mixture of individual patient and aggregate out-
June 2017.
paper are available online. To come data with an overlap of included randomised
Eligibility criteria for selecting studies  SRs of
view these files, please visit controlled trials across reviews.
the journal online (http://​dx.​doi.​ randomised control trials (RCTs) or individual patient data
►► Some relevant reviews may have been excluded
org/​10.​1136/​bmjopen-​2018-​ (IPD) from RCTs, examining the effectiveness of statins
because we could not ascertain the proportion of
023085). versus placebo or no treatment on all-cause mortality,
primary prevention participants within the system-
coronary heart disease, CVD (including stroke) and
Received 22 March 2018
atic review.
composite endpoints, with stratification by baseline risk
Revised 10 December 2018 ►► Limited data reporting reductions in various out-
and gender.
Accepted 12 December 2018 comes based on baseline risk and some studies
Data extraction and synthesis  Two independent
included participants considered risk equivalent to
reviewers extracted data and assessed methodological
those with established cardiovascular disease.
quality. A narrative synthesis was conducted.
Results  Three SRs were included. Quality of included
SRs was mixed, and none reported on the risk of bias of
included trials.  We found trends towards reduced all- factors for cardiovascular disease (CVD).1
cause mortality in all SRs (RR 0.91 [95% CI 0.85 to 0.97]), Statins, or 3-hydroxy-3-methylglutaryl co-en-
(RR 0.91 [95% CI 0.83 to 1.01]) and (RR 0.78 [95% CI zyme A reductase inhibitors, are a class of
0.53 to 1.15]) though it was not statistically significant in lipid-lowering drugs and are first choice
two SRs. When stratified by baseline risk, the effect on agents for reducing plasma LDL cholesterol
all-cause mortality was no longer statistically significant and thereby reducing CVD risk.2 Statins may
except in one medium risk category. One review reported be used for the primary or secondary preven-
© Author(s) (or their significant reductions (RR 0.85 [95% CI 0.77 to 0.95]) tion of CVD. Primary prevention comprises
employer(s)) 2019. Re-use in vascular deaths and non-significant reductions in
permitted under CC BY-NC. No treating people who do not have estab-
non-vascular deaths (RR 0.97 [95% CI 0.88 to 1.07]).
commercial re-use. See rights lished CVD but who may be at risk of future
There were significant reductions in composite outcomes
and permissions. Published by CVD events, whereas secondary prevention
BMJ. overall, but mixed results were reported in these when
stratified by baseline risk. These reviews included studies involves treating those with established CVD.3
1
J.E. Cairnes School of Business The last 30 years have seen a large increase
and Economics, National with participants considered risk equivalent to those with
University of Ireland Galway, established CVD. in the utilisation of statins,4–7 which is consis-
Galway, Ireland Conclusions  There is limited evidence on the tent with changes in recommendations
2
Department of Pharmacology effectiveness of statins for primary prevention with mixed in clinical guidelines.8 However, there is
and Therapeutics, University of findings from studies including participants with widely ongoing debate about these changes,9–11 as
Dublin Trinity College, Dublin, ranging baseline risks. Decision making for the use of they have tended to expand the number of
Ireland statins should consider individual baseline risk, absolute
3
HRB Centre for Primary Care
people eligible for treatment, particularly in
risk reduction and whether risk reduction justifies potential primary prevention.12 13 Our previous anal-
Research and Department of
harms and taking a daily medicine for life.
General Practice, Royal College ysis found that almost two-thirds of people
Trial registration number CRD42017064761.
of Surgeons in Ireland, Dublin, who were taking statins did so for primary
Ireland prevention.14
Correspondence to Introduction   Several SRs investigating the use of statins
Ms Paula Byrne; Raised total cholesterol (TC) and low-den- for the primary prevention of CVD have been
​pbyrne82@​gmail.​com sity lipoprotein (LDL) cholesterol are risk published reaching varying conclusions.15–20

Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085 1

Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
However, most published SRs reported on trials that
Box 1  Patients, Interventions, Comparators, Outcomes,
included a proportion of participants with a history of
Timing, Setting and Study design
CVD.21–23 In addition, the primary prevention population
is heterogeneous, ranging from those at very low risk of Patients
CVD to those considered ‘risk equivalent’ to those in the ►► Adults >18 years of age.
secondary prevention category. The latter, for example, ►► Without established cardiovascular disease (CVD).
includes people with diabetes mellitus exhibiting target Interventions
organ damage or people with chronic kidney disease.1 ►► Statins (3-hydroxy-3-methylglutaryl co-enzyme A reductase
The net benefit or absolute risk reduction achieved with inhibitors).
statin therapy is critically dependent on the baseline risk. Comparators
Therefore, the outcomes reported in the SRs that were ►► Placebo.
►► Control.
stratified by baseline risk or by gender were of particular
Outcomes
interest, these data being the most pertinent to clinical
►► All-cause mortality.
decision making with individual patients. ►► Fatal and non-fatal coronary heart disease (CHD), CVD and stroke
To address this evidence gap and support decision events.
making, we undertook an overview of SRs that reported ►► Combined endpoint (fatal and non-fatal CHD, CHD and stroke
on exclusively primary prevention trials or individual events).
patient data (IPD) of trial participants using only data ►► Any of the above outcomes stratified by a calculation of future risk
from patients without established CVD. of CVD or by gender.
Timing
►► Studies of any duration.
Setting and study design
Methods
►► Systematic reviews of randomised controlled trials.
This overview was conducted according to the methods of
►► Systematic reviews of individual patient data.
the Cochrane Handbook for Systematic Reviews of Inter-
ventions.24 The protocol for the overview was published
on PROSPERO.25
We included any SR of RCTs or IPD from RCTs, in effectiveness of statins versus placebo or no treatment
any language, which examined the effectiveness of exclusively in those without prior CVD. We extracted data
statins versus placebo or no treatment exclusively in on outcomes of relevance; all-cause mortality; CHD, CVD
those without prior CVD. We searched for the following and stroke events; composite endpoints; and any of these
outcomes: (1) all-cause mortality; (2) fatal and non-fatal outcomes stratified by a calculation of future risk of CVD
coronary heart disease (CHD) and CVD events including or by gender. Both reviewers independently assessed the
stroke; (3) composite endpoints; and (4) any of these methodological quality of the included reviews using the
outcomes stratified by a calculation of future risk of CVD R-AMSTAR tool.28 One of the included SRs, published
or by gender (see box 1). by the Cholesterol Treatment Trialists’ Collaboration
The search strategy, terms and databases were chosen (CTT), consisted of analyses of IPD and was reported in
with the assistance of a health sciences librarian and are two publications based on the same trials, one reporting
described in online supplementary appendix 1, tables 1 results overall29 and one that included analyses stratified
and 2. We searched the Cochrane Database of System- by gender.30 Some of the methods were described not in
atic Reviews, MEDLINE, Embase, PubMed, Scopus and the SR itself but in other referenced CTT papers and in
PROSPERO from the date of the first statin’s approval in the CTT protocol, which was published in 1995.31–33 As we
198726 to June 2017. found some limitations in using R-AMSTAR in the context
Two overview authors (PB and AS) independently of IPD, we further assessed CTT and the secondary
screened search results by title and abstract and obtained papers describing their methodology, using the Preferred
full-text versions of the articles identified by both as Reporting Items for Systematic Reviews and Meta-analyses
potentially relevant. PB and AS selected relevant articles of Individual Participant Data (PRISMA-IPD) checklist.34
by reading the full texts and applying the inclusion and We searched the SRs for any assessments of the quality
exclusion criteria. PB and AS used Covidence systematic of evidence of included trials such as GRADE, Cochrane
review software to manage the searches and extraction Risk of Bias or the Jadad Scale. We undertook a narrative
of data.27 Any differences of opinion on inclusion were synthesis of the included reviews and summarised their
resolved by consulting another overview author (SMS or main results on the effectiveness of statins regarding the
JC). outcomes of relevance and those outcomes stratified by
baseline risk and gender.
Data collection and analysis
PB performed data extraction, while AS independently Patient and public involvement
checked the extracted data. SRs were extracted that There was no patient or public involvement in this over-
fulfilled inclusion criteria, that is, SRs of RCTs or IPD view, and patient and public involvement was not reported
from RCTs, in any language, which examined the in the included systematic reviews.

2 Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085

 Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
Figure 1  Flow chart of included systematic reviews.  CTT, Cholesterol Treatment Trialists’ Collaboration; IPD, individual patient
data.

Results that had been included. Some primary prevention SRs

Search results included trials with up to 50% of participants with CVD,35
Our initial searches yielded 1462 results of which 181 were while the most up-to-date Cochrane review included trials
evaluated as full-text articles following title and abstract with up to 10% of participants having CVD.21
screening. Thirty-one full-text articles were analysed in
detail. We supplemented the electronic search by scan- Characteristics of included studies
ning the reference lists of the full-text articles we read. Of the three included SRs, two comprised analyses of
No further SRs were identified. Reasons for exclusion and aggregate data from RCTs,15 36 and one29 30 presented
references to 27 final excluded articles are presented in analyses of IPD from the CTT. Table 1 describes
online supplementary appendix 2. On the basis of our the characteristics of the included reviews. Table 2
search and extraction strategy, three SRs were included in describes the population demographics reported in
this overview. One of the included SRs was reported in two each SR.
separate publications but included the same trial data, so The CTT analyses included IPD from 22 trials of statin
was treated as a single systematic review for the purpose versus control as well as five of more versus less statin.
of this overview (see figure 1). Many excluded SRs did However, the trials of more versus less statin did not
not specify the proportion of participants without CVD include any primary prevention participants; therefore,

Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085 3

4
Open access

Table 1  Characteristics of the included systematic reviews

Quality of
included
Included studies reviews
Mean Quality of
follow-up evidence of Revised- Subgroup Last
duration included trials AMSTAR Total Baseline risk of Outcomes analysis of search
Review (year) Population Intervention Comparator RCTs (n) (years) GRADE score participants participants reported outcomes Synthesis date

CTT Men and Statin Control 22 4.8 NR 27 1 34 537 <5%; Any deaths. All outcomes IPD meta- 2011
2012 and 2015 women ≥5%–<10%; Any vascular by baseline risk analysis
without ≥10%–<20%; death. profile.
prevalent CVD ≥20%–<30% and Non-vascular Major vascular
at baseline. ≥30%. death. events by
Major coronary gender.
events. Major
vascular events.
Mora et al 2010 Women Statin Control 3 >1 NR 19 13 154 NA Total CVD in RCT meta- 2009
without women. Total analysis.
prevalent CVD mortality in
at baseline. women.
Ray et al 2010 Men and Statin Control 11 3.7 NR 32 65 229 NA All-cause RCT meta- 2009
women mortality. analysis.
without
prevalent CVD
at baseline.

CTT, Cholesterol Treatment Trialists’ Collaboration; CVD, cardiovascular disease; IPD, individual patient data; NA, not applicable; NR, not reported; RCT, randomised control trial.

Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
 Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
Table 3  Trial overlap of included systematic reviews

‡This proportion was calculated from Mora et al. Statins for the primary prevention of cardiovascular events in women with elevated high-sensitivity C-reactive protein or dyslipidaemia results
†This proportion was calculated from Fulcher et al. Cholesterol Treatment Trialists’ (CTT) Collaboration. Efficacy and safety of LDL-lowering therapy among men and women: meta-analysis of
LDL-C mmol/L*
Included trials CTT Mora Ray

3.6 (women) 3.6

from the Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER) and meta-analysis of women from primary prevention trials. Circulation.
AFCAPS/TexCAPS ✓ ✓ ✓*
ALERT ✓

(men)
NR
3.6
ALLHAT-LLT ✓ ✓†
ALLIANCE ✓
ASCOT-LLA ✓ ✓†
141.2 (women)

ASPEN ✓ ✓†
139.3 (men)
Mean SBP

AURORA
mm Hg

✓
CARDS
141

✓ ✓
NR

CARE ✓
15.4 (women) 9.4

CORONA ✓
CTT, Cholesterol Treatment Trialists’ Collaboration; LDL-C, low-density lipoprotein cholesterol; NR, not reported; SBP, systolic blood pressure. 4D ✓
Smokers

GISSI-HF ✓
(men)

GISSI-P ✓
NR
23
%

HPS ✓
24.5 (women) 19.4

HYRIM ✓*
With diabetes

JUPITER ✓ ✓ ✓
LIPID ✓
individual data from 1 74 000 participants in 27 randomised trials. Lancet. 2015;385 (9976):1397–405 of table 2.
(men)

LIPS ✓
7‡
19
%

MEGA ✓ ✓ ✓
Post-CABG ✓
PREVEND-IT ✓†
65.3 (women)

PROSPER ✓ ✓†
62.0 (men)
Mean age

4S ✓
64‡

WOSCOPS
62

✓ ✓
*Provided hitherto unpublished tabular data on all-cause mortality.
†Shared tabular data on subset of participants without CVD.
% Male

CTT, Cholesterol Treatment Trialists’ Collaboration; CVD,

cardiovascular disease.
0
59†

65

*SI conversion factor: to convert LDL-C to mg/dL divided by 0.0259.

Women and men

Women and men

we assume, although it is not stated in the SR, that all anal-

yses in those with ‘no known history of vascular disease’
are taken from only the statin versus control trials, and we
Gender

Women

included only these analyses in our overview as per our

protocol. Online supplementary appendix 3 describes
analyses from the included SRs that were not included in
Table 2  Population demographics

Participants (n)

this overview.
Ray et al15 included 11 trials with additional unpub-
2010;121(9) :1069–77 of figure 2.

lished data from authors that provided data on primary

13 154
65 229
134 537

prevention participants. Mora et al36 included three exclu-

sively primary prevention trials in women only. Two types
of data therefore are included in our overview, IPD and
aggregate data from trials.
Mora et al 2010

The overlap of included reviews is reported in table 3.

Ray et al 2010
CTT 2012 and

While all SRs included overlapping RCTs, IPD were used

by CTT, Ray et al reported previously unpublished data,
Review

2015

and outcomes were reported differently across the three

included systematic reviews.

Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085 5

Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
Outcomes reported

Total
All three SRs reported outcomes for all-cause mortality,

27
19
32
one of which, Mora et al, was in women only. CTT reported
outcomes for vascular deaths, non-vascular deaths, major

conflict of

included?
Was the

interest
coronary events (defined as non-fatal myocardial infarc-
tion [MI] or coronary death) and major vascular events

2
3
2
(defined as major coronary events, coronary revasculari-

likelihood of
publication

assessed?
sation and stroke).

Was the
In addition, all outcomes reported by CTT were strat-

bias
ified by the participants’ 5-year vascular risk at baseline.

4
1
1
The risk categories reported in CTT describe a person’s

methods used
estimated 5-year risk of having a major vascular event and

appropriate?
the findings
to combine
were stratified as follows: <5%; ≥5%–<10%; ≥10%–<20%;

of studies
Were the
≥20%–<30%; and ≥30%. This method of calculating risk
was modelled by CTT, and how these categories relate to

3
4
3
more commonly used methods such as SCORE, QRISK

methods used

the findings
and Framingham is unclear. However, Robinson et al37 esti-

appropriate
to combine

of studies
Were the
mated CTT’s ≥5%–<10% 5-year vascular risk to be nearly
identical to the more standard US 10-year atherosclerotic

4
2
2
CVD (ASCVD) event rate. The outcome major vascular

Was the scientific

events was also stratified by gender by CTT, and Mora et

included studies

documented?
assessed and
al reported results for the composite outcome total CVD

quality of the
events in women. (See online supplementary appendix 3
for outcomes that were reported in the included SRs but
which we did not include in this overview.)

3
2
2
characteristics
of the included
Quality assessment of the included systematic reviews
R-AMSTAR

provided?
Were the

studies
The CTT SR received a R-AMSTAR score of 27. The
reviews by Mora et al and Ray et al were assigned ratings

3
4
1
of 19 and 32, respectively, out of a maximum score of 44

(included and
(see table 4). of studies

excluded)
Was a list

provided
In general, the included SRs scored lowest in criteria
describing search strategies, excluded studies and reasons

2
for exclusion. Neither Mora et al nor CTT clearly reported
2
1
their search strategy, methods of study selection or
comprehensive literature) used
literature search as an inclusion
Was the status
of publication

extraction or provided lists and characteristics of excluded

studies; nor how disagreements among extractors were
criterion?
(ie, grey

resolved or if two independent researchers extracted data

Table 4  R-AMSTAR assessment per systematic review

from the included studies. Across the included SRs, the

1
2
2

highest scoring criterion was for the statistical methods of

combining included studies.
performed?

Preferred Reporting Items for Systematic Reviews and Meta-

Was a

analyses of Individual Participant Data

CTT, Cholesterol Treatment Trialists’ Collaboration.
3
2
1

Using PRISMA-IPD as well as R-AMSTAR, we were

study selection

still unable to assess potential risk of bias in the CTT

extraction?
R-AMSTAR score per review
Was there
duplicate

and data

SR including how IPD were collected, requested and

managed, including whether IPD were sought and not
available from other trials, the full electronic search
4
1
1
R-AMSTAR criteria

strategy, details of databases searched, methods for

be established

conduct of the
criteria should
priori’ design

The research
question and

resolving disagreements between those extracting studies,

before the
provided?
Was an ‘a

inclusion

lists and characteristics of excluded studies, assessment

review.

of risk of bias in included RCTs and assessment of publi-

cation bias. It should be noted that this checklist was
3
4
Mora et al 3

published in 2015, which is after the publication of CTT’s

Ray et al
Review

SR. However, it would seem reasonable to expect that all

CTT

items included in the PRISMA-IPD checklist would be

6 Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085

 Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
incorporated in IPD analysis. The CTT protocol notes on gender and baseline cardiovascular risk. The main
that publication bias can be avoided by ‘prospectively plan- outcomes reported were all-cause mortality, vascular and
ning an overview based on individual patient data from all non-vascular deaths and the composite outcomes of total
relevant randomized trials’. However, no tests for publica- CVD events, major coronary events and major vascular
tion bias were reported. Ray et al was the only included events. CTT reported a significant reduction in all-cause
SR that reported on publication bias and assessed this mortality, but no significant reductions were found in the
using a funnel plot and Egger’s test and found no strong two other systematic reviews for this outcome. Though
evidence of publication bias (p=0.50). point estimates are very similar, the difference in statistical
significance may be due to the numbers of participants
Risk of bias from the primary RCTs included in SRs included in each analysis, suggesting an issue with statis-
None of the SRs reported on the risk of bias in their tical power. It may be the case that the smaller analysis did
included primary trials using GRADE, Cochrane Risk of not have sufficient statistical power though, conversely, it
Bias or the Jadad Scale. can be argued that when an analysis includes very large
numbers, minimal clinical effects can reach statistical
Effectiveness of statins significance.38 39 However, when CTT stratified results
Due to the variability in the reviews and different by baseline risk profile, non-significant reductions were
outcomes reported, we have not attempted to combine reported for all-cause mortality in all but one level of risk.
or re-meta-analyse results and have presented a narrative As noted, the overall reduction in any deaths reported
synthesis as an overview of all results. We found a trend in the reviews includes participants who, although cate-
towards reduced all-cause mortality in the three system- gorised as ‘primary prevention’, may include those who
atic reviews, though only one of the three showed a statis- are risk equivalent to people with established vascular
tically significant difference. CTT reported statistically disease, such as people with diabetes and chronic kidney
significant relative risk (RR) reductions in ‘any deaths’ disease. Because of this limitation, arguably the results
(RR 0.91 [95% CI 0.85 to 0.97]). Ray et al conducted both relevant to low-risk people are those specific to their base-
fixed and random effects meta-analyses, with and without line risk category (<5% and ≥5%–<10%), rather than the
two trials reporting results for people with diabetes. aggregate results reported.
There were no statistically significant reductions in In an attempt to specify risk reductions in the lowest
all-cause mortality in any of the meta-analyses: random risk people included in the CTT analysis, that is, in
effects models (RR 0.91 [95% CI 0.83 to 1.01]) including those for whom statins were not already recommended
diabetes trials; fixed effects models (RR 0.93 [95% CI 0.86 because of CHD risk equivalence, Abramson et al12
to 1.00]) including diabetes trials; random effects models reanalysed data from this review for those whose 5-year
(RR 0.92 [95% CI 0.84 to 1.02]) excluding diabetes trials; risk was <5% and for those whose risk was ≥5%–<10%.
fixed effects models (RR 0.94 [95% CI 0.86 to 1.01]) They found there was no significant effect on mortality
excluding diabetes trials. Mora et al found no significant in this group of patients (RR 0.95 [95% CI 0.86 to 1.04]).
reduction in total mortality (RR 0.78 [95% CI 0.53 to However, Abramson et al’s analysis included participants
1.15]) in women. with and without vascular disease. It could be expected
CTT reported significant reductions in any vascular that the effect of statins would be seen most clearly in the
death (RR 0.85 [95% CI 0.77 to 0.95]) and non-signifi- outcome of vascular deaths. However, while an overall
cant reductions in non-vascular death (RR 0.97 [95% CI significant reduction was reported for this outcome by
0.88 to 1.07]). the CTT, non-significant reductions were reported at
In addition, CTT reported results stratified by base- all levels of risk when stratified by baseline risk profile.
line risk category. There were non-significant reductions Significant reduction in major vascular events and major
reported by CTT in ‘any deaths’ except at one level of risk. coronary events were reported and mixed results for these
Non-significant results for vascular deaths were reported outcomes when stratified by gender and baseline risk
in all risk categories. Three risk categories were found profile. It should be noted, however, that CTT’s reporting
to have non-significant increases in non-vascular deaths, of the composite outcome major vascular events is an
while two had non-significant reductions in the outcome. addition to the outcomes prespecified in their protocol.
There were reductions reported in all risk categories for No stroke outcomes were reported in any of the included
major coronary events and major vascular events when SRs except as part of composite outcomes.
stratified by baseline risk category, but these were not
statistically significant in the two highest risk categories Strengths and limitations of the overview
(see table 5). As far as we are aware, this is the first overview of SRs
that investigates statins in an exclusively primary preven-
tion population. We synthesised evidence from SRs,
Discussion which are considered the highest quality evidence for
Principal findings healthcare interventions. We used a transparent search
Three SRs were included in this overview reporting a mix strategy and followed a published protocol25 to guide
of aggregate and IPD data and a range of reporting based our search, extraction and analysis. We used validated

Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085 7

8
Table 5  Reported results of the included systematic reviews
Review Reported results
All cause mortality Any vascular death Non vascular death Major coronary events Major vascular events Total CVD
Open access

CTT 2012 and 2015 Overall RR 0.91 (95% CI 0.85 to RR 0.85 (95% CI 0.77 RR 0.97 (95% CI 0.88 to RR 0.71 (95% CI 0.65 RR 0.75 (95% CI 0.70 to 0.80),
0.97), p=0.007 to 0.95), p=0.04 1.07), p=0.6 to 0.77), p=0.0001 p=0.0001
Stratified by baseline
risk profile:
  <5% RR 0.94 (95% CI 0.71 RR 0.80 (95% CI 0.43 RR 1.13 (95% CI 0.76 RR 0.59 (95% CI 0.37 RR 0.61 (95% CI 0.45 to 0.81)
to 1.26) to 1.47) to 1.69) to 0.96)
  ≥5%–<10% RR 0.83 (95% CI 0.69 RR 0.75 (95% CI 0.55 RR 0.87 (95% CI 0.67 RR 0.58 (95% CI 0.48 RR 0.66 (95% CI 0.57 to 0.77)
to 0.99) to 1.04) to 1.11) to 0.72)
  ≥10%–<20% RR 0.88 (95% CI 0.76 RR 0.84 (95% CI 0.67 RR 0.94 (95% CI 0.76 RR 0.78 (95% CI 0.65 RR 0.82 (95% CI 0.72 to 0.93)
to 1.02) to 1.05) to 1.15) to 0.93)
≥20%–<30% RR 1.06 (95% CI 0.86 RR 0.97 (95% CI 0.72 RR 1.13 (95% CI 0.81 RR 0.80 (95% CI 0.60 RR 0.81 (95% CI 0.65 to 1.01)
to 1.32) to 1.32) to 1.57) to 1.06)
  ≥30% RR 0.94 (95% CI 0.70 RR 0.88 (95% CI 0.59 RR 1.07 (95% CI 0.68 RR 0.76 (95% CI 0.50 RR 0.83 (95% CI 0.58 to 1.18)
to 1.25) to 1.33) to 1.69) to 1.17)
Stratified by gender:
  Men RR 0.72 (95% CI 0.66 to 0.80)
  Women RR 0.85 (95% CI 0.72 to 1.00)
Mora et al
Stratified by gender
Women RR 0.78 (95% CI 0.53 RR 0.63 (95% 
to 1.15) CI 0.49  to  0.82)
Ray et al 2010
 
Stratified by inclusion
criteria
  Including diabetes trials Random effects model:
RR 0.91 (95% CI 0.83
to 1.01)
Fixed effects model:
RR 0.93 (95% CI 0.86
to 1.00)
  Excluding diabetes Random effects model:
trials RR 0.92 (95% CI 0.84
to 1.02)
Fixed effects model:
RR 0.94 (95% CI 0.86
to 1.01)

CTT, Cholesterol Treatment Trialists’ Collaboration; CVD, cardiovascular disease.

Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
 Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
instruments to assess the methodological quality of Clinical implications
included reviews. The main limitation of the overview is Some studies have shown that, in absolute terms, the
the need to synthesise evidence from SRs with a mixture majority of statins users are in the primary prevention
of IPD and aggregate outcome data and the high level of category.50 A higher proportion of women who take
overlap of included RCTs across reviews. Some trials are statins fall into the primary prevention category than
part of two or three of the included systematic reviews, men,50 51 and the distribution of statin prescribing has
whereas many others contribute only to the results of one shifted from secondary to primary prevention partic-
systematic review. Therefore, the results of the analysis ularly among women.52–54 Given the ongoing debate
presented here may be driven mostly by these ‘over-rep- on the appropriateness of statin use in primary preven-
resented’ populations. Some relevant reviews may have tion,51 55 it is surprising that so few systematic reviews
been excluded from this overview because we could not of exclusively primary prevention data exist. Clinical
ascertain the proportion of primary prevention partic- guidelines do not inform the physician whether recom-
ipants within the SR. Only one SR reported on risk of mended thresholds represent valid demarcation lines in
bias of the included trials, and we found some of the terms of the individual patient,56 and decisions to take
R-AMSTAR criteria ambiguous and difficult to answer. In or prescribe a medicine involve a trade-off between the
addition, trials that fail to find significant benefit from an perceived benefits and harms of that medicine for the
intervention are often not published,40 and it is possible individual. This trade-off is particularly salient for low-risk
that, as a result, our overview may be affected by publi- people choosing to take a statin for primary prevention of
cation bias. As we did not retrieve data from primary CVD as the patient often feels healthy and may perceive
trials, we were limited to the information and judge- the medicine as unnecessary, with uncertain benefits and
ments of the included SR authors. For example, the potential side effects. Conversely, it may be the case that
more recent primary prevention study, Heart Outcomes clinicians and patients would desire a reduction in CVD,
Prevention Evaluation-3 (HOPE-3), reported significant regardless of how small, if they can tolerate statins. There-
reductions in composite cardiovascular outcomes for fore, for people at low-risk of CVD, it is important that the
those at ‘intermediate risk’ (defined as an annual risk of decision to prescribe or take statins is considered in terms
Major Vascular Events (MVEs) of approximately 1%).41 of absolute risk reduction to ensure the potential bene-
fits outweigh the potential harms in the context of that
Inclusion of this trial could have influenced the results
patient’s preferences. Unfortunately, some of the gaps
reported in the included reviews in those in the inter-
in the data we have presented here cannot be overcome.
mediate risk category. Two of the included reviews had
Only one included review stratified patients by risk and
important methodological issues, notably the absence
gender and one by gender only. The question remains
of clear reporting of search strategy and methods of
for the clinician: what is the relevant information for the
study selection. This in fact is the core definition of a
individual patient?
systematic review and warrants consideration of whether
Arguably, clinical decisions should be based on ‘hard’
the individual included studies were simply reviews, as
endpoints such as cardiovascular death, MI and stroke
opposed to systematic reviews. In addition, the use of
because these are least subject to bias in adjudication.12
the composite outcomes by CTT was not prespecified in As these outcomes were not reported separately in the
their 1995 protocol, and this change from the original overview, ‘all-cause mortality’ is the most reliable outcome
protocol is not acknowledged or justified in the paper, on which to base decisions. The use and reporting of
which may introduce bias.42 43 composite outcomes has been criticised as they may be
Despite calls to make the provision of clinical trial unreasonably combined, inconsistently defined and
data a legal, regulatory or ethical requirement,44 45 and inadequately reported.57 Reported risk reductions of
specifically for the publication of CTT’s IPD from statin composite outcomes may be driven by large reduction in
trials,46 47 CTT’s data, as well as much of the data from the less serious components of the outcome rather than
trials from the other two included reviews, remain unavail- the more serious. For example, if a composite outcome
able for independent analysis, and thus, the goal of fully comprises a larger proportion of ‘less serious’ outcomes
informed shared decision making cannot be achieved. In such as angina and revascularisations compared with MIs
addition, while CTT analyses include data from ‘almost or stroke, this may result in misleading impressions of the
all of the relevant randomised trials’,48 members of the impact of treatment.36 Mora et al analysed the components
BMJ expert advisory group on statins stated that they of the composite outcome total CVD events in women in
intended to contact the authors of 183 statin trials for one large trial included in their review and found that
additional published and unpublished data.49 The inclu- women had a significant reduction in revascularisations
sion of such data in systematic reviews and meta-analyses and unstable angina but not in other components of
may alter reported results. However, though the gaps in the composite outcome, including stroke. Patients or
the data cannot be overcome, such as the lack of trans- prescribers may alter their decision making about the
parency in the primary data, we believe this overview pres- potential benefits of statin use even though larger treat-
ents to patients and clinicians the best, although limited, ment effects may be associated with less important compo-
data available. nents.58 Data for each component of each composite

Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085 9

Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
outcome in the included reviews were not supplied and risk reduction, the next consideration should be the
some of the meta-analyses included composite outcomes, potential harms from taking statins. Collins et al have
which may be inappropriate.57 The details of composite reported that treating 10 000 patients for 5 years would
outcomes in the included SRs are described in online cause about five cases of myopathy, 50–100 new cases of
supplementary appendix 4. diabetes and 5–10 haemorrhagic strokes. The authors
For the individual patient and clinician, there are argue that the harmful effects of statin therapy can
three considerations in the process of informed decision usually be reversed without residual effect by stopping the
making. First, what is the RR reduction according to the statin therapy, whereas ‘harmful effects of heart attacks
baseline risk of the individual. Second, what is the abso- or strokes that occur because statin therapy has not been
lute risk reduction in risk for that person and, finally, used can be devastating’.60 Even if there are side effects
what are potential side effects from taking statins in the in a lower risk person, they may derive long-term benefit
context of that patient’s preferences. by stabilising or slowing the progression of subclinical
As we have outlined in this overview, the outcomes vascular disease. However, the scale of these benefits may
reported in the SRs that were stratified by baseline risk or not justify the potentially wider effect of medicalisation of
by gender may be the most pertinent to clinical decision low-risk individuals. In addition, this definition of myop-
making. For example, for a woman whose risk is <5%, athy, as described by Armitage et al,61 may be high bar
which results are relevant? Should she be presented for diagnosing muscle symptoms among real people who
with the relevant non-gender-specific results reported by may simply define myopathy as any muscle symptom, and
CTT, such as the overall results for relative reduction in observational data suggest that the frequency of statin
all-cause mortality (RR 0.91 [95% CI 0.85 to 0.97]) or the myopathy may be higher.62 63 Indeed, there is a large
non-significant relative reductions for those at her rele- difference between the quantification of muscle side
vant baseline risk (RR 0.94 [95% CI 0.71 to 1.26])? Or effects between Collins et al and Buettner et al, the former
should she be presented with the overall non-significant reporting five cases of myopathy per 10  000 treated
relative reduction presented by Mora et al for women (RR patients over 5 years, while the latter, a difference of 5.3%
0.78 [95% CI 0.53 to 1.15])? The same dilemma would between statin and placebo groups, which is the equiv-
arise for a high-risk woman, for example, one whose alent of 530 cases of ‘musculoskeletal pain’ per 10 000
baseline risk is 30% or greater. In this case, the relevant patients treated. Thus, the estimates vary by a factor of
risk reduction reported at her baseline risk was RR 0.94 100. However, a recent systematic review of observational
(95% CI 0.70 to 1.25). In a discussion on how to apply studies on statins use and new-onset diabetes noted that
results of systematic reviews to patient care, Murad et this association may be limited due to ‘indication bias’64;
al40 suggest that clinicians consider the upper and lower that is, the risk of an adverse event is related to the indica-
bounds of CIs. They can then consider how they would tion for medication use but not the use of the medication
advise their patients were the upper boundary to repre- itself. In an observational study, one can only observe the
sent the truth and how they would advise their patients effect of an exposure, in this case to statins. Groups are
were the lower boundary to represent the truth. not randomly assigned to treatment or placebo groups.
The included SRs reported reductions in risk of CVD Therefore, observational studies are much more vulner-
outcomes as RR reductions, but for an individual patient, able to confounding bias, and their results may be less
knowing their absolute risk reduction is more relevant robust than those of a randomised controlled trial.65 For
when making a decision to take a statin.40 Sun et al59 give example, prediabetes, the most important risk factor
a good example of two people for comparison. One is a for type 2 diabetes, is associated with dyslipidaemia, and
65-year-old man who smokes, does not have heart disease this increases both the chances that people with predia-
but who has high total cholesterol levels and elevated betes will be treated with statins and that these subjects
blood pressure. The second is a 45-year-old woman who will develop type 2 diabetes.66 Moreover, those with type
does not smoke, has elevated total cholesterol levels and 2 diabetes are considered risk equivalent to those in the
slightly elevated blood pressure. Based on the Amer- secondary prevention category and their treatment with
ican College of Cardiology/American Heart Association statins recommended by clinical guidelines.1 Consider-
(ACC/AHA) risk calculator, the man has a 38% absolute ation of the potential risk of developing type 2 diabetes
risk of having a major coronary event in the next 10 years; from statin use is complex. The small risk of developing
the woman has a 1.4% absolute risk. According to the diabetes may be favourably balanced by the cardiovascular
risk reductions reported by CTT in this overview,29 statin benefit.67 In addition, although those with diabetes have
therapy would reduce the man’s RR of major coronary a higher cardiovascular event rate than those without, it
events by 24% and the woman’s RR by 41%. However, the may be the case that the event rate in those with new-onset
man could expect an absolute risk reduction of about 9% diabetes is lower than those with established diabetes at
(number needed to treat of 11) and the woman of 0.6% baseline.68 This would strengthen the argument that any
(number needed to treat of 166) (online supplementary potential risk of new-onset diabetes is outweighed by the
appendix 5). lowering of cardiovascular risk.
Having considered which RR reduction is most relevant A recent systematic review by Albarqouni et al69
to the particular individual and the associated absolute attempted to quantify the minimum acceptable risk

10 Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085

 Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
reduction that patients say is necessary to justify a daily authors had full access to the data and had final responsibility for the decision to
intake of medication to prevent CVD events. Our forth- submit this publication.
coming analysis (Byrne et al)70 of those considered low, Competing interests  None declared.
medium, high and very high risk according to the most Patient consent for publication  Not required.
recent 2016 guideline found that only some of those at Provenance and peer review  Not commissioned; externally peer reviewed.
high or very high risk would reach an acceptable level of Data sharing statement  No additional data are available.
risk reduction to justify taking a medicine for life. In addi- Open access This is an open access article distributed in accordance with the
tion, Albarqouni et al reported that in one study only 3% Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which
of community living older people would agree to CVD permits others to distribute, remix, adapt, build upon this work non-commercially,
preventative medicines if that medication had adverse and license their derivative works on different terms, provided the original work is
properly cited, appropriate credit is given, any changes made indicated, and the use
effects that could affect their activities of daily living and is non-commercial. See: http://​creativecommons.​org/​licenses/​by-​nc/​4.​0/.
half would not agree to take the medication if it was asso-
ciated with even mild fatigue or nausea.70

Comparison with other studies References

We are aware of only one other overview of systematic 1. Catapano AL, Graham I, De Backer G, et al. 2016 ESC/EAS
Guidelines for the Management of Dyslipidaemias. Eur Heart J
reviews on the topic of statins for the primary prevention 2016;37:2999–3058.
of CVD by Karmali et al.71 This overview considered drugs, 2. Taylor F, Huffman MD, Macedo AF, et al. Statins for the primary
prevention of cardiovascular disease. Cochrane Database Syst Rev
including statins, for the primary prevention of CVD and 2013;1.
reported statistically significant reductions in the risk of 3. Perk J, De Backer G, Gohlke H, et al. European Association
CVD for statins (RR 0.75 [95% CI 0.70 to 0.81]). The for Cardiovascular Prevention & Rehabilitation (EACPR); ESC
Committee for Practice Guidelines (CPG). European Guidelines on
Karmali et al overview included three of the four reviews cardiovascular disease prevention in clinical practice (version 2012).
from our overview.15 29 30 However, no attempt was made The Fifth Joint Task Force of the European Society of Cardiology
and Other Societies on Cardiovascular Disease Prevention in Clinical
to disaggregate exclusively primary prevention data, and Practice (constituted by representatives of nine societies and by
therefore, it cannot be compared with our overview. invited experts). Eur Heart J 2012;33:1635–701.
4. Walley T, Folino-Gallo P, Stephens P, et al. Trends in prescribing and
utilization of statins and other lipid lowering drugs across Europe
1997-2003. Br J Clin Pharmacol 2005;60:543–51.
Conclusion 5. Feely J, McGettigan P, Kelly A. Growth in use of statins after trials
is not targeted to most appropriate patients. Clin Pharmacol Ther
This overview suggests there is mixed evidence on the 2000;67:438–41.
effectiveness of stains in primary prevention populations; 6. Kildemoes HW, Støvring H, Andersen M. Driving forces
however, this population is heterogeneous with widely behind increasing cardiovascular drug utilization: a dynamic
pharmacoepidemiological model. Br J Clin Pharmacol
ranging baseline risks. For the individual patient and 2008;66:885–95.
clinician, there are three considerations in the process 7. DeWilde S, Carey IM, Bremner SA, et al. Evolution of statin
prescribing 1994-2001: a case of agism but not of sexism? Heart
of informed decision making. First, what is the RR reduc- 2003;89:417–21.
tion according to the baseline risk of the individual. 8. Moynihan RN, Cooke GP, Doust JA, et al. Expanding disease
Second, what is the absolute risk reduction in risk for that definitions in guidelines and expert panel ties to industry: a cross-
sectional study of common conditions in the United States. PLoS
person, and finally, what are potential side effects from Med 2013;10:e1001500.
taking statins in the context of that patient’s preferences. 9. Greenland P, Bonow RO. Interpretation and use of another statin
guideline. JAMA 2016;316:1977–9.
The gaps in the data, outlined in this overview, cannot be 10. Redberg RF, Katz MH. Statins for primary prevention: the debate is
overcome without patient-level data being made available intense, but the data are weak. JAMA 2016;316:1979–81.
for independent analysis. However, this overview clearly 11. Unruh L, Rice T, Rosenau PV, et al. The 2013 cholesterol guideline
controversy: Would better evidence prevent pharmaceuticalization?
elucidates the information that is available at present Health Policy 2016;120.
allowing patients and clinicians to know that this is the 12. Abramson JD, Rosenberg HG, Jewell N, et al. Should people at low
risk of cardiovascular disease take a statin? BMJ 2013;347:f6123.
best, although limited, data available. 13. Goldacre B, Smeeth L. Mass treatment with statins. Br Med J
2014;349:g4745.
Acknowledgements  We would like to thank health sciences librarian Ms Jane 14. Byrne P, Cullinan P, Murphy C, et al. Cross-sectional analysis of
Mulligan and Ms Edel Murphy of the HRB Clinical Research Facility and the Primary the prevalence and predictors of statin utilisation in Ireland with a
Care Clinical Trials Network Ireland in NUI Galway. Thanks also to biostatistician focus on primary prevention of cardiovascular disease. BMJ Open
2018;8:e018524.
Dr Finbarr Lacey of RCSI, Dublin, for his help with the quality assessment tools. 15. Ray KK, Seshasai SR, Erqou S, et al. Statins and all-cause mortality
Contributors  PB was the lead researcher and involved in the design, in high-risk primary prevention: a meta-analysis of 11 randomized
implementation and analysis and reporting of the overview. PB conducted the initial controlled trials involving 65,229 participants. Arch Intern Med
search for systematic reviews. PB and AS undertook title and abstract screening, 2010;170:1024–31.
16. de Lorgeril M, Rabaeus M. Beyond confusion and controversy can
full-text screening, extraction of data and quality assessment. SMS arbitrated any we evaluate the real efficacy and safety of cholesterol-lowering with
disagreements between PB and AS on inclusion and quality assessment of reviews. statins?  JCBMR 2016;1:67–92.
JC and SMS provided substantial contributions to the conception, design, analysis 17. Ijioma N, Robinson JG. Statins and primary prevention of
and reporting of the work. All authors have read and approved the final manuscript cardiovascular disease in women a critical appraisal of the evidence.
and agree to be accountable for all aspects of the work. PB is the guarantor of this Am J Lifestyle Med 2015;9:114–29.
paper. 18. Petretta M, Costanzo P, Perrone-Filardi P, et al. Impact of gender in
primary prevention of coronary heart disease with statin therapy: a
Funding  This overview is part of PB’s PhD, which is funded by the SPHeRE HRB meta-analysis. Int J Cardiol 2010;138:25–31.
structured PhD programme (HRB SPHeRE/2013/1). AS is funded by the Irish Cancer 19. Brugts JJ, Yetgin T, Hoeks SE, et al. The benefits of statins in people
Society Collaborative Cancer Research Centre BREAST-PREDICT (CCRC13GAL). All without established cardiovascular disease but with cardiovascular

Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085 11

 Open access

BMJ Open: first published as 10.1136/bmjopen-2018-023085 on 23 April 2019. Downloaded from http://bmjopen.bmj.com/ on September 9, 2019 by guest. Protected by copyright.
risk factors: meta-analysis of randomised controlled trials. BMJ 43. Kirkham JJ, Altman DG, Williamson PR. Bias due to changes in
2009;338:b2376. specified outcomes during the systematic review process. PLoS One
20. de Vries FM, Denig P, Pouwels KB, et al. Primary prevention of major 2010;5:e9810.
cardiovascular and cerebrovascular events with statins in diabetic 44. Krumholz HM. Why data sharing should be the expected norm. BMJ
patients: a meta-analysis. Drugs 2012;72:2365–73. 2015;350:h599.
21. Taylor FC, Huffman M, Ebrahim S. Statin therapy for primary 45. Lo B. Sharing clinical trial data: maximizing benefits, minimizing risk.
prevention of cardiovascular disease. JAMA 2013;310:2451–2. JAMA 2015;313:793–4.
22. Chou R, Dana T, Blazina I, et al. Statins for prevention of 46. Krumholz HM. Statins evidence: when answers also raise questions.
cardiovascular disease in adults: evidence report and systematic BMJ 2016;354:354.
review for the US Preventive services task force. JAMA 47. Godlee F. Statins: we need an independent review. BMJ
2016;316:2008–24. 2016;354:i4992.
23. Thavendiranathan P, Bagai A, Brookhart MA, et al. Primary prevention 48. Collins C. In: Godlee F, ed. CTSU: University of Oxford, 2014.
of cardiovascular diseases with statin therapy: a meta-analysis of 49. Parish E, Bloom T, Godlee F. Statins for people at low risk. 2015.
randomized controlled trials. Arch Intern Med 2006;166:2307–13. 50. Byrne P, Cullinan J, Murphy C, et al. Cross-sectional analysis of
24. Higgins JP, Green S. Cochrane handbook for systematic reviews of the prevalence and predictors of statin utilisation in Ireland with a
interventions: John Wiley & Sons, 2011. focus on primary prevention of cardiovascular disease. BMJ Open
25. Byrne P, Cullinan J, Smith A, et al. Statins for the primary prevention 2018;8:e018524.
of cardiovascular disease: an overview of systematic reviews. 2017. 51. Wallach-Kildemoes H, Stovring H, Holme Hansen E, et al. Statin
26. Endo A. The Origin of the Statins. International Congress Series: prescribing according to gender, age and indication: what about the
Elsevier, 2004. benefit-risk balance? J Eval Clin Pract 2016;22:235–46.
27. Veritas Health Innovation. www.​covidence.​org. 52. Wallach Kildemoes H, Vass M, Hendriksen C, et al. Statin
28. Kung J, Chiappelli F, Cajulis OO, et al. From systematic reviews to utilization according to indication and age: a Danish cohort study
clinical recommendations for evidence-based health care: validation on changing prescribing and purchasing behaviour. Health Policy
of revised assessment of multiple systematic reviews (r-amstar) for 2012;108:216–27.
grading of clinical relevance. Open Dent J 2010;4:84-91. 53. Feely J, Bennett K. Epidemiology and economics of statin use. Ir
29. Mihaylova B, Emberson J, Blackwell L, et al. The effects of lowering Med J 2008;101:188–91.
LDL cholesterol with statin therapy in people at low risk of vascular 54. Wallach Kildemoes H, Hendriksen C, Andersen M. Drug utilization
disease: meta-analysis of individual data from 27 randomised trials. according to reason for prescribing: a pharmacoepidemiologic
Lancet 2012;380:581–90. method based on an indication hierarchy. Pharmacoepidemiol Drug
30. Fulcher J, O’Connell R, Voysey M, et al. Cholesterol Treatment Saf 2012;21:1027–35.
Trialists’(CTT) Collaboration. Efficacy and safety of LDL-lowering 55. Godlee F. Statins and The BMJ. BMJ 2014;349:g5038.
therapy among men and women: meta-analysis of individual 56. Getz L, Kirkengen AL, Hetlevik I, et al. Ethical dilemmas arising
data from 174,000 participants in 27 randomised trials. Lancet from implementation of the European guidelines on cardiovascular
2015;385:1397–405. disease prevention in clinical practice. A descriptive epidemiological
31. Baigent C, Keech A, Pa K, et al. Cholesterol Treatment Trialists’(CTT) study. Scand J Prim Health Care 2004;22:202–8.
Collaborators. Efficacy and safety of cholesterol-lowering treatment: 57. Cordoba G, Schwartz L, Woloshin S, et al. Definition, reporting, and
prospective meta-analysis of data from 90,056 participants in 14 interpretation of composite outcomes in clinical trials: systematic
randomised trials of statins. Lancet 2005;366:1267–78. review. BMJ 2010;341:c3920.
32. Baigent C, Blackwell L, Emberson J, et al. Efficacy and safety of 58. Ferreira-González I, Busse JW, Heels-Ansdell D, et al. Problems
more intensive lowering of LDL cholesterol: a meta-analysis of with use of composite end points in cardiovascular trials: systematic
data from 170,000 participants in 26 randomised trials. Lancet review of randomised controlled trials. BMJ 2007;334:786.
2010;376:1670–81. 59. Sun X, Ioannidis JP, Agoritsas T, et al. How to use a subgroup
33. Downs J, Gotto A, Clearfield M, et al. Protocol for a prospective analysis: users' guide to the medical literature. JAMA
collaborative overview of all current and planned randomized trials of 2014;311:405–11.
cholesterol treatment regimens. Am J Cardiol 1995;75:1130–4. 60. Collins R, Reith C, Emberson J, et al. Interpretation of the evidence
34. Stewart LA, Clarke M, Rovers M, et al. Preferred reporting items for for the efficacy and safety of statin therapy. Lancet 2016;388.
systematic review and meta-analyses of individual participant data: 61. Armitage J. The safety of statins in clinical practice. The Lancet
the PRISMA-IPD statement. JAMA 2015;313:1657–65. 2007;370:1781–90.
35. Mills EJ, Rachlis B, Wu P, et al. Primary prevention of cardiovascular 62. Buettner C, Davis RB, Leveille SG, et al. Prevalence of
mortality and events with statin treatments: a network meta- musculoskeletal pain and statin use. J Gen Intern Med
analysis involving more than 65,000 patients. J Am Coll Cardiol 2008;23:1182–6.
2008;52:1769–81. 63. Fernandez G, Spatz ES, Jablecki C, et al. Statin myopathy: a
36. Mora S, Glynn RJ, Hsia J, et al. Statins for the primary prevention common dilemma not reflected in clinical trials. Cleve Clin J Med
of cardiovascular events in women with elevated high-sensitivity 2011;78:393.
C-reactive protein or dyslipidemia: results from the Justification for 64. Strom BL. Pharmacoepidemiology: John Wiley & Sons, 2006.
the Use of Statins in Prevention: An Intervention Trial Evaluating 65. Jepsen P, Johnsen SP, Gillman MW, et al. Interpretation of
Rosuvastatin (JUPITER) and meta-analysis of women from primary observational studies. Heart 2004;90:956–60.
prevention trials. Circulation 2010;121:1069–77. 66. Casula M, Mozzanica F, Scotti L, et al. Statin use and risk of new-
37. Robinson JG. Starting primary prevention earlier with statins. Am J onset diabetes: a meta-analysis of observational studies. Nutr Metab
Cardiol 2014;114:1437–42. Cardiovasc Dis 2017;27:396–406.
38. Hair JF, Black WC, Babin BJ, et al. Multivariate data analysis. 6th edn. 67. Sattar N, Preiss D, Murray HM, et al. Statins and risk of incident
Uppersaddle River: Pearson Prentice Hall, 2006. diabetes: a collaborative meta-analysis of randomised statin trials.
39. Figueiredo Filho DB, Paranhos R, ECd R, et al. When is statistical Lancet 2010;375:735–42.
significance not significant? Brazilian Political Science Review 68. Waters DD, Ho JE, DeMicco DA, et al. Predictors of new-onset
2013;7:31–55. diabetes in patients treated with atorvastatin: results from 3 large
40. Murad MH, Montori VM, Ioannidis JP, et al. How to read a systematic randomized clinical trials. J Am Coll Cardiol 2011;57:1535–45.
review and meta-analysis and apply the results to patient care: users' 69. Albarqouni L, Doust J, Glasziou P. Patient preferences for
guides to the medical literature. JAMA 2014;312:171–9. cardiovascular preventive medication: a systematic review. Heart.
41. Yusuf S, Bosch J, Dagenais G, et al. Cholesterol lowering in 2017.
intermediate-risk persons without cardiovascular disease. N Engl J 70. Byrne P, Cullinan J, Gillespie P. Statins for primary prevention of
Med 2016;374:2021–31. cardiovascular disease: modelling guidelines and patient preferences
42. Page M, McKenzie J, Dwan K, et al. Bias due to selective inclusion based on an Irish cohort. Br J Gen Prac 2019. Forthcoming.
and reporting of outcomes and analyses in systematic reviews of 71. Karmali KN, Lloyd-Jones DM, Berendsen MA, et al. Drugs for primary
randomised trials of healthcare interventions (Protocol). Cochrane prevention of atherosclerotic cardiovascular disease: an overview of
Database Syst Rev 2012:5. systematic reviews. JAMA Cardiol 2016;1:341-9.

12 Byrne P, et al. BMJ Open 2019;9:e023085. doi:10.1136/bmjopen-2018-023085