H

Te

LAa

3b

c

d

a

ARRA

KCCCCPL

ba

(A

0o

Health Policy 120 (2016) 797–808

Contents lists available at ScienceDirect

Health Policy

journa l h om epa ge: www.elsev ier .com/ locate /hea l thpol

ealth Reform Monitor

he 2013 cholesterol guideline controversy: Would bettervidence prevent pharmaceuticalization?�

ynn Unruha,∗, Thomas Riceb,1, Pauline Vaillancourt Rosenauc,2,ndrew J. Barnesd,3

Health Services Administration, Department of Health Management & Informatics, University of Central Florida, Orlando, FL2816-2200, United StatesHealth Policy and Management, UCLA Fielding School of Public Health, Los Angeles, CA 90095-1772, United StatesSchool of Public Health, University of Texas Health Science Center Houston, TX 77030, United StatesHealthcare Policy and Research, Virginia Commonwealth University, United States

r t i c l e i n f o

rticle history:eceived 28 June 2015eceived in revised form 10 May 2016ccepted 12 May 2016

eywords:ardiovascular diseaseholesterol guidelineslinical guidelinesonflict of interestharmaceuticalizationifestyle changes

a b s t r a c t

Cardiovascular disease (CVD) remains the leading cause of death globally. A class of med-ications, known as statins, lowers low-density lipoprotein cholesterol levels, which areassociated with CVD. The newest 2013 U.S. cholesterol guideline contains an assessmentof risk that greatly expands the number of individuals without CVD for whom statins arerecommended. Other countries are also moving in this direction. This article examines thecontroversy surrounding these guidelines using the 2013 cholesterol guidelines as a casestudy of broader trends in clinical guidelines to use a narrow evidence base, expand theboundaries of disease and overemphasize pharmaceutical treatment.

We find that the recommendation in the 2013 cholesterol guidelines to initiate statinsin individuals with a lower risk of CVD is controversial and there is much disagreementon whether there is evidence for the guideline change. We note that, in general, clinicalguidelines may use evidence that has a number of biases, are subject to conflicts of interestat multiple levels, and often do not include unpublished research. Further, guidelines may

contribute to the “medicalization” or “pharmaceuticalization” of healthcare.

Specific policy recommendations to improve clinical guidelines are indicated: theseinclude improving the evidence base, establishing a public registry of all results, includingunpublished ones, and freeing the research process from pharmaceutical sector control.

© 2016 The Author(s). Published by Elsevier Ireland Ltd. This is an open access articleY-NC-N

under the CC B

� Open Access for this article is made possible by a collaborationetween Health Policy and The European Observatory on Health Systemsnd Policies.∗ Corresponding author. Tel.: +1 407 823 4237.

E-mail addresses: lynn.unruh@ucf.edu (L. Unruh), trice@ucla.eduT. Rice), Pauline.Rosenau@uth.tmc.edu (P.V. Rosenau),barnes3@vcu.edu (A.J. Barnes).1 Tel.: +1 310 206 1824.2 Tel.: +1 713 500 9491.3 Tel.: +1 804 827 4361.

http://dx.doi.org/10.1016/j.healthpol.2016.05.009168-8510/© 2016 The Author(s). Published by Elsevier Ireland Ltd. This is an operg/licenses/by-nc-nd/4.0/).

D license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Cardiovascular disease (CVD) remains the leading causeof death globally, and prevention of CVD is a priority inworld health systems. Prevention focuses on the reduc-tion of risk factors such as an unhealthy diet, inadequateexercise, obesity, and smoking, as well as reduction in totalblood cholesterol and the low density lipoprotein (LDL)portion of cholesterol.

Several decades ago a class of pharmaceuticals, HMG-CoA reductase inhibitors – commonly known as statins –were found to have a significant impact on cholesterol,particularly LDL. Statins have been recommended for both

n access article under the CC BY-NC-ND license (http://creativecommons.

h Policy

798 L. Unruh et al. / Healt

secondary prevention of CVD (individuals with known CVD,so prevention focuses on reducing further development ofthe disease and complications), and primary prevention(those without CVD but with risk factors for the disease,including certain levels of cholesterol or LDL). Since theirentry into the U.S. market in 1987, the utilization of statinshas skyrocketed. Between 2007 and 2010 statins were themost commonly prescribed therapeutic class in the U.S.[143]. According to the U.S. Centers for Disease Control andPrevention, in the last two decades statin use in the U.S.increased seven-fold among adults age 45–64 [1].

Statin utilization has been guided by clinical protocols.In the U.S., the National Cholesterol Education Programissued Adult Treatment Protocol (ATP) reports in 1988,1993 and 2001, each one recommending a successivelybroader application of statins to the population [2]. Thenewest guidelines in the U.S., issued in 2013 by theAmerican College of Cardiology and the American HeartAssociation (ACC/AHA), simplified some of the older guide-lines and added risk categories [3]. More importantly,however, they contain a controversial new threshold andcalculation of risk that would greatly expand the numberof individuals that should be placed on statins [4].

A thorough understanding of this controversy hasbecome even more important given two new lipid-lowering medications, Repatha (Evolocumab) and Praluent(Alirocumab), approved by the U.S. Food and Drug Adminis-tration (FDA) for marketing. Medications in this new class(PCSK-9 inhibitors) appear to be capable of lowering LDLbelow even that of statins, and (in a post hoc study) ofreducing the incidence of cardiovascular disease [5]. Theyare being promoted for the care of patients whose LDLsare not adequately lowered by statins, and for those whodo not stay on statins due to their side effects. However,to date, the effects of the new medications are based onjust a few trial studies with surrogate outcomes [5], andthe medications are far more expensive than statins [6].If the 2013 guidelines become widely accepted and lipid-lowering continues to be called for in low-risk individuals,it is very possible that the new class of medications will benecessary to achieve that lowering and could become partof the next generation of lipid-lowering guidelines.

This article examines the controversy surrounding the2013 cholesterol guidelines. We use cholesterol guidelinesas a case study to address the broader issues of clinicalguideline development, including the quality of evidenceand conflicts of interest embedded in guideline recommen-dations. The evidence presented suggests a link betweenthe clinical guideline evidence-base and the expansion ofdisease categories, “medicalization” and “pharmaceutical-ization” of health and illness in the U.S. and other countries.Policies are recommended for future clinical guidelinedevelopment.

1. 2013 U.S. guideline controversy

The 2013 ACC/AHA cholesterol guidelines recommend

statins for both secondary and primary prevention of CVD.For secondary prevention the guidelines recommend statinuse in anyone who has had heart or peripheral vasculardisease, angina, heart bypass or angioplasty, and stroke or

120 (2016) 797–808

transient stroke (TIA) [3]. Statins are recommended for pri-mary prevention for those: between 40 and 75 years ofage with type 1 or 2 diabetes; over 21 years of age withLDL of 190 mg/dl or more; and 40–75 years of age with a7.5 percent or higher risk of developing CVD (heart attackor stroke) within 10 years. The risk calculator is based onage, gender, race, total cholesterol, High Density Lipopro-tein (HDL), systolic blood pressure level, and current bloodpressure medication status [3].

It is mainly the calculation of risk and the recommen-dation for starting statin therapy for primary preventionat ≥7.5 per cent risk of CVD within 10 years that has beencontroversial. Several studies show that the risk calcula-tor overestimates the risk of CVD by 50% or more [7–10].As far as the threshold, some consider it to be “aggres-sive,” and point out that guidelines in other countries havehigher thresholds, as for example the most recent ones inthe U.K. and Australia, which are set at 10% over 10 yearsand 10–15% over 5 years respectively [11]. Estimates indi-cate that this aspect of the U.S. guidelines would broadenthe use of statins for those between the ages of 40 and 75 by25–30% [4,12,13]. In practical terms, it is estimated that thenew guidelines would recommend statin therapy for nearlyall men > 60, women > 69 years [7,14], and all African Amer-ican men over the age of 65 with normal blood pressure andcholesterol levels [15].

Proponents of the 2013 guidelines point out that the rec-ommendations for statin use in lower risk populations arebased on randomized controlled trials (RCTs) of statin effi-cacy. Two large meta-analyses of statin RCTs, both before[16] and after [17] the guideline change, also support thechange. The first review, published in 2012 by the Choles-terol Treatment Trialists’ Group (CTT), found that in low riskindividuals (five-year risk of major vascular events <10%)a reduction in LDL due to statins was associated with anabsolute reduction in the risk of major vascular events andall-cause mortality. The second review, published in 2014in the Cochrane Collection of Systematic Analyses, usedstudy-level results, including those from the CTT analysis,and found that non-fatal CVD events and all-cause mor-tality were reduced with the use of statins in individualswith no prior history of CVD. Some post-guideline studieshave found the new guidelines to be better at predictingindicators of CVD such as blood vessel plaque [18,19] andcoronary artery calcification [20], and to be more accuratein identifying increased risk of CVD incidents, particularlyin intermediate-risk participants [20].

These studies represent a degree of support for thenew guidelines, but critics of the guidelines charge thatthere are a number of issues with this evidence. First, therisk categories in the 2013 guidelines are not the same asthose studied in RCTs, including the meta-analyses, so itis impossible to apply outcomes in the studies to individ-uals under the new guidelines. Second, while RCTs havefound statins to be efficacious for those with CVD or highrisk of CVD, there is less evidence that they are effective inlower risk populations, especially older adults [2,21–24].

This is especially true for the most important outcomesof statin therapy – lower all-cause mortality, few sideeffects or adverse events, and positive patient-reportedoutcomes such as good quality of life [22,25]. CTT claims of

h Policy

shlmac[ifua[Rrs

gtrwAni

CtclaCbki

plttYaebfftdlrmdewehfAid

di

L. Unruh et al. / Healt

tatins reducing all-cause mortality in low risk individualsave been challenged by other researchers who recalcu-

ated their results and did not find this [22,25]. Neithereta-analysis examined patient-reported quality of life or

dequately reported common side effects such as mus-le disorders [22,26–31] or less common adverse reactions22,32–41]. These issues may not be adequately discoveredn clinical trials because they are designed more to testor efficacy than to catch adverse reactions, and individ-als who are less likely to benefit and more likely to havedverse reactions may be excluded from the RCT sample22,28,42]. Further, individuals experiencing side effects inCTs tend to drop out, so their negative experience is notecorded [22], and the length of many RCTs may be toohort to observe side effects [28].

A third concern regarding the evidence for the 2013uidelines is that, in general, the effectiveness of statinsends to be overstated. A common way this occurs is toeport relative risk, rather than absolute risk [2,21], asas the case with the Cochrane Collection meta-analysis.lso, RCTs examine efficaciousness rather than effective-ess; that is, they show the effect on a narrower sample of

ndividuals than occurs in the “real world.”A fourth concern involves data transparency. The

ochrane review analyzed published studies available tohem, and did not include any unpublished data from thelinical trials. Therefore, the review could suffer from pub-ication bias – negative studies that were never publishednd therefore were not included in the analysis [43]. TheTT study used patient level data from the team’s studies,ut whether that included unpublished study results is notnown and the data have not been made available to othernvestigators [44].

Another concern is that while the 2013 guidelines forrimary prevention of CVD instruct providers to promote

ifestyle management in the ≥7.5% risk group as part ofheir discussion with patients, there is no recommenda-ion for a trial period of just lifestyle changes in this group.et, there is evidence that a healthy lifestyle should ben important part, if not the focus of primary prevention,specially in lower risk categories [45–56]. While it maye difficult to achieve compliance with this approach (andor this reason, it is understandable that physicians mayeel that the best thing for their patients is to go straighto statins), a number of motivational programs have beeneveloped that have shown success in weight loss, other

ifestyle improvements, and/or lowering of cardiovascularisk [57–64]. The 2013 cholesterol guidelines also do notention the importance of socioeconomic status in the

evelopment of CVD. Socioeconomic (SE) factors such asducation, employment (or unemployment), occupation,ork hours, stress, income, neighborhood environment,

arly childhood health, and racial and income inequalities,ave been found to affect lifestyle behaviors and CVD risk

actors [65–80], as well as CVD directly [81–83,78,84]. TheHA declares that addressing socioeconomic determinants

s the most significant opportunity for reducing death and

isability from CVD in the U.S. [85].

Finally, there were conflicts of interest (COI) in theevelopment of the guidelines, from the individual stud-

es, to the meta-analyses, to the guideline committee itself.

120 (2016) 797–808 799

In the CTT meta-analysis most of the RCTs were funded bythe manufacturers, members of the CTT writing committeereceived funds from the pharmaceutical industry for meet-ings, and two authors also received speaking honorariafrom the industry [16]. In the Cochrane Collection study,industry-sponsorship of all of the trials was acknowledgedin post-report comments but the meta-analysis authorsclaimed no COI [17]. As far as the guideline committee,eight of 15 panelists, including the chair and two co-chairs,had past or current ties to pharmaceutical companies[15].

For these reasons, pro and con, practitioners andresearchers are divided on the appropriateness of the pri-mary prevention portion of the 2013 guidelines. Somebelieve that there is sufficient evidence to proceed with theimplementation of these primary prevention guidelines inthe population. Others disagree, and call for more researchwhich would improve upon the methodological limitationsof existing evidence, and for reforms in the clinical guide-line development process.

2. Expenditures on statins

Under the new guidelines expenditures related tostatins will increase. Abramson and colleagues [22] esti-mate that expanding statin use would result in up to $1USD per person per day in pharmaceutical drug costs.Expenditures will also rise due to an increase in visits tophysicians to obtain new prescriptions and to follow up onthe progress of statin regimens. Also, the aforementionedside effects occurring in some individuals will add to healthcare costs. Unfortunately, at this time data are not availableto accurately estimate the total costs of the increased statinuse.

These costs must be weighed against the health benefitsand cost savings from a decrease in cardiovascular events.Pencina and et al. [4] estimate that the new guidelines, iffully implemented, would reduce such events by 475,000,but they admit that this is based on a number of assump-tions, including “the independence of the relative benefitof statin use from the levels of LDL cholesterol or abso-lute risk” (p. 1430) [4]. Other researchers are skeptical thatthe new guidelines will result in nearly so great a benefit.Abramson et al. [22] find that over five years only one caseof non-fatal heart attack or stroke would be averted out of140 lower risk patients. Given the 12.8 million newly eligi-ble, and allowing for a take-up rate of 58%, that translatesinto only 50,000 cases averted. Cost savings would also bereduced due to adverse events.

To date we are aware of one cost-effectiveness analy-sis (CEA) of the 2013 guidelines [86]. This CEA found thatthe threshold of 7.5% or higher (with an estimated 48% ofadults treated) had an incremental cost-effectiveness ratio(ICER) of $37,000/QALY compared with a 10% or higherthreshold. Thresholds of 4.0% or higher (61% of adultstreated) and 3.0% or higher (67% of adults treated) had

ICERs of $81,000/QALY and $140,000/QALY, respectively.Cost-effectiveness results were sensitive to statin price andthe risk of statin-induced diabetes. The CEA results of thecurrent threshold were considered to be acceptable.

h Policy

800 L. Unruh et al. / Healt

The degree to which pharmaceutical companies willprofit from the 2013 cholesterol guidelines is not certain.Some believe that it will not be significant since patent pro-tections have run out on the medications. However, newdrugs in the class may be on the horizon that could essen-tially renew patent protection. Additionally, as Kaplan [2]points out, the use of drugs for chronic diseases is not likethat of antibiotics where one takes them for a short periodof time. With chronic diseases, the medications are takenfor a lifetime. The more people that are put on drugs for pre-vention of chronic diseases and the younger they are puton those medications, the more pharmaceutical companiesprofit.

In addition, if other countries follow suit and lower theirguideline thresholds, markets will expand even more. Forexample, Muntner et al. [87] estimate that if China adoptedthe 2013 U.S. cholesterol guidelines, 33.4 more million Chi-nese adults should begin statin therapy [87]. If all countriesadopted the U.S. guidelines and CVD risk were similar to theU.S., close to 1 billion people would be eligible for statintherapy for primary prevention [88].

Although expenditures to prevent and treat diseaseshould never be the decisive factor in the developmentof clinical guidelines, when accompanied by questionsregarding the efficacy and safety of initiating statins inlow-risk populations, the appropriateness of the 2013guidelines should come under closer scrutiny.

3. Cholesterol guidelines internationally

Examining the cholesterol guideline policies in othercountries puts US guidelines in perspective. Table 1

presents the primary prevention portion of current choles-terol guidelines from seven developed countries. Theevidence base and criteria for statin eligibility differ acrosscountries with most employing all types of studies in their

Table 1Comparison of statin guidelines in seven developed countries.

Country (latest guideline) Evidence for guidelines Threspreve

Australia, (HF, 2013) RCTs and meta-analyses ofRCTs

>10–1

Canada (CCS, 2012) Mostly RCTs of statin efficacy Frami10–20

Europe (ESC, 2011) All evidence (RCTs,epidemiological studies, etc.)

≥5% rcalcul

Japan (JAS, 2002) All evidence (RCTs,epidemiological studies, etc.)

A couLDL: 0or ≥h

Netherlands (CVRM, 2011) Unable to ascertain for currentguidelines, but policy analystsadvocate utilizing all evidence

Risk o>2.5 m15 ye

United Kingdom (NICE, 2014) Unable to ascertain ≥10%calculCVD (

United States (ACC/AHA, 2013) RCTs and meta-analyses ofRCTs

≥7.5%calcul

Sources: Australia [161]; Canada [162,163]; Europe [147]; Japan [148]; Netherlan

120 (2016) 797–808

development. Risk thresholds vary from as low as 5% forthe European (ESC) guidelines to 20% for the Canadian andDutch guidelines. The degree to which lifestyle changes areemphasized also differ, with the Japanese and Europeanguidelines recommending lifestyle changes before initia-tion of statins, and the U.S. and Canada recommending adiscussion (but not necessarily a trial of lifestyle changesfirst).

It is admittedly difficult to rank these guidelines in termsof agressiveness since each differs in a number of com-ponents, including calculations of risk. Still, based on therisk thresholds, the European and U.S. guidelines appearto be among the most aggressive. Vaucher et al. [89] findthat when applied to the Swiss population both the ESCand ACC/AHA guidelines would increase the number ofindividuals eligible for statin therapy, the ACC/AHA guide-lines more so (it would more than double the prevalenceof high risk individuals) [89]. Another recent study notedthat ACC/AHA guidelines applied to Europe would result innearly all men and two-thirds of women age 55 years orolder being candidates for statin treatment [90].

4. How evidence based are clinical guidelines?

The 2013 cholesterol guidelines were developed usingprotocols similar to other clinical guidelines. Developed byboth non-governmental professional societies and govern-mental agencies [92], clinical guidelines are “systematicallydeveloped statements to assist practitioner and patientdecisions about appropriate health care for specific clinicalcircumstances” [91]. In the U.S., guidelines are predomi-nantly developed by professional societies, such as the ACC

and AHA,[93] while governmental agencies, such as theAgency for Healthcare Research and Quality (AHRQ), playmore of a sponsorship and oversight role [94]. In othercountries governmental agencies, such as the National

hold for initiation for primaryntion

Importance of lifestyle changein primary prev.

5% risk within 5 years 10–15% risk: encouragelifestyle change first andreview progress in 6–12months. >15% risk: Lifestylechange and pharmacologicaltherapy

ngham risk score ≥20% or% and LDL ≥ 3.5 mm

Discuss w/those meetingthreshold

isk CVD in 10 years. SCORE Riskation, validated

Lifestyle interventions shouldbe tried first

nt of major risk factors other than = low risk; 1–2 = intermediate; 3

igh

Statins should be consideredonly after 3–6 months lifestylechange

f CVD ≥20% in 10 years & LDLmol/l (97 mg dl). If diabetic, add

ars to age for risk calculation

Unable to ascertain

risk of CVD in 10 years. Riskation of factors contributing toQRISK2)

Discuss w/those meetingthreshold. Consider thebenefits of lifestyle changes indecision to initiate statins.

risk CVD in 10 years. Riskation may overestimate risk

Discuss w/those meetingthreshold

ds [164]; U.K. [165]; U.S. [3].

L. Unruh et al. / Health Policy

Box 1: Types of bias in clinical research

1. Individual study research design and conduct• Choice of sample

• Likelihood of fewer comorbidities[22,23,28,42,92,98]

• Higher probability of positive response to inter-vention [22,28,42,98]

• May not include women, minorities, disadvan-taged, older individuals, children [100]

• Choice of outcomes• Narrow [92]• Surrogate endpoints [22,92,100,105]

• Changing outcomes after starting study/datadredging [92,152]

• Short time horizon [28,92,97]• Participant attrition [22,97,98]

2. Individual study interpretation and publication• Selective reporting [98,100]• Reporting relative risk [2,21,22,100]• Report significance but not clinical effect

[22,92,100,101]• Failure to report side effects [22,100]

3. Poor external validity – study population is nar-row or excludes population the intervention will begiven to [21,23,92,98,100,101,103–105]

4. Systematic review or meta-analysis• Use of studies with biases [98]

IU

tworaHt

4

Jsobwetc

ttostilf

• Failure to include unpublished results[42,92,98,101]

nstitute for Health and Clinical Excellence (NICE) in the.K., play a bigger development role [92].

Clinical guidelines “have become deeply rooted inhe principles of evidence-based medicine“[93] (p. 706),ith expectations that guideline committees be composed

f clinical and research experts that utilize high-qualityesearch – optimally systematic reviews and meta-nalyses of RCTs – in their recommendations [92,93].owever, clinical guidelines are not without issues related

o their evidence base [93,95].

.1. Bias in research evidence

First is the problem of bias in the research evidence.adad and Enkin [96] list 60 and Warden [97] lists 150 pos-ible sources of bias in clinical research. Some of the majornes are listed in Box 1. We mentioned several of theseiases when discussing the 2013 cholesterol guidelines andill not review them again. Suffice it to say that the pres-

nce in a study of any one of these will call into questionhe true impact of the intervention, yet studies used to formlinical guidelines often suffer from several of these biases.

The biases common to individual studies carry overo the systematic reviews and meta-analyses that utilizehem. These syntheses intend to improve upon the validityf the individual studies, but they are only as good as thetudies they synthesize: if the individual studies are biased,

he syntheses will be biased as well [92,98,99]. In addition,f the syntheses do not detail individual study biases andimitations, the final synthesis appears to be more bias-ree than it really is. Finally, these syntheses may not track

120 (2016) 797–808 801

down or be able to get at unpublished data and results.Since unpublished results are likely to be those showinginsignificance or even negative effects, syntheses that omitthese may not reflect the true impact of the intervention[92,98].

These systematic reviews and meta-analyses, in turn,are relied upon by clinical guideline committees that typ-ically follow a hierarchy of evidence in which systematicreviews and meta-analyses of RCTs are at the top [92].Other types of studies will often not be considered or willbe given less consideration, as causation cannot be deter-mined from those studies. Although RCTs are thought to bethe “gold standard” of clinical research, they have signifi-cant limitations (in fact, all of the biases in Box 1 may bepresent in them) [98,100–104]. On the other hand, what arelimitations with RCTs tend to be strengths of observationaland other non-RCT research designs [98,100–106]. Table 2outlines how non-RCT studies could offset some of thelimitations of RCTs. Silverman [103] summarizes that non-RCTs can compliment RCTs by assessing an interventions’effect in the real world, “by considering efficacy outcomes,safety, tolerability, and patient compliance in large patientpopulations and by using clinically relevant long-term out-comes” (p. 114). The frequent failure to include non-RCTstudies in the evidence-base for clinical guidelines is prob-lematic.

4.2. Conflicts of interest

The second major issue in clinical guidelines is conflictsof interest (COI), which have been found in all levels of clin-ical research and the guidelines that utilize that research.At the regulatory level, experts and voting members onadvisory committees for the FDA body that reviews newdrug applications – the Center for Drug Evaluation andResearch (CDER) – have connections with pharmaceuticalcompanies [107,108]. Lurie et al. [108] report that COIs are“common, often of considerable monetary value, and rarelyresult in recusal of advisory committee members” (1921).The CDER itself receives over one-half of its funding frompharmaceutical companies [107].

At the trial level, pharmaceutical companies fundapproximately 75% of clinical drug trials published in topmedical journals [107]. A third of this funding goes to uni-versities and academic medical centers, while the rest goesto for-profit research companies.

Pharmaceutical companies not only fund clinicalresearch, they manage it from design conception throughpublication – what has come to be known as “ghostmanagement” of drug research [109–113]. Pharmaceuti-cal company statisticians design the trial and perform theanalysis, but are not usually listed as an author. Hired med-ical writers produce first drafts and edit papers. Academicsare asked or paid to author the final product. Multiplemanuscripts will be written from one study, and com-munication experts help move manuscripts through thepublication process. An example is the publication record

of Zoloft, in which 85 manuscripts, which were written byone medical communications company between 1998 and2001, were more predominantly published and cited thanother manuscripts on Zoloft in that period [113].

802 L. Unruh et al. / Health Policy 120 (2016) 797–808

Table 2Strengths and limitations of RCT and non-RCT research designs.

RCT Non-RCT

Strength Limitation Strength Limitation

Causation Possible to ascertain withsome confidence.

True causal relationshipsubject to othermethodological issueslisted below.

Weak support for causation iflongitudinal, instrumentalvariable, propensity score, orother special design is used.

Generally not possible toascertain d/t selectionbias and confounders.

Choice of sample Randomization allows forassumption of equivalencebetween groups.

Use of narrow and selectsample in order todetermine efficacy.

Can be epidemiological samplefrom entire populationreceiving the intervention. Canbe a large and broadpopulation.

Lack of randomizationmeans that groupscannot be assumed to beequivalent.

Choice of outcomes Some use of clinicalendpoints.

Frequent use of surrogateend points.

Outcomes usually focus onclinical endpoints such asmortality, quality of life andadverse events.

Consistency of researchdesign & outcomes

With ethical oversight andtransparency, the originaldesign & outcomesmeasures can bemaintained.

Research design &outcomes may changeduring the study.

With ethical oversight andtransparency, data dredgingcan be avoided.

Data dredging mayoccur.

Time horizon May includepre-intervention period.

Studies commonly runfor short time.

Population can be studiedhaving received theintervention for a long timeperiod.

Often entails the use ofsecondary data with nopre-intervention data.

Participation Attrition of participantscan be addressed withspecial designs.

Attrition of participantsis a common problem.

Attrition of participants can beaddressed with special designsand statistical methods.

Attrition of participantsis a common problem.

Reporting of results Reporting of all results canoccur with ethicaloversight andtransparency.

Negative or insignificantresults may not bepublished.

Reporting of all results canoccur with ethical oversightand transparency.

Negative or insignificantresults may not bepublished.

Reporting of risk Absolute risk can bereported.

Results are commonlyreported as relative risk.

Absolute risk can be reported. Results are commonlyreported as relative risk.

due tople.

External validity Can be improved with amore representativesample.

Can be weaknarrow sam

At the meta-analysis level, COI can enter into the anal-ysis, first, by using individual studies that have COIs, andsecond, by COIs among meta-analysis authors themselves.We saw that this was the case for the meta-analyses onstatins discussed earlier. Meta-analyses often do not iden-tify COIs of the studies in the analysis, but even if theydo, they do not necessarily remove the studies from theanalysis [114,115].

At the guideline level, a number of studies report thata significant percentage of guideline authors have finan-cial ties to the pharmaceutical industry [116–120], even ashigh as 91% and 94% in two U.S. guidelines [120]. Financialties can be financial support for research, stock ownership,employment, being a speaker for, or consultation with apharmaceutical company [112,116,118,119]. Authors canhave ties with a number of different pharmaceutical com-panies (Choudhry et al. [116] found 10.5 on average; [116]Norris et al. [119] found two to ten or more), including withcompanies whose drugs are recommended in the guide-lines (59% in Choudhry et al. [116]). Norris et al. [120]reported that of the guidelines they examined, 56% of themanufacturers of recommended drugs had authors with afinancial interest in the company [119]. It should be noted

that the actual amount of COI may be more since COI isnot always identified, and if authors are queried, not allrespond.

Can be strong if data includes arepresentative sample frompopulation of interest.

Often dependent uponexisting data which maynot be representative.

Guideline COIs exist internationally on many differentconditions. For example, guidelines for type 2 diabetes andAlzheimer’s disease in France were withdrawn in 2011 dueto conflicts of interest that had been undeclared [121].Nearly half of all authors of 297 German guidelines in 2011had financial COI [122]. In a study of psychiatric guidelinesby Cosgrove et al. [117] 100% of members had financial tiesto pharmaceutical companies. Each author had an averageof 20.5 relationships [117]. A study of U.S. and Canadianguidelines for diabetes and hyperlipidemia found that only29% of panelists potentially had no financial ties [123].

Cardiovascular disease guidelines fall into this pattern.We already noted that the 2013 cholesterol guidelines hadsignificant COI, but this is not a new phenomenon. Manyof the authors and sponsors (including the ACC and AHA)of prior CVD guidelines have had financial ties to pharma-ceutical and other companies [124,125]. Eight of the nineauthors of the 2004 cholesterol lowering guidelines hadfinancial ties to statin makers [107]. A study by Mendelsonet al. [118] of 17 ACC/AHA guidelines through 2008 foundthat 56% of 498 guideline committee members disclosedCOIs. The connections with industry were not just aboutresearch: a significant number of committee members

engaged in promotional speaking for, and had stock owner-ship in, companies producing products recommended bythe committee [118]. As Nissen [126] comments:

h Policy

mCgp[ifancbmdugq

5

ilbc[sit

bcwwldtagh

gtbcom

L. Unruh et al. / Healt

. . .no conceivable logic can defend the practice ofincluding promotional speakers and stockholders onCPG writing committees. Participants in speaker’sbureaus essentially become temporary employees ofindustry, whose duty is the promotion of the company’sproducts. . . To allow such individuals to write CPGsdefies logic (584–585) [126].

COI is not independent of research biases, but insteaday contribute to those biases. There is evidence that

OI impacts individual study results, synthesis results, anduideline recommendations. Clinical drug trials funded byharmaceutical companies are less likely to be published127], and if published more likely to have outcomes favor-ng the company [127–130]. A recent Cochrane reviewound that industry-supported clinical trials and meta-nalyses of those trials are significantly more likely thanon-commercially funded studies to report positive effi-acy and safety results [131]. A correlation has been foundetween guideline committee members with ties to phar-aceutical companies and guidelines that have expanded

isease categories [132]. Cosgrove, et al. [117] found thatnder conditions of significant COI less than half of theuidelines for psychiatric conditions met criteria for highuality [117].

. Pharmaceuticalization of clinical guidelines?

When commercial interests influence the design of theresearch protocol, the selection of research subjects, theconduct of the trial, the collection of data, the interpre-tation of results, and the publication of the outcome,there is good reason to worry about the integrity of theprocess that produces medical evidence ([133], p. 2698).

It has been noted that the biases and conflicts ofnterest inherent in the evidence base of clinical guide-ines support an approach to healthcare that expands theoundaries of “disease” and emphasizes the use of pharma-euticals and other medical products to treat those diseases109,134,135]. Environmental, socio-economic, and life-tyle components of maintaining health and preventingllness are often not part of this research or the guidelineshat emerge from it.

An example of this trend can be found in a 2013 studyy Moynihan et al. [132]. Out of 16 clinical guidelines on 14ommon conditions from 2000 to 2013, ten recommendedidening and one narrowing the definitions of disease. Fiveere unclear. Widening included creating “predisease,”

owering diagnostic thresholds, and proposing earlier orifferent diagnostic methods [132]. None of the guidelineshat proposed widening the disease categories examinedny adverse outcomes coming from the change. Of the 14uidelines that disclosed COI, 75% of the authors on averagead ties to industry.

Specific examples of this trend include the clinicaluidelines for diabetes, depression, and CVD in the U.S. Inhe diabetes guidelines normal blood glucose targets have

een lowered over time so that more individuals now areonsidered diabetic or prediabetic and need to be placedn medication [124,136]. Depression guidelines recom-end that mild depression be treated with antidepressants

120 (2016) 797–808 803

even though there is not enough evidence to support thisrecommendation [134]. CVD guidelines have over time rec-ommended lower and lower LDL or total cholesterol levelsuntil the 2013 guideline, which switched to a risk algo-rithm at a low level of risk that resulted in an even broaderpopulation considered to be at risk [136,137].

The trend to redefine the boundaries of disease andmedical care has engendered a number of labels. The con-cept of “disease-mongering” or “selling of illness” describesthe tendency to move formerly healthy biological param-eters and conditions into the realm of disease [137], whilethe related labels of “medicalization” and “pharmaceu-ticalization” describe the processes of applying medicalproducts and drugs to treat those diseases [138–141].With the concept of disease mongering, normal char-acteristics of aging and health problems amenable tonon-pharmaceutical interventions (such as low risk fac-tors for chronic diseases) are considered to be “widespread,severe and treatable with pills [137]. “Examples includehigh cholesterol, osteoporosis, sexual dysfunctions, andattention deficit hyperactivity disorder [142].

With the concept of medicalization, an aspect of healthis defined in medical terms, such as an illness, and istreated with a medical intervention, often pharmaceuticals[139,140]. Widening diagnostic criteria for ADHD and bi-polar disorders are examples in which concerns exist thatthe thresholds indicating abnormalities have been set toolow [138].

Pharmaceuticalization is the “process by which social,behavioral or bodily conditions are treated or deemed to bein need of treatment with medical drugs”(604) [138]. Theexpanding markets for antidepressants, mood stabilizersand antipsychotics, and their off-label utilization (market-ing the drugs for conditions they have not been approvedfor), have been cited as examples of this trend [138]. Mostgermane to our case study of cholesterol guidelines, Pol-lock and Jones [143] discuss whether the growth in statinprescriptions is a pharmaceuticalization of coronary arterydisease, but they also bring into the picture the possibil-ity of a “surgicalization” of CVD, with the increasing use ofbypass surgery and angioplasty [143].

When it comes to the role of clinical guidelines in allthis, the issue may be more complex than a simple guide-line reinforcement of these tendencies. Guideline changesmay be instrumental to these developments in that they rec-ommend redefining disease to exist at low levels of clinicalsymptoms and to need medical intervention, the simplestand most efficacious to be pharmaceuticals [135,143]. Sinceproviders rely upon guidelines for their practice, thesedevelopments in clinical guidelines could play a significantrole in perpetrating the selling of illness and the medical-ization/pharmaceuticalization of medical care.

Clearly, not all medical progress over the past decadescan be placed under these labels. A number of medical pro-fessionals speak of the changes in medical care in termsof advances in medicine and the benefits of new drugs andtechnology [109]. Under this perspective advances in medi-

cal research enables medical care to be better at diagnosing,preventing and treating disease. Busfield [109] calls thisperspective the “progressive” model. Similar to the con-cept of the changes as progressive, others believe that the

h Policy

804 L. Unruh et al. / Healt

issue is one of “biomedicalism” [138]. That is, healthcareis proceeding in a positive direction in which health isimproved by the increasing use of technology to preventand treat conditions. The more technology, the better.

The reality of clinical research and guidelines no doubtreflects a mixture of these tendencies. The ability to identifyrisks of chronic disease and begin preventive interventionsfor those diseases before symptoms are felt and serious ill-ness develops is an important step forward in medical care.However placing large numbers of individuals on pharma-ceuticals because they are at a questionably determinedsmall risk of developing a condition in the future is not aprogressive step forward.

6. Policies for future guideline development

Several policies could be implemented to make clinicalguidelines more evidence based and freer from indus-try influence, and therefore reduce the tendencies towarddisease mongering, medicalization and pharmaceutical-ization. First, the guideline evidence base should becomposed of well-designed and conducted RCTs and non-RCT studies that assess the treatment effectiveness inall eligible populations (e.g., low risk, older adults, eth-nic and racial groups, women, etc.) and over time. RCTsneed to be designed to test drugs on all target popula-tions, with clinical health outcomes that matter, such asall-cause mortality, patient-reported outcomes, and sideeffects/adverse events [144–146]. Well-conducted obser-vational studies should be seen as complementary to RCTs,as in the European and Japanese guidelines [147,148],providing information about how the recommended treat-ment affects the population over time. A national orinternational registry could be utilized to collect data oneffects and side effects on the population. Medical jour-nals could devote regular space for the purpose of criticallyreviewing the results of published trials [138]. Sharing ofanonymous individual patient data with non-trial inves-tigators would also aid transparency [44]. Additionally,studies of lifestyle and socio-economic approaches topreventing illnesses should be considered in guideline for-mation.

Many of the biases encountered in clinical research (nar-row and select samples, changing outcomes during thestudy, unpublished negative results) could be minimizedif all trials were required to register their research objec-tives, outcomes measures and design at the start of thetrial, and all trial results were reported in the registry[92,149–151]. To some extent this is already a require-ment of drug regulators in several countries (including theU.S. FDA) [152], of the International Committee of MedicalJournal Editors [153], and of the signatories to the Dec-laration of Helsinki for medical research [154]. However,it has yet to be enforced [152–155]. Negative results ofwell designed studies need to be embraced and openlypublished, which means that journal editors and review-

ers must be more open to this [149,150]. Many scholarsare also asking for the release of primary source data sothat others can replicate trials and perform patient-levelmeta-analyses [131,149,156,157].

120 (2016) 797–808

A major focus of policies should be to make guidelines,and the clinical research they draw from, freer from indus-try influence. In the past, COI has been addressed throughdeclarations, but this is not enough [125,158]. It does notnecessarily eliminate the COI, and in fact may make mattersworse since it provides the illusion that by declaring COIthe research has not been influenced by it [159]. Lexchinand O’Donovan [160] recommend prohibiting conflict ofinterest altogether, but do not discuss how this is to beaccomplished. Short of prohibition, clinical trial investiga-tors contact with the pharmaceutical industry should belimited to obtaining the funding for their trial; in otherwords, no speaking engagements, gifts, stock ownership,employment, or consultation with a pharmaceutical com-pany. Importantly, the research design, conduct, results,interpretation and dissemination should remain indepen-dent of industry management.

7. Conclusion

In summary, there is ongoing controversy over the 2013cholesterol guidelines in terms of their evidence base forstatin initiation in primary prevention for low risk individ-uals and their conflicts of interest with the pharmaceuticalindustry. The cholesterol guidelines are typical of otherclinical guidelines, in which reliance on RCTs and con-flicts of interest call into question the evidence base for therecommendations. Based on this problematic evidence, anumber of clinical guidelines, and cholesterol guidelinesin particular, have been expanding disease boundaries andcalling for greater pharmaceutical intervention, calling intoquestion whether the guidelines are a gain in preven-tion or in “pharmaceuticalization.” More appropriate andopen development of evidence and guidelines is neededto ensure that human conditions are properly defined andtreated with the best balance of pharmaceutical, lifestyle,and other interventions.

References

[1] Center for Disease Control & Prevention. Health, United States. Fig-ure 17, Centers for Disease Control and Prevention; 2010. Retrievedfrom http://www.cdc.gov/nchs/data/hus/hus10.pdf.

[2] Kaplan RM. Disease, diagnoses, and dollars. NY: Copernicus Books;2009.

[3] Stone NJ, Robinson J, Lichtenstein AH, Bairey Merz CN, BlumCB, Eckel RH, et al. ACC/AHA Guideline on the treatment ofblood cholesterol to reduce atherosclerotic cardiovascular risk inadults: a report of the American College of Cardiology/AmericanHeart Association task force on practice guidelines. Circulation2013;129:S1–45.

[4] Pencina M, Navar-Boggan A, D’Agostino Sr R, Williams K, NeelyB, Sniderman A, et al. Application of new cholesterol guidelinesto a population-based sample. New England Journal of Medicine2014;370(15):1422–31.

[5] Stone NJ, Lloyd-Jones DM. Lowering LDL cholesterol is good,but how and in whom? The New England Journal of Medicine2015;372(16):1564–5.

[6] Kolata G. F.D.A. approves Repatha, a second drug for cholesterol ina potent new class. New York Times 2015. http://www.nytimes.com/2015/08/28/health/fda-approves-another-in-a-new-class-of-cholesterol-drugs.html? r=1.

[7] Andersson C, Enserro D, Larson MG, Xanthakis V, Vasan RS. Implica-tions of the US cholesterol guidelines on eligibility for statin therapyin the community: comparison of observed and predicted risksin the Framingham Heart Study Offspring Cohort. Journal of theAmerican Heart Association 2015;4(4):e001888.

h Policy

2014;35(3):507–13.

L. Unruh et al. / Healt

[8] Cook NR, Ridker PM. Further insight into the cardiovascular riskcalculator: the roles of statins, revascularizations, and underascer-tainment in the Women’s Health Study. Journal of the AmericanMedical Association Internal Medicine 2014;174(12):1964–71.

[9] Khalili D, Asgari S, Hadaegh F, Steyerberg EW, Rahimi K, FahimfarN, et al. A new approach to test validity and clinical usefulness ofthe 2013 ACC/AHA guideline on statin therapy: a population-basedstudy. International Journal of Cardiology 2015;184:587–94.

[10] Ridker PM, Cook NR. Statins: new American guidelines for preven-tion of cardiovascular disease. The Lancet 2013;382:1762–4.

[11] Miedema MD, Lopez FL, Blaha MJ, Virani SS, Coresh J, BallantyneCM, et al. Eligibility for statin therapy according to new cholesterolguidelines and prevalent use of medication to lower lipid levels inan older US cohort: the atherosclerosis risk in communities studycohort. JAMA Internal Medicine 2015;175(1):138–40.

[12] Schoen MW, Salas J, Scherrer JF, Buckhold FR. Cholesterol treatmentand changes in guidelines in an academic medical practice. TheAmerican Journal of Medicine 2015;128(4):403–9.

[13] Tran JN, Caglar T, Stockl KM, Lew HC, Solow BK, Chan PS. Impactof the new ACC/AHA Guidelines on the treatment of high bloodcholesterol in a managed care setting. American Health & DrugBenefits 2014;7(8):430–43.

[14] Karmali KN, Goff DC, Ning H, Lloyd-Jones DM. A systematic exam-ination of the 2013 ACC/AHA pooled cohort risk assessment toolfor atherosclerotic cardiovascular disease. Journal of the AmericanCollege of Cardiology 2013;64:959–68.

[15] Lenzer J. Majority of panelists on controversial new cholesterolguideline have current or recent ties to drug manufacturers. BritishMedical Journal 2013;347:f6989.

[16] Mihaylova B, Emberson J, Blackwell L, Keech A, Simes J, Barnes EH,et al. The effects of lowering LDL cholesterol with statin therapy inpeople at low risk of vascular disease: meta-analysis of individualdata from 27 randomised trials. Lancet 2012;380(9841):581–90.

[17] Taylor F. Statins for the primary prevention of cardiovascular dis-ease. Cochrane Database of Systematic Reviews 2013;1. CD004816.

[18] Johnson KM, Dowe DA. Accuracy of statin assignment using the2013 AHA/ACC Cholesterol Guideline versus the 2001 NCEP ATP IIIGuideline: correlation with atherosclerotic plaque imaging. Journalof the American College of Cardiology 2014;64(9):910–9.

[19] Rhee E, Park SE, Oh HG, Park C, Oh K, Park S, et al. Statin eligibilityand cardiovascular risk burden assessed by coronary artery cal-cium score: Comparing the two guidelines in a large Korean cohort.Atherosclerosis 2015;240(1):242–9.

[20] Pursnani A, Massaro JM, D’Agostino RS, O’Donnell CJ, Hoffmann U.Guideline-based statin eligibility, coronary artery calcification, andcardiovascular events. Journal of the American Medical Association2015;314(2):134–41.

[21] Abramson J, Wright J. Are lipid-lowering guidelines evidence-based? Lancet 2007;369(9557):168–9.

[22] Abramson JD, Rosenberg HG, Jewell N, Wright JM. Should peopleat low risk of cardiovascular disease take a statin? British MedicalJournal 2013;347:1–5.

[23] McAlister FA, van Diepen S, Padwal RS, Johnson JA, Majumdar SR.How evidence-based are the recommendations in evidence-basedguidelines? PLoS Medicine 2007;4(8):e250.

[24] Petersen LK, Christensen K, Kragstrup J. Lipid-lowering treat-ment to the end? A review of observational studies and RCTson cholesterol and mortality in 80+-year olds. Age and Ageing2010;39(6):674–80.

[25] Donzelli A. Statins for people at low risk of cardiovascular disease.Lancet 2012;380(9856):1814–5.

[26] Buettner C, Davis RB, Leveille SG, Mittleman MA, Mukamal KJ. Prev-alence of musculoskeletal pain and statin use. Journal of GeneralInternal Medicine 2008;23(8):1182–6.

[27] Cohen JD, Brinton EA, Ito MK, Jacobson TA. Understanding StatinUse in America and Gaps in Patient Education (USAGE): an internet-based survey of 10,138 current and former statin users. Journal ofClinical Lipidology 2012;6(3):208–15.

[28] Fernandez G, Spatz ES, Jablecki C, Phillips PS. Statin myopathy: acommon dilemma not reflected in clinical trials. Cleveland ClinicJournal of Medicine 2011;78(6):393–403.

[29] Majeed A. Increasing the use of statins in people at low cardiovas-cular risk is difficult. British Medical Journal 2013;347:f6901.

[30] Hippisley-Cox J, Coupland C. Unintended effects of statins inmen and women in England and Wales: population based cohortstudy using the QResearch database. British Medical Journal

2010;340:c2197.

[31] Rallidis LS, Anastasiou-Nana M. Is myopathy the achilles’ heelof statins? Differences between the new cholesterol treatment

120 (2016) 797–808 805

guidelines and everyday clinical practice. Journal of the AmericanCollege of Cardiology 2014;63(21):2300–1.

[32] Preiss D, Seshasai SR, Welsh P, Murphy SA, Ho JE, Waters DD,et al. Risk of incident diabetes with intensive-dose comparedwith moderate-dose statin therapy: a meta-analysis. Journal of theAmerican Medical Association 2011;305(24):2556–64.

[33] Ridker PM, Pradhan A, MacFadyen JG, Libby P, Glynn RJ. Car-diovascular benefits and diabetes risks of statin therapy inprimary prevention: an analysis from the JUPITER trial. Lancet2012;380(9841):565–71.

[34] Culver AL, Ockene IS, Balasubramanian R, Olendzki BC, SepavichDM, Wactawski-Wende J, et al. Statin use and risk of diabetes mel-litus in postmenopausal women in the women’s health initiative.Archives of Internal Medicine 2012;172(2):144–52.

[35] Sattar N, Preiss D, Murray HM, Welsh P, Buckley BM, deCraen AJ, et al. Statins and risk of incident diabetes: a col-laborative meta-analysis of randomised statin trials. Lancet2010;375(9716):735–42.

[36] Leuschen J, Mortensen EM, Frei CR, Mansi EA, Panday V, MansiI. Association of statin use with cataracts: a propensity score-matched analysis. JAMA Ophthalmology 2013;131(11):1427–34.

[37] Machan CM, Hrynchak PK, Irving EL. Age-related cataract is asso-ciated with type 2 diabetes and statin use. Optometry & VisionScience 2012;89(8):1165–71.

[38] McDougall JA, Malone KE, Daling JR, Cushing-Haugen KL, Porter PL,Li CI. Long-term statin use and risk of ductal and lobular breastcancer among women 55 to 74 years of age. Cancer Epidemiology,Biomarkers and Prevention 2013;22(9):1529–37.

[39] de Denus S, Spinler SA, Miller K, Peterson AM. Statins and livertoxicity: a meta-analysis. Pharmacotherapy 2004;24(5):584–91.

[40] Kraft R, Kahn A, Medina-Franco JL, Orlowski ML, Baynes C, Lopez-Vallejo F, et al. A cell-based fascin bioassay identifies compoundswith potential anti-metastasis or cognition-enhancing functions.Disease Models & Mechanisms 2013;6(1):217–35.

[41] Strom BL, Schinnar R, Karlawish J, Hennessy S, Teal V, BilkerWB. Statin therapy and risk of acute memory impairment. JAMAInternal Medicine 2015, http://dx.doi.org/10.1001/jamainternmed.2015.2092 [online first].

[42] Light DW, Lexehin J, Darrow JJ. Institutional corruption of pharma-ceuticals and the myth of safe and effective drugs. Journal of Law,Medicine & Ethics 2013;41(3):590–600.

[43] Sinatra S, Teter B, Bowden J, Houston M, Martinez-GonzalezM. The cholesterol and statin controversy: the new 2013statin-cholesterol guidelines. Alternative Therapies in Health andMedicine 2014;20(5):14–7.

[44] Parish E, Bloom T, Godlee F. Statins for people at low risk. BMJ(Clinical Research Ed.) 2015;351:h3908.

[45] Ahmed HM, Blaha MJ, Blumenthal RS. Modifiable lifestyle risks, car-diovascular disease, and all-cause mortality. International Journalof Cardiology 2014;172(1):e199–200.

[46] Lazin MM, Jalil RA, Wan Muda WM, Wan Nik WS, Zakaria R. USMbehavioural lifestyle modification program reduces lipid-basedcardiovascular risk in obese adults: a pilot study. InternationalMedical Journal 2014;21(6):558–61.

[47] Lin JS, O’Connor E, Evans CV, Senger CA, Rowland MG, GroomHC. Behavioral counseling to promote a healthy lifestyle in per-sons with cardiovascular risk factors: a systematic review for theU.S. Preventive Services Task Force. Annals of Internal Medicine2014;161(8):568–78.

[48] Opie LH, Dalby AJ. Cardiovascular prevention: lifestyle and statins– competitors or companions? South African Medical Journal2014;104(3):168–73.

[49] Schwingshackl L, Dias S, Hoffmann G. Impact of long-term lifestyleprograms on weight loss and cardiovascular risk factors in over-weight/obese participants: a systematic review and networkmeta-analysis. Systematic Reviews 2014:3130.

[50] Cuenca-García M, Ortega FB, Ruiz JR, González-Gross M, Labayen I,Jago R, et al. Combined influence of healthy diet and active lifestyleon cardiovascular disease risk factors in adolescents. ScandinavianJournal of Medicine & Science in Sports 2014;24(3):553–62.

[51] Flynn SE, Gurm R, DuRussel-Weston J, Aaronson S, Gaken-heimer L, Smolarski J, et al. High-density lipoprotein cholesterollevels in middle-school children: association with cardiovas-cular risk factors and lifestyle behaviors. Pediatric Cardiology

[52] Saffi ML, Polanczyk CA, Rabelo-Silva ER. Lifestyle interventionsreduce cardiovascular risk in patients with coronary artery disease:a randomized clinical trial. European Journal of CardiovascularNursing 2014;13(5):436–43.

h Policy

806 L. Unruh et al. / Healt

[53] Akesson A, Larsson SC, Discacciati A, Wolk A. Low-risk diet andlifestyle habits in the primary prevention of myocardial infarctionin men: a population-based prospective cohort study. Journal of theAmerican College of Cardiology 2014;64:1299–306.

[54] Del Gobbo LC, Kalantarian S, Imamura F, Lemaitre R, Siscovick DS,Psaty BM, et al. Contribution of major lifestyle risk factors for inci-dent heart failure in older adults: the cardiovascular health study.JACC, Heart Failure 2015;3(7):520–8.

[55] Rippe JM, Angelopoulos TJ. Lifestyle strategies for cardiovascularrisk reduction. Current Atherosclerosis Reports 2014;16(10):444.

[56] Petersen KN, Johnsen NF, Olsen A, Albieri V, Olsen LH, Dragsted LO,et al. The combined impact of adherence to five lifestyle factorson all-cause, cancer and cardiovascular mortality: a prospectivecohort study among Danish men and women. British Journal ofNutrition 2015;113(5):849–58.

[57] Crowe C, Gibson I, Cunningham K, Kerins C, Costello C, WindleJ, et al. Effects of an eight-week supervised, structured lifestylemodification programme on anthropometric, metabolic and car-diovascular risk factors in severely obese adults. BMC EndocrineDisorders 2015;15(1):37.

[58] Dutton GR, Laitner MH, Perri MG. Lifestyle interventions for cardio-vascular disease risk reduction: a systematic review of the effectsof diet composition, food provision, and treatment modality onweight loss. Current Atherosclerosis Reports 2014;16(10):442.

[59] Goyer L, Dufour R, Janelle C, Blais C, L’Abbé C, Raymond É, et al. Ran-domized controlled trial on the long-term efficacy of a multifaceted,interdisciplinary lifestyle intervention in reducing cardiovascularrisk and improving lifestyle in patients at risk of cardiovasculardisease. Journal of Behavioral Medicine 2013;36(2):212–24.

[60] Greaves C, Gillison F, Stathi A, Bennett P, Reddy P, Dunbar J, et al.Waste the waist: a pilot randomized controlled trial of a primarycare based intervention to support lifestyle change in people withhigh cardiovascular risk. The International Journal of BehavioralNutrition and Physical Activity 2015;12:1.

[61] Jarl J, Tolentino JC, James K, Clark MJ, Ryan M. Supporting cardiovas-cular risk reduction in overweight and obese hypertensive patientsthrough DASH diet and lifestyle education by primary care nursepractitioners. Journal of the American Association of Nurse Practi-tioners 2014;26(9):498–503.

[62] Kramer MK, Molenaar DM, Arena VC, Venditti EM, Meehan RJ, MillerRG, et al. Improving employee health: evaluation of a worksitelifestyle change program to decrease risk factors for diabetes andcardiovascular disease. Journal of Occupational and EnvironmentalMedicine 2015;57(3):284–91.

[63] Rautio N, Jokelainen J, Pölönen A, Oksa H, Peltonen M, Vanhala M,et al. Changes in lifestyle modestly reduce the estimated cardio-vascular disease risk in one-year follow-up of the Finnish diabetesprevention program (FIN-D2D). European Journal of CardiovascularNursing 2015;14(2):145–52.

[64] Stephenson J. USPSTF: lifestyle counseling advised for overweight,obese adults with other cardiovascular risk factors. Journal of theAmerican Medical Association 2014;312(11):1085.

[65] Liburd LC, Jack LJ, Williams S, Tucker P. Intervening on the socialdeterminants of cardiovascular disease and diabetes. AmericanJournal of Preventive Medicine 2005;29(5 (Suppl. 1)):18–24.

[66] Kreatsoulas C, Anand SS. The impact of social determinantson cardiovascular disease. The Canadian Journal of Cardiology2010;26(Suppl.):C8C–13C.

[67] Damiani G, Federico B, Bianchi CB, Ronconi A, Basso D, FiorenzaS, et al. Socio-economic status and prevention of cardiovasculardisease in Italy: evidence from a national health survey. EuropeanJournal of Public Health 2011;21(5):591–6.

[68] Davis SK, Gebreab S, Quarells R, Gibbons GH. Social determinantsof cardiovascular health among black and white women residingin Stroke Belt and Buckle regions of the South. Ethnicity & Disease2014;24(2):133–43.

[69] Dunton GF, Kaplan J, Wolch J, Jerrett M, Reynolds KD. Physical envi-ronmental correlates of childhood obesity: a systematic review.Obesity Reviews: An Official Journal of The International Associ-ation for the Study of Obesity 2009;10(4):393–402.

[70] Greiser KH, Tiller D, Kuss O, Kluttig A, Schumann B, Werdan K, et al.034 Association of neighbourhood socio-economic status and indi-vidual socio-economic status with cardiovascular risk factors in anEastern German population—The CARLA Study 2002–2006. Journal

of Epidemiology & Community Health 2010;64:A13–64A14.

[71] Kristensen PL, Wedderkopp N, Møller NC, Andersen LB, Bai CN,Froberg K. Tracking and prevalence of cardiovascular disease riskfactors across socio-economic classes: a longitudinal substudy ofthe European Youth Heart Study. BMC Public Health 2006;6:20.

120 (2016) 797–808

[72] McGibbon E, Waldron I, Jackson J. The social determinants of car-diovascular disease: time for a focus on racism. Diversity & Equalityin Health & Care 2013;10(3):139–42.

[73] Palomo L, Félix-Redondo F, Lozano-Mera L, Pérez-Castán J,Fernández-Berges D, Buitrago F. Cardiovascular risk factors,lifestyle, and social determinants: a cross-sectional populationstudy. The British Journal of General Practice 2014;64(627):e627–33.

[74] Panagiotakos DB, Pitsavos C, Manios Y, Polychronopoulos E, Chryso-hoou CA, Stefanadis C. Socio-economic status in relation to riskfactors associated with cardiovascular disease, in healthy individ-uals from the ATTICA study. European Journal of CardiovascularPrevention and Rehabilitation 2005;12(1):68–74.

[75] Quarells RC, Liu J, Davis SK. Social determinants of cardiovasculardisease risk factor presence among rural and urban Black and Whitemen. Journal of Men’s Health 2012;9(2):120–6.

[76] Scholes S, Bajekal M, Raine R, O’Flaherty M, Capewell S. Socio-economic trends in cardiovascular risk factors in England,1994–2008. Journal of Epidemiology & Community Health2010;64:A13.

[77] Solovieva S, Lallukka T, Virtanen M, Viikari-Juntura E. Psychoso-cial factors at work, long work hours, and obesity: a systematicreview. Scandinavian Journal of Work, Environment & Health2013;39(3):241–58.

[78] Stafford M, Soljak M, Pledge V, Mindell J. Socio-economic differ-ences in the health-related quality of life impact of cardiovascularconditions. European Journal Of Public Health 2012;22(3):301–5.

[79] Vescio MF, Smith GD, Giampaoli S. Socio-economic position andcardiovascular risk factors in an Italian rural population. EuropeanJournal of Epidemiology 2001;17(5):449–59.

[80] Zaman MJ, Patel A, Jan S, Hillis GS, Raju PK, Neal B, et al.Socio-economic distribution of cardiovascular risk factors andknowledge in rural India. International Journal of Epidemiology2012;41(5):1302–14.

[81] Harper S, Lynch J, Smith GD. Social determinants and the decline ofcardiovascular diseases: understanding the links. Annual Review ofPublic Health 2011:3239–69.

[82] Kinra S, Johnson M, Kulkarni B, Rameshwar Sarma KV, Ben-Shlomo Y, Smith GD. Socio-economic position and cardiovascularrisk in rural Indian adolescents: evidence from the AndhraPradesh children and parents study (APCAPS). Public Health2014;128(9):852–9.

[83] Meyer U, Schindler C, Bloesch T, Schmocker E, Zahner L, Puder JJ,et al. Combined impact of negative lifestyle factors on cardiovas-cular risk in children: a randomized prospective study. Journal ofAdolescent Health 2014;55(6):790–5.

[84] Van den Berg GJ, Doblhammer-Reiter G, Christensen K. Being bornunder adverse economic conditions leads to a higher cardiovas-cular mortality rate later in life: evidence based on individualsborn at different stages of the business cycle. Demography2011;48(2):507–30.

[85] Havranek EP, Mujahid MS, Barr DA, Blair IV, Cohen MS, Cruz-FloresS, et al. Social determinants of risk and outcomes for cardiovasculardisease: a scientific statement from the American Heart Associa-tion. Circulation 2015;132(9):873–98.

[86] Pandya A, Sy S, Cho S, Weinstein MC, Gaziano TA. Cost-effectivenessof 10-year risk thresholds for initiation of statin therapy for pri-mary prevention of cardiovascular disease. Journal of the AmericanMedical Association 2015;314(2):142–50.

[87] Muntner P, Gu D, Reynolds R, Wu X, Chen J, Whelton P, et al.Therapeutic lifestyle changes and drug treatment for high bloodcholesterol in China and application of the Adult TreatmentPanel III guidelines. American Journal of Cardiology 2005;96(9):1260–5.

[88] Ioannidis JA. More than a billion people taking statins? Potentialimplications of the new cardiovascular guidelines. Journal of theAmerican Medical Association 2014;311(5):463–4.

[89] Vaucher J, Marques-Vidal P, Preisig M, Waeber G, Vollenweider P.Population and economic impact of the 2013 ACC/AHA guidelinescompared with European guidelines to prevent cardiovascular dis-ease. European Heart Journal 2014;35:958–9.

[90] Kavousi M, Leening M, Nanchen D, Greenland P, Graham I, Steyer-berg E, et al. Comparison of application of the ACC/AHA guidelines,Adult Treatment Panel III guidelines, and European Society of

Cardiology guidelines for cardiovascular disease prevention in aEuropean cohort. Journal of the American Medical Association2014;311(14):1416–23.

[91] Institute of Medicine. Clinical practice guidelines: directions for anew program. Washington, DC: National Academies Press; 1990.

h Policy

L. Unruh et al. / Healt

[92] Pilling S. History, context, process, and rationale for the devel-opment of clinical guidelines. Psychology and Psychotherapy2008;81(Pt 4):331–50.

[93] Sniderman AD, Furberg CD. Why guideline-making requiresreform. Journal of the American Medical Association 2009;301(4):429–31, http://dx.doi.org/10.1001/jama.2009.15.

[94] AHRQ. National Guideline Clearing House. Agency for HealthcareResearch and Quality; 2016. http://www.guideline.gov/.

[95] Pyon E. Primer on clinical practice guidelines. Journal of PharmacyPractice 2013;26(2):103–11.

[96] Jadad A, Enkin M. Randomized controlled trials: questions, answersand musings., 2nd ed. Oxford: Blackwell; 2007.

[97] Warden G.Definition of bias in clinical research. In: Parfrey PS, Bar-rett BJ, editors. Clinical epidemiology: practice and methods, 2nded. N.Y.: Humana Press, (Springer); 2013.

[98] Borgerson K. Valuing evidence: bias and the evidence hierarchyof evidence-based medicine. Perspectives in Biology and Medicine2009;52(2):218–33.

[99] Rosen M, Axelsson S, Lindblom J. Observational studies versusRCTs: what about socioeconomic factors? Lancet 2009;373(9680):2026.

[100] Booth CM, Tannock IF. Evaluation of treatment benefit: random-ized controlled trials and population-based observational research.Journal of Clinical Oncology 2013;31(26):3298–9.

[101] de Simone G, Izzo R, Verdecchia P. Are observational studies moreinformative than randomized controlled trials in hypertension? Proside of the argument. Hypertension 2013;62(3):463–9.

[102] Kovesdy CP, Kalantar-Zadeh K. Observational studies versusrandomized controlled trials: avenues to causal inference innephrology. Advances in Chronic Kidney Disease 2012;19(1):11–8,8p.

[103] Silverman SL. From randomized controlled trials to observa-tional studies. The American Journal of Medicine 2009;122(2):114–20.

[104] Sniderman AD, Furberg CD. Pluralism of viewpoints as the antidoteto intellectual conflict of interest in guidelines. Journal of ClinicalEpidemiology 2012:705–7.

[105] Rothwell P. External validity of randomised controlled trials: ‘towhom do the results of this trial apply?’. Lancet (North AmericanEdition) 2005;365(9453):82–93.

[106] Shrier I, Boivin J, Steele RJ, Platt RW, Furlan A, Kakuma R,et al. Should meta-analyses of interventions include observationalstudies in addition to randomized controlled trials? A critical exam-ination of underlying principles. American Journal of Epidemiology2007;166(10):1203–9.

[107] Abramson J, Starfield B. The effect of conflict of interest on biomed-ical research and clinical practice guidelines: can we trust theevidence in evidence-based medicine? The Journal of the AmericanBoard of Family Practice 2005;18(5):414–8.

[108] Lurie P, Almeida CM, Stine N, Stine AR, Wolfe SM. Financial con-flict of interest disclosure and voting patterns at Food and DrugAdministration Drug Advisory Committee meetings. Journal of theAmerican Medical Association 2006;295(16):1921–8.

[109] Busfield J. ‘A pill for every ill’: explaining the expansion in medicineuse. Social Science & Medicine 2010;70(6):934–41.

[110] Gotzsche PC, Hrobjartsson A, Johansen HK, Haahr MT, Altman DG,Chan AW. Ghost authorship in industry-initiated randomised trials.PLoS Med 2007;4(1):e19.

[111] Moffatt B, Elliott C. Ghost marketing: pharmaceutical companiesand ghostwritten journal articles. Perspectives in Biology andMedicine 2007;50(1):18–31.

[112] Sismondo S. How pharmaceutical industry funding affects trialoutcomes: causal structures and responses. Social Science andMedicine 2008;66:1904–14.

[113] Sismondo S, Doucet M. Publication ethics and the ghost manage-ment of medical publication. Bioethics 2010;24(6):273–83.

[114] Fava GA. Meta-analyses and conflict of interest. CNS Drugs2012;26(2):93–6.

[115] Roseman M, Milette K, Bero LA, Coyne JC, Lexchin J, Turner EH,et al. Reporting of conflicts of interest in meta-analyses of trialsof pharmacological treatments. Journal of the American MedicalAssociation 2011;305(10):1008–17.

[116] Choudhry NK, Stelfox HT, Detsky AS. Relationships betweenauthors of clinical practice guidelines and the pharmaceu-

tical industry. Journal of the American Medical Association2002;287(5):612–7.

[117] Cosgrove L, Wheeler EE. Drug firms, the codification of diagnos-tic categories, and bias in clinical guidelines. The Journal of Law,Medicine & Ethics 2013;41(3):644–53.

120 (2016) 797–808 807

[118] Mendelson TB, Meltzer M, Campbell EG, Caplan AL, Kirkpatrick JN.Conflicts of interest in cardiovascular clinical practice guidelines.Archives of Internal Medicine 2011;171(6):577–84.

[119] Norris SL, Holmer HK, Ogden LA, Burda BU. Conflict of interest inclinical practice guideline development: a systematic review. PLoSONE 2011;6(10):e25153.

[120] Norris SL, Holmer HK, Ogden LA, Burda BU, Fu R. Conflicts of interestamong authors of clinical practice guidelines for glycemic controlin type 2 diabetes mellitus. PLoS ONE 2013;8(10):e75284.

[121] Lenzer J. French guidelines are withdrawn after court finds poten-tial bias among authors. British Medical Journal 2011;342:d4007.

[122] Langer T, Conrad S, Fishman L, Gerken M, Schwarz S, Weikert B,et al. Conflicts of interest among authors of medical guidelines.Deutsches Arzteblatt International 2012;109(48):836–42.

[123] Neuman J, Korenstein D, Ross JS. Conflicts of interest among guide-line panel members in Canada and United States: cross sectionalstudy. British Medical Journal 2011;343:d5621.

[124] Hunt LM, Kreiner M, Brody H. The changing face of chronic illnessmanagement in primary care: a qualitative study of underlyinginfluences and unintended outcomes. Annals of Family Medicine2012;10(5):452–60.

[125] Lenzer J, Hoffman JR, Furberg CD, Ioannidis JP. Guideline PanelReview Working Group Ensuring the integrity of clinical practiceguidelines: a tool for protecting patients. British Medical Journal2013;347:f5535.

[126] Nissen SE. Can we trust cardiovascular practice guidelines? Com-ment on “Conflicts of interest in cardiovascular clinical practiceguidelines.”. Archives of Internal Medicine 2011;171(6):584–5.

[127] Lexchin J, Bero LA, Djulbegovic B, Clark O. Pharmaceutical industrysponsorship and research outcome and quality: systematic review.British Medical Journal 2003;326(7400):1167–70.

[128] Bekelman J, Li Y, Gross C. Scope and impact of financial conflicts ofinterest in biomedical research: a systematic review. Journal of theAmerican Medical Association 2003;289:454–65.

[129] Lexchin JR. Implications of pharmaceutical industry fund-ing on clinical research. The Annals of Pharmacotherapy2005;39(1):194–7.

[130] Sismondo S. Pharmaceutical company funding and its conse-quences: a qualitative systematic review. Contemporary ClinicalTrials 2008;29(2):109–13.

[131] Bero L. Industry sponsorship and research outcome: a Cochranereview. JAMA Internal Medicine 2013;173(7):580–1.

[132] Moynihan RN, Cooke GP, Doust JA, Bero L, Hill S, Glasziou PP.Expanding disease definitions in guidelines and expert panel tiesto industry: a cross-sectional study of common conditions in theUnited States. PLoS Med 2013;10(8):e1001500.

[133] De Vries R, Lemmens T. The social and cultural shaping of medicalevidence: case studies from pharmaceutical research and obstetricscience. Social Science & Medicine 2006;62(11):2694–706.

[134] Cosgrove L, Shaughnessy AF, Wheeler EE, Krimsky S, PetersSM, Freeman-Coppadge DJ, et al. From caveat emptor to caveatvenditor: time to stop the influence of money on practiceguideline development. Journal of Evaluation in Clinical Practice2014;20(6):809–12.

[135] Dumit J. Drugs for life: how pharmaceutical companies define ourhealth. Durham: Duke University Press; 2012.

[136] Brody H, Light DW. Efforts to undermine public health: theinverse benefit law: how drug marketing undermines patientsafety and public health. American Journal of Public Health2011;101(3):399–404.

[137] Moynihan R, Henry D. The fight against disease mongering. PLoSMedicine 2006;3:425–8.

[138] Abraham J. Pharmaceuticalization of society in context: theoretical,empirical and health dimensions. Sociology 2010;44(4):603–22.

[139] Bell S, Figert A. Medicalization and pharmaceuticalization at theintersections: looking backward, sideways and forward. Social Sci-ence & Medicine 2012;75(5):775–83 [in English].

[140] Conrad P, Leiter V. Medicalization, markets and consumers. Journalof Health and Social Behavior 2004;45(Suppl.):158–76.

[141] Williams SJ, Martin P, Gabe J. The pharmaceuticalisation of soci-ety? A framework for analysis. Sociology of Health & Illness2011;33(5):710–25.

[142] Moynihan R, Cassels A. Selling sickness. London: Allen & Unwin;2005.

[143] Pollock A, Jones DS. Coronary artery disease and the contoursof pharmaceuticalization. Social Science & Medicine 2015;131:221–7.

[144] Anker SD, Agewall S, Borggrefe M, Calvert M, Jaime Caro J, CowieMR, et al. The importance of patient-reported outcomes: a call for

h Policy

[165] National Clinical Guideline Centre. Lipid modification: cardiovas-cular risk assessment and the modification of blood lipids for the

808 L. Unruh et al. / Healt

their comprehensive integration in cardiovascular clinical trials.European Heart Journal 2014;35:2001–9.

[145] Calvert M, Blazeby J, Altman DG, Revicki DA, Moher D, BrundageMD, et al. Reporting of patient-reported outcomes in randomizedtrials: the CONSORT PRO extension. Journal of the American Medi-cal Association 2013;309:814–22.

[146] Golomb BA. The importance of monitoring adverse events in statin,and other, clinical trials. Clinical Investigations 2013;3(10):913–6.

[147] Perk J, Backer G, Gohlke H, Graham I, Reiner Z, Verschuren W,et al. European guidelines on cardiovascular disease prevention inclinical practice (Version 2012). International Journal of BehavioralMedicine 2012;19(4):403–88.

[148] Teramoto T, Sasaki J, Ueshima H, Egusa G, Kinoshita M, ShimamotoK, et al. Executive summary of Japan Atherosclerosis Society (JAS)guideline for diagnosis and prevention of atherosclerotic car-diovascular diseases for Japanese. Journal of Atherosclerosis andThrombosis 2007;14(2):45–50.

[149] All Trials. All trials registered, all results reported; 2015. http://www.alltrials.net/.

[150] Song F, Parekh S, Hooper L, Loke YK, Ryder J, Sutton AJ, et al. Dissem-ination and publication of research findings: an updated review ofrelated biases. Health Technology Assessment 2010;14(8):iii.

[151] Ewart R, Lausen H, Millian N. Undisclosed changes in outcomesin randomized controlled trials: an observational study. Annals ofFamily Medicine 2009;7(6):542–6, 5p.

[152] Prayle AP, Hurley MN, Smyth AR. Compliance with mandatoryreporting of clinical trial results on ClinicalTrials.gov: cross sec-tional study. BMJ (Clinical Research Ed.) 2012;344:d7373.

[153] De Angelis C, Drazen JM, Frizelle FA, Haug C, Hoey J, Horton R. Clini-cal trial registration: a statement from the International Committeeof Medical Journal Editors. The Lancet 2004;364(9438):911–2.

[154] World Medical Association. World Medical Association Declara-tion of Helsinki: Ethical principles for medical research involvinghuman subjects. Journal of the American Medical Association

2013;310(20):2191–4.

[155] Mathieu S, Boutron I, Moher D, Altman DG, Ravaud P. Compari-son of registered and published primary outcomes in randomizedcontrolled trials. Journal of the American Medical Association2009;302(9):977–84.

120 (2016) 797–808

[156] Doshi P, Dickersin K, Healy D, Vedula SS, Jefferson T. Restoring invis-ible and abandoned trials: a call for people to publish the findings.BMJ (Clinical Research Ed.) 2013;346:f2865.

[157] Jefferson T, Doshi P, Thompson M, Heneghan C. Ensuring safe andeffective drugs: who can do what it takes? BMJ (Clinical ResearchEd.) 2011;342:c7258.

[158] Harrington RA, Califf RM. Controversies in cardiovascular medicine:there is a role for industry-sponsored education in cardiology. Cir-culation 2010;121:2221–7.

[159] Loewenstein G, Sah S, Cain DM. The unintended consequences ofconflict of interest disclosure. Journal of the American MedicalAssociation 2012;307(7):669–70.

[160] Lexchin J, O’Donovan O. Prohibiting or ‘managing’ conflict of inter-est? A review of policies and procedures in three European drugregulation agencies. Social Science & Medicine 2010;70:643–7.

[161] National Heart Foundation of Australia. The National HeartFoundation of Australia’s summary of the recommendations forCholesterol Management; 2013. http://new-allirs.rhcloud.com/australian-national-heart-foundation-lipid-management/.

[162] Anderson TJ, Grégoire J, Hegele RA, Couture P, Mancini GJ, McPher-son R, et al. 2012 update of the Canadian Cardiovascular Societyguidelines for the diagnosis and treatment of dyslipidemia for theprevention of cardiovascular disease in the adult. The CanadianJournal of Cardiology 2013;29(2):151–67.

[163] Anderson TJ, Grégoire J, Hegele RA, Couture P, Mancini GJ, McPher-son R, et al. Are the ACC/AHA guidelines on the treatment of bloodcholesterol a game changer? A perspective from the Canadian Car-diovascular Society Dyslipidemia Panel. The Canadian Journal ofCardiology 2014;30(4):377–80.

[164] Smulders YM, Sern EH, Stehouwer CDA. Divergent paradigm shiftsin national, European and American cardiovascular preventionguidelines. Netherlands Journal of Medicine 2014;72(5):289–93.

primary and secondary prevention of cardiovascular disease, NICEclinical guideline 181, guidance.nice.org.uk/cg181; 2014. Retrievedfrom http://www.nice.org.uk/guidance/CG181.