Atherosclerosis Supplements 14 (2013) 1e5www.elsevier.com/locate/atherosclerosis

Hyperlipoproteinemia(a): Clinical significance and treatment options

Heiner K. Berthold a,b,*, Ioanna Gouni-Berthold c

aCharite University Medicine Berlin, Evangelical Geriatrics Center Berlin (EGZB), Berlin, GermanybCharite University Medicine Berlin, Virchow Clinic Campus, Lipid Clinic at the Interdisciplinary Metabolism Center, Berlin, Germany

cUniversity of Cologne, Center for Endocrinology, Diabetes and Preventive Medicine, Cologne, Germany

Abstract

Recent epidemiologic and Mendelian randomization studies together have provided evidence that lipoprotein(a) (Lp(a)) plays a causalrole in the pathogenesis of atherosclerosis and cardiovascular disease (CVD). The risk association with CVD is weak but seems continuousin shape and without an obvious threshold for Lp(a) levels. A plasma concentration of 60 mg/dl compared to usual levels is associated withan odds ratio for coronary heart disease of about 1.5 after adjustment for other cardiovascular risk factors. Niacin (nicotinic acid) is thepharmacologic means of choice for decreasing elevated Lp(a) levels but the drug is often poorly tolerated due to adverse reactions. Dietarymeasures, exercise and other lipid-lowering drugs, especially statins, fibrates and ezetimibe, are without effect. In patients with severeprogressive cardiovascular disease and very high Lp(a) levels, lipoprotein apheresis may be used to effectively decrease Lp(a) concen-trations. The method is expensive and impractical for most patients and its feasibility depends by and large on the healthcare reimbursementsystem of the respective country. No established treatment, however, selectively reduces Lp(a) without influencing other lipoproteins.Moreover, despite the clear association of hyperlipoproteinemia(a) with cardiovascular risk, no rigorously designed study to date hasdemonstrated that lowering Lp(a) concentrations has beneficial effects on cardiovascular endpoints. Randomized trials to this effect areurgently needed.� 2012 Elsevier Ireland Ltd. All rights reserved.

1. Introduction

Lipoprotein(a) (Lp(a)) has been discovered in 1963 byBerg [1]. It is a low-density lipoprotein-like particlesynthesized by the liver that consists of an apolipoprotein-B100 (apoB-100) molecule covalently linked to a verylarge glycoprotein known as apolipoprotein(a) (apo(a)),which has structural homologies with plasminogen [2,3].In vitro and animal studies have shown that Lp(a) canpromote thrombosis, inflammation, and foam cell forma-tion [4e6]. The overall physiological effects of Lp(a)remain elusive, but Lp(a) has been shown to enter thearterial intima of humans and thus to be involved in thepathogenesis of atherosclerosis [7]. Many prospective

* Corresponding author. Charite University Medicine Berlin, Evangelical

Geriatrics Center Berlin (EGZB), Reinickendorfer Str. 61, 13347 Berlin,

Germany. Tel.: þ49 30 450 553167; fax: þ49 30 8187 8889.

E-mail address: heiner.berthold@charite.de (H.K. Berthold).

1567-5688/$ - see front matter � 2012 Elsevier Ireland Ltd. All rights reserved.

http://dx.doi.org/10.1016/j.atherosclerosissup.2012.10.037

epidemiological studies have reported positive associationsbetween baseline Lp(a) concentrations and coronary heartdisease (CHD) risk [8e10].

The current status of Lp(a) as a cardiovascular riskfactor has been reviewed recently by a European Athero-sclerosis Society Consensus Panel [11]. The panel suggeststhat mechanistically elevated Lp(a) levels may eitherinduce a prothrombotic/anti-fibrinolytic effect as apo(a)resembles both plasminogen and plasmin but has no fibri-nolytic activity, or may accelerate atherosclerosis because,like LDL, the Lp(a) particle is rich in cholesterol, or both.

2. Association of cardiovascular disease risk with Lp(a)

2.1. Epidemiology

Recent epidemiologic evidence supports that elevatedplasma levels of Lp(a) are robustly and specifically asso-ciated with an increased risk for cardiovascular disease.

2 H.K. Berthold, I. Gouni-Berthold / Atherosclerosis Supplements 14 (2013) 1e5

The association is continuous in shape and seems withoutthreshold. It is independent of high levels of LDL choles-terol or on the levels or presence of other classic cardio-vascular risk factors [11].

Analyses of previous prospective trials have suggestedthat the risk for CHD is increased with elevated Lp(a) levelsbut it was not clear whether Lp(a) is associated with CHDthroughout the concentration range (similar to blood pres-sure and LDL cholesterol) or whether it is particularlyatherothrombogenic in specific subgroups of individuals.

The largest epidemiological study to date investigatingthe association of the risk of cardiovascular disease withelevated Lp(a) concentrations included 36 prospectivestudies with >126.000 participants [12]. For CHD,assuming a log-linear Lp(a) concentration/risk relationship,the adjusted relative risk is 1.16 (95% CI 1.09e1.18) perone standard deviation (¼3.5-fold) higher than usual Lp(a)levels [12]. The relative risk did not vary significantly bysex, non-HDL and HDL cholesterol, triglycerides, bloodpressure, diabetes, or body mass index. Most importantly,the risk association is not dependent on LDL cholesterol[13]. The adjusted relative risk for ischemic stroke is 1.10(95% CI 1.02e1.18) per 3.5-fold higher than usual Lp(a)levels and is therefore weaker than the one for CHD [12].

2.2. Genetics

Together with the described epidemiological evidence,findings from recent genetics studies suggest that elevatedLp(a) levels are causally related to the risk for atheroscle-rosis development and cardiovascular disease. Variability inthe plasma levels of Lp(a) is determined to a large extent byvariation in the apo(a) gene [2]. Apo(a) is polymorphic insize and contains a variable number from 2 to >40 of thekringle IV type 2 repeats, which is inversely associatedwith plasma Lp(a) levels [2].

A large Mendelian randomization study on the basis ofthree population studies from Copenhagen was published in2009 [14]. The data show that multivariably adjustedhazard ratios for myocardial infarction are elevateddepending on Lp(a) levels up to an HR of 2.6 (95% CI1.6e4.1) for Lp(a) levels >95th percentile (correspondingto Lp(a) concentrations >117 mg/dl). The number of KIVtype 2 repeats explained between 20 and 30 percent of thevariation in Lp(a) concentrations.

A further publication investigated in >8000 CHD casesand >8000 controls the association of Lp(a) levels, CHDand genetic variants [15]. The authors analyzed almost50.000 single nucleotide polymorphisms (SNPs) in 2100candidate genes for CHD and found that 2 SNPs combinedexplained 36% of the variation in plasma Lp(a) levels. Bothgene loci were closely associated with elevated Lp(a),a decreased number of KIV type 2 repeats and a smallerLp(a) mass. The odds ratio for CHD in the subjects havingboth SNPs was 2.57 (95% CI 1.80e3.67) and, mostimportantly, after adjusting for Lp(a) concentrations, the

association between genotype and CHD risk disappeared.This was suggesting there is a causal link between Lp(a)and CHD.

3. Treatment options for reducing elevated Lp(a)concentrations

3.1. Drug treatment

The pharmacologic options for decreasing Lp(a)concentrations have been reviewed recently [16]. The mosteffective drug to decrease Lp(a) concentrations, byapproximately 20e30%, is niacin. Also acetylsalicylic acidand L-carnitine have been shown to decrease Lp(a). Somedrugs in clinical development have been shown to decreaseLp(a), among them mipomersen, PCSK-9 inhibitors, CETPinhibitors, and eprotirome. The currently most commonlyused lipid-lowering drugs such as statins, fibrates or ezeti-mibe are without effect. Diet and exercise also do notappear to significantly influence Lp(a) concentrations.

3.1.1. Niacin (nicotinic acid)Niacin is currently considered the pharmacologic treat-

ment of choice for elevated Lp(a). It decreases Lp(a) ina dose-dependent manner (1e3 g/day) usually by about20% and in some cases up to 30e40% [17]. The mecha-nism of decreasing Lp(a) involves inhibition of its synthesisrather than catabolism [17]. Niacin has additional positiveeffects on other lipoproteins by increasing HDL cholesteroland decreasing LDL cholesterol and triglycerides. A recentmetaanalysis of randomized trials has shown that niacinreduced significantly major coronary events, stroke and anycardiovascular events by about 25% [18]. Whether theLp(a)-lowering effects of niacin contribute to its risk-reducing effects on cardiovascular events remains unclear.There are no studies showing that decreasing isolatedhyperlipoproteinemia(a) with niacin reduces cardiovascularrisk. Niacin is often poorly tolerated due to adverse effectssuch as cutaneous vasodilation with flushing (by stimu-lating biosynthesis of prostaglandins), which is a majorreason for discontinuation of the drug. The formulation asextended-release niacin or the combination with lar-opiprant, a prostaglandin D2 receptor-1 antagonist reducingflushing, may improve drug adherence.

3.1.2. Other drugsStatins, the most frequently used class of drugs in

treating lipid disorders, have been shown to variablyinfluence Lp(a) but do not substantially lower its concen-trations, presumably because the LDL receptor is notinvolved in the clearance of Lp(a). Similarly, studies usingfibrates showed no significant effects of this drug class onLp(a) levels. Also the cholesterol absorption inhibitorezetimibe does not influence Lp(a) levels.

There is some, albeit limited, evidence that acetylsali-cylic acid decreases elevated Lp(a) levels. The effects seem

3H.K. Berthold, I. Gouni-Berthold / Atherosclerosis Supplements 14 (2013) 1e5

to be greater when baseline levels are higher. The mecha-nism may involve the apo(a) synthesis process at the levelof the apo(a) gene transcription. It is safe to assume thatmost of the patients at increased cardiovascular risk willtake acetylsalicylic acid for general reasons so that this isno specific therapy to lower elevated Lp(a) levels.

L-Carnitine, which is a naturally occurring compoundplaying a role in fatty acid metabolism, seems to moder-ately decrease Lp(a) concentrations by about 20%. It is welltolerated.

Sex hormones have an influence on Lp(a) concentra-tions. While systemic estrogen alone or in combinationwith progestins decreases Lp(a), transdermal hormonereplacement therapy and the selective estrogen receptormodulator raloxifene do not affect Lp(a) levels. Tibolone,a synthetic steroid with weak estrogenic, progestagenic,and androgenic properties, has been shown to decreaseLp(a) levels. Also androgenic anabolic steroids have beenshown to decrease Lp(a) levels. The mechanism throughwhich sex hormones influence Lp(a) levels is most likelythe downregulation of the apo(a) gene expression. None ofthese substances is suitable for routine use to decreaseelevated Lp(a) concentrations.

3.1.3. New approachesA number of new approaches with novel mechanisms of

action are under development and may be promising inmodulating Lp(a) levels. The apoB synthesis inhibitormipomersen has been shown to significantly decrease Lp(a)concentrations by about 30% [19]. The mechanism seemsto involve inhibition of apoB synthesis. Inhibitors of thecholesteryl ester transfer protein (CETP) are a new class oflipid-modifying drugs. Anacetrapib has been shown todecrease Lp(a) concentrations by about 40% [20], whiledalcetrapib, whose clinical development has been stopped,had no effect on Lp(a). The mechanism through whichCETP inhibitors may influence Lp(a) is not known. Thethyroid hormone analog eprotirome was found to decreaseLp(a) by about 40% [21]. However, further clinical devel-opment of the drug has been stopped recently due to sideeffects. Other novel mechanisms to influence Lp(a)concentrations may involve the inhibition of interleukin-6(IL-6) receptor signaling with the IL-6 receptor antibodytocilizumab [22].

3.2. Lipoprotein apheresis

In rare cases of hyperlipoproteinemia(a), lipoproteinapheresis may be a therapeutic option, as it is in patientswith severe familial hypercholesterolemia to remove LDLparticles from the circulation [23]. Lipoprotein apheresis isable to decrease Lp(a) concentrations by 50e70% asmeasured at the end of a weekly or 2-weekly treatmentinterval but partial rebound to higher concentrations occurs.Several systems have been developed for the extracorporealelimination of LDL and Lp(a) from plasma, among them

various LDL immunoadsorption techniques using immo-bilized monoclonal or polyclonal antibodies to apoB-100,LDL binding by dextran sulfate attached to cellulose,heparin-induced extracorporeal LDL precipitation, anddirect adsorption of lipoproteins from whole blood. There isalso an immunosorption column containing specific anti-bodies against Lp(a) available. Recent reviews have sug-gested that LDL and Lp(a) apheresis are safe and effectiveprocedures [23e26].

There is no randomized trial yet that showed a decreasein cardiovascular endpoints by treating elevetated Lp(a)levels with apheresis. A recent non-randomized retrospec-tive cohort study in 120 apheresis subjects in Germanyshowed that decreasing Lp(a) levels decreased the annualrate of major adverse cardiovascular events (MACE) butthe study design was flawed [27]. Another recent smallrandomized trial showed that in 21 patients with angio-graphically documented CHD apheresis decreased Lp(a) by57% but showed no difference in the rate of cardiovascularevents in a follow-up of one year [28].

In Germany the Federal Joint Committee (GemeinsamerBundesausschuss) decided in 2008 to authorize the use ofLp(a) apheresis for patients with progressive cardiovasculardisease (CVD) and Lp(a) levels greater than 60 mg/dl (inthe presence of LDL cholesterol in the normal range) inpatients where all other measures failed to stop diseaseprogression. This reimbursement decision was coupled tomandating a controlled trial which would definitely prove(or disprove) the benefit of Lp(a)-lowering by apheresis inthose patients. A study protocol for a hybrid trial (ELAILatrial NCT01064934), comprising a randomized studyand an observational study, was filed but is currently onhold due to a negative vote of the competent ethicscommittee [29].

The difficulty of randomizing to conventional therapypatients whose disease has proved refractory to the latter isone of the main reasons why much of the evidence con-cerning the therapeutic benefits of apheresis is regarded asflawed [30]. Thus, it is generally assumed that patients withsevere CVD benefit from lipoprotein apheresis. The methodis expensive and impractical for most patients and itsfeasibility depends mainly on the healthcare reimbursementsystem of the respective country [11]. However, thescientific data to support its use are sparse. In particular,there is no convincing evidence that lipoprotein apheresis todecrease Lp(a) affects cardiovascular outcomes. This iscompounded by the decision in Germany to reimburse theprocedure, thus suggesting to both medical professionalsand patients alike, that such evidence exists. Approving thereimbursement of a procedure and ex post mandatinga clinical trial to prove its efficacy is unusual. However, thedecision was made, among other reasons, in considerationof the ‘rare disease’ status of the patients concerned.Reimbursement has to be approved case-by-case by specialapheresis commissions. On the other hand, there is ongoingdebate whether efficacy criteria required for reimbursement

4 H.K. Berthold, I. Gouni-Berthold / Atherosclerosis Supplements 14 (2013) 1e5

decisions should be different between rare and commondiseases. Moreover, it seems that in the studies supportingthe reimbursement decision, conventional therapies(‘control group’) were often not maximal and thus theeffects of apheresis may be overestimated.

There are not enough data to determine cost-effectiveness of the procedure but it can be assumed thatacross countries there are geographical inequities inpatients’ access to this treatment.

Without an available treatment that selectively reducesLp(a) concentrations, the methodological difficulty remainsthat in many patients there is an overlap of increased Lp(a)and LDL cholesterol concentrations. It would be hard todifferentiate in an endpoint trial to what extent the benefitwas derived from Lp(a)- or from LDL-lowering. This issueresembles the discussion as to what extent in statin trials theclinical benefit came from decreasing LDL and to whatextent from their pleiotropic effects.

4. Screening, measurement and desirable levels

Plasma levels of Lp(a) have a skewed distribution in thepopulation and vary interindividually by a factor ofw1000.Caucasians have lower median plasma levels than Asiansand Blacks. Several laboratory assays are available and it isimportant to use kits whose immunoreactivity with Lp(a) isinsensitive to the isoform of apo(a). Lp(a) concentrationsshould be expressed as total Lp(a) protein and referencevalues should be given in ranges and percentiles for therespective ethnicities. This would enable to calculate race-specific estimates of risk thresholds [11].

Intraindividual long-term variation of Lp(a) is very lowdue to genetic influence, so that single determinations areusually sufficient to determine the general risk. Repeatmeasurements may be required if treatment is initiated [31].The National Academy of Clinical Biochemistry has issuedrecommendations for screening in 2009 [31]. Combiningthese recommendations with the ones of the EASConsensus Panel [11], Lp(a) should be determined insubjects with premature CVD, familial hypercholesterol-emia, a family history of premature CVD and/or elevatedLp(a), recurrent CVD despite statin treatment and very highCVD according to current guidelines. Population routinescreening is not generally recommended.

There is no general agreement as to what the “normal”values are or what the treatment target levels of Lp(a)should be. Some laboratories advise a limit of 30 mg/dl asupper normal range, however, no scientific evidencesupports such cut-off value, which is rather arbitrary. Sincethe risk association is continuous in shape and without anobvious threshold, normal ranges or treatment targets aredifficult to establish. The EAS Consensus Panel recom-mends as desirable Lp(a) levels concentrations <50 mg/dl,which represents values below the 80th percentile [11]. Asit is the case with cardiovascular risk management ingeneral, risk factors have to be considered in total and the

ones that can be treated should be approached accordinglybased on an individual’s overall risk. Recommendationsproposing desirable levels should preferably be based onevidence from a metaanalysis of randomized controlledintervention trials documenting benefit of treatment. Thesetrials are completely lacking for Lp(a). Needless to say thatLp(a) reduction therefore remains a secondary prioritywhile reduction in LDL cholesterol concentrations remainsthe primary goal of lipid-lowering treatment [32].

5. Conclusions

Elevated Lp(a) concentrations are causally associatedwith increased cardiovascular risk. There is no treatmentavailable that selectively decreases Lp(a) levels. Niacin isthe drug of choice to treat high concentrations. In indi-vidual high-risk patients with progressive CVD lipoproteinapheresis may be a therapeutic option. Randomizedcontrolled trials examining whether selective reduction ofLp(a) levels is able to improve cardiovascular outcomes arelacking and urgently needed.

Disclosure

Professor Berthold has nothing to disclose. ProfessorGouni-Berthold has received honoraria and/or researchgrants from MSD Sharp & Dohme, Genzyme, Otsuka,Bayer Health Care, and Novartis.

References

[1] Berg K. A new serum type system in manethe Lp system. Acta

Pathol Microbiol Scand 1963;59:369e82.

[2] Utermann G. The mysteries of lipoprotein(a). Science 1989;246:

904e10.[3] Hobbs HH, White AL. Lipoprotein(a): intrigues and insights. Curr

Opin Lipidol 1999;10:225e36.

[4] Nielsen LB, Juul K, Nordestgaard BG. Increased degradation of

lipoprotein(a) in atherosclerotic compared with nonlesioned aortic

intima-inner media of rabbits: in vivo evidence that lipoprotein(a)

may contribute to foam cell formation. Arterioscler Thromb Vasc

Biol 1998;18:641e9.

[5] Poon M, Zhang X, Dunsky KG, Taubman MB, Harpel PC. Apoli-

poprotein(a) induces monocyte chemotactic activity in human

vascular endothelial cells. Circulation 1997;96:2514e9.

[6] Boffa MB, Marcovina SM, Koschinsky ML. Lipoprotein(a) as a risk

factor for atherosclerosis and thrombosis: mechanistic insights from

animal models. Clin Biochem 2004;37:333e43.

[7] Nielsen LB, Gronholdt ML, Schroeder TV, Stender S,

Nordestgaard BG. In vivo transfer of lipoprotein(a) into human

atherosclerotic carotid arterial intima. Arterioscler Thromb Vasc Biol

1997;17:905e11.

[8] Bennet A, Di Angelantonio E, Erqou S, et al. Lipoprotein(a) levels

and risk of future coronary heart disease: large-scale prospective data.

Arch Intern Med 2008;168:598e608.

[9] Danesh J, Collins R, Peto R. Lipoprotein(a) and coronary heart

disease. Meta-analysis of prospective studies. Circulation 2000;102:

1082e5.

5H.K. Berthold, I. Gouni-Berthold / Atherosclerosis Supplements 14 (2013) 1e5

[10] Craig WY, Neveux LM, Palomaki GE, Cleveland MM, Haddow JE.

Lipoprotein(a) as a risk factor for ischemic heart disease: meta-

analysis of prospective studies. Clin Chem 1998;44:2301e6.

[11] Nordestgaard BG, Chapman MJ, Ray K, et al. Lipoprotein(a) as

a cardiovascular risk factor: current status. Eur Heart J 2010;31:

2844e53.

[12] Erqou S, Kaptoge S, Perry PL, et al. Lipoprotein(a) concentration and

the risk of coronary heart disease, stroke, and nonvascular mortality.

JAMA 2009;302:412e23.

[13] Kamstrup PR, Benn M, Tybjaerg-Hansen A, Nordestgaard BG.

Extreme lipoprotein(a) levels and risk of myocardial infarction in the

general population: the Copenhagen City Heart Study. Circulation

2008;117:176e84.

[14] Kamstrup PR, Tybjaerg-Hansen A, Steffensen R, Nordestgaard BG.

Genetically elevated lipoprotein(a) and increased risk of myocardial

infarction. JAMA 2009;301:2331e9.

[15] Clarke R, Peden JF, Hopewell JC, et al. Genetic variants associated

with Lp(a) lipoprotein level and coronary disease. N Engl J Med

2009;361:2518e28.[16] Gouni-Berthold I, Berthold HK. Lipoprotein(a): current perspectives.

Curr Vasc Pharmacol 2011;9:682e92.

[17] Seed M, O’Connor B, Perombelon N, O’Donnell M, Reaveley D,

Knight BL. The effect of nicotinic acid and acipimox on lip-

oprotein(a) concentration and turnover. Atherosclerosis 1993;101:

61e8.

[18] Bruckert E, Labreuche J, Amarenco P. Meta-analysis of the effect of

nicotinic acid alone or in combination on cardiovascular events and

atherosclerosis. Atherosclerosis 2010;210:353e61.

[19] Gouni-Berthold I, Berthold HK. Antisense oligonucleotides for the

treatment of dyslipidemia. Curr Pharm Des 2011;17:950e60.[20] Cannon CP, Shah S, Dansky HM, et al. Safety of anacetrapib in

patients with or at high risk for coronary heart disease. N Engl J Med

2010;363:2406e15.

[21] Ladenson PW, Kristensen JD, Ridgway EC, et al. Use of the thyroid

hormone analogue eprotirome in statin-treated dyslipidemia. N Engl

J Med 2010;362:906e16.

[22] Berthold HK, Laudes M, Krone W, Gouni-Berthold I. Association

between the interleukin-6 promoter polymorphism -174G/C and

serum lipoprotein(a) concentrations in humans. PLoS One 2011;6:

e24719.

[23] Thompson GR. Recommendations for the use of LDL apheresis.

Atherosclerosis 2008;198:247e55.

[24] Keller C. Apheresis in coronary heart disease with elevated Lp(a):

a review of Lp(a) as a risk factor and its management. Ther Apher

Dial 2007;11:2e8.

[25] Thompsen J, Thompson PD. A systematic review of LDL apheresis in

the treatment of cardiovascular disease.Atherosclerosis 2006;189:31e8.

[26] Julius U, Frind A, Tselmin S, Kopprasch S, Poberschin I, Siegert G.

Comparison of different LDL apheresis methods. Expert Rev Car-

diovasc Ther 2008;6:629e39.

[27] Jaeger BR, Richter Y, Nagel D, et al. Longitudinal cohort study on

the effectiveness of lipid apheresis treatment to reduce high lip-

oprotein(a) levels and prevent major adverse coronary events. Nat

Clin Pract Cardiovasc Med 2009;6:229e39.

[28] Stefanutti C, Vivenzio A, Di Giacomo S, Mazzarella B, Ferraro PM,

Abbolito S. Treatment of symptomatic hyperLp(a)lipidemia with

LDL-apheresis vs. usual care. Transfus Apher Sci 2010;42(1):21e6.

[29] Kassner U, Vogt A, Rosada A, et al. Designing a study to evaluate the

effect of apheresis in patients with elevated lipoprotein(a). Athe-

roscler Suppl 2009;10:85e8.

[30] Thompson GR. Lipoprotein apheresis. Curr Opin Lipidol 2010;21:

487e91.

[31] Myers GL, Christenson RH, Cushman M, et al. National Academy of

Clinical Biochemistry laboratory medicine practice guidelines:

emerging biomarkers for primary prevention of cardiovascular

disease. Clin Chem 2009;55:378e84.[32] Graham I, Atar D, Borch-Johnsen K, et al. European guidelines on

cardiovascular disease prevention in clinical practice: executive

summary: fourth Joint Task Force of the European Society of

Cardiology and other societies on cardiovascular disease prevention

in clinical practice (constituted by representatives of nine societies

and by invited experts). Eur Heart J 2007;28:2375e414.