logy 138 (2010) 281–289www.elsevier.com/locate/ijcard

International Journal of Cardio

Levosimendan for the treatment of acute severe heart failure: Ameta-analysis of randomised controlled trials

Anthony Delaney a,b,⁎, Celia Bradford a,b, John McCaffrey a,Sean M. Bagshaw c,d, Richard Lee a,b

a Intensive Care Unit, Royal North Shore Hospital, St. Leonards, NSW, Australiab Northern Clinical School, University of Sydney, St. Leonards, NSW, Australia

c Division of Critical Care Medicine, University of Alberta Hospital, Edmonton, AB, Canadad Faculty of Medicine and Dentistry, University of Alberta, Edmonton, AB, Canada

Received 30 October 2007; received in revised form 27 June 2008; accepted 10 August 2008Available online 24 September 2008

Abstract

Background: The objective of this study was to critically review the literature to evaluate whether levosimendan compared to standard therapy,in patients with acute severe heart failure, is associated with improved clinical outcomes.Methods: Medline, EMBASE, and the Cochrane central register of clinical trials were searched. We also searched clinical trials registries,bibliographies of included studies and review articles and contacted themanufacturers of levosimendan to identify unpublished studies. Randomisedclinical trials comparing levosimendan to standard therapy or placebo, in adult patientswith acute severe heart failure, reporting at least one outcomeof interest were included. Data were extracted regarding the characteristics, methodological quality and clinical outcomes, and combined using afixed-effect meta-analysis.Results:We identified 19RCTs enrolling 3650 patients, only two studies fulfilled all of the validity criteria. Therewas a non-significant reduction inmortalitywith levosimendan comparedwith placebo (OR0.83, 95%CI, 0.62–1.10, p=0.20). Levosimendanwas associatedwith reducedmortalitycompared to dobutamine (OR 0.75, 95%CI, 0.61–0.92, p=0.005). Levosimendan was associated with improvements in haemodynamicparameters when compared to either placebo or dobutamine.Conclusions: Levosimendan improved haemodynamic parameters when compared with placebo, without showing evidence of survivalbenefit. Levosimendan improved both haemodynamics and survival when compared with dobutamine.Crown Copyright © 2008 Published by Elsevier Ireland Ltd. All rights reserved.

Keywords: Heart failure; Drugs; Meta-analysis; Levosimendan; Inotropic agents

1. Introduction

Recent international studies have highlighted the scopeand impact of acute heart failure [1,2]. Heart failure is one ofthe most important causes of morbidity and mortality in thedeveloped countries [3], with the numbers of patients hos-pitalised for acute heart failure increasing due to ageing

⁎ Corresponding author. Intensive Care Unit, Royal North Shore Hospital,Pacific Highway, St. Leonards, NSW, 2065, Australia. Tel.: +61 2 99268656; fax: +61 2 9439 8418.

E-mail address: adelaney@med.usyd.edu.au (A. Delaney).

0167-5273/$ - see front matter. Crown Copyright © 2008 Published by Elsevierdoi:10.1016/j.ijcard.2008.08.020

populations and improvements in the treatments of ischaemicheart disease and chronic heart failure [4,5]. Patients withacute severe heart failure have an estimated one-year morta-lity of 30–50% [6]. Acute heart failure accounts for 2–3% ofall hospital admissions, or more than 6.5million hospital daysannually in the United States alone [6]. Clearly the manage-ment of acute severe heart failure is an important issue withconsiderable health resource implications.

After correcting the precipitant and treating the under-lying cause of the acute deteriorations in cardiac function,current management largely focuses on the use of diureticsand vasodilators [3], combined with non-invasive positive

Ireland Ltd. All rights reserved.

Fig. 1. Study flow diagram.

282 A. Delaney et al. / International Journal of Cardiology 138 (2010) 281–289

pressure ventilation for those with acute pulmonary oedema[7]. When heart failure is severe and, not responding to firstline treatments, positive inotropic agents such as dobutamineor milrinone may be indicated [3]. However, there is littlehigh quality evidence for the efficacy of these agents toimprove long-term cardiac performance or survival [8].

Levosimendan is a novel inotropic agent. It acts on cardiactroponin C, stabilizing the bound Ca2+, prolonging the inter-action between of actin and myosin, and thus enhancingcardiac contractility [9]. Preclinical studies demonstrated theefficacy of levosimendan in enhancing myocardial functionwith improvements in a number of measured haemodynamicvariables and neurohumoral marker [10–14]. Levosimendanalso exerts a significant vasodilatory action that may have animportant clinical effect [15,16].

Early clinical trials suggested that levosimendan may be auseful adjunct in the treatment of acute heart failure [17,18].These studies, however, were not designed or powered todetect reductions in clinically relevant and patient-centredoutcomes such as mortality. Yet, on the basis of these trials,levosimendan has been recommended for the treatment ofpatients with symptomatic low cardiac output heart failure[3]. More recently, a number of larger clinical trials have nowbeen completed. Accordingly, the principal objective of thisstudy was to critically review the literature to evaluate whe-ther levosimendan compared to standard therapy, in patientswith acute severe heart failure, is associated with improvedclinical outcomes, in particular survival.

2. Methods

2.1. Search

The primary search for randomised clinical trials (RCTs)was conducted using the Medline (via the PubMED interface)and EMBASE databases, as well as the Cochrane CentralRegistry of Clinical Trials (using the OVID interface). Searchterms used were levosimendan or calcium sensit⁎. This searchwas combined with sensitive filters to identify RCTs in theMedline database [19] and the EMBASE database [20]. Thesearch was limited to human subjects, with no language res-triction placed on the search. Two authors (AD and CB)independently conducted the search. The search was finalisedon 18th June, 2007. We also searched the metaRegister ofControlled Trials (http://www.controlled-trials.com/mrct/), in-cluding the Medical Editors Trial Amnesty and other archivedregistries using the search term “levosimendan”. We contactedthemanufacturers of levosimendan (no additional studies wereidentified), and reviewed the bibliographies of included RCTsand review articles, to identify otherwise unrecognised orunpublished RCTs [21].

2.2. Study selection

Two authors (AD and CB) reviewed all abstracts toidentify potentially eligible RCTs. If the abstract described a

study that could potentially meet the inclusion criteria, thenthat study was considered for further review. Where possible,full text articles were retrieved and reviewed, by the sameauthors, to determine if they met the eligibility criteria. Dis-agreements were resolved by discussion, with resort to a thirdauthor (RL) if needed. Reports of RCTs of levosimendancompared to any other agent, for acute heart failure in adulthumans, and that reported at least one of the outcomes ofinterestwere considered eligible for inclusion. These outcomesincluded mortality, changes in haemodynamic parameters (i.e.ejection fraction (EF), cardiac index (CI), pulmonary capillarywedge pressure (PCWP)) and b-type natureitic peptide (BNP).

2.3. Validity assessment and data abstraction

All included articles were assessed for validity by two au-thors (AD CB), independently, with disputes resolved by dis-cussion. A component approach was used [22]. Each report wasassessed for the adequacy of the allocation concealment (definedas attempts to shield those who admit participants to a trial fromknowing the upcoming treatment assignments [23]), blinding(defined as attempts to keep trial participants, investigators orassessors unaware of the assigned treatment [24]) and the per-formance of an intention to treat analysis. When it was unclearwhether a criterion had been met it was adjudicated as absent[25].

Data were abstracted onto specific data collection forms bytwo authors (AD and CB) and entered into a spreadsheet whereit was checked for accuracy (JM). Data collected included:patient baseline characteristics, study definition of acute heartfailure, details of delivery of levosimendan, specific of com-parator therapy and haemodynamic and clinical outcomes (i.e.mortality for each group for the longest period of follow-up).When the report did not contain sufficient detail to adjudicate thevalidity of the study, or outcome data were missing, attempts tocontact the authors, by email and in writing were made.

Table 1Summary of the characteristics of randomised clinical trials of levosimendan in patients with acute severe heart failure.

Study Year Population N Definition ofheart failure

Mean age(years)

% Male Levo bolus(µg/kg)

Levo infusion(µg/kg/min)

Levo duration(h)

Control dose Controlduration (h)

Duration offollow-up (days)

Levosimendan v placebo

Nieminen [37] 2000 Heart failure of ischaemic origin 151 EFb40% 63.2 86.3 3 to 36 0.05 to 0.6 24 NA 24 9RUSSLAN [15] 2002 AMI with evidence of LVF on

CXR and need for inotropes504 LVF on CXR 67.2 51.6 6 to 24 0.1 to 0.4 6 NA 6 180

REVIVE I [41] 2003 Hospitalised for acutedecompensated heart failure

100 NA NR NR 12 0.1 to 0.1 24 NA 24 90

Parissis [13] 2004 NYHA III/IV 27 EFb30%, CIb2.5 70.4 NR 6 0.1 to 0.4 24 NA 24 2CASINO [40] 2004 Hospitalised with EFb35% 199 LV EFb35% NR NR 24 NA 24 180Parissis [14] 2005 NYHA III/IV hospitalised 34 EFb35%, CIb2.5 67 91.2 6 0.1 to 0.4 24 NA 24 150De Luca [33] 2005 AMI with LV dysfunction 26 EFb40%, PCWPN

18 and CIb2.557.2 65.6 12 0.1 24 NA 24 1

REVIVE II [39] 2006 Hospitalised for Heart failure 600 LV EFb35% 63 72 12 0.1 to 0.2 24 NA 24 90Tziakis [38] 2006 NYHA III/IV admitted to CCU 60 NYHA III/IV 66 55 6 0.1 24 NA 24 1Adamopolous [12] 2006 NYHA III/IV admitted with

decompensated heart failure46 EFb30%, CIb2.5 69.7 84.1 6 0.1 24 NA 24 120

Levosimendan v dobutamineFollath [34] 1999 Low output heart failure or

cardiogenic shock19 CIb2.5 58.8 78.9 12 0.2 to 0.6 24 8 to 16 µg/kg/min 24 14

Nieminen [37] 2000 Heart failure of ischaemicorigin

EFb40% 63 87.8 3 to 36 0.05 to 0.6 24 6 µg/kg/min 24 9

Follath [16] 2002 Admitted to hospital withlow output heart failure

203 EFb35% CIb2.5,PCWPN15

59 86.7 24 0.1 to 0.2 24 5 to 10 24 180

CASINO [40] 2004 Hospitalised with EFb35% 200 LVEFb35% NR NR NR NR 24 NR 24 180Avgeropolou [31] 2005 NYHA IV hospitalized 29 NYHA IV 70.5 75.9 12 0.1 24 5 to 10 24 5Adamopolous [12] 2006 NYHA III/IV admitted with

decompensated heart failure46 EFb30% CIb2.5 69.7 84.1 6 0.1 24 5 to 10 24 120

SURVIVE [28] 2006 Hospitalised for heart failure 1327 LVEFb30% 67 72 12 0.1 to 0.2 24 5 to 40 Titrated 180Alvarez [30] 2006 Low cardiac output after

heart surgery withextracorporeal circulation

41 CIb2.2, PCWPN15 69.1 48 12 0.2 24 7.5 24 15

Morelli [36] 2005 Septic shock with EFb45% 30 EFb45% 61.9 70 0 0.2 24 5 24 30

Levosimendan v milrinoneAl-Shawaf [29] 2006 Type II diabetics with

EFb35% post CABG30 CIb2.2 59.2 93.3 12 0.1 to 0.2 24 50 µg/kg then

0.3 µg/kg/min24 ICU LOS

De Hert [32] 2007 Elective cardiac surgerywith LVEFb30%

30 LVEFb30% 68 66.7 0 0.1 19 +/− 4 0.1 µg/kg 24 Hospital LOS

Levosimendan v PGE1Moertl [35] 2005 Admitted with decompensated

heart failure73 NYHA III or IV,

EFb35%55.3 80.8 12 0.1 24 2.5 – 10 ng/kg 24 7

N = total number of patients, Levo = levosimendan, EF = ejection fraction, NA= not applicable, AMI = acutemyocardial infarction, LVF = left ventricular failure, CXR= chest X-ray, NR= not reported, NYHA=NewYorkHeart Association classification, CI = cardiac index, LV = left ventricular, PCWP = pulmonary capillary wedge pressure, CABG = coronary artery bypass graft surgery, ICU = intensive care unit, LOS = length of stay.

283A.Delaney

etal.

/International

Journalof

Cardiology

138(2010)

281–289

Table 2Summary of the validity assessments of randomised clinical trials oflevosimendan in patients with acute severe heart failure.

Study Allocationconcealment

Intention totreat analysis

Blinding

Follath [37] No Yes NoNieminen [40] No Yes NoRUSSLAN [17] Yes No YesFollath [18] Yes Yes YesREVIVE I [44] No No YesParrisis [13] No Yes NoCASINO [43] No No YesDeLuca [36] No Yes NoMoertl [38] No No NoParrisis [14] No Yes NoAvgeropolou [34] No Yes NoMorelli [39] No No YesAdamopolous [12] No Yes NoTziakis [41] No Yes NoREVIVE II [42] No No YesAl-Shawaf [32] Yes Yes NoAlvarez [33] No No NoSURVIVE [31] Yes Yes YesDeHert [35] Yes Yes No

284 A. Delaney et al. / International Journal of Cardiology 138 (2010) 281–289

2.4. Data synthesis

Agreement on RCT inclusion was assessed using thekappa statistic. Heterogeneity was assessed using the chi-square statistic and quantified using the I2 statistic, with anI2 of N50% taken as representing at least moderate hetero-geneity [26]. The potential for bias was assessed by the

Fig. 2. The effect of levosimendan com

inspection of funnel plots and the use of Eggers statistic [27].Studies that compared levosimendan to dobutamine werecombined separately from studies that compared levosimen-dan to placebo. As there were three studies that had threegroups (i.e. levosimendan, dobutamine and placebo), we cal-culated a pooled estimate of the effect of dobutamine com-pared to placebo. Data were extracted from the primarystudies to allow calculation of the odds ratio (OR) for eachstudy. The primary analyses were performed using a fixed-effect model [28] to produce a pooled OR [29]. Sensitivityanalyses were conducted using the random effects model ofDer Simonian and Laird. The effect of trial quality on out-come was examined by pooling studies adjudicated to haveadequate allocation concealment, blinding and an intention totreat analysis separately, to examine for between group hete-rogeneity. Changes in haemodynamic parameters and BNPwere pooled using weighted mean differences [30]. All ana-lyses were performed using STATA 8.2 (Statcorp, CollegeStation, Tx). The authors had full access to the data and takeresponsibility for its integrity. All authors have read and agreeto the manuscript as written.

3. Results

3.1. Study characteristics

There were 393 reports identified by the search, 41 fulltext articles were retrieved for in depth review. In total, 19studies enrolling 3650 participants fulfilled all eligibilitycriteria. Fig. 1 shows the flow of studies and the reasons forexclusion. Of the trials included, 16 [12–14,17,18,31–41]

pared to placebo on mortality.

Fig. 3. The effect of levosimendan compared to dobutamine on mortality.

285A. Delaney et al. / International Journal of Cardiology 138 (2010) 281–289

were available as full reports and 3 [42–44] in abstract formonly. Agreement on study inclusion was reached in 39 of 41cases (kappa=0.90). The trial characteristics are shown inTable 1. The validity assessments are shown in Table 2. Onlytwo studies fulfilled all of the validity criteria.

3.2. Levosimendan v placebo

There were six studies [12,17,40,42–44] including at totalof 1578 participants that compared the effect of levosimendan

Fig. 4. The effect of dobutamine com

to placebo and reported mortality. There was no evidence ofsignificant heterogeneity (χ2 p=0.24, I2=25.7%). Neitherinspection of the funnel plot (Appendix) nor Egger's biasstatistic (bias=−0.41, p=0.07) revealed evidence of bias. Thefixed-effect pooled estimate of the OR for mortality was 0.83(95%CI 0.62 to 1.10, p=0.20), suggesting that levosimendanwas no better than placebo for reducing mortality in acutecardiac failure (Fig. 2). This result was consistent when thestudies were pooled using a random effects model, the pooledOR=0.80 (95%CI 0.54 to 1.17, p=0.24).

pared to placebo on mortality.

Table 3The effect of study quality indicators on odds of mortality for levosimendan compared with control for patients with acute heart failure.

Control group Quality indicator Number of studies Estimate of OR 95%CI Test for between subgroup heterogeneity

Placebo Allocation concealment Yes 1 0.64 0.40–1.03 p=0.20No 5 0.95 0.66–1.36

Blinding Yes 4 0.84 0.63–1.13 p=0.64No 2 0.57 0.13–2.6

Intention to treat analysis Yes 2 0.57 0.13–2.6 p=0.64No 4 0.84 0.63–1.13

Dobutamine Allocation concealment Yes 2 0.85 0.68–1.07 p=0.005No 6 0.38 0.23–0.63

Blinding Yes 4 0.75 0.61–0.93 p=0.56No 4 0.56 0.19–1.62

Intention to treat analysis Yes 5 0.84 0.67–1.04 p=0.006No 3 0.35 0.20–0.63

OR = odds ratio, CI = confidence interval.

286 A. Delaney et al. / International Journal of Cardiology 138 (2010) 281–289

3.3. Levosimendan v dobutamine

There were eight studies [12,18,31,33,37,39,40,43] inclu-ding a total of 1979 participants that compared levosimendanto dobutamine and reported mortality. There was no evi-dence of significant bias on visual inspection of the funnelplot (Appendix) or by examination of Egger's bias statistic(bias=−1.16, p=0.17). There was evidence of some hetero-geneity, with theχ2 p=0.08 and the I2=44.6%. Pooled analysisshowed levosimendan was associated with a significant reduc-tion in mortality compared with dobutamine (OR 0.75, 95%CI0.61 to 0.92, p=0.005) (Fig. 3). This finding was similar whendata was pooled using a random effects model (OR 0.58, 95%CI, 0.37 to 0.91, p=0.02).

3.4. Dobutamine v placebo

There were three studies [12,40,43] that randomised301 participants to dobutamine or placebo and reportedmortality. There was no evidence of significant bias(Egger's statistic, bias=0.34, p=0.73), funnel plot asym-metry or statistical heterogeneity (χ2 p=0.72, I2 =0%). Thepooled estimate of the OR for mortality for dobutaminecompared to placebo was 1.82 (95%CI 1.06 to 3.12,

Table 4Summary of the haemodynamic effects of levosimendan compared to control.

Control group Haemodynamic parameter Number of stu

Placebo EF (%) 3Cardiac index (l/min/m2) 2PAOP (mm Hg) 2BNP (pg/ml) 2

Dobutamine EF (%) 1Cardiac index (l/min/m2) 3PAOP (mm Hg) 3BNP (pg/ml) 2

Milrinone Cardiac index (l/min/m2) 1PAOP (mm Hg 1

PGE1 PAOP (mm Hg) 1BNP (pg/ml) 1

WMD = weighted mean difference, CI = confidence interval, EF = ejection fraction, P

p=0.03), indicating a significant increase in mortality fordobutamine compared to placebo in patients with acutesevere acute heart failure (Fig. 4). This finding was similarwhen pooled using a random effects model (OR 1.81, 95%CI1.06 to 3.11, p=0.03).

3.5. Levosimendan compared to other agents

Two studies compared levosimendan tomilrinone [32,35].The OR for mortality for levosimendan compared to mil-rinone was 0.33 (95%CI 0.05 to 2.1, p=0.24) indicating nostatistical evidence of an effect of levosimendan compared tomilrinone on mortality.

One study compared levosimendan to PGE1 in patientsadmitted to hospital with acute decompensated heart failure[38]. Mortality was not reported in this study.

3.6. The effect of study quality on estimates of efficacy

The effect of each component of study qualitywas assessed,by pooling studies adjudicated to be higher quality separately.The results of this assessment are shown in Table 3. Whenlevosimendan was compared to dobutamine, there was anexaggerated estimate of efficacy in studies without adequate

dies WMD 95%CI p value

3.2 0.3 to 6.0 0.030.15 −0.1 to 0.4 0.26−3.8 −6.3 to −1.3 0.003−329 −617 to −42 0.0253.0 −1.4 to 7.4 0.180.33 0.24 to 0.43 b0.0005−2.91 −3.3 to −2.5 b0.0005−595 −1000 to −180 0.0050.1 −0.48 to 0.68 0.73−2.0 −4.5 to 0.5 0.12−1.0 −3.4 to 1.4 0.41−210 −611 to 191 0.31

AOP = pulmonary artery occlusion pressure, BNP = B-type natureitic peptide.

287A. Delaney et al. / International Journal of Cardiology 138 (2010) 281–289

allocation concealment, and those that did not perform anintention to treat analysis.

3.7. Effects of levosimendan on haemodynamic parameters

The effect of levosimendan on the various measuredhaemodynamic parameters and BNP are shown in Table 4.Levosimendan was associated with improvements in all thehaemodynamic parameters evaluated, however, broad infer-ences are limited due to a small number of RCTs reportingdata for each parameter.

4. Discussion

We performed a systematic review and meta-analysis ofRCTs evaluating the effect of levosimendan compared withstandard therapies or placebo on survival and haemody-namics parameters in patients presenting with acute severeheart failure. We found levosimendan, when compared toplacebo, was associated with a non-significant reduction inthe odds for mortality, and was associated with significantimprovements in several haemodynamic parameters alongwith a reduction in serum BNP. More interestingly, we foundthat levosimendan was superior to dobutamine, and asso-ciated with improved haemodynamics, lower serum BNP,and a significant survival benefit. Our data suggest that thisfinding may, in part, be due to a higher mortality associatedwith the use of dobutamine, rather than only a reduction inmortality associated with levosimendan. This finding is sup-ported by evidence of increased mortality with use of dobu-tamine when compared to only placebo.

We believe our study has a number of notable strengths.First, we have applied the recommended methods for theperformance of systematic reviews [45]. The use of duplicatesearches reduces the likelihood that significant studies weremissed, and the thorough methods for applying the inclusioncriteria, extracting and verifying the data, avoids errors andbiases. Second, our study is the most comprehensive and up-to-date evaluation of levosimendan performed to date. Finally,our study has benefited from the recent completion of anumber of larger RCTs.

As with all meta-analyses, there are also potential weak-nesses. The methodological quality of the studies included inthis analysis was less than optimal, with only two studiesmeeting all of the validity criteria. Some of these deficienciesmay relate to the reporting of the studies rather than theconduct of the studies [46]. Certainly, as some of the largermore recent studies are yet to appear in peer reviewed jour-nals, these deficiencies may be corrected. Not all trials re-ported all haemodynamic parameters or measured BNP,so these estimates are based on relatively small numbers ofmeasurements and should be interpreted with caution. Simi-larly, the pooled estimate for effect of dobutamine comparedto placebo was also based on few trials and, as an explora-tory analysis should be considered more hypothesisgenerating rather than definitive. Finally, there is a huge

variation in follow-up time, varying between 2 and 180 days.Clearly this limits the strengths of the conclusions in thisanalysis.

With these limitations in mind, we believe our findingsare important and relevant for clinicians. While the estimateof the OR for mortality favoured levosimendan over placebo,this result did not reach statistical significance. This maysuggest that the use of levosimendan in a general populationof patients with acute severe heart failure does not offer amortality benefit. However, even the pooled analysis couldhave been insufficiently powered to detect a clinically rele-vant reduction in mortality in a general population of patientswith acute heart failure. The other potential role for levo-simendan in patients with acute heart failure may be intargeted populations. Non randomised studies have providedcautious support for the use of levosimendan in patients withacute severe heart failure that is expected to recover withtime; for example those with peripartum cardiomyopathy[47], early graft failure following heart transplantation [48]and those with left ventricular failure following cardiacsurgery [49]. There is also some evidence to support the useof levosimendan in patients with right heart failure [50].These patientsmay benefit from the short-term haemodynamicimprovements seen with levosimendan. It is also possible thatlevosimendan may have a role in subgroups of patients withacute heart failure, such as those already taking beta-blockersor those with hypotension. It was not possible to examine theseissues with the data available in this review, and so furtherstudies are warranted.

The overall benefit seen with the use of levosimendancompared to dobutamine also warrants closer scrutiny. Theoverall estimate of treatment effect may be exaggerated bythe inclusion of lower quality RCTs. This observation issupported by the evidence from the sensitivity analysis thatshows a greater treatment effect in studies without allocationconcealment and those that did not perform an intention totreat analysis. It is also plausible that the dosing regime ofdobutamine used was less than optimal. As dobutamine is adrug with a short half-life, it is generally titrated to clinicaleffect [51], and is used for as long as clinically indicated.Levosimendan has metabolites that exert a clinical effect forup to a week [9]. In most RCTs, dobutamine was only givenfor 24 h, and therefore equipotent doses of dobutamine andlevosimendan were not compared. Interestingly, in theSURVIVE study [31], where dobutamine was given accor-ding to clinical need, no mortality benefit was found. Theother possibility is that the difference in mortality was duenot to a decrease in mortality with levosimendan, but ratheran increase in mortality with the use of dobutamine.

While the finding of an increase in mortality with dobu-tamine should be viewed as preliminary, it is in keepingwith other evidence for adverse effects of dobutamine inpatients with heart failure [8], and with evidence that β-blocking agents are associated with a beneficial effect inpatients with more chronic heart failure [52]. The results ofthis review would certainly give rise to concerns about the

288 A. Delaney et al. / International Journal of Cardiology 138 (2010) 281–289

continued use of dobutamine outside the confines of a well-designed RCT.

The fact that the findings of this meta-analysis are at oddswith those of the most recent large clinical trials also givesrise to some concerns. It is recognised that the results ofmeta-analyses are often discordant with those of large cli-nical trials [53]. In these cases, it is prudent to examine theindividual studies to resolve these discrepancies. The pub-lication of studies such as CASINO, REVIVE I and II in peerreviewed journals, as well as other ongoing studies, will shedfurther light on the role of levosimendan in the treatment ofpatients with acute severe heart failure.

There are still some important questions that require fur-ther investigation. The role of levosimendan in targeted po-pulations, such as those waiting for cardiac transplantation,those with acute heart failure following cardiac surgery, andother potentially reversible causes of cardiac failure are stillneeded. In addition, the role of dobutamine in the manage-ment of acute heart failure requires further investigation.

5. Conclusion

In summary, this systematic review of 19 randomisedclinical trials found that there was no evidence of survivalbenefit when levosimendan was compared to placebo. On theother hand, levosimendan was associated with an improve-ment in both haemodynamics and survival when comparedwith dobutamine. This difference may have been due to eitheran increase in mortality with the use of dobutamine or areduction in mortality with use of levosimendan. We believeadditional high quality studies of adjunctive inotropic therapy,in particular to further delineate the role of dobutamine, forpatients with acute heart failure are clearly needed.

Acknowledgements

The authors would like to thank Dr Doreen Rabi for herinsightful comments on the manuscript. The authors of thismanuscript have certified that they comply with thePrinciples of Ethical Publishing in the International Journalof Cardiology [54].

Appendix A. Supplementary data

Supplementary data associatedwith this article can be found,in the online version, at doi:10.1016/j.ijcard.2008.08.020.

References

[1] NieminenMS,BrutsaertD,DicksteinK, et al. EuroHeart Failure Survey II(EHFS II): a survey on hospitalized acute heart failure patients: descriptionof population. Eur Heart J 2006;27:2725–36.

[2] Adams Jr KF, Fonarow GC, Emerman CL, et al. Characteristics andoutcomes of patients hospitalized for heart failure in the United States:rationale, design, and preliminary observations from the first 100,000cases in the Acute Decompensated Heart Failure National Registry(ADHERE). Am Heart J 2005;149:209–16.

[3] Nieminen MS, Bohm M, Cowie MR, et al. Executive summary of theguidelines on the diagnosis and treatment of acute heart failure: the TaskForce on Acute Heart Failure of the European Society of Cardiology.Eur Heart J 2005;26:384–416.

[4] Krum H, Liew D. New and emerging drug therapies for the managementof acute heart failure. Intern Med J 2003;33:515–20.

[5] McCullough PA, Philbin EF, Spertus JA, et al. Confirmation of a heartfailure epidemic: findings from theResourceUtilization amongCongestiveHeart Failure (REACH) study. J Am Coll Cardiol 2002;39:60–9.

[6] Nieminen MS, Harjola VP. Definition and epidemiology of acute heartfailure syndromes. Am J Cardiol 2005;96:5G–10G.

[7] Peter JV, Moran JL, Phillips-Hughes J, Graham P, Bersten AD. Effectof non-invasive positive pressure ventilation (NIPPV) on mortality inpatients with acute cardiogenic pulmonary oedema: a meta-analysis.Lancet 2006;367:1155–63.

[8] Thackray S, Easthaugh J, Freemantle N, Cleland JG. The effectivenessand relative effectiveness of intravenous inotropic drugs acting throughthe adrenergic pathway in patients with heart failure—a meta-regressionanalysis. Eur J Heart Fail 2002;4:515–29.

[9] Toller WG, Stranz C. Levosimendan, a new inotropic and vasodilatoragent. Anesthesiology 2006;104:556–69.

[10] De Luca L, Colucci WS, Nieminen MS, Massie BM, Gheorghiade M.Evidence-based use of levosimendan in different clinical settings. EurHeart J 2006;27:1908–20.

[11] Parissis JT, Adamopoulos S, Farmakis D, et al. Effects of serial levo-simendan infusions on left ventricular performance and plasma bio-markers of myocardial injury and neurohormonal and immune activationin patients with advanced heart failure. Heart 2006.

[12] Adamopoulos S, Parissis JT, Iliodromitis EK, et al. Effects oflevosimendan versus dobutamine on inflammatory and apoptoticpathways in acutely decompensated chronic heart failure. Am J Cardiol2006;98:102–6.

[13] Parissis JT, Adamopoulos S, Antoniades C, et al. Effects of levo-simendan on circulating pro-inflammatory cytokines and solubleapoptosis mediators in patients with decompensated advanced heartfailure. Am J Cardiol 2004;93:1309–12.

[14] Parissis JT, Panou F, Farmakis D, et al. Effects of levosimendan onmarkersof left ventricular diastolic function and neurohormonal activation inpatients with advanced heart failure. Am J Cardiol 2005;96:423–6.

[15] Montes FR, Echeverri D, Buitrago L, et al. The vasodilatory effects oflevosimendan on the human internal mammary artery. Anesth Analg2006;103:1094–8.

[16] Yildiz O. Vasodilating mechanisms of levosimendan: involvement ofK+ channels. J Pharmacol Sci 2007;104:1–5.

[17] Moiseyev VS, Poder P, Andrejevs N, et al. Safety and efficacy of a novelcalcium sensitizer, levosimendan, in patients with left ventricular failuredue to an acute myocardial infarction. A randomized, placebo-controlled,double-blind study (RUSSLAN). Eur Heart J 2002;23:1422–32.

[18] Follath F, Cleland JG, Just H, et al. Efficacy and safety of intravenouslevosimendan compared with dobutamine in severe low-output heartfailure (the LIDO study): a randomised double-blind trial. Lancet2002;360:196–202.

[19] Haynes RB, McKibbon KA, Wilczynski NL, Walter SD, Werre SR.Optimal search strategies for retrieving scientifically strong studies oftreatment from Medline: analytical survey. BMJ 2005;330:1179.

[20] Wong SS, Wilczynski NL, Haynes RB. Developing optimal searchstrategies for detecting clinically sound treatment studies in EMBASE.J Med Libr Assoc 2006;94:41–7.

[21] Hopewell S, McDonald S, Clarke M, Egger M. Grey literature in meta-analyses of randomized trials of health care interventions. CochraneDatabase Syst Rev 2007 MR000010.

[22] Juni P, Altman DG, Egger M. Systematic reviews in health care:assessing the quality of controlled clinical trials. BMJ 2001;323:42–6.

[23] Schulz KF, Grimes DA. Allocation concealment in randomised trials:defending against deciphering. Lancet 2002;359:614–8.

[24] Schulz KF, Grimes DA. Blinding in randomised trials: hiding who gotwhat. Lancet 2002;359:696–700.

289A. Delaney et al. / International Journal of Cardiology 138 (2010) 281–289

[25] Schulz KF, Chalmers I, Hayes RJ, Altman DG. Empirical evidence ofbias. Dimensions of methodological quality associated with estimatesof treatment effects in controlled trials. JAMA 1995;273:408–12.

[26] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuringinconsistency in meta-analyses. BMJ 2003;327:557–60.

[27] Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34.

[28] Villar J, Mackey ME, Carroli G, Donner A. Meta-analyses in systematicreviews of randomized controlled trials in perinatalmedicine: comparisonof fixed and random effects models. Stat Med 2001;20:3635–47.

[29] Deeks JJ. Issues in the selection of a summary statistic for meta-analysisof clinical trials with binary outcomes. Stat Med 2002;21:1575–600.

[30] Deeks JJ, Higgins JP, Altman DG. Summarising effects across studies.In: Higgins JP, Green S, editors. Cochrane Handbook for SystematicReviews of Interventions 426; 2006.

[31] Mebazaa A, Nieminen MS, Packer M, et al. Levosimendan vs dobu-tamine for patients with acute decompensated heart failure: theSURVIVE Randomized Trial. JAMA 2007;297:1883–91.

[32] Al-Shawaf E,AyedA,Vislocky I, et al. Levosimendan ormilrinone in thetype 2 diabetic patient with low ejection fraction undergoing electivecoronary artery surgery. J Cardiothorac Vasc Anesth 2006;20:353–7.

[33] Alvarez J, Bouzada M, Fernandez AL, et al. Hemodynamic effects oflevosimendan compared with dobutamine in patients with low cardiacoutput after cardiac surgery. [Spanish]Revista Espanola de Cardiolo-gia 2006;59(4):338–45.

[34] Avgeropoulou C, Andreadou I, Markantonis-Kyroudis S, et al. The Ca2+-sensitizer levosimendan improves oxidative damage, BNP and pro-in-flammatory cytokine levels in patients with advanced decompensatedheart failure in comparison to dobutamine. Eur J Heart Fail 2005;7:882–7.

[35] De Hert SG, Lorsomradee S, Cromheecke S, Van der Linden PJ. Theeffects of levosimendan in cardiac surgery patients with poor leftventricular function. Anesth Analg 2007;104:766–73.

[36] De Luca L, Proietti P, Celotto A, et al. Levosimendan improveshemodynamics and coronary flow reserve after percutaneous coronaryintervention in patients with acute myocardial infarction and left ven-tricular dysfunction. Am Heart J 2005;150:563–8.

[37] Follath F, Hinkka S, Jager D, et al. Dose-ranging and safety withintravenous levosimendan in low-output heart failure: experience inthree pilot studies and outline of the levosimendan infusion versusdobutamine (LIDO) trial. Am J Cardiol 1999;83:21I–5I.

[38] Moertl D, Berger R,HuelsmannM,Bojic A, Pacher R. Short-term effectsof levosimendan and prostaglandin E1 on hemodynamic parameters andB-type natriuretic peptide levels in patients with decompensated chronicheart failure. Eur J Heart Fail 2005;7:1156–63.

[39] Morelli A, De Castro S, Teboul JL, et al. Effects of levosimendan onsystemic and regional hemodynamics in septic myocardial depression.Intensive Care Med 2005;31:638–44.

[40] Nieminen MS, Akkila J, Hasenfuss G, et al. Hemodynamic and neuro-humoral effects of continuous infusion of levosimendan in patients withcongestive heart failure. J Am Coll Cardiol 2000;36:1903–12.

[41] Tziakas DN, Chalikias GK, Hatzinikolaou HI, et al. Levosimendan usereduces matrix metalloproteinase-2 in patients with decompensatedheart failure. Cardiovasc Drugs Ther 2005;19:399–402.

[42] Cleland JG, Freemantle N, Coletta AP, Clark AL. Clinical trials updatefrom the American Heart Association: REPAIR-AMI, ASTAMI, JELIS,MEGA, REVIVE-II, SURVIVE, and PROACTIVE. Eur J Heart Fail2006;8:105–10.

[43] Coletta AP, Cleland JG, Freemantle N, Clark AL. Clinical trials updatefrom the EuropeanSociety of CardiologyHeart Failuremeeting: SHAPE,BRING-UP 2 VAS, COLA II, FOSIDIAL, BETACAR, CASINO andmeta-analysis of cardiac resynchronisation therapy. Eur J Heart Fail2004;6:673–6.

[44] Packer M, Colucci WS, Fisher L, et al. Development of a comprehensivenew endpoint for the evaluation of new treatments for acute decom-pensated heart failure: results with levosimendan in the REVIVE 1 study.J Card Fail 2003;9:S61.

[45] Moher D, CookDJ, Eastwood S, et al. Improving the quality of reports ofmeta-analyses of randomised controlled trials: the QUOROM statement.Quality of reporting of meta-analyses. Lancet 1999;354:1896–900.

[46] Hill CL, LaValley MP, Felson DT. Discrepancy between publishedreport and actual conduct of randomized clinical trials. J Clin Epidemiol2002;55:783–6.

[47] Benlolo S, Lefoll C, KatchatouryanV, PayenD,MebazaaA. Successfuluse of levosimendan in a patient with peripartum cardiomyopathy.Anesth Analg 2004;98:822–4.

[48] Petaja LM, Sipponen JT, Hammainen PJ, et al. Levosimendan rever-sing low output syndrome after heart transplantation. Ann Thorac Surg2006;82:1529–31.

[49] Raja SG, Rayen BS. Levosimendan in cardiac surgery: current bestavailable evidence. Ann Thorac Surg 2006;81:1536–46.

[50] Parissis JT, Paraskevaidis I, Bistola V, et al. Effects of levosimendan onright ventricular function in patients with advanced heart failure. Am JCardiol 2006;98:1489–92.

[51] Lehtonen LA, Antila S, Pentikainen PJ. Pharmacokinetics and pharma-codynamics of intravenous inotropic agents. Clin Pharmacokinet2004;43:187–203.

[52] Foody JM, Farrell MH, Krumholz HM. beta-Blocker therapy in heartfailure: scientific review. JAMA 2002;287:883–9.

[53] LeLorier J, Gregoire G, Benhaddad A, Lapierre J, Derderian F.Discrepancies between meta-analyses and subsequent large rando-mized, controlled trials. N Engl J Med 1997;337:536–42.

[54] Coats AJ. Ethical authorship and publishing. Int J Cardiol 2009;131:149–50.