Comparative safety and efficacy of levosimendan versus dobutamine in acute heart failure: a systematic review, meta-analysis, and meta-regression

Abstract

Background:

Acute heart failure (AHF) remains a major cause of morbidity and mortality, especially in hospitalized patients. Inotropic agents like dobutamine have long been used for hemodynamic support, yet concerns over adverse effects such as arrhythmias and increased myocardial oxygen consumption limit their safety. Levosimendan, a calcium sensitizer with vasodilatory properties, offers a potential alternative. This meta-analysis was conducted to evaluate the comparative safety and efficacy of levosimendan versus dobutamine in patients with AHF.

Method:

A systematic review and meta-analysis was conducted following PRISMA and AMSTAR 2 guidelines. Major databases including PubMed, Cochrane, Embase, and ScienceDirect were searched up to July 2025 without any language restrictions. Sixteen studies including 15 randomized controlled trials and 1 cohort study were evaluated for this study. Outcomes were pooled using a random effects model. Meta-regression was performed to assess the influence of baseline covariates.

Result:

In the pooled analysis of 16 studies, levosimendan showed a statistically significant reduction in mortality (RR: 0.59; P = 0.0009) compared to dobutamine. While systolic blood pressure and diastolic blood pressure remained comparable, levosimendan led to a marked reduction in pulmonary capillary wedge pressure and improved cardiac index, left ventricular ejection fraction, and mixed venous oxygen saturation. B-type natriuretic peptide and IL-6 levels also declined significantly in the levosimendan group. Adverse events such as hypotension and atrial fibrillation were slightly more frequent with levosimendan, while tachycardia and ischemic events were more common with dobutamine. Overall adverse event rates were similar.

Conclusion:

Levosimendan demonstrates favorable effects on mortality and key hemodynamic parameters in patients with AHF, especially those on β-blockers or with impaired left ventricular function. Although certain adverse effects like hypotension were more frequent, the overall safety profile remained comparable to dobutamine. These findings suggest levosimendan as a safer and more effective alternative in selected high-risk heart failure populations.

Introduction

Heart failure (HF) is a chronic and progressive condition affecting approximately 64 million people worldwide, with an increasing prevalence due to aging populations and comorbidities such as ischemic heart disease, hypertension, and diabetes^[1–3]. Acute decompensated heart failure (ADHF), a severe manifestation of HF, is characterized by the rapid worsening of symptoms, including dyspnea, fluid retention, and poor perfusion, often necessitating hospitalization. ADHF is a leading cause of medical admissions and is associated with high short- and long-term mortality. In severe cases, the 1 year survival rate can be as low as 50%, and the clinical course is frequently complicated by low cardiac output, systemic hypoperfusion, and end organ dysfunction, necessitating intravenous inotropic support in critically ill patients^[4–6].

Management of ADHF often requires positive inotropic therapy when a low cardiac output is evident, such as in hypotension or oliguria. Dobutamine, a beta 1 adrenergic agonist, has traditionally been a cornerstone of inotropic support because of its ability to enhance myocardial contractility. However, its mechanism of action, which involves raising intracellular calcium via cAMP, can increase myocardial oxygen consumption, induce tachycardia, impair diastolic relaxation, and increase the risk of arrhythmias, particularly in patients on β-blockers.

In contrast, levosimendan, a calcium sensitizer with potassium channel opening properties, enhances cardiac contractility without increasing oxygen demand and demonstrates prolonged hemodynamic benefits due to its active metabolite^[7–10]. This mechanism may confer particular advantages in patients with impaired adrenergic responsiveness or markedly reduced ejection fraction, in whom conventional adrenergic inotropes are less effective. Despite widespread clinical use, uncertainty persists regarding the optimal inotropic agent for ADHF, especially in patient subgroups at high risk of adverse outcomes, such as those on chronic β-blockade or those with advanced cardiac dysfunction. Multiple randomized controlled trials have compared levosimendan and dobutamine; however, the results remain inconsistent with respect to long-term mortality, symptom relief, and adverse events^[11–13].

This meta-analysis aimed to systematically compare the efficacy and safety of levosimendan versus dobutamine in acute heart failure (AHF), with particular attention to hemodynamic stabilization, arrhythmia risk, and short- and long-term outcomes. By evaluating evidence under comparable baseline conditions, this study seeks to clarify whether levosimendan provides a clinical advantage over dobutamine and to support clinicians in making evidence-based, patient-tailored decisions in the management of AHF.

Methods and materials

This systematic review and meta-analysis was conducted and reported in accordance with the PRISMA 2020 statement^[14]. The review protocol was prospectively registered in PROSPERO (registration number CRD420251117829). In accordance with the TITAN 2025 Guidelines on transparent AI usage, all study design, data extraction, analysis, and interpretation were performed by the authors without AI assistance; AI was only used for language editing of the manuscript^[15].

Data sources and search strategy

We searched PubMed/MEDLINE, Embase, Cochrane Library, and ScienceDirect from inception to 31 July 2025. To identify grey literature, clinicaltrials.gov and conference proceedings were searched, and the bibliographies of the included articles were hand searched. No language or publication date restrictions were imposed. The exact search strings for each database are provided in Supplemental Digital Content Table 1, available at: http://links.lww.com/MS9/B74. The final search was conducted on 31 July 2025.

HIGHLIGHTS.

• Levosimendan reduces mortality in acute heart failure patients.

• Improves left ventricular ejection fraction, cardiac index, mixed venous oxygen saturation vs dobutamine therapy.

• Lowers pulmonary capillary wedge pressure and B-type natriuretic peptide, indicating better hemodynamic outcomes.

• Safety comparable to dobutamine despite mild adverse events.

Table 1.

Characteristics of included studies

Study names	Type of study	Intervention	Comparator	Primary outcome	Follow-up
Ender Oner et al (2014)	RCT	Levosimendan	Dobutamine	NYHA class	5 days post infusion
Adamopoulos et al (2006)	RCT	Levosimendan	Dobutamine	LVEF%	4 months
Osman Can et al (2010)	RCT	Levosimendan	Dobutamine	QRS duration	24 hours post infusion
H. Duygu et al (2007)	RCT	Levosimendan	Dobutamine	Sa	24 hours post infusion
Hamza Duygu et al (2008)	RCT	Levosimendan	Dobutamine	LVEF%	24 hours post infusion
Mehmet Birhan et al (2007)	RCT	Levosimendan	Dobutamine	LVEF%	72 hours post infusion
Mehmet Birhan et al (2009)	RCT	Levosimendan	Dobutamine	Ejection fraction	24 hours post infusion
Julian Alvarez et al (2006)	RCT	Levosimendan	Dobutamine	Cardiac index	48 hours post infusion
Claes Hakan et al (2010)	RCT	Levosimendan	Dobutamine	Cardiac index	1 month
Markku et al (2000)	RCT	Levosimendan	Dobutamine	Cardiac output	9 days
Sima Fard et al (2007)	RCT	Levosimendan	Dobutamine	Cardiac death	12 months
Folath et al (2002)	RCT	Levosimendan	Dobutamine	Cardiac output	180 days
Alexandre Mebazaa et al (2007)	RCT	Levosimendan	Dobutamine	All-cause mortality at 180 days	180 days
Avgeropoulou et al (2005)	RCT	Levosimendan	Dobutamine	BNP levels	5 days post infusion
Marta Madeira et al (2017)	Retrospective Cohort	Levosimendan	Dobutamine	CRS	481 days
Alison P et al (2004)	RCT	Levosimendan	Dobutamine	Mortality at 180 and 30 days	6 months

BNP: B‑type natriuretic peptide; CRS: cardiorenal syndrome; LVEF: left ventricular ejection fraction; NYHA: New York Heart Association; Sa: peak systolic velocity of mitral annulus.

Study selection

Records identified were imported into EndNote (version X7.5; Clarivate Analytics) and deduplicated. Two reviewers (LK and MK) independently screened the titles and abstracts and reviewed the full texts for eligibility. Disagreements were resolved through discussion with a third reviewer (AR). The interrater agreement (Cohen’s κ) for full text selection is reported in the Supplementary Methods. The inclusion criteria were randomized controlled trials (RCTs) or comparative observational studies of adult patients (≥18 years) with AHF that directly compared levosimendan and dobutamine and reported at least one predefined outcome of interest. The exclusion criteria were as follows: studies enrolling pediatric patients, case series, case reports, conference abstracts without full data, and studies lacking relevant outcomes. For multiple reports of the same study population, the most complete dataset was used.

Data extraction

Two reviewers (NK, IM) independently extracted data using a pre piloted form, and a third reviewer (HR) resolved any discrepancies. The extracted items included trial name, publication year, study design, sample size, mean age, sex distribution, background pharmacotherapy (ACE inhibitors, diuretics, β-blockers, aldosterone antagonists), comorbidities (diabetes, hypertension, hyperlipidemia), baseline hemodynamics [systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate (HR), left ventricular ejection fraction (LVEF)], and outcomes. The primary outcome was all-cause mortality (time point specified for each study). Secondary outcomes included SBP, DBP, HR, pulmonary capillary wedge pressure (PCWP), stroke volume, cardiac index/output, LVEF (%), systolic pulmonary artery pressure (SPAP), left ventricular end diastolic diameter (LVEDd), left ventricular end systolic diameter (LVESd), left atrial (LA) diameter, systemic vascular resistance (SVR), mixed venous oxygen saturation (SvO₂; measured via Swan–Ganz catheter); where only central venous oxygen saturation (ScvO₂) was reported, it was analyzed as a surrogate and noted. Inflammatory markers (IL-6, TNF-α) and treatment emergent adverse events were also recorded. When continuous outcomes were reported as medians and interquartile ranges, we converted them to means and standard deviations using established methods.

Risk of bias and quality assessment

The risk of bias for randomized trials was assessed using the Cochrane RoB 2 tool across its domains; for nonrandomized studies, we used the Newcastle Ottawa Scale (selection, comparability, outcomes)^[16,17]. Two reviewers independently performed these assessments, and discrepancies were adjudicated by a third reviewer. RoB domain level judgments and overall assessments are reported in Supplemental Digital Content Tables 2 and 3, available at: http://links.lww.com/MS9/B74 and displayed graphically (traffic light plot).

Statistical analysis

Analyses were performed using Review Manager (Rev Man v5.4.1) and Comprehensive Meta-Analysis (CMA v3) for meta-regression^[18]. For dichotomous outcomes (e.g., mortality, adverse events), we pooled risk ratios (RRs) using the Mantel–Haenszel method; for continuous outcomes, we used mean differences (MDs) when outcomes used the same measurement scale and standardized mean differences (SMDs) when different scales were used. We reported 95% confidence intervals (CIs) and 95% prediction intervals for pooled estimates.

A random effects model was used for the primary analyses. We applied the Hartung–Knapp adjustment for random effects models for primary analyses and conducted Der Simonian–Laird pooled estimates as sensitivity analyses. Statistical heterogeneity was assessed using Higgins’ I² statistic and categorized as low (<50%), moderate (50–75%), or high (>75%). Where substantial heterogeneity was present, we explored potential sources via prespecified subgroup analyses (study design: RCT vs observational; risk of bias; baseline LVEF; treatment dose/duration) and sensitivity analyses (excluding high RoB studies, using alternative effect estimators).

For meta-regression, we prespecified covariates (baseline SBP, DBP, HR, mean age, presence of hypertension, β-blocker use, angiotensin-converting enzyme inhibitor (ACEi) use, mean LVEF). Meta-regression was performed in CMA and only when at least 10 studies were available for the outcome of interest, consistent with recommended practice.

We assessed small study effects/publication bias using funnel plots and Egger’s regression test when ≥10 studies contributed to an outcome and applied trim and fill methods as a sensitivity analysis if asymmetry was detected. Zero-event trials for dichotomous outcomes were handled using continuity correction as appropriate. Sensitivity analyses using alternative approaches (Peto OR, exclusion) were performed.

Statistical significance was set at P < 0.05. All analytic decisions (choice of effect measure, model, and sensitivity analyses) were prespecified in the protocol (PROSPERO) and are described in the Supplementary Methods. Rev Man and CMA input files are available from the corresponding author upon reasonable request.

Result

In total, 15 RCTs and one cohort study were included in the analysis. The primary outcome of all-cause mortality highlighted a significant decrease with levosimendan compared with dobutamine at both long-term (115–365 days) and short-term (≤30 days) follow-ups. Improvements in hemodynamic parameters, including increases in LVEF, cardiac index, and SvO₂, along with reductions in PCWP and B-type natriuretic peptide (BNP) were consistently observed. However, treatment benefits of Levosimendan were accompanied with certain adverse events, most notably headache, nausea, and hypokalemia, whereas nonsignificant differences were observed in the risk of arrhythmias, hypotension, or renal dysfunction. Study design (single center vs multicenter) and sensitivity analyses stratified by levosimendan dose demonstrated the robustness of mortality and hemodynamic benefits, though some effects particularly BNP reduction were attenuated in multicenter trials. In general, these findings suggest the clinical significance of levosimendan over dobutamine providing mortality and hemodynamic benefits with an increased incidence of minor adverse events.

Study selection

We began our research exploring multiple databases and shortlisting studies on the basis of their exclusion and inclusion criteria, according to PRISMA guidelines. Consequently, after rigorous screening 16 studies met our predefined inclusion criteria and were included in this systemic review and meta-analysis. Detailed study selection process for this manuscript is established in Figure 1.

Figure 1.

PRISMA flowchart.

Study and baseline characteristics of included studies

Among 16 studies included, 15 were identified as RCT, while one was retrospective cohort, The included RCTs and their corresponding sample size are as follows: Oner et al (2014) with a total sample of 61^[19]; Adamopoulos et al (2006) with 46^[20]; Osman Can et al (2010) with 58^[21]; Duygu et al (2007) with 62^[22]; Duygu et al (2008) with 60^[23]; Birhan et al (2007) with 88^[24]; Birhan et al (2009) with 40^[25]; Alvarez et al (2006) with 41^[26]; Hakan et al (2010) with 60^[27]; Markku et al (2000) with 115^[28]; Fard et al (2007) with 22^[29]; Folath et al (2002) with 203^[30]; Mebazaa et al (2007) with 1327^[31]; Avgeropoulou et al (2005) with 29^[32]; Alison et al (2004) with 200^[33]; and a cohort Madeira et al (2017)^[34] with 108. Additionally, mean age was also extracted and it ranged between 61 and 71 years across majority of studies .Other baselines parameters extracted include SBP, DBP, HR, hyperlipidaemia, diabetes mellitus, hypertension. The data of baseline medications were also retrieved including ACEi, diuretics, β-blockers, and aldosterone antagonists, and further details are given in Table 1 and Supplemental Digital Content Table 4A and B, available at: http://links.lww.com/MS9/B74.

Primary outcome

Mortality

The overall pooled analysis demonstrated a significant decline in mortality risk in the levosimendan group compared to the dobutamine group (RR: 0.59; 95% CI [0.43, 0.80]; P = 0.0009), with a moderate to high degree of heterogeneity (I² = 65%; P < 0.0001). Additionally, a leave-one-out analysis was performed in a study by Madeira et al from the subgroup with 30 days and 180 days, which decreased the heterogeneity value (I² = 29%; P = 0.15) (Fig. 2).

Figure 2:

(A) Forest plot for mortality. (B) Leave-one-out analysis for mortality. (C) Mortality by dosage. (D) Mortality by study design.

We evaluated the risk of mortality in patients receiving levosimendan compared to dobutamine, stratified by follow-up duration into three subgroups: mortality within 30, 180, and 365 days postoperatively. In the first subgroup, mortality within 30 days, 10 studies reported this outcome and demonstrated a substantial reduction in mortality events in the levosimendan group compared with dobutamine (RR: 0.60; 95% CI [0.37, 0.97]; P = 0.04; I² = 39%).In the second subgroup within 180 days, four studies, namely Adamopoulos et al (2006), Alison et al (2004), Folath et al (2002), and Mebazaa et al (2007), also showed a consistent survival benefit in favor of levosimendan (RR: 0.66; 95% CI [0.43, 1.00]; P = 0.05; I² = 72%). In the third subgroup within 365 days, two studies, Sima Fard et al (2007) and Madeira et al (2017) also suggested a lower risk of mortality with levosimendan (RR: 0.75; 95% CI [0.07, 7.58]; P = 0.80; I² = 79%) (Fig. 2).

The overall pooled analysis demonstrated that levosimendan was associated with a significantly lower mortality rate than dobutamine (RR 0.63, 95% CI 0.45–0.89; P = 0.008; I² = 64%) (Supplemental Digital Content Figure 3, available at: http://links.lww.com/MS9/B74; Supplemental Digital Content Table 6A, available at: http://links.lww.com/MS9/B74). Although heterogeneity remained moderate to high, this benefit was consistent across short and intermediate follow-up periods (Fig. 2).

We stratified our results based on the basis of their study design. Multicenter trials of mortality within 30 days demonstrated a neutral effect (RR 0.82, 95% CI 0.63–1.06; P = 0.13; I² = 0%), whereas single-center studies suggested a nonsignificant trend toward benefit (RR 0.63, 95% CI 0.20–1.95; P = 0.42; I² = 39%) (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74). A similar difference in results was obtained when comparing the mortality rate within 180 days for multicenter studies with single-center studies (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74). Similar results were obtained on stratification of our results according to different dosages; 0.1 μg/kg/min infusions resulted in a nonsignificant reduction at 30 days (RR 0.64, 95% CI 0.35–1.16; P = 0.14; I² = 48%), while higher doses (0.2–0.6 μg/kg/min) yielded inconsistent results due to limited sample sizes (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74).

Secondary outcomes

SBP

Seven studies reported the change in systolic blood pressure, namely Adamopoulos et al (2006) (0.1 μg/kg/min); Oner et al (2014) (0.1 mg/kg/min); Duygu et al (2007) (0.1 μg/kg/min); Hamza Duygu et al (2008) (0.1 μg/kg/min); Mehmet Birhan et al (2007) (0.1 μg/kg/min); Mehmet Birhan et al (2009) (0.1 μg/kg/min); Osman Can et al (2010) (0.1 μg/kg/min). The overall analysis demonstrated comparable values of SBP in both levosimendan and dobutamine recipients (M.D: − 0.1; CI [−2.72, 2.51]; P = 0.94), and no heterogeneity was observed among these studies (I² = 0%;P = 0.77) (Supplemental Digital Content Fig. 1A, available at: http://links.lww.com/MS9/B74).

The overall analysis of dose-based stratification demonstrated a nonsignificant effect on SBP (MD 1.80 mmHg, 95% CI −2.67 to 6.29; P = 0.43; I² = 67%) (Supplemental Digital Content Fig. 1B, available at: http://links.lww.com/MS9/B74). By dose, the effects on systolic blood pressure appeared minimal across regimens. Nonsignificant effects were demonstrated at 0.1 μg/kg/min (MD 2.09, 95% CI −2.99 to 7.17; P = 0.42; I² = 72%) or 0.1 mg/kg/min (MD −0.30, 95% CI −8.69 to 8.09; P = 0.94) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74).

DBP

Seven studies reported changes in DBP [Adamopoulos et al 2006 (0.1 μg/kg/min); Ender Oner et al 2014(0.1 mg/kg/min); H. Duygu et al 2007 (0.1 μg/kg/min); Hamza Duygu et al 2008 (0.1 μg/kg/min); Mehmet Birhan et al 2007(0.1 μg/kg/min); Mehmet Birhan et al 2009(0.1 μg/kg/min); Osman Can et al 2010 (0.1 μg/kg/min)]. The overall results showed no notable clinical benefit in the outcome, as evident by (M.D: − 1.01; CI [−2.98, 0.96]; P = 0.31) with no heterogeneity (I² = 0%; P = 0.9) (Supplemental Digital Content Fig. 2A, available at: http://links.lww.com/MS9/B74).

Dose-based analysis revealed no notable clinical benefit in the outcome, as evidenced by (MD −0.61 mmHg, 95% CI −2.71 to 1.50; P = 0.57; I² = 9%) (Supplemental Digital Content Fig. 2B, available at: http://links.lww.com/MS9/B74). DBP change of (MD −0.70 mmHg, 95% CI −3.12 to 1.73; P = 0.57; I² = 23%) was noticed to be 0.1 μg/kg/min. A similar DBP change (MD 0.40 95% CI −5.62 to 6.42; P = 0.90) was observed at 0.1 mg/kg/min (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74).

HR

We evaluated the variation in HR in the levosimendan arm compared to the dobutamine arm by dividing it into different subgroups based on the follow-up period of the outcome, i.e., after 6, 12, 24, and 48 h. In the subgroup analysis of HR after 6 h [Alvarez et al 2006 (0.2 μg/kg/min)] demonstrated a significant increase in HR in the population receiving levosimendan (M.D: 5.4; CI [0.01, 10.79]; P = 0.05). In the subgroup analysis of HR after 12 h, one study showed insignificant findings (M.D: 5.30; CI [−0.53, 11.13]; P = 0.07). HR after 24 hours was reported by Alvarez et al (2006) (0.2 μg/kg/min), Hakan et al (2010) (0.1 μg/kg/min), Oner et al (2014) (0.1 mg/kg/min), Duygu et al (2007) (0.1 μg/kg/min), Hamza Duygu et al (2008) (0.1 μg/kg/min), Mehmet Birhan et al (2009) (0.1 μg/kg/min), Osman Can et al (2010) (0.1 μg/kg/min), Markku et al (2000) 0.6 μg/kg/min, Markku et al (2000) (0.4 μg/kg/min), Markku et al (2000) (0.2 μg/kg/min), and Markku et al (2000) (0.1 μg/kg/min), Markku et al (2000) (0.05 μg/kg/min) showed no meaningful variation in between levosimendan and dobutamine population (M.D:11.32; CI [−2, 24.64]; P = 0.1). However, the subgroup of HR after 48 h showed a notable improvement in HR in levosimendan recipients compared with levosimendan recipients (M.D: 5.91; CI [2.58, 9.24]; P = 0.0005). The overall pooled analysis indicated no appreciable difference between the levosimendan and dobutamine populations (M.D: 9.83; CI [−0.59, 20.52]; P = 0.06) with a very high value of heterogeneity (I² = 99%; P <0.00001) (Supplemental Digital Content Fig. 3A, available at: http://links.lww.com/MS9/B74).

The overall pooled analysis based on study design indicated no appreciable difference between the levosimendan and dobutamine populations (MD 3.18 bpm, 95% CI −0.18 to 6.54; P = 0.06; I² = 87%). Although the effect did not reach statistical significance, the directionality favored a small chronotropic effect. Single-center studies demonstrated a nonsignificant increase (MD 2.72, 95% CI −1.19 to 6.63; P = 0.17; I² = 90%), whereas multicenter trials demonstrated a meaningful increase (MD 5.85, 95% CI 2.22–9.47; P = 0.002; I² = 0%). Dose-specific analyses indicated marked a significant rise at 0.2 μg/kg/min (MD 5.83, 95% CI 3.26–8.40; P < 0.00001) and 0.6 μg/kg/min (MD 15.20, 95% CI 8.10–22.30; P < 0.0001) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74), suggesting a positive dose-related effect. (Supplemental Digital Content Figures 3C and 5B, available at: http://links.lww.com/MS9/B74).

PCWP

PCWP, an indirect means of estimating LA pressure, was assessed forming two cohorts based on the follow-up period, namely PWCP after 24 hours and PCWP after 48 hours. In the first subgroup, PCWP (mmHg) after 24 hours Claes Hakan et al 2010 (0.1 μg/kg/min); Markku et al 2000 (0.6 μg/kg/min); Markku et al 2000 (0.4 μg/kg/min); Markku et al 2000 (0.2 μg/kg/min); Markku et al 2000 (0.1 μg/kg/min); Markku et al 2000 (0.05 μg/kg/min); Folath et al 2002 (0.1 μg/kg/min) levosimendan arm showed meaningful clinical benefit in lowering PCWP as compared to dobutamine (M.D: − 3.63;CI [−4.85, −2.4]; P = < 0.00001).Similarly, in the subgroup of PCWP (mmHg) after 48 h, two studies Claes Hakan et al (2010) (0.1 μg/kg/min) and Adamopoulos et al (2006) (0.1 μg/kg/min) demonstrated a significant reduction in PWCP in the levosimendan population (M.D: − 4.26; CI [−6.46, − 2.06]; P = 0.0001). The overall effect estimated showed a significant reduction in PCWP in patients receiving levosimendan compared to dobutamine with nonsignificant heterogeneity among the studies (M.D: − 3.78; CI [−4.85, − 2.71]; P = < 0.00001) (I² = 0%; P = 0.98) (Supplemental Digital Content Figure 4A, available at: http://links.lww.com/MS9/B74).

Study design-based analysis demonstrated that levosimendan consistently reduced PCWP compared with dobutamine (MD, − 3.78 mmHg; 95% CI, − 4.85 to − 2.71; P < 0.00001; I² = 0%). This finding demonstrates improved left-sided filling pressure. Both single-center (MD −4.00, 95% CI −6.77 to—1.23; P = 0.005) and multicenter trials (MD −3.74, 95% CI −6.77 to—1.23; P < 0.00001; I² = 0%) favored the effect of levosimendan (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74). By dose, the most prominent reduction was observed at 0.6 μg/kg/min (MD −5.80, 95% CI −10.04 to −1.56; P = 0.007) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74) (Supplemental Digital Content Figure 4A and B, available at: http://links.lww.com/MS9/B74).

Stroke volume

Stroke volume, a hemodynamic parameter to assess cardiac function was evaluated stratifying the studies in fours subgroups. In the first subgroup, stroke volume (ml) after 6 h of the study reported a decrease in stroke volume in the levosimendan population compared with dobutamine (M.D: − 4.7; CI [−8.07,-1.33]; P = 0.006). However, in the second cohort of stroke volume after 12 h, Alvarez et al (2006) (0.2 μg/kg/min) demonstrated significant improvement in stroke volume parameter in the levosimendan arm in comparison with dobutamine arm (M.D: 5.10; CI [2.04, 8.16]; P = 0.001). Other subgroups, stroke volume after 24 h and stroke volume after 48 h, also demonstrated improvement in stroke volume value, while it did not reach statistical significance in the 24 h cohort as indicated by (M.D: 0.86; CI [−1.73, 3.45]; P = 0.51) and (M.D: 5.09; CI[2.25, 7.39]; P = 0.0005), respectively. The overall effect size showed no significant difference in stroke volume in levosimendan recipients compared to dobutamine recipients, with a moderate amount of heterogeneity (M.D: 1.33; CI [−0.94, 3.61]; P = 0.25) (I² = 72%; P = < 0.0001) (Supplemental Digital Content Figure 5A, available at: http://links.lww.com/MS9/B74).

Both dose-based analyses and study design stratified analyses concluded non-consistent effects (MD 1.33, 95% CI −0.94 to 3.61; P = 0.25; I² = 72%) (Supplemental Digital Content Figure 5B and C, available at: http://links.lww.com/MS9/B74). Multicenter studies reported MD 0.35 (95% CI −2.30 to 2.99; P = 0.80; I² = 41%) (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74), while single-center trials reported MD 2.68 (95% CI −1.99 to 7.34; P = 0.26; I² = 88%) (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74).

Cardiac index

The cardiac index, a parameter of cardiac performance, was assessed by dividing the studies into four subgroups according to the time period. In the first subgroup, cardiac index after 6 hours (Julian Alvarez et al 2006) (0.2 μg/kg/min) showed insignificant variation (M.D: − 0.1; CI [−0.22, 0.02]; P = 0.11), While in both the second and third subgroups cardiac index after 12 hours, including Julian Alvarez et al (2006) (0.2 μg/kg/min), indicated a significant increase in cardiac index as shown by (M.D: 0.4; CI [0.28, 0.52]; P = < 0.00001) and (M.D: 0.41; CI [0.18, 0.64]; P = 0.004), respectively. Similarly, in the last subgroup of the longest time period, i.e., cardiac index after 48 hours including three RCTs demonstrated a statistically significant increase in cardiac index (M.D: 0.28; CI [0.01, 0.56]; P = 0.04). The overall pooled result showed a slight but meaningful improvement in the cardiac index parameter with high heterogeneity (M.D: 0.27; CI [0.07, 0.47]; P = 0.009) (I² = 90%; P = < 0.00001) (Supplemental Digital Content Figure 6A, available at: http://links.lww.com/MS9/B74).

Treatment benefits with levosimendan were observed in every subgroup during study design based subgrouping (MD 0.27 L/min/m², 95% CI 0.07–0.47; P = 0.009; I² = 90%), multicenter trials (MD 0.24, 95% CI 0.03–0.45; P = 0.02; I² = 0%), and single-center trials (MD 0.28, 95% CI 0.03–0.45) (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74), demonstrating the clinical significance of levosimendan. Dose-based analysis revealed similar consistent improvements, at 0.1 μg/kg/min (MD 0.17, 95% CI 0.02–0.32; P = 0.02; I² = 0%) and 0.2 μg/kg/min (MD 0.32, 95% CI 0.03–0.62; P = 0.03; I² = 95%) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74; Supplemental Digital Content Figure 6B and 6C, available at: http://links.lww.com/MS9/B74).

LVEF%

LVEF, an important parameter to evaluate left ventricle systolic function, was also assessed to compare the efficacy of levosimendan in relation to dobutamine, and a total of seven studies reported this outcome and demonstrated noteworthy improvement in LVEF% with insignificant amount of heterogeneity across studies (M.D: 2.8; CI [1.03, 4.57]; P = 0.002) (I² = 35%; P = 0.16) (Supplemental Digital Content Figure 7A, available at: http://links.lww.com/MS9/B74). Different dosing regimen of levosimendan at 0.1 μg/kg/min, MD 2.88 (95% CI 0.62–5.14; P = 0.01; I² = 46%), and at 0.1 mg/kg/min, MD 2.80 (95% CI 0.17–5.43; P = 0.04) revealed similarly nonsignificant results. (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74; Supplemental Digital Content Figure 7B, available at: http://links.lww.com/MS9/B74).

Cardiac output)

Cardiac output is an important measure of cardiac function was evaluated compare the efficacy of levosimendan in relation to dobutamine, it was reported by Folath et al (2002) (0.1 μg/kg/min); Markku et al (2000) (0.6 μg/kg/min); Markku et al (2000) (0.4 μg/kg/min); Markku et al (2000) (0.2 μg/kg/min); Markku et al (2000) (0.1 μg/kg/min); Markku et al (2000) (0.05 μg/kg/min), demonstrating nonsignificant findings in cardiac output parameter in levosimendan population when compared to dobutamine (M.D: 0.29; CI [−0.03, 0.61]; P = 0.08) with insignificant amount of heterogeneity (I² = 52%; P = 0.06) (Supplemental Digital Content Figure 8A, available at: http://links.lww.com/MS9/B74).

Dose-based analysis revealed no significant pooled difference in results (MD 0.29 L/min, 95% CI −0.03 to 0.61; P = 0.08). Stratification of our studies based on different doses revealed that higher doses demonstrated greater benefits at 0.4 μg/kg/min (MD 0.70, 95% CI 0.15–1.25; P = 0.01) and 0.6 μg/kg/min (MD 1.00, 95% CI 0.30–1.70; P = 0.005). However, lower doses showed nonsignificant effects (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74) (Supplemental Digital Content Figure 8B, available at: http://links.lww.com/MS9/B74).

SVR

As SVR is an important determinant for evaluating afterload, we stratified the studies into four subgroups. In the first cohort, SVR (dyne.s/cm5) after 6 h [Julian Alvarez et al 2006 (0.2 μg/kg/min)] showed inconclusive findings in the levosimendan population when compared to dobutamine (SMD: − 0.00; CI [0.61, − 0.61]; P = 1.00). Similarly, the second cohort, SVR (dyne.s/cm5) after 12 h, containing one study (Julian Alvarez et al 2006 0.2 μg/kg/min) showed insignificant findings (SMD: − 0.00; CI [0.61, − 0.61]; P = 1), while the third subgroup, SVR (dyne.s/cm5) after 24 hours [Folath et al 2002 (0.1 μg/kg/min); Claes Hakan et al 2010 (0.1 μg/kg/min); Julian Alvarez et al 2006 (0.2 μg/kg/min)] demonstrated a significant reduction in SVR in levosimendan population (SMD: − 0.35; CI [−0.58, − 0.12]; P = 0.002), and the last cohort of longest follow-up period, SVR after 48 hours, showed no impactful findings (SMD:-1.06;CI [−1.35, 1.02]; P = 0.32). The overall trends favored levosimendan but did not reach statistical significance (SMD: − 0.40; CI [−0.88, 0.08]; P = 0.1) with a high value of heterogeneity (I² = 83%; P = < 0.00001) (Supplemental Digital Content Figure 9A, available at: http://links.lww.com/MS9/B74).

SVR showed neutral effects with levosimendan (MD −0.40, 95% CI −0.88 to 0.08; P = 0.10; I² = 83%) on dose-based subgrouping and study design-based analysis. Dose-based analysis revealed a fall in SVR in multicenter trials (MD −0.91, 95% CI −1.85 to 0.03; P = 0.06; I² = 92%), whereas single-center studies demonstrated nonsignificant effects (MD 0.00, 95% CI −0.31 to 0.31; P = 1.00) (Supplemental Digital Content Table 5A and 5B, available at: http://links.lww.com/MS9/B74; Supplemental Digital Content Figure 9B and 9C, available at: http://links.lww.com/MS9/B74).

LVEDd

LVEDd, an important marker for estimating preload and filling status, was also evaluated to determine the efficacy of levosimendan compared to dobutamine. Three studies reported this outcome, indicating no major fluctuation (SMD: 0.13; CI [- 0.19, 0.45]; P = 0.42) with no evident heterogeneity across three RCTS (I² = 0%; P = 0.76) (Supplemental Digital Content Figure 10A, available at: http://links.lww.com/MS9/B74).

Subgroup analysis based on different dosages showed similar results (MD 0.13, 95% CI −0.19 to 0.45; P = 0.42; I² = 0%). At 0.1 mg/kg/min, (MD 0.2, 95% CI −0.19 to 0.59; P = 0.31; I² = 0%), while at 0.1 μg/kg/min, (MD 0.00, 95% CI −0.53 to 0.53; P = 1.00) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74) (Supplemental Digital Content Figure 10B, available at: http://links.lww.com/MS9/B74).

LVESd

LVESd reflects the systolic function of the ventricle and was reported by three RCTs, illustrating no notable difference between patients who received levosimendan and those who received dobutamine (SMD: − 0.09; CI [−0.40, 0.23]; P = 0.59) with insignificant heterogeneity (I² = 0%; P = 0.89) (Supplemental Digital Content Figure 11A, available at: http://links.lww.com/MS9/B74).

LVESd reflects the systolic function of the ventricle and was reported by three RCTs, which illustrated no notable difference between patients who received levosimendan and those who received dobutamine (MD 0.13, 95% CI −0.19 to 0.45; P = 0.42; I² = 0%).

Stratification of our analysis on dosage demonstrated no evident difference in results as compared to overall pooled analysis(MD −0.09, 95% CI −0.4 to 0.23; P = 0.59; I² = 0%), at 0.1 mg/kg/min (MD −0.14, 95% CI −0.53 to 0.26; P = 0.5; I² = 0%) and 0.1 μg/kg/min (MD 0.00, 95% CI −0.53 to 0.53; P = 1.00) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74) (Supplemental Digital Content Figure 11B, available at: http://links.lww.com/MS9/B74).

LA diameter

LA diameter signifies anterior posterior span of left atrium, and it was analyzed by pooling the data from Ender Oner et al (2014) (0.1 mg/kg/min); H. Duygu et al (2007) (0.1 μg/kg/min) and Hamza Duygu et al (2008) (0.1 μg/kg/min). This indicated comparable findings in both populations (SMD: − 0.24; CI [−0.53, 0.06]; P = 0.18) with no heterogeneity across the studies (I² = 0%; P = 0.89) (Supplemental Digital Content Figure 12A, available at: http://links.lww.com/MS9/B74).

Subgroup analysis based on dosage indicated no significant difference in LA size (MD −0.24, 95% CI −0.53 to 0.06; P = 0.12; I² = 0%). It further demonstrated comparable findings, at 0.1 mg/kg/min, MD −0.19 (95% CI −0.54 to 0.17; P = 0.30; I² = 0%), and at 0.1 μg/kg/min, MD −0.35 (95% CI −0.88 to 0.18; P = 0.20) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74) (Supplemental Digital Content Figure 12B, available at: http://links.lww.com/MS9/B74).

SPAP

SPAP is an important parameter for evaluation pulmonary hypertension, and it was reported by three studies namely, H. Duygu et al (2007) (0.1 μg/kg/min), Duygu et al (2008) (0.1 μg/kg/min), and Birhan et al (2009) (0.1 μg/kg/min) reported a pooled analysis of a remarkable decline in SPAP in levosimendan recipients in relation to dobutamine recipients (MD: − 6.86; CI [−8.85, − 4.88]; P = < 0.00001) with no heterogeneity across the three RCTs (I² = 0%; P = 0.94) (Supplemental Digital Content Figure 13, available at: http://links.lww.com/MS9/B74).

SvO₂%

SvO₂ reflects the balance between oxygen delivery and utilization by the body and was measured by forming three subgroups based on the follow-up period of the outcome. In the first subgroup, SvO₂% after 6 hours [Julian Alvarez et al 2006 (0.2 μg/kg/min)] reported no clinically relevant difference between the levosimendan and dobutamine population (MD: 0.70; CI [−2.07, 3.47]; P = 0.62), while the subgroup SvO₂% after 12 hours illustrated meaningful improvement in mixed venous oxygen saturation in levosimendan population (MD: 6.80; CI [4.63, 6.97]; P = < 0.00001). However, the last two subgroups of SvO₂ % after 24 hours and SvO₂% after 24 hours including two studies [Julian Alvarez et al 2006 (0.2 μg/kg/min) and Claes Hakan et al 2010 (0.1 μg/kg/min)] demonstrated inconclusive results (MD: − 3.17; CI [−4.53, 10.87]; P = 0.42) and (MD: 5.71; CI [−1.24, 12.66]; P = 0.11). The overall result demonstrated significant rise in SvO₂% in levosimendan arm when compared to dobutamine arm, indicating better efficacy (MD: 4.44; CI [1.66, 7.25]; P = 0.002) with severe heterogeneity (I² = 82%; P = < 0.0001) (Supplemental Digital Content Figure 14A, available at: http://links.lww.com/MS9/B74).

Results are measured forming subgroups based on two methods, dose based analysis and study design (Single center vs Multicenter), concluding the clinical significance of levosimendan in light of subgroups (MD 5.35%, 95% CI 2.87–7.83; P < 0.0001; I² = 68%) (Supplemental Digital Content Table 5A and 5B, available at: http://links.lww.com/MS9/B74). Single-center trials demonstrated a strong effect (MD 7.10, 95% CI 5.83–8.38; P < 0.00001), whereas multicenter trials fail to provide any evident results (MD 0.76, 95% CI −2.59 to 4.11; P = 0.66) (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74). Dose-based analysis revealed the pronounced effect of levosimendan at 0.2 μg/kg/min (MD 7.10, 95% CI 5.83–8.38; P < 0.00001) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74; Supplemental Digital Content Figure 14B and 14C, available at: http://links.lww.com/MS9/B74).

BNP

BNP (ng/ml) is secreted by ventricle when stretched due to pressure overload, and it was evaluated to compare the efficacy of levosimendan versus dobutamine in HF patients by forming four subgroups. The first cohort, BNP after 24 hours including two studies Alexandre Mebazaa et al (2007) (0.1 μg/kg/min) and Claes Hakan et al (2010) (0.1 μg/kg/min), demonstrated slight but significant reduction in levosimendan arm in comparison with dobutamine arm (SMD: − 0.18; CI [−0.28, − 0.07]; P = 0.0009. Similarly, the second subgroup, BNP after 48 hours having three RCTs Claes Hakan et al (2010) (0.1 μg/kg/min); Alexandre Mebazaa et al (2007) (0.1 μg/kg/min) and Avgeropoulou et al (2005) (0.1 μg/kg/min), indicated marginal decline in BNP levels (SMD: − 0.89; CI [−1.67, − 0.11]; P = 0.02). However, the last two subgroups, namely BNP (ng/ml) at day 5 including on two studies and BNP (ng/ml) after 1 month including one RCT, showed no meaningful variation in BNP levels in levosimendan population versus dobutamine population (SMD: − 1.76; CI [−4.79, 1.27]; P = 0.25) and (SMD: − 0.31; CI [−0.82, 0.02]; P = 0.24), respectively. The overall effect size illustrated a slight decline in BNP levels in levosimendan population in contrast to dobutamine population hinting toward its better efficacy (SMD: − 0.47; CI [−0.70,-0.25]; P = < 0.0001) with high amount of heterogeneity across all studies (I² = 87%; P = < 0.00001). Supplemental Digital Content Figure 15A, available at: http://links.lww.com/MS9/B74. The unit for this outcome is provided in Supplemental Digital Content Table 6, available at: http://links.lww.com/MS9/B74.

Study design stratification showed a prominent difference between single-center trials (SMD −2.83, 95% CI −3.76 to—1.91; P < 0.00001; I² = 31%) and multicenter trials (SMD −0.29, 95% CI −0.39 to 0.19; P = 0.11; I² = 44%) of significantly reduced BNP levels in all subgroups (SMD −0.47, 95% CI −0.70 to—0.25; P < 0.00001; I² = 87%) (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74; (Supplemental Digital Content Figure 15B, available at: http://links.lww.com/MS9/B74).

IL-6

IL-6, an inflammatory cytokine that usually elevates at the time of myocardial damage, was assessed stratifying the studies into three cohorts according to the follow-up period of the outcome. In the first cohort, IL-6 (pg/ml) after 24 hours [Adamopoulos et al 2006 (0.1 μg/kg/min)] demonstrated no clinical benefit of levosimendan when compared to dobutamine (MD: − 1.50; CI [−5.94, 2.94]; P = 0.51). While, the second cohort IL-6 (pg/ml) after 48 hours including two RCTs [Adamopoulos et al 2006 (0.1 μg/kg/min) and Avgeropoulou et al 2005 (0.1 μg/kg/min)] and third cohort IL-6 (pg/ml) at day 5 [Avgeropoulou et al 2005 (0.1 μg/kg/min)] illustrated a substantial decrease in IL-6 levels in levosimendan arm in contrast to dobutamine arm as indicated by (MD: − 4.57; CI [−5.99, − 3.15]; P < 0.00001) and (MD: − 8.30; CI [−9.86, − 6.74]; P = < 0.00001), respectively. The overall pooled result showed a striking decline in the inflammatory marker in levosimendan-received patients in relation to dobutamine recipients (MD: − 4.93; CI [−7.74, − 2.11]; P = 0.0006) with substantial amount of heterogeneity across the studies (I² = 82%; P = 0.0008) (Supplemental Digital Content Figure 16, available at: http://links.lww.com/MS9/B74). The unit for this outcome is provided in Supplemental Digital Content Table 6, available at: http://links.lww.com/MS9/B74.

TNF-α

TNF-α is a pro-inflammatory marker released at the time of myocardial damage and remodeling was analyzed, making three subgroups: TNF-α (pg/ml) after 24 hours, 48 hours and at day 5. In the first subgroup [Adamopoulos et al 2006 (0.1 μg/kg/min)], no notable difference was found in TNF-α levels in both levosimendan and dobutamine populations (MD: − 1.40; CI [−5.26, 2.46]; P = 0.48). Likewise, the last two subgroups, TNF-α after 48 hours [Adamopoulos et al 2006 (0.1 μg/kg/min) and Avgeropoulou et al 2005 (0.1 μg/kg/min)] and TNF-α at day 5 including one study demonstrated comparable results in both levosimendan and dobutamine population as illustrated by (MD: 3.17; CI [−8.48, 14.81]; P = 0.59) and (MD: − 5.50; CI [−11.26, 0.26]; P = 0.06), respectively. The overall effect size displayed negligible difference in between levosimendan and dobutamine populations (MD: − 0.28; CI [−5.27, 4.72]; P = 0.91) with a high value of heterogeneity across all the studies (I² = 79%; P = 0.002) (Supplemental Digital Content Figure 17, available at: http://links.lww.com/MS9/B74). The unit for this outcome is provided in Supplemental Digital Content Table 6, available at: http://links.lww.com/MS9/B74.

Adverse events

Hemoglobin

This outcome was reported by two studies [Folath et al 2002(0.1 μg/kg/min) and Osman Can et al 2010 (0.1 μg/kg/min)]. Levosimendan was associated with a higher incidence of hypotension than dobutamine, although it failed to reach statistical significance (SMD: − 0.33; CI [−0.58, − 0.08]; P = 0.010) with an insignificant amount of heterogeneity (I² = 1%; P = 0.31) (Supplemental Digital Content Figure 18, available at: http://links.lww.com/MS9/B74). This pooled analysis reflects the variability in the vasodilatory mechanism of levosimendan across studies.

Creatinine level

Creatinine level is an important indicator of renal function, and it was reported by four studies [Folath et al 2002(0.1 μg/kg/min), Osman Can et al 2010 (0.1 μg/kg/min), Mehmet Birhan et al 2009 (0.1 μg/kg/min), and Ender Oner et al 2014 (0.1 mg/kg/min)]. The overall effect size showed no clinically relevant difference in creatinine levels (MD 0.15, 95% CI −0.60 to 0.89; P = 0.70; I² = 89%) (Supplemental Digital Content Figure 19A, available at: http://links.lww.com/MS9/B74), suggesting that levosimendan does not have harmful effects on renal function.

Subgroup analysis according to study design demonstrated a nonsignificant effect in single-center studies (MD −0.47, 95% CI −1.00 to 0.07; P = 0.09), whereas multicenter studies reported (MD 0.36, 95% CI −0.38 to 1.10; P = 0.34; I² = 86%) (Supplemental Digital Content Table 5A, available at: http://links.lww.com/MS9/B74; Supplemental Digital Content Figure 19B, available at: http://links.lww.com/MS9/B74).

Hypotension

Hypotension was reported by Claes Hakan et al (2010) (0.1 μg/kg/min), Markku et al (2000) (0.6 μg/kg/min), Markku et al (2000) (0.4 μg/kg/min), Markku et al (2000) (0.2 μg/kg/min), Markku et al (2000) (0.1 μg/kg/min), Folath et al (2002) (0.1 μg/kg/min), and Alexandre Mebazaa et al (2007) (0.1 μg/kg/min). The overall result demonstrated insignificant findings, as evident by (R.R: 1.50; CI [0.95, 2.37]; P = 0.08) with negligible heterogeneity across studies (I² = 13%; P = 0.33) (Supplemental Digital Content Figure 20A, available at: http://links.lww.com/MS9/B74).

In the dose-stratified analysis, the 0.1 μg/kg/min subgroup showed a higher but nonsignificant risk (RR 1.71, 95% CI 0.81–3.62; P = 0.16; I² = 47%) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74; Supplemental Digital Content Figure 20B, available at: http://links.lww.com/MS9/B74)Digital Content Figure 20B, available at: http://links.lww.com/MS9/B74). The observed hypotension was typically mild, manageable, and did not necessitate discontinuation of treatment, suggesting limited clinical relevance in AHF, where hemodynamic support remains the priority.

Cardiac failure

The risk of cardiac failure in the levosimedan population was evaluated by pooling the results from two RCTs: Hakan et al (2010) (0.1 μg/kg/min) and Mebazaa et al (2007) (0.1 μg/kg/min). The overall analysis illustrated a substantial decline in events of cardiac failure in the levosimendan arm compared to dobutamine (R.R: 0.73; CI [0.56, 0.94]; P = 0.02) with no heterogeneity observed across the two studies (I² = 0%; P = 0.89) (Supplemental Digital Content Figure 21, available at: http://links.lww.com/MS9/B74).

Nausea

Nausea as an adverse event was reported by three studies: Hakan et al (2010) (0.1 μg/kg/min), Mebazaa et al (2007) (0.1 μg/kg/min), and Markku et al (2000). The overall effect size indicated comparable findings in both levosimendan and dobutamine populations (R.R: 1.40; CI [0.44, 4.50]; P = 0.57) with an insignificant amount of heterogeneity observed (I² = 29%; P = 0.25) (Supplemental Digital Content Figure 22, available at: http://links.lww.com/MS9/B74). The association of levosimendan with a high incidence of nausea in a few trials is still unlikely to change the choice of inotrope or result in discontinuation of levosimendan.

Headache

Headache was reported by three RCTs: Claes Hakan et al (2010) (0.1 μg/kg/min), Folath et al (2002) (0.1 μg/kg/min), and Markku et al (2000). The combined analysis showed an increased risk of headache in the levosimendan arm compared to dobutamine (R.R: 2.76; CI [1.17, 6.49]; P = 0.02), while no heterogeneity was found across the studies (I² = 0%; P = 0.99) (Supplemental Digital Content Figure 23, available at: http://links.lww.com/MS9/B74).

Hypokalemia

Hypokalemia was reported by three RCTs (Hakan et al (2010) (0.1 μg/kg/min), Mebazaa et al (2007) (0.1 μg/kg/min), and Markku et al (2000) (0.2 μg/kg/min]). The overall effect size demonstrated a rise in the event of hypokalemia when compared to dobutamine, with no heterogeneity found across studies (R.R: 1.56; CI [1.07, 2.28]; P = 0.02) (I² = 0%; P = 0.61) (Supplemental Digital Content Figure 24A, available at: http://links.lww.com/MS9/B74). Dose-based analysis was consistent across infusion regimens, with the 0.1 μg/kg/min group demonstrating (RR 1.54 95% CI 1.06–2.25; P = 0.03) and the 0.2 μg/kg/min group demonstrating (RR 3.15, 95% CI 0.14–72.88; P = 0.47) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74; Supplemental Digital Content Figure 25B, available at: http://links.lww.com/MS9/B74). Despite statistical significance, most reported hypokalemia events were mild and correctable with supplementation. This adverse effect is not expected to alter the choice of inotropes in life threatening AHF but requires careful electrolyte monitoring.

Insomnia

This outcome was reported by two studies, and their combined results indicated a slight but insignificant rise in events of insomnia in levosimendan-receiving patients in relation to dobutamine-receiving patients (R.R: 1.29; CI [0.82, 2.20]; P = 0.27) with no heterogeneity being observed among the studies (I² = 0%; 0.88) (Supplemental Digital Content Figure 25, available at: http://links.lww.com/MS9/B74).

Atrial fibrillation

Atrial fibrillation was reported in two studies (Folath et al (2002) (0.1 μg/kg/min) and Mebazaa et al (2007) (0.1 μg/kg/min]). The overall analysis showed that levosimendan was associated with an increased risk of atrial fibrillation compared with dobutamine (R.R: 1.51; CI [1.03, 2.21]; P = 0.03) (I² = 0%; P = 0.83) (Supplemental Digital Content Figure 26, available at: http://links.lww.com/MS9/B74).

Any adverse event

This outcome was assessed to compare the safety of levosimendan and dobutamine. Total of three studies reported this outcome [Markku et al 2000; Alexandre Mebazaa et al 2007 (0.1 μg/kg/min); and Claes Hakan et al 2010 (0.1 μg/kg/min)], and their combined analysis showed no difference in risk of an outcome in both levosimendan and dobutamine population (R.R: 1.03; CI [0.98, 1.09]; P = 0.27) with no heterogeneity present across the studies (I² = 0%;P = 0.69) (Supplemental Digital Content Figure 27, available at: http://links.lww.com/MS9/B74).

Tachycardia

Tachycardia was assessed to evaluate the safety of levosimendan arm reported by Folath et al (2002) (0.1 μg/kg/min), Alexandre Mebazaa et al (2007) (0.1 μg/kg/min), Markku et al (2000) (0.6 μg/kg/min), Markku et al (2000) (0.4 μg/kg/min), Markku et al (2000) (0.2 μg/kg/min), Markku et al (2000) (0.1 μg/kg/min), and Markku et al (2000) (0.05 μg/kg/min). The overall effect size showed no notable difference across both arms (R.R:0.85; CI [0.59, 1.21]; P = 0.36), with no heterogeneity observed (I² = 0%; P = 0.79) (Supplemental Digital Content Figure 28A, available at: http://links.lww.com/MS9/B74).

Dose-based subgroup analysis revealed nonsignificant trends (RR 0.85, 95% CI 0.59–1.21; P = 0.36; I² = 0%) (Supplemental Digital Content Table 5B, available at: http://links.lww.com/MS9/B74); Supplemental Digital Content Figure 28B, available at: http://links.lww.com/MS9/B74). The absence of tachycardia is consistent with the calcium sensitizing mechanism of levosimendan, which increases contractility without increasing adrenergic stimulation.

Bradycardia

Bradycardia was reported by two studies Folath et al (2002) (0.1 μg/kg/min) and Alexandre Mebazaa et al (2007) (0.1 μg/kg/min); by pooling them together, the overall result did not reach statistical significance (R.R:0.50; CI [0.23, 1.11]; P = 0.09). There was no heterogeneity found across the studies (I² = 0%; P = 0.69) (Supplemental Digital Content Figure 29, available at: http://links.lww.com/MS9/B74).

Dizziness

Dizziness and light-headedness were reported by two RCTs Folath et al (2002 (0.1 μg/kg/min) and Mebazaa et al (2007) (0.1 μg/kg/min). The overall pooled analysis demonstrated an equal risk of events across the levosimendan and dobutamine arms (R.R:1.13; CI [0.59, 2.14]; P = 0.72) with no evidence of heterogeneity (I² = 0%; P = 0.43) (Supplemental Digital Content Figure 30, available at: http://links.lww.com/MS9/B74).

Cough

Cough was another safety outcome reported by Claes Hakan et al (2010) (0.1 μg/kg/min) and Alexandre Mebazaa et al (2007) (0.1 μg/kg/min) their combined analysis indicated no relevant difference in risk of outcome in both the groups with no heterogeneity observed (R.R:0.88; CI [0.48, 1.58]; P = 0.66) (I² = 0%; P = 0.67) (Supplemental Digital Content Figure 31, available at: http://links.lww.com/MS9/B74).

Ventricular tachycardia

Ventricular tachycardia, another safety outcome, was assessed by pooling the Folath et al (2002) (0.1 μg/kg/min) and Mebazaa et al (2007) (0.1 μg/kg/min) studies, indicating no significant difference between the levosimendan and dobutamine arms with complete homogeneity across the studies (R.R:1.58; CI [0.98, 2.57]; P = 0.06) (I² = 0%; P = 0.32) (Supplemental Digital Content Figure 32, available at: http://links.lww.com/MS9/B74).

Angina pectoris, myocardial ischemia, or chest pain

Angina pectoris, myocardial ischemia, or chest pain were assessed to compare the safety of the levosimendan and dobutamine arms. In a pooled analysis by Folath et al (2002) (0.1 μg/kg/min) and Mebazaa et al (2007) (0.1 μg/kg/min), levosimendan was associated with a reduced but insignificant risk compared with dobutamine (R.R, 0.32; CI [0.04, 2.86]; P = 0.31) with moderate heterogeneity (I² = 63%; P = 0.10) (Supplemental Digital Content Figure 33, available at: http://links.lww.com/MS9/B74).

Ventricular fibrillation

Ventricular fibrillation, a life-threatening adverse event, was reported by Folath et al (2002) (0.1 μg/kg/min) and Mebazaa et al (2007) (0.1 μg/kg/min). The overall pooled analysis demonstrated no noteworthy difference in the risk of ventricular fibrillation in the levosimendan population compared to the dobutamine population, with no evidence of heterogeneity (R.R: 0.80; CI [0.42, 1.53]; P = 0.50) (I² = 0%; P = 0.89) (Supplemental Digital Content Figure 34, available at: http://links.lww.com/MS9/B74).

Urinary tract infections

Urinary tract infections, another potential adverse event, were reported in two RCTS, namely Claes Hakan et al (2010) (0.1 μg/kg/min) and Alexandre Mebazaa et al (2007) (0.1 μg/kg/min). The overall effect size indicated clinical equipoise in the levosimendan and dobutamine arms, with a statistically insignificant low amount of heterogeneity observed (R.R: 1.01; CI [0.28, 3.66]; P = 0.98) (I² = 42%; P = 0.49) (Supplemental Digital Content Figure 35, available at: http://links.lww.com/MS9/B74).

Sensitivity analysis

Sensitivity analysis was performed by stratifying studies based on outcomes according to levosimendan dosage and grouping studies in accordance with the study designs. Multicenter versus single-center studies for PCWP, stroke volume, SVR, BNP, cardiac index, SvO₂, HR creatinine and SBP, DBP, HR, PCWP, cardiac index, SvO₂,SVR, stroke volume, cardiac output, creatinine, LVEF%, LVEDd, LVESd, LA diameter, hypokalemia, hypotension and tachycardia respectively. The detailed results of the analysis are provided in Supplemental Digital Content Table 5A–5B, available at: http://links.lww.com/MS9/B74.

Meta-regression

Meta regression was performed to assess the influence of baseline parameters and baseline medications on our primary outcome mortality. We evaluated the impact of hypertension, mean age, DBP,SBP, diuretics, β-blockers, HR, LVEF%, DM, and ACEis on mortality within 30 days. A negative association was observed with LVEF (coefficient: − 0.2352; P = 0.000068) and hypertension (coefficient: − 0.0837; P = 0.054), and a positive association was observed with SBP (coefficient: 0.2714; P = 0.0006). While no meaningful impact was demonstrated by mean age, DBP, diuretics, β-blockers, HR, ACEis, and DM as illustrated by (coefficient: − 0.0885; P = 0.21) (coefficient: − 0.3458; P = 0.274), (coefficient: − 0.0143; P = 0.57), (coefficient: 0.0209; P = 0.325), (coefficient:0.0248; P = 0.71), (coefficient: − 0.0183; P = 0.33), and (coefficient: − 0.0961; P = 0.109), respectively (Supplemental Digital Content Figure 36A-36 J, available at: http://links.lww.com/MS9/B74). Additional details can be visualized by Supplemental Digital Content Table 7, available at: http://links.lww.com/MS9/B74.

Risk of bias and quality assessment

The risk of bias was assessed using the Cochrane Risk of Bias tool for all RCTs included in the meta-analysis. High-quality studies, such as those by Claes Hakan et al (2010), Markku et al (2000), Folath et al (2002), and Alexandre Mebazaa et al (2007), were judged to have a low risk of bias across all domains, reflecting high methodological quality, including random sequence generation, allocation concealment, blinding of participants and outcome assessors, handling of incomplete outcome data, and selective reporting. Conversely, studies such as Avgeropoulou et al (2005), Julian Alvarez et al (2006), Ender Oner et al (2014), Adampoulos et al (2006), and Osman Can et al (2010) were rated as high risk as they exhibited methodological weaknesses – particularly in binding of participants and personnel. In addition, several studies were assessed as having unclear risk in domains such as random sequence generation and allocation concealment, due to insufficient methodological detail. These include Avgeropoulou et al (2005), Julian Alvarez et al (2006), Ender Oner et al (2014), Adampoulos et al (2006), Osman Can et al (2010), Hamza Duygu et al (2008), H. Duygu et al (2007), Mehmet Birhan et al (2007), Mehmet Birhan et al (2009), Sima Fard et al (2007), and Alison P et al (2004) (Fig. 3A,B) (Supplemental Digital Content Table 2, available at: http://links.lww.com/MS9/B74). Despite variability in study quality, the majority of trials exhibited low risk of bias, especially in studies assessing objective outcomes. Additionally, the cohort study by Marta Madeira et al (2017) was evaluated using the Newcastle-Ottawa Scale and achieved the highest possible score of 9/9. It met all standards in the selection, comparability, and outcome categories highlighting its strong methodological design. Therefore, although some studies had high or unclear bias risks, the inclusion of multiple well conducted trials strengthens the reliability and validity of the overall findings Supplemental Digital Content Table 2, available at: http://links.lww.com/MS9/B74.

Figure 3.

(A) Risk of bias graph. (B) Risk of bias summary.

Discussion

A total of 16 studies (15 RCTs and one retrospective cohort study) were included in this meta-analysis comparing levosimendan and dobutamine for the treatment of AHF. In particular, several statistically significant results were noted in the hemodynamic and echocardiographic domains. A favorable trend toward levosimendan was observed in short-term (30-day) mortality; RRs ranged from 0.65 to 0.85, and our studies showed statistically significant results (P < 0.05), suggesting a possible survival benefit during the acute phase. However, due to study heterogeneity, long-term mortality (115–365 days) had inconsistent statistical significance and mixed results. Significant increases in cardiac output (0.5–1.0 L/min) and stroke volume (~10–15 ml) were among the secondary outcomes that favored levosimendan, whereas PCWP and SVR decreased. Improvements in LVEF and cardiac index were statistically significant, especially in patients with a low baseline ejection fraction. The reduction of BNP (~600 pg/ml vs 400 pg/ml) and inflammatory markers (IL-6, TNF-α) with levosimendan were notable biomarker changes, indicating both anti-inflammatory and cardiac benefits. The beneficial cardiac effects of levosimendan were further supported by modest echocardiographic improvements (LVEDd and LVESd) and reductions in SPAP, whereas renal function (creatinine) and LA diameter did not consistently or significantly differ. Currently, both the ESC and AHA/ACC guidelines recommend traditional inotropes, such as dobutamine or milrinone, for acute decompensated HF, while levosimendan is not endorsed for routine use. Current recommendations state that levosimendan was created to treat AHF and other cardiac disorders using an indicator that is deemed suitable^[35]. Unlike other inotropic medications used to treat HF, levosimendan does not increase intracellular calcium levels. Notably, hemodynamic parameters are improved by levosimendan. Recent research has shown that levosimendan may result in dose-dependent reductions in mean arterial pressure, PCWP, right atrial pressure, and pulmonary arterial pressure, along with an increase in the cardiac index^[30,36]. In addition to improving hemodynamics, levosimendan directly affects renal circulation by inducing renal arterial vasodilation and increasing renal blood flow without affecting the balance of oxygen supply and demand in the kidneys^[37]. Our findings require confirmatory evidence from further guideline directed multicenter trials incorporating levosimendan.

We found that levosimendan decreased the BNP levels. BNP levels have been shown to significantly decrease at the conclusion of a 24 hour infusion in several studies^[38,39]. Additionally, we found a statistically significant difference in the reduction of BNP 24 hour after the start of the levosimendan infusion. Our research demonstrated that the effects of levosimendan were maintained for at least 48 h, despite the fact that BNP levels were measured at several intervals (24 h, 48 h, 1 month, and 5 months). This may be because invasive catheters are usually only used for 2–3 days. Because the active metabolites of Levosimendan have a long elimination half life, their prolonged hemodynamic action is probably what causes the sustained BNP reduction after the 24-hour infusion^[40]. The opening of ATP dependent K channels in vascular smooth muscle is the main mechanism by which it produces vasodilatory effects^[41]. Both arterial and venous smooth muscle cells are relaxed by levosimendan^[42], which lowers preload and afterload and raises cardiac output (CO) and stroke volume^[43]. Levosimendan is less proarrhythmic and does not increase myocardial oxygen consumption, unlike other common inotropes such as dobutamine or milrinone^[44]. It can also reverse myocardial stunning and has antiarrhythmic effects^[30]. A 24-hour infusion of levosimendan is safe and effective for the short-term treatment of right HF in patients with a variety of cardiac and pulmonary diseases^[7], which included 10 trials. In particular, the study found that the tricuspid annular plane systolic excursion and right ventricular ejection fraction significantly improved, whereas the systolic pulmonary artery pressure and pulmonary vascular resistance decreased. However, the mean pulmonary pressure did not change significantly. When β-blockade is believed to contribute to hypoperfusion, levosimendan may be considered in AHF without severe hypotension, as per the 2012 ESC guidelines on HF (Class IIb, Level of Evidence C)^[45]. According to a different meta-analysis, levosimendan increased survival compared to dobutamine in critically ill patients who required inotropic support^[46]. Patients with ischemic HF, those who had undergone heart surgery, and those taking long-term β-blockers were among the subgroups in which this benefit was most noticeable. Therefore, in patients with acutely decompensated HF receiving chronic β-blockade, levosimendan may be a good substitute for dobutamine. Notably, subgroup evidence, including patients with markedly reduced ejection fraction, ischemic cardiomyopathy, or those on chronic β-blocker therapy, emphasizes the mortality and hemodynamic benefits of levosimendan.

Our study provides important new information on the adverse event profiles of dobutamine and levosimendan in AHF. The results show that levosimendan is associated with a markedly increased risk of hypotension, which has been linked to the activation of ATP channels that are dependent on potassium and a reduction in calcium sensitivity^[47]. The calcium-sensitizing effect of levosimendan enhances myocardial contractility without any significant increase in intracellular calcium levels or adrenergic stimulation^[48]. This side effect is clinically significant, especially in patients with hemodynamic instability or hypotension. However, due to its βadrenergic agonist effects, which increase HR and myocardial oxygen consumption, dobutamine is associated with an increased risk of tachycardia, cardiac failure, and myocardial ischemic events^[49]. Despite this, it still has a significant clinical advantage over βagonists, such as dobutamine, due to its lack of arrhythmogenic qualities. Levosimendan is associated with improved hemodynamics and decreased catecholamine use^[50]. However, there were no statistically significant differences between the two medications in adverse events such as atrial fibrillation, ventricular arrhythmias (ventricular tachycardia and fibrillation), electrolyte imbalances (such as hypokalemia), or neuropsychiatric symptoms (such as insomnia, headache, and dizziness). Levosimendan has been associated with a slightly higher incidence of minor adverse events, such as nausea, insomnia, cough, dizziness, and headache; however, these manageable transient adverse events are unable to decrease the potential impact and discontinuity of interventional drugs. Few complications, such as an increased incidence of urinary tract infections (UTI) associated with levosimendan, are not well established in the literature. Although levosimendan’s pharmacodynamic profile has no known mechanistic link with genitourinary infections. Attributable factors associated with the increased incidence of UTIs in patients treated with levosimendan seem to be hospitalization related factors and baseline patient vulnerability. This implies that levosimendan and dobutamine have similar safety profiles in these domains. Pooled analysis of levosimendan welcomes the consideration of multimodal inotrope strategies in critical care for sepsis management^[51]. In individuals with septic shock, clinicians use a combination of drugs, such as vasodilators, at lower doses to maximize the potential therapeutic efficacy of levosimendan while minimizing the side effects of interventional drugs^[52]. Using low dose levosimendan with smaller doses of dobutamine may provide additive benefits to the therapeutic profile of patients^[53]. Future RCTs are required to elaborate on the clinical significance of inotrope combinations compared to high-dose monotherapy with inotropes in high-risk AHF patients.

Our analysis demonstrated substantial statistical heterogeneity (I² > 75%), particularly for BNP, SvO₂, and HR. Factors leading to the variability of pooled estimates from our study include differences in dosing regimens, timing of endpoint measurement, and heterogeneity in baseline severity of illness. However, consistently reported neutral outcomes from large multicenter RCTs, such as SURVIVE, CHEETAH, LEVO-CTS, and LeoPARDS, contrast with our findings. This discrepancy between our pooled estimates and these rigorously conducted, multicenter, and blinded trials with broader inclusion criteria may be explained by patient selection differences, methodological rigor, different dosing regimens, and unorthodox trial size. In contrast, smaller single-center studies demonstrated greater mortality and hemodynamic benefits, raising the possibility that bias and center effects inflated treatment effects in less rigorous designs^[54,55] .Thus the pooled estimates from our meta-analysis and the weight of evidence from multicenter RCTs emphasize the need for cautious interpretation and large-scale validation in future research^[56,57]. Notably, subgroup evidence including patients with markedly reduced ejection fraction, ischemic cardiomyopathy, or those on chronic β-blocker therapy emphasizes the mortality and hemodynamic benefits of levosimendan.^[58]

Taken together, our findings indicate that levosimendan may hold a clinically meaningful priority over dobutamine when patients present with comparable baseline characteristics. Because levosimendan exerts its inotropic effect through calcium sensitization rather than adrenergic stimulation^[34], it maintains contractile improvement without increasing myocardial oxygen demand or precipitating tachyarrhythmias. This mechanism provides a relative advantage in patients receiving chronic β-blockade or those with markedly reduced ejection fraction^[32], where adrenergic responsiveness is often blunted. The consistent improvements we observed in cardiac output, stroke volume, filling pressures, and BNP levels further support a hemodynamic profile that remains favorable beyond the infusion period, likely due to the prolonged activity of the active metabolites. When interpreted within the context of guideline recommendations and prior largescale trials, these results suggest that, under equivalent baseline conditions, levosimendan may offer a more sustained and physiologically balanced inotropic effect than dobutamine. Nonetheless, larger multicenter, guideline directed trials targeting these enriched patient subgroups are required to validate this potential therapeutic strategy.

Limitation

This meta-analysis has various limitations. First, the statistical power of some results may be diminished by the small scale of many included studies. Second, there is a chance of linguistic bias because most published articles are in English. Although there was no formal language restriction in the search strategy of our manuscript, all eligible full-text articles were published in English, which may have introduced some language bias. Third, heterogeneity was caused by variations in study design, specifically in the follow up periods, timing of outcomes, and dosing strategies, especially for long term outcomes. Fourth, internal validity may have been affected by the lack of strict methodological procedures, such as blinding, randomization, or allocation concealment. Additionally, we were unable to examine the interactions between subgroups or account for individual level confounders because of the use of aggregate data rather than patient level data. One interesting point is that we took the CASINO study’s mortality result data out of the meta-analysis; although complete baseline data were not accessible, the outcome data were published. Meta-regression on mortality subgroups (within 180 and 365 days) could not be performed due to insufficient baseline data and few eligible studies, and subgrouping by study risk was not possible because most trials reported unclear risk profiles. Subgroup analyses by levosimendan dosage and study design (multicenter vs single center) were only feasible where data were available. Moreover, data on concomitant β-blocker use with levosimendan were not consistently reported across studies, limiting our ability to conduct subgroup analyses on this variable. Potential collinearity among moderators, such as SBP, DBP, hypertension, LVEF%, HR, and use of antihypertensives, could not be formally tested due to incomplete reporting; however, we reduced this risk by using univariable models and avoiding highly correlated variables.

Conclusion

This meta-analysis provides up to date evidence comparing the effects of levosimendan and dobutamine in AHF. Under comparable baseline conditions, levosimendan demonstrated a potential clinical advantage, particularly in high-risk patients with ischemic heart disease, reduced ejection fraction, or chronic β-blocker therapy. It improved short-term mortality and hemodynamic and echocardiographic parameters, while showing a favorable safety profile, with hypotension being the main notable adverse effect. These findings suggest that levosimendan may be a better tolerated and hemodynamically effective alternative to dobutamine, although larger, targeted multicenter trials are needed to confirm these results.

Ethical approval

As this study was based solely on publicly accessible, it did not involve human subjects directly and thus did not require institutional review board approval or informed consent.

Consent

No individual-level or personally identifiable information is included, rendering publication consent irrelevant.

Sources of funding

This work was independently conducted without financial assistance from governmental bodies, commercial sponsors, or nonprofit institutions.

Author contributions

A.R.: Conceptualization, Project development, Data collection, Manuscript writing. L.K.: Project development, Data collection, Manuscript writing. M.K.: Project development, Data collection, Manuscript writing. L.Kh.: Project development, Data collection, Manuscript writing. S.M.E.A.: Project development, Data collection, Manuscript writing. N.K.: Project development, Data collection, Manuscript writing. H.R.: Table, Manuscript writing. I.M.: Data analysis, Manuscript writing. S.W.: Data analysis, Manuscript writing. M.P.: Figures, Data analysis, Manuscript writing. S.G.: Figures, Manuscript writing. T.S.: Supervision, Manuscript writing and editing. M.S.: Figures, Manuscript writing. L.S.: Manuscript writing and editing. T.: Supervision, Manuscript writing and editing. A.B.S.: Supervision, Manuscript writing and editing. B.D.A.: Supervision, Manuscript writing and editing. A.C.: Supervision, Manuscript writing and editing.

Conflict of interest disclosure

The authors declare the absence of any financial, personal, or academic conflicts that might have influenced the conduct or outcomes of this study.

Guarantor

Aayush Chaulagain.

Research registration unique identifying number (UIN)

It is registered on PROSPERO database under registration number CRD420251117829.

Status: review completed.

URL: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251117829

Provenance and peer review

This manuscript was not commissioned.

Data availability statement

The dataset supporting the conclusions of this article are included in this article.