WEANILEVO randomized 80 adults ready for VA-ECMO weaning to levosimendan or placebo. Weaning failure occurred in 25% of both groups. Levosimendan was associated with more acute kidney injury and longer ICU stays, while mortality differences were not statistically significant. Recruitment stopped at 80 of 210 planned patients after funding withdrawal, leaving substantial uncertainty. Exploratory Bayesian mortality signals require confirmation and do not establish a survival benefit.
Why the timing of VA-ECMO weaning matters
Veno-arterial extracorporeal membrane oxygenation (VA-ECMO) supports patients with severe circulatory failure while clinicians address the underlying cause and assess recovery. Removing that support creates a demanding transition: the native heart must sustain circulation, the right ventricle must tolerate its workload, and adequate perfusion must reach vulnerable organs. A drug that improves contractility could theoretically make this transition easier, but vasodilation and increased requirements for blood pressure support may complicate the same period.
Levosimendan is a calcium-sensitizing inodilator. WEANILEVO examined whether a planned infusion before liberation could reduce weaning failure. Its timing is central to interpreting the results. This was a trial in patients already considered ready to come off VA-ECMO, rather than a test of treatment at cannulation or across every stage of refractory cardiogenic shock. The findings therefore address a specific clinical decision in a selected recovering population.
A randomized trial in patients ready for liberation
Investigators conducted a double-blind, randomized, placebo-controlled superiority trial at five French centers. Eighty adults were assigned equally to levosimendan or placebo. Medical cardiogenic shock accounted for approximately 44% of cases and postcardiotomy indications for approximately 40%. Patients had received a median of four days of ECMO support.
Eligibility required evidence of recovery, including a vasoactive-inotropic score below 10, left ventricular ejection fraction above 20%, subaortic velocity-time integral above 10 cm, and lactate below 2 mmol/L. Right ventricular and oxygenation criteria were also specified, and clinicians expected liberation within 48 hours. Allocation used minimization incorporating center, age, and medical versus surgical indication. These criteria helped standardize enrollment, but they should be understood as trial selection criteria rather than a universal prescription for ECMO removal.
Treatment and the primary endpoint
The intervention was a fixed levosimendan infusion of 0.2 micrograms/kg/min for 24 hours, without a loading bolus or dose adjustment, approximately 24 hours before planned weaning. The placebo was color matched using a riboflavin-containing multivitamin preparation. Preparation was performed by an unblinded individual while patients, treating staff, and investigators remained blinded.
The primary composite outcome counted failure to remove ECMO within 48 hours, a subsequent need for another circulatory assist device, or death during the seven days after liberation. Additional support included VA-ECMO, an Impella device, or an intra-aortic balloon pump. Secondary outcomes included kidney injury, lengths of stay, vasoactive support, and mortality at 30 days and one year. This combination assessed whether initial removal translated into a sustained short-term transition.
No demonstrated improvement in weaning success
Weaning failure occurred in 10 of 40 patients in each group, or 25% with levosimendan and 25% with placebo. The odds ratio was 1.00, with a 95% confidence interval of 0.36–2.75 and a P value of 1.00. Adjusted and per-protocol analyses remained neutral. A sensitivity analysis addressing logistical delays likewise did not demonstrate superiority.
Equal observed event rates do not establish equivalent treatments. WEANILEVO was designed to test superiority, and its confidence interval was wide enough to include clinically meaningful benefit or harm. The accurate conclusion is that this trial did not demonstrate improved weaning success with this levosimendan regimen. It cannot settle whether a smaller treatment effect exists or whether other populations, dosing strategies, or timing might yield different results.
Early termination limits certainty
The planned sample size was 210 patients, based on an anticipated failure rate of 50% with placebo and 30% with levosimendan. Enrollment ended after 80 patients, approximately 38% of the target, because sponsor funding was withdrawn. This was not a stopping decision based on a data safety monitoring recommendation of efficacy or futility.
The observed failure rate was also lower than the placebo rate assumed in planning. Both fewer participants and fewer events reduced the trial’s ability to identify a treatment difference. Selection of patients already meeting recovery criteria may help explain the relatively high success rate, although that explanation does not prove a causal relationship. These limitations are especially relevant when applying the findings to less recovered patients or earlier phases of ECMO support.
Kidney injury and resource use warrant attention
Acute kidney injury by day 30 occurred in 25 of 40 patients receiving levosimendan (62.5%) and 15 of 39 receiving placebo (38.5%). The odds ratio was 2.67, with a 95% confidence interval of 1.08–6.62 and a P value of 0.03. A sensitivity analysis excluding deaths retained an association. Median ICU stay was 12 days with levosimendan versus six days with placebo (P = 0.02).
Vasoactive-inotropic scores after weaning were higher with levosimendan on the first two days. Norepinephrine use during treatment was numerically more frequent, at 47.5% versus 28.2%, but this comparison was not statistically significant (P = 0.07). It should not be described as a confirmed increase.
These secondary findings raise safety questions without establishing a mechanism. Baseline creatinine differed numerically between groups, multiple outcomes were examined, and death can influence the interpretation of renal endpoints. Proposed explanations involving vasodilation and renal vulnerability remain hypotheses. For perfusion and critical care teams, the practical implication is careful attention to renal function and circulatory support during the transition, rather than attributing every renal event directly to the study drug.
Mortality signals remain exploratory
Thirty-day mortality was 18.0% with levosimendan and 33.3% with placebo, but the difference did not reach statistical significance (P = 0.07). One-year mortality was 31.6% versus 51.4%, also without a statistically significant difference (P = 0.08). These numerical differences are of interest, but the limited sample prevents a firm survival conclusion.
Post hoc Bayesian analyses using a noninformative prior estimated a 95% probability of mortality superiority at 30 days and 96% at one year. However, the corresponding credible intervals included the null value. These estimates depend on the analysis model and prior assumptions and were exploratory rather than prespecified confirmation. They should inform future research questions, not be presented as proof that levosimendan improves survival or as a reason to disregard the neutral primary endpoint.
Implications for perfusion practice
WEANILEVO does not support a claim that a routine single infusion improves liberation from VA-ECMO in adults already meeting the trial’s weaning criteria. At the same time, early termination and wide confidence intervals prevent a definitive exclusion of benefit or harm. The renal and ICU signals deserve consideration alongside the uncertain mortality observations.
The study contributes an important controlled comparison to an area in which treatment enthusiasm can outpace randomized evidence. Its multicenter design, blinding, standardized selection, and clinically relevant outcomes strengthen the methods. Its incomplete enrollment and exploratory secondary analyses limit the confidence of its conclusions. A larger, adequately powered trial would be needed to clarify the overall balance of efficacy, renal safety, and survival. Until then, application should remain attentive to the population and timing actually studied.
Keywords: WEANILEVO, Levosimendan, VA-ECMO, ECMO Weaning, Cardiogenic Shock, Acute Kidney Injury, Mechanical Circulatory Support, Randomized Controlled Trial.
Source: Guinot PG, Berthoud V, Soudry-Faure A, et al., for the WEANILEVO group. Anesthesiology. 2026;145(3):676–688. doi:10.1097/ALN.0000000000006143. ClinicalTrials.gov: NCT04158674.
Funding and disclosures: The study received Orion and institutional funding. Orion was reported to have no role in study conception, design, conduct, or statistical analysis. Funding withdrawal caused early termination. The authors report industry relationships, including lecture fees and consulting, as detailed in the original paper.
This independent IPA educational summary accompanies the unchanged original open-access PDF, distributed under its CC BY-NC-ND 4.0 license.
