Intermittent inotropic support with levosimendan in advanced heart failure as destination therapy: The LEVO-D registry

Authors, Journal, Affiliations, Type, DOI

• David Dobarro, Víctor Donoso-Trenado, Eduard Solé-González, Carlos Moliner-Abós, José Manuel Garcia-Pinilla, Silvia Lopez-Fernandez, and 19 additional authors
• ESC Heart Failure 2023; 10:1193–1204
• 23 Spanish tertiary hospitals (lead: Hospital Álvaro Cunqueiro, Vigo; Hospital Universitari de Bellvitge, Barcelona)
• Multicentre retrospective registry
• DOI: 10.1002/ehf2.14278
• COI: D.D., J.G.C., J.J.B., J.F. received speaker fees from Orion Pharma; Orion funded the registry web-based database only; design, data collection, and analysis performed independently

Overview

The largest real-world multicentre registry of AHF patients treated with ambulatory intermittent levosimendan as destination therapy (n=403; 23 Spanish hospitals; 2015–2020). Restricted to patients not candidates for heart transplant or LVAD, on optimal medical therapy. Demonstrates a large reduction in HF hospitalisations and unplanned visits in the year after vs before levosimendan, without an increase in ICD therapies or VT/VF episodes. Introduces the LEVO-D score — a 5-variable prediction tool for 1-year responder status (AUC 0.71, superior to MAGGIC score AUC 0.60). No dosing strategy or dose per administration was independently associated with clinical response.

Keywords

Inotropes; Levosimendan; Palliative care; Advanced heart failure

Key Takeaways

Study Population and Design

• n=403; multicentre retrospective; 23 Spanish tertiary hospitals; patients 1 Jan 2015 – 1 Sep 2020
• Inclusion: AHF, NOT HT or LVAD candidates; on optimal medical therapy; at least one dose of ambulatory levosimendan
• Exclusion: de novo HF; any procedure that could improve prognosis post-levosimendan (coronary revasc, valve repair, CRT implant)
• Pre-SGLT2i era; OMT did not include SGLT2 inhibitors
• Median follow-up: 15.3 months (IQR 7.5–28.2)
• Baseline: mean age 71.5 years; 79.5% male; 60.9% permanent atrial arrhythmia; mean LVEF 27.5%; median MAGGIC score 27 (high risk); mean SBP 106.6 mmHg; 78.7% NYHA III

Outcomes

Mortality

• 52.6% died during follow-up; 73.9% of deaths cardiovascular; 17.5% non-cardiovascular; 8.5% unknown
• 1-year mortality: 26.8%; median survival 24.7 months (95% CI 20.4–26.9)
• CV death rate: 24.1 per 100 patient-years; HF hospitalisation rate: 30.1 per 100 patient-years

HF Event Reduction (Year Before vs After Levosimendan)

• HF hospitalisations: 77.9% → 38.7% (P<0.0001)
• Unplanned HF visits: 43.7% → 22.7% (P<0.0001)
• Combined event (death + HF hospitalisation + unplanned visit): 81.4% → 56.3% (P<0.0001)
• Total HF admissions per patient: 1.69 ± 1.7 → 1.12 ± 1.8 (P<0.001) — 33% reduction
• Total unplanned HF visits: 1.37 ± 3.2 → 0.68 ± 1.7 (P<0.0001) — 50% reduction

Responder Status

• 43.7% were "responders" — alive at 1 year without urgent HF admission or unplanned HF visit
• Non-responders included 26.8% who died (so 56.3% of non-responders were alive with HF events)
• Age NOT related to probability of response

Predictors of Response (Multivariate)

• Positive: beta-blockers (HR 1.44, 95% CI 1.03–2.06; P=0.04), haemoglobin >12 g/dL (HR 1.52, 95% CI 1.11–2.08; P=0.008), TEER (HR 2.06, 95% CI 1.01–4.27; P=0.04)
• Negative: previous unplanned HF visit (HR 1.97, 95% CI 1.43–2.71; P<0.0001), HR >70 bpm (HR 1.43, 95% CI 1.04–1.96; P=0.02), amiodarone use (HR 1.45, 95% CI 1.01–2.11; P=0.04)

Overall Mortality Predictors (Multivariate)

• Ischaemic HD (HR 1.84; P<0.0001), ICD therapies prior year (HR 2.10; P=0.006), DM2 (HR 1.49; P=0.02), age, HR per bpm, NT-proBNP, lower cholesterol, lower LVEF
• Neurohormonal blockade NOT related to reduced overall or 1-year mortality in this population

Safety

• ICD therapies: 16.2% vs 14.9% year after vs before (P=0.39) — NO increase
• VT/VF episodes: 0.91 ± 7.1 vs 0.87 ± 3.1 per year (P=0.94) — NO increase
• Levosimendan actively withdrawn in 40.4% of patients: 48.7% due to clinical improvement; 43.4% due to futility; only 7.9% due to adverse events or side effects (7 ventricular arrhythmias, 1 hypotension, 7 unspecified = 3.7% of whole cohort)
• Survival after withdrawal: clinical improvement group best (median 54.6 months); continued levosimendan (24.7 months); futility/side effects worst (10.8–11.8 months)

Levosimendan Administration Strategy and Dosing

• Strategies: bailout (40.2%), fixed number (33.3%), sine die (26.5%)
• Sine die patients showed lower sodium/albumin/cholesterol — marker of more advanced disease/frailty
• Strategy NOT related to outcomes (mortality, 1-year survival, HF hospitalisation, responder status)
• Dose: 6.25 mg (38%) and 0.2 µg/kg/min for 6h (28%) most common; 61.5% received "high dose" (≥6.25 mg)
• High dose better survival at 1 year and long-term in univariate analysis only; NOT significant in multivariate for response or mortality (HR 0.70, P=0.05 for overall mortality — trend only)
• Median doses received: 6 (IQR variable by strategy); sine die patients received most doses
• Response rate at 1 year NOT related to dosing — even low doses may be beneficial for HF event reduction

LEVO-D Score

• 5-variable prediction score for 1-year responder status (alive + no HF event)
• Scoring: TEER (+2), beta-blockers (+1.5), Hgb >12 g/dL (+1.5), amiodarone (−1.5), unplanned HF visit year before (−1.5), HR >70 bpm (−2)
• Range: −5 to +5; three risk groups:
  • Low probability (score ≤−1.5, 24.3% of cohort): 21.5% response rate
  • Intermediate (−1 to +1.5, 60.2%): 50.9% response rate
  • High probability (≥2, 15.4%): 68.4% response rate
• AUC 0.71 — superior to MAGGIC score (AUC 0.60) in this cohort

Contextual Comparisons

• VICTORIA trial high-risk cohort (NT-proBNP >8,000): CV death 22.9–26.5/100 patient-years and HF hosp 51.6–60.7/100 patient-years — LEVO-D showed similar death rate but substantially lower HF hosp (30.1/100), supporting a real drug effect rather than surveillance bias
• RELEVANT-HF registry (similar concept but smaller, younger, 40% HT/LVAD candidates): comparable 1-year death/HT/LVAD rate ~24%; LEVO-D older and higher risk
• MedaMACS (true AHF medical management): 1-year death/transplant/LVAD 34% vs LEVO-D 26.8%; but MedaMACS patients >10 years younger, lower LVEF, less comorbid

Limitations of the Document

• Retrospective design; no control group; inherent selection bias
• Pre-SGLT2i era — current optimal medical therapy may differ
• Heterogeneous dosing protocols across 23 centres; no pre-specified follow-up schedule
• Comparison with year before levosimendan is an uncontrolled before-after design (regression to mean, Hawthorne effect, surveillance bias cannot be excluded)
• COI: 4 authors received Orion speaker fees; Orion funded the database

Key Concepts Mentioned

• entities/Levosimendan — primary subject; real-world destination therapy outcomes and LEVO-D score
• concepts/Vasoactive-Agents-in-CS — inotrope context

Key Entities Mentioned

• LEVO-D Score — novel 5-variable prediction tool for levosimendan response
• LION-HEART trial — referenced as context; LEVO-D consistent with its positive results
• LevoRep trial — referenced; LEVO-D discussion attributes LevoRep failure to low dose count
• LAICA trial — referenced as context

Wiki Pages Updated

• wiki/entities/Levosimendan.md — LEVO-D registry added to pulsed infusion section; LEVO-D Score section added; arrhythmia section updated with real-world ICD/VT safety data; practical dosing updated with strategy/dosing findings; source_count 3→4
• wiki/sourceindex.md — new entry prepended
• wiki/wikiindex.md — Levosimendan entry updated