Right ventricular dysfunction in cardiogenic shock on temporary mechanical circulatory support Authors: Jan Estoppey, Iulia Maia Muresan, Matthias Bossard, Gregor Leibundgut, Gregor Fahrni, Stefan Blöchlinger, Roger Ludwig, Anna S Messmer, Valentin Köchli, Thomas Seiler, Armin Spahr, Alain M Bernheim, Anna Schmidt, Florim Cuculi, Lukas Hunziker, Monika Fürholz, Christian Jung Published: 2026 ABSTRACT AbstractAimsRight ventricular dysfunction (RVD) is common in cardiogenic shock (CS) due to left ventricular failure and is associated with increased mortality, but its impact in patients requiring temporary mechanical circulatory support (tMCS) remains unclear. We explored RVD and its association with clinical outcomes in CS patients under tMCS.Methods and resultsIndividuals with CS from the SWISS Circulatory Support Registry (NCT04117230) were included. RVD was defined by ≥1 of the following parameters: tricuspid annular plane systolic excursion (TAPSE) < 14 mm, TAPSE/systolic pulmonary artery pressure (sPAP) < 0.34, right atrial pressure (RAP) > 15 mmHg, RAP/pulmonary capillary wedge pressure >0.63 or pulmonary artery pulsatility index (PAPi) < 1.85. The primary endpoint was 6-month mortality; secondary endpoints were hemodynamic stabilization markers. 745 patients were included: 580 supported with left-side-Impella-devices© (with 39 requiring concomitant second device) and 165 with VA-ECMO. RVD was more prevalent in patients on VA-ECMO compared to Impella (60% vs. 33%, P < 0.001). In Impella patients, RVD was associated with higher 6-month mortality (43% vs. 30%, P = 0.015), higher lactate levels (1.6 mmol/L vs. 1.3 mmol/L; P = 0.01), lower central venous saturation (55% vs. 61%; P = 0.005) and reduced mean arterial pressure (55 mmHg vs. 58 mmHg; P = 0.002). In VA-ECMO patients with RVD, 6-month mortality was numerically higher, without statistical significance (61% vs. 45%, P = 0.13).ConclusionRVD was linked to increased mortality and worse haemodynamic stabilization in patients supported with Impella, whereas this association was not demonstrated in those under VA-ECMO. RV function is a key prognostic marker in CS patients under tMCS and should be assessed to guide device selection, management, and weaning.Keywords: Right ventricular dysfunction, Cardiogenic shock, Temporary mechanical circulatory support, Percutaneous microaxial flow pump, Venoarterial extracorporeal membrane oxygenation Introduction Mortality in cardiogenic shock (CS) remains high (27–51%) despite advances in early reperfusion therapy, aetiological treatment and temporary mechanical circulatory support (tMCS).1 Right ventricular dysfunction (RVD) is common in CS secondary to left ventricular (LV) failure and may result from increased afterload, elevated preload, or impaired RV contractility.2,3 RVD has been reported in 48% of patients presenting with acute decompensation of advanced chronic heart failure4 and in up to 38% of patients with CS due to acute myocardial infarction (AMI).5 While some studies in CS patients reported an association between RVD and mortality6 others did not.5 Early detection and treatment of RVD have been associated with reduced in-hospital mortality.7 However, little is known on the prevalence and prognostic impact of RVD in CS patients under tMCS. tMCS is increasingly utilized in CS to restore cardiac output without increasing cardiac workload.8 Despite associated potential complications such as bleeding, infections, and limb ischaemia, tMCS in addition to standard of care has been associated with increased survival in selected patients, particularly those with AMI-related CS and/or refractory cardiac arrest9–11. Assessing RV function in CS is important as it may influence the choice of tMCS device. The percutaneous microaxial flow pump (Impella, J&J Medtech) requires adequate right heart function for optimal LV filling.12 In refractory right ventricular failure (RVF), a secondary device to support the right ventricle (RV) may be necessary. Venoarterial extracorporeal membrane oxygenation (VA-ECMO) can bypass the RV and maintain systemic output.12,13 However, VA-ECMO increases LV afterload and, in the presence of a nonejecting heart, escalation to a second device to unloads the LV and ensure pulmonary flow may become necessary. Diagnosis of RVD under tMCS is poorly defined and represents a challenge in CS patients.14 Common RV function parameters include tricuspid annular plane systolic excursion (TAPSE), right atrial pressure (RAP) or central venous pressure (CVP), RV-pulmonary arterial coupling (TAPSE/systolic pulmonary arterial pressure (sPAP)), cardiac filling pressures (RAP/pulmonary capillary wedge pressure (PCWP)) and pulmonary artery pulsatility index (PAPi = ((sPAP-diastolic PAP (dPAP))/mRAP).13 TAPSE<14 mm was associated with increased mortality in CS due to acute coronary syndrome (ACS).6 The other parameters have a wide range of pathologic thresholds depending on the type of disease. For example, RAP/CVP >15 mmHg,15 TAPSE/sPAP<0.34,16 RAP/PCWP >0.63,15 and PAPi <1.8517 were reported to be predictors of RVF after permanent left ventricular assist device (LVAD). However, data on evaluation of these parameters in patients supported with tMCS is scarce. RVD among patients requiring tMCS represents a key issue hindering negative fluid balance, optimal device function, and device weaning.18 This study aimed to explore the impact of RVD in patients with CS receiving tMCS with Impella or VA-ECMO. We evaluated the distribution and correlation of various RV function parameters and tested whether RVD would be associated with worse haemodynamic stabilization and survival. Methods Patient population and clinical follow up Consecutive CS patients were selected from the SWISS Circulatory Support Registry, CARDSUP (NCT04117230), an ongoing multicentric registry in Switzerland, including adults (> 18 years) receiving tMCS. Five Swiss tertiary cardiology centres participate in the registry: Bern University Hospital, Cantonal Hospital of Lucerne, University Hospital Basel, Hospital Winterthur, and Triemli Hospital Zurich. Device selection, peri-interventional management, and patient care are left at the discretion of the treating physicians. Follow-up assessment at 6 months was conducted by phone call, clinic visits, as well as reviewing outpatient clinic records, referring physicians' documentation, and hospital discharge summaries. All data are entered into a dedicated, secure online database (REDCap©, Vanderbilt University, Tennessee) maintained by the Clinical Trial Unit at the University of Bern, Switzerland. Data from the years 2012 to 2018 were retrospectively collected in some centres using the same inclusion criteria as for the prospective group from 2019 upwards, when the registry was formally established and approved by the local ethics committee. Inclusion and exclusion criteria for this analysis are highlighted in the results section in detail and Figure 1. For simplicity, patients supported with LV percutaneous microaxial flow pump are referred to as ‘Impella’, and those supported with VA-ECMO as ‘ECMO’. Patients in the Impella group who required an additional combined device were labelled ‘Impella + 2nd device.’ They were included in the Impella group, as RV function was mostly assessed under Impella support alone, before second device escalation. To minimize misclassification, subjects (n = 18) who died under ongoing tMCS from a probable noncardiac cause were excluded because the available information was insufficient to reliably classify the cause of death as cardiac or noncardiac. Flow chart. Flow chart illustrating patient selection from the Swiss Circulatory Support (CARDSUP) registry, including inclusion and exclusion criteria. Impella RP = right ventricular Impella, IABP = intra-aortic balloon pump, ProtekDuo = Right ventricular support device. RV function parameters and definition of RV dysfunction The following RV function parameters with abnormal thresholds commonly accepted as predictors of RV failure were selected, although no fixed cut-offs are established for tMCS and CS: (i) TAPSE <14 mm19 (lowest value within 3 days after device implantation); (ii) RAP or CVP >15 mmHg15 (highest value within 24 h after device implantation); (iii) TAPSE/sPAP <0.3416 (lowest value within 24 h after device implantation); (iv) RAP/PCWP >0.6315 (highest value within 24 h after device implantation); and (v) PAPi <1.8517 (lowest value within 24 h after device implantation). Due to the observational nature of this study, some RV function parameters were incompletely reported. A composite RV dysfunction group (named ‘RVD’), defined by the presence of at least one of the aforementioned pathological RV function parameters, was chosen. Normal RV function required all available parameters to be within normal ranges, while RV function was classified as unknown when no parameters were available. Distribution of unknown RVD population according to device type and CS stages is presented in supplementary material online, Figures S1 and S2. In addition to this RVD group, correlations among the different RV function parameters were investigated. Endpoint assessment The primary endpoint was all-cause mortality at six months. Secondary endpoints were categorized into (i) haemodynamic stabilization parameters, including lactate levels, mixed venous oxygen saturation, arterial blood pressure, and cardiac output; (ii) treatment dependency, including duration of inotrope use, device support time and duration of mechanical ventilation; and (iii) resource utilization outcomes, including intensive care unit (ICU) length of stay. Statistics Statistical analyses were performed using R software (version 4.5.1). Descriptive statistics were reported as frequencies and percentages for categorical variables, and as median with first and third quartiles [Q1–Q3]. Group comparisons were conducted using Pearson’s Chi-squared test or Fisher’s exact test (used when expected cell counts were less than 5) for categorical variables, and the Wilcoxon rank-sum test for continuous variables that were not normally distributed. Correlations were assessed using the nonparametric Spearman rank correlation, with P-values adjusted for multiple comparisons using the Bonferroni correction. Time-to-death was displayed as cumulative event curves using the Kaplan–Meier method, with differences between groups assessed using the log-rank test. A two-sided P-value of <0.05 was considered statistically significant. Results The analysed population was selected from 1377 consecutively enrolled patients in the CARDSUP registry who received a tMCS device between January 2012 and 30th June 2024 (Figure 1). Inclusion criteria were CS stage C or higher according to the Society for Cardiovascular Angiography and Intervention,20 and the use of a single left ventricular support device (all left-sided Impella types) or ECMO. Intra-aortic balloon pumps were excluded due to low numbers, and isolated right ventricular support devices (Impella RP, ProtekDuo) were excluded as these are typically used in isolated RVD only. In the Impella group, escalation to a concomitant second device supporting the RV was allowed, as it likely reflected RV failure. In contrast, patients in the ECMO group who required an additional device for left ventricular unloading were excluded, as this mainly reflected further LV failure and because our objective was to assess RV function parameters under isolated ECMO support. The final cohort comprised 745 patients, of whom 580 (77.9%) received an Impella, including 39 requiring escalation to a second device (ECMO in 100% of cases), and 165 (22.1%) received ECMO support. Patient characteristics and RVD prevalence The baseline characteristics are depicted in Table 1, while clinical and haemodynamic data during and after device implantation, along with outcomes, are provided in Table 2. The median age of the cohort was 66 years, with 26% female representation. Overall mortality was 42% at 1 month and 47% at 6 months. Mortality was higher in the ECMO group at both 1 month (56%) and 6 months (61%), compared to 49% and 54%, respectively, in the Impella+2nd device group, and 37% and 42% in the Impella-only group. Table 1Baseline characteristicsVariablesAllImpella, N = 580ECMO P-valuec N = 745aImpella only, N = 541a,bImpella+2nd Device, N = 39a,b N = 165a,b Demographic Age (years)66 (56, 74), n = 74569 (58, 75), n = 54161 (51, 67), n = 3961 (50, 68), n = 165 <0.001Gender (female)196/745 (26%)139/541 (26%)10/39 (26%)47/165 (28%)0.5BMI [kg/m2]25.9 (24.0, 29.1), n = 72225.8 (23.9, 29.0), n = 52524.9 (23.9, 30.0), n = 3626.2 (24.0, 29.1), n = 1610.3 Cardiovascular risk factors Arterial hypertension374/743 (50%)292/540 (54%)10/39 (26%)72/164 (44%)0.062Diabetes mellitus201/743 (27%)164/540 (30%)4/39 (10%)33/164 (20%) 0.024Hypercholesterolaemia or dyslipidaemia322/744 (43%)251/541 (46%)10/39 (26%)61/164 (37%)0.075Smoker220/743 (30%)166/540 (31%)11/39 (28%)43/164 (26%)0.3Family history of coronary artery disease114/741 (15%)82/539 (15%)3/39 (7.7%)29/163 (18%)0.3 Cardiovascular history Previous myocardial infarction139/744 (19%)105/540 (19%)6/39 (15%)28/165 (17%)0.5Previous PCI155/743 (21%)113/540 (21%)6/38 (16%)36/165 (22%)0.7Previous CABG surgery47/743 (6.3%)33/540 (6.1%)1/39 (2.6%)13/164 (7.9%)0.3Previous congestive heart failure124/743 (17%)85/541 (16%)6/39 (15%)33/163 (20%)0.2Known renal failure (GFR < 30 mL/min)31/744 (4.2%)27/541 (5.0%)1/39 (2.6%)3/164 (1.8%)0.090Peripheral artery disease83/744 (11%)68/541 (13%)1/39 (2.6%)14/164 (8.5%)0.2Previous stroke50/743 (6.7%)39/539 (7.2%)4/39 (10%)7/165 (4.2%)0.15Chronic lung disease57/743 (7.7%)44/541 (8.1%)1/38 (2.6%)12/164 (7.3%)0.8Atrial fibrillation90/744 (12%)69/541 (13%)4/39 (10%)17/164 (10%)0.4 Haemodynamic parameters before device implantation Heart rate [per minute]101 (80, 125), n = 480100 (80, 120), n = 412111 (76, 132), n = 31130 (90, 150), n = 37 <0.001Mean arterial pressure [mmHg]64 (50, 78), n = 46665 (52, 79), n = 40260 (55, 70), n = 2949 (40, 67), n = 35 <0.001 Cardiogenic shock Stage of cardiogenic shock <0.001 C Classic cardiogenic shock 258/745 (35%)230/541 (43%)13/39 (33%)15/165 (9.1%) D Deteriorating 211/745 (28%)173/541 (32%)19/39 (49%)19/165 (12%) E Extremis 276/745 (37%)138/541 (26%)7/39 (18%)131/165 (79%)Onset of cardiogenic shock (> 12 h)62/575 (11%)53/399 (13%)2/31 (6.5%)7/145 (4.8%) 0.008Use of catecholamines before implantation590/745 (79%)402/541 (74%)30/39 (77%)158/165 (96%) <0.001Respiratory support, intubated before implantation476/735 (65%)298/532 (56%)22/39 (56%)156/164 (95%) <0.001 Reasons for cardiogenic shock Acute coronary syndrome454/745 (61%)349/541 (65%)22/39 (56%)83/165 (50%) 0.002RCA77/454 (17%)54/349 (15%)3/22 (14%)20/83 (24%)0.055LM/prox. LAD302/454 (67%)239/349 (68%)12/22 (55%)51/83 (61%)0.3LCX36/454 (7.9%)28/349 (8.0%)3/22 (14%)5/83 (6.0%)0.5Post-ECC9/745 (1.2%)5/541 (0.9%)1/39 (2.6%)3/165 (1.8%)0.4Resuscitation358/745 (48%)213/541 (39%)11/39 (28%)134/165 (81%) <0.001 Type of impella Impella 2.5/CP/CP smart assist534/541 (99%)39/39 (100%)Impella 5.0/5.5 smart assist7/541 (1.3%)0 (0%) Impella+2nd device Type of 2nd device used: VA-ECMO39/39 (100%)Cardiogenic shock as indication for second device36/39 (92%) aValues are median with first and third quartiles, [Q1–Q3], or n (%). bImpella only: Impella in the left ventricle only. Impella + 2nd device: Impella in the left ventricle followed by concomitant VA-ECMO. ECMO: VA-ECMO as the first device. c P-values indicate the statistical significance of the comparison between two groups: Impella (Impella only and Impella+2nd device) and ECMO.BMI = body mass index; GFR = glomerular filtration rate, PCI = percutaneous coronary intervention; CABG = coronary artery bypass grafting, RCA = right coronary artery; LM = left main; LAD = left anterior descending artery; LCX = left circumflex artery; ECC = Extracorporeal circulation, LVEDP = left ventricular end diastolic pressure. Baseline characteristics aValues are median with first and third quartiles, [Q1–Q3], or n (%). bImpella only: Impella in the left ventricle only. Impella + 2nd device: Impella in the left ventricle followed by concomitant VA-ECMO. ECMO: VA-ECMO as the first device. c P-values indicate the statistical significance of the comparison between two groups: Impella (Impella only and Impella+2nd device) and ECMO. BMI = body mass index; GFR = glomerular filtration rate, PCI = percutaneous coronary intervention; CABG = coronary artery bypass grafting, RCA = right coronary artery; LM = left main; LAD = left anterior descending artery; LCX = left circumflex artery; ECC = Extracorporeal circulation, LVEDP = left ventricular end diastolic pressure. Table 2Clinical and haemodynamic variables during and after device implantation and outcomesVariableAll cardiogenic shock patientsImpella, N = 580ECMO P-valuec N = 745aImpella only, N = 541a,bImpella+2nd Device, N = 39a,b N = 165a,b Laboratory Lactate [mmol/L] during implantation4.9 (2.1–10.4), n = 6363.4 (1.7–7.3), n = 4537.2 (2.8–12.2), n = 3211.3 (6.9–16.0), n = 154 <0.001Lactate 24 h after implantation [mmol/L]1.50 (1.10–2.60), n = 4951.40 (1.10–2.30), n = 3563.55 (1.60–4.50), n = 261.90 (1.10–2.90), n = 113 0.035Creatinine at baseline [umol/L]98 (75–130), n = 66097 (74–133), n = 485104 (80–125), n = 35100 (76–128), n = 1400.8Peak creatinine [umol/L]130 (93–218), n = 654134 (93–222), n = 481152 (108–288), n = 32122 (91–189), n = 1410.13Lowest GFR (CKD-EPI) [mL/min]46 (26–72), n = 61344 (26–71), n = 44538 (21–61), n = 3452 (33–75), n = 134 0.016Haemoglobin at baseline [g/L]131 (114–147), n = 701134 (116–148), n = 504144 (129–159), n = 37126 (103–138), n = 160 <0.001HS troponin T at baseline [ng/L]329 (77–1601), n = 486399 (102–1800), n = 352793 (84–2200), n = 19141 (46–1200), n = 115 0.002Peak HS troponine T [ng/L]3863 (1083–12 307), n = 5243677 (1090–10 385), n = 3977201 (1225–18 919), n = 215067 (764–22 375), n = 1060.15CK at baseline [U/I]262 (121–907), n = 558296 (126–928), n = 408229 (105–1882), n = 22225 (103–749), n = 1280.2Peak creatine kinase [U/L]2645 (809–6005), n = 6352311 (725–5567), n = 4646884 (3241–10 471), n = 303451 (994–6792), n = 1410.079ASAT [U/L]268 (112–676), n = 484226 (98–559), n = 336466 (185–1577), n = 31408 (145–759), n = 117 0.014ALAT [U/L]107 (47–284), n = 54286 (44–245), n = 386140 (72–450), n = 34147 (65–339), n = 122 0.017NTpro-BNP [pg/mL] highest value after device implantation4410 (2001–11 461), n = 1744429 (1711–11 353), n = 1518762 (2966–20 253), n = 112843 (1182–12 488), n = 120.4 Haemodynamic data and support Systolic arterial pressure [mmHg]70 (60–81), n = 54075 (64–85), n = 37660 (51–68), n = 2560 (53–70), n = 139 <0.001Diastolic arterial pressure [mmHg]49 (42–55), n = 53950 (43–56), n = 37548 (40–54), n = 2548 (42–54), n = 1390.065Mean arterial pressure [mmHg]55 (50–61), n = 50856 (51–63), n = 34452 (45–59), n = 2553 (48–59), n = 139 <0.001Cardiac output index [l/min/m2]1.60 (1.10–2.10), n = 2341.80 (1.50–2.20), n = 1511.30 (0.85–2.10), n = 120.80 (0.60–1.20), n = 71 <0.001Mixed venous saturation or central venous saturation [%]58 (50–65), n = 29955 (48–62), n = 21050 (45–66), n = 1566 (58–71), n = 74 <0.001 Cardiac LV function LVEF on first day (echo- or angiographic determination)25 (15–35), n = 66428 (20–40), n = 47917 (10–26), n = 2810 (0–25), n = 157 <0.001Left ventricular end-diastolic diameter [mm]51 (45–57), n = 21450 (45–56), n = 18145 (36–53), n = 755 (49–62), n = 26 0.045 Cardiac RV function TAPSE [mm]18 (14–22), n = 23619 (15–22), n = 19014 (12–17), n = 913 (9–19), n = 37 <0.001TAPSE/sPAP0.41 (0.29–0.56), n = 840.43 (0.31–0.54), n = 510.39 (0.30–0.63), n = 40.34 (0.28–0.60), n = 290.5RAP or CVP [mmHg]14 (11–18), n = 35715 (11–18), n = 21117 (14–19), n = 1713 (11–16), n = 129 0.027RAP/PCWP0.94 (0.75–1.19), n = 1890.94 (0.73–1.18), n = 1151.00 (0.76–1.07), n = 90.94 (0.75–1.25), n = 570.7PAPi1.20 (0.80–1.80), n = 2191.40 (1.00–1.80), n = 1180.80 (0.60–1.30), n = 101.00 (0.60–1.60), n = 91 <0.001 Right ventricular dysfunction <0.001 No 171/745 (23%)134/541 (25%)4/39 (10%)33/165 (20%) Yes 288/745 (39%)172/541 (32%)17/39 (44%)99/165 (60%) Unknown 286/745 (38%)235/541 (43%)18/39 (46%)33/165 (20%) Risk factors for RVF Length of mechanical ventilation [hours]39 (3–119), n = 65724 (0–97), n = 474192 (72–312), n = 3453 (20–115), n = 149 <0.001Systolic pulmonary pressure [mmHg]39 (31–47), n = 22743 (36–50), n = 12638 (33–42), n = 1033 (26–40), n = 91 <0.001Diastolic pulmonary arterial pressure [mmHg]20 (16–25), n = 22122 (17–25), n = 12024 (20–26), n = 1018 (14–24), n = 91 0.004Mean pulmonary arterial pressure [mmHg]27 (21–34), n = 24030 (24–35), n = 13728 (25–34), n = 1224 (18–29), n = 91 <0.001Pulmonary capillary wedge pressure [mmHg]15.0 (12.0–19.0), n = 19616.0 (13.0–20.0), n = 12918.5 (15.0–24.0), n = 1014.0 (11.0–16.0), n = 57 0.005Severe tricuspid regurgitation12/459 (2.6%)7/338 (2.1%),1/26 (3.8%)4/95 (4.2%)0.3 In hospital follow-up Device support time [hours] 37 (14–72), n = 69536 (12–72), n = 50196 (36–218), n = 3531 (15–72), n = 1590.4 Length of ICU stay [days] 4 (2–8), n = 7354 (2–8), n = 53810 (4–17), n = 384 (1–7), n = 159 0.024 Heart transplantation 1/745 (0.1%)001/165 (0.6%)0.2 LVAD 16/745 (2.1%)10/541 (1.8%)4/39 (10.2%)2/165 (1.2%)0.5 6-month follow-up MACCE 482/745 (65%)342/541 (63%)26/39 (67%)114/165 (69%)0.18 All-cause death 348/745 (47%)226/541 (42%)21/39 (54%)101/165 (61%) <0.001Cardiac death283/745 (38%)186/541 (34%)16/39 (41%)81/165 (49%) <0.001Non-cardiac death65/745 (9%)40/541 (7%)5/39 (13%)20/165 (12%)0.08 a,b,cAs in Table 1.HS troponin = high sensitivity troponin; NTpro-BNP = N-terminal pro b-type natriuretic peptide; CK = creatinine kinase; ASAT = aspartate aminotransferase; ALAT = alanine aminotransferase; mPCWP = mean pulmonary capillary wedge pressure; LVAD = permanent left ventricular assist device. Other abbreviations as in Table 1.Timing of the following variables:Lowest value within the 24 h after device implantation: all variables under ‘Hemodynamic data and support’, TAPSE/sPAP, PAPi.Highest value within the 24 h after device implantation: all: RAP or CVP, RAP/PCWP, all variables under ‘Risk factors for RVF’.Lowest value within 3 days after device implantation: TAPSE.Highest value within 3 days after device implantation: ASAT, ALAT, Left ventricular end-diastolic diameter.Highest value within 5 days after device implantation: Peak creatinine.Lowest value within 5 days after device implantation: Lowest GFR. Clinical and haemodynamic variables during and after device implantation and outcomes a,b,cAs in Table 1. HS troponin = high sensitivity troponin; NTpro-BNP = N-terminal pro b-type natriuretic peptide; CK = creatinine kinase; ASAT = aspartate aminotransferase; ALAT = alanine aminotransferase; mPCWP = mean pulmonary capillary wedge pressure; LVAD = permanent left ventricular assist device. Other abbreviations as in Table 1. Timing of the following variables: Lowest value within the 24 h after device implantation: all variables under ‘Hemodynamic data and support’, TAPSE/sPAP, PAPi. Highest value within the 24 h after device implantation: all: RAP or CVP, RAP/PCWP, all variables under ‘Risk factors for RVF’. Lowest value within 3 days after device implantation: TAPSE. Highest value within 3 days after device implantation: ASAT, ALAT, Left ventricular end-diastolic diameter. Highest value within 5 days after device implantation: Peak creatinine. Lowest value within 5 days after device implantation: Lowest GFR. Acute coronary syndrome (ACS) accounted for 61% of CS cases. Impella was used more frequently than ECMO in CS stage C (42% vs. 9%), while ECMO was used more frequently than Impella in CS stage E (79% vs. 25%). Compared to the Impella group, the ECMO group experienced higher lactate levels, lower mean arterial pressure, higher heart rate, and worse left ventricular ejection fraction (LVEF) at the time of implantation (Tables 1, 2). RVD was observed in 39% of the patients overall, increasing from 32% in the Impella-only group to 44% in the Impella+2nd device group, and 60% in the ECMO group. Distribution, correlations and prognostic value of individual RV function parameters Figure 2 and Table 2 show RV function parameters within the first hours after tMCS implantation. Overall, TAPSE, RAP and PAPi were higher in the Impella group compared to ECMO group, while TAPSE/sPAP and RAP/PCWP were similar. Pulmonary pressures, including wedge pressure, were higher in the Impella group. There were no differences in severe tricuspid regurgitation between groups. Distribution of RV function parameters. RV function parameter distribution in Impella vs. ECMO patients (median [Q1–Q3]): TAPSE (19 [15–22] vs. 13 [9–19], P = 0.001), TAPSE/sPAP (0.43 [0.31–0.54], vs. 0.34 [0.28–0.6], P = 0.5), RAP (15 [11–18], vs. 13 [11–16], P = 0.027), RAP/PCWP (0.938 [0.742–1.168], vs. 0.938 [0.75–1.24], P = 0.7), PAPi (1.4 [0.9–1.8] vs. 1 [0.6–1.6], P=<0.001). Figure 3 illustrates correlations between RV function parameters. In the Impella group, RAP correlated with all other RV function parameters. In the ECMO group, RAP correlated with all other parameters except TAPSE. Correlations were stronger in the Impella group than in the ECMO group. Correlations between RV function parameters. Correlations between RV function parameters are illustrated; for Impella group and ECMO group. Significant P-values are marked with asterisks: *** for P < 0.001, ** for P < 0.01, and * for P < 0.05. Correlation coefficients are shown numerically and depicted by circles whose size and colour reflect correlation strength, as indicated by the colour gradient. In the Impella group, pathological TAPSE, TAPSE/sPAP and PAPi were associated with increased 6-month mortality, whereas abnormal RAP and RAP/PCWP were not (Table 3). In the ECMO group, pathological RAP and TAPSE/sPAP were associated with increased mortality, while abnormal TAPSE, RAP/PCWP and PAPi were not. Table 3Six-month mortality according to normal vs. pathological RV function parametersImpellaECMORV function parametersNormalaPathologicala P-valuebNormalaPathologicala P-valueb TAPSE 32/166 (19%)16/33 (48%) <0.0015/18 (28%)10/19 (53%)0.12 TAPSE/sPAP 6/35 (17%)10/20 (50%) 0.0103/15 (20%)8/14 (57%) 0.039 RAP 45/123 (37%)51/105 (49%)0.06848/95 (51%)24/34 (71%) 0.043 RAP/PCWP 5/13 (38%)45/111 (41%)0.94/7 (57%)22/50 (44%)0.7 PAPi 6/29 (21%)47/99 (47%) 0.0107/17 (41%)43/74 (58%)0.2 aPathological RV function parameters are defined as in the methods section. Normal values are as follows: TAPSE ≥14, TAPSE/sPAP ≥0.34, RAP ≤15, RAP/PCWP ≤0.63, PAPi: ≥1.85. b P-values indicate the statistical significance of the comparison between normal and pathologic RV function parameters.Abbreviations as in Tables 1 and 2. Six-month mortality according to normal vs. pathological RV function parameters aPathological RV function parameters are defined as in the methods section. Normal values are as follows: TAPSE ≥14, TAPSE/sPAP ≥0.34, RAP ≤15, RAP/PCWP ≤0.63, PAPi: ≥1.85. b P-values indicate the statistical significance of the comparison between normal and pathologic RV function parameters. Abbreviations as in Tables 1 and 2. Clinical outcomes according to composite RVD status Figure 4 illustrates all-cause mortality and Table 4 shows clinical and haemodynamic outcomes by RVD status. In the Impella group, patients with RVD exhibited higher 6-month all-cause mortality (43% vs. 30%, P = 0.015). No differences were observed in MACCE (66% vs. 59%, P = 0.21). LVEF was lower in patients with RVD in both Impella (25 [15–35] % vs. 29 [20–43] %; P = 0.031) and ECMO (10 [0–20] % vs. 15 [10–35] %; P = 0.009) groups. RVD was linked to worse haemodynamic stabilization after Impella implantation: higher lactate (1.6 [1.1–2.8] mmol/L vs. 1.3 [1.0–1.9] mmol/L; P = 0.01), lower mixed/central venous saturation (55 [47–62] % vs. 61 [53–66] %; P = 0.005), and lower mean arterial pressure (55 [50–60] mmHg vs. 58 [52–66] mmHg; P = 0.002). No significant differences were observed in indexed cardiac output (1.8 [1.4–2.1] L/min/m2 vs. 1.9 [1.7–2.5] L/min/m2; P = 0.17). All-cause mortality according to presence or absence of RVD. Kaplan–Meier all-cause death curves stratified by RVD status within each device group. Numbers at risk are displayed below the plots. Patients alive were censored at their last known follow-up or administratively at 180 days and are shown as tick marks on the plots. Shaded areas represent 95% confidence intervals. Table 4Haemodynamic and clinical outcomes according to presence or absence of RVDVariableImpella N = 580ECMO N = 165RVD—No, N = 138a,bRVD—YES, N = 189a,b P-valuecRVD—No, N = 33a,bRVD—YES, N = 99a,b P-valuecLactate during implantation [mmol/L]2.4 (1.5–5.7), n = 1274.7 (2.6–9.3), n = 188 <0.0017.4 (3.2–13.8), n = 2911.3 (7.8–16.0), n = 97 0035 Lactate 24 h after implantation [mmol/L]1.30 (1.00–1.90), n = 1001.60 (1.10–2.80), n = 167 0.011.50 (1.10–2.80), n = 231.95 (1.20–2.80), n = 780,31Mixed venous saturation or central venous saturation [%]61 (53–66), n = 4255 (47–62), n = 155 0005 68 (62–70), n = 565 (57–73), n = 680.78Lowest GFR (CKD-EPI) [mL/min]49 (26–74), n = 11641 (26–68), n = 1810.1758 (38–79), n = 2752 (33–74), n = 890,41ALAT [U/L]71 (41–208), n = 96112 (57–376), n = 155 0009 156 (72–333), n = 30155 (50–333), n = 770,62Systolic arterial pressure [mmHg]80 (69–93), n = 10868 (59–76), n = 179 <0.00166 (53–70), n = 3360 (53–69), n = 960,26Mean arterial pressure [mmHg]58 (52–66), n = 10055 (50–60), n = 177 0002 57 (49–63), n = 3252 (48–57), n = 98 0,03 Cardiac output index [l/min/m2]1.90 (1.70–2.50), n = 161.80 (1.40–2.10), n = 1300,171.05 (0.90–1.70), n = 40.70 (0.60–1.20), n = 670,13Duration of inotrope support [h]50 (16–96), n = 5445 (18–96), n = 1520,8730 (15–65), n = 2534 (17–84), n = 850,5Device support time [hours]26 (12–65), n = 13440 (20–70), n = 185007448 (13–96), n = 3137 (20–71), n = 980.86Length of mechanical ventilation [hours]8 (0–96), n = 12948 (12–123), n = 184 <0.00170 (28–125), n = 3266 (24–115), n = 940.77Length of ICU stay [days]4.0 (2.0–7.0), n = 1374.0 (2.0–8.0), n = 1890.154.0 (2.0–7.0), n = 334.0 (2.0–7.0), n = 970.70 LVEF on first day 29 (20–43), n = 13525 (15–35), n = 182 0031 15 (10–35), n = 3110 (0–20), n = 97 0009 MACCE at 6 months 82/138 (59%)125/189 (66%)0.2118/33 (55%)70/99 (71%)0.088 All-cause death at 6 months 41/138 (30%)81/189 (43%) 0.01515/33 (45%)60/99 (61%)0.13Cardiac death at 6 months34/138 (25%)61/189 (32%)0.1313/33 (39%)46/99 (46%)0.48Non-cardiac death at 6 months7/138 (5%)20/189 (11%)0.0742/33 (6%)14/99 (14%)0.36 aValues are median with first and third quartiles, [Q1–Q3], or n (%). bRVD = right ventricular dysfunction—yes/no. c P-values indicate the statistical significance of the comparison between RVD-No and RVD-Yes groups.Abbreviations and timing of the variables as in Table 2. Haemodynamic and clinical outcomes according to presence or absence of RVD aValues are median with first and third quartiles, [Q1–Q3], or n (%). bRVD = right ventricular dysfunction—yes/no. c P-values indicate the statistical significance of the comparison between RVD-No and RVD-Yes groups. Abbreviations and timing of the variables as in Table 2. Among patients supported with ECMO, mortality was numerically higher in patients with RVD (61% vs. 45%, P = 0.13) without statistical significance. No differences in MACCE or consistent haemodynamic outcomes were observed by RVD status in the ECMO group. Discussion In this real-world CS cohort managed with tMCS, 6-month overall mortality was 47%, consistent with reported rates ranging from 40% to 60%.21,22 Impella devices were more frequently employed as first devices, compared to ECMO (77.9% vs. 22.1%), highlighting a trend towards less invasive microaxial devices. Our population reflects a severe CS phenotype, with 79% of ECMO patients and 25% of Impella patients classified as CS stage SCAI E. Patient characteristics and indication for device implantation differed between groups, with ECMO patients presenting with more advanced CS stages, higher haemodynamic instability, and greater multiorgan failure. Our data show that the prevalence of RVD was significantly higher in the ECMO group (60%) compared to the Impella group (33%), aligning with previous reports.4,5,23 Patient selection and device-specific haemodynamics probably influenced the distribution of individual RV parameters: TAPSE, RAP, and PAPi were higher in the Impella than in the ECMO group, whereas TAPSE/sPAP and RAP/PCWP were comparable. In the Impella group, pathological TAPSE, TAPSE/sPAP, and PAPi were linked with increased mortality. Abnormal RAP showed a similar trend, while no association was observed for RAP/PCWP, likely due to the Impella-mediated reduction of filling pressures, attenuating its discriminative value. In the ECMO group, elevated RAP and TAPSE/sPAP were associated with increased mortality. Interpretation is limited by the lack of data on ECMO support level and the complex interaction between RAP, volume status, and RV function under this device. Under full ECMO support, RAP primarily reflects volume status rather than RV function.24 In contrast, during partial support with an ejecting ventricle, elevated RAP may reflect both RV dysfunction and volume overload. The limited ECMO sample size precludes a firm conclusion. Correlations among RV parameters were stronger in the Impella group. RV function parameters may remain physiologically more meaningful under Impella support, where blood continues to flow through the native cardiac chambers and pulmonary circulation, in contrast to ECMO, which bypasses the native circulation. RAP was the only parameter correlating with all other RV function parameters in the Impella group and correlating with all except TAPSE in the ECMO group. This supports its clinical utility as an indicator of RV compromise, even when direct RV function parameters are unavailable.23 Elevated RAP may represent such a robust marker because it integrates impaired RV forward output, increased venous backpressure and right sided congestion.7 The absence of systematic association between RV metrics and mortality, alongside their inconsistent intercorrelation, underscores the physiological limitations of individual RV parameters: TAPSE is load-dependent; PAPi is sensitive to pulmonary flow and pressure, TAPSE/sPAP depends on both; and RAP is influenced by volume status.16 This reinforces the value of the composite RVD definition used here. The main finding of this study is that in the Impella group, RVD was associated with increased 6-month mortality (43% vs. 30, P = 0.015), worse haemodynamic stabilization, and organ failure markers. The group with RVD was associated with higher initial lactate levels and worse LVEF compared to the group without RVD, indicating more critically ill patients with more advanced heart failure. RVD has been reported to identify patients at increased risk of death in CS.23 To the best of our knowledge, this is the first study to demonstrate this association specifically in a tMCS population on left-sided Impella support. The impact of Impella on RV function is multifaceted: it may impair RV function by increasing preload and septal shift, but it can also improve RV function by reducing afterload and improving LV function, thus improving ventricular interdependence.9,11,25,26 Despite RVD, only 6.7% of Impella patients required escalation to ECMO, consistent with preclinical evidence that LV unloading may partially mitigate RVD, though likely insufficient in severe cases.27,28 In the ECMO group, 6-month mortality was numerically higher with RVD (61% vs. 45%, P = 0.13), without statistical significance. Conventional RV parameters may have limited prognostic value under ECMO, as device flow substantially alters RV loading conditions.29 We nevertheless believe that the assessment of RV function parameters remains clinically meaningful, especially under partial VA-ECMO support. Finally, when haemodynamically feasible, a dynamic assessment during ECMO weaning, integrating multiple RV function parameters, may appear most informative.18,30 Limitations This registry-based study has several limitations. The cohort was heterogeneous in terms of CS aetiology, shock severity, and management across centres. Competing risks and noncardiac causes of death may have confounded the mortality analysis, despite efforts to exclude patients who died from noncardiac causes while the device was in situ. Missing RV function data predominantly affected patients under Impella support (43.6% vs. 20% in the ECMO group) and those with less severe CS stages (see supplementary material online, Figures S1 and S2). Lower prevalence of RVD in this subgroup is likely, suggesting a slight overestimation of overall RVD rates in our study. However, this limitation is unlikely to have substantially distorted the observed association between RVD and mortality, as it would tend to attenuate the association by selecting less severely ill patients. Our broad definition of RVD (≥1 pathological RV parameter) might also have overestimated RVD prevalence, though individual parameter analysis supports the overall conclusion. The ECMO group had limited statistical power, with only 33 patients classified as having normal RV function. The absence of a statistically significant mortality association should therefore be interpreted as inconclusive. Generalizability may be limited to other cohorts and health care systems, since management can differ. Future studies with standardized RV-function-monitoring protocols, with serial RV function assessments and larger cohorts are needed to confirm our findings, exclude possible confounding factors, and clarify the prognostic role of RVD in ECMO-supported patients. Conclusion In this multicentre registry, RVD was present in one-third of Impella and 60% of ECMO-supported CS patients. Among patients on Impella, RVD was associated with increased 6-month mortality and worse haemodynamic stabilization—a novel finding in this specific population. The prognostic value of conventional RV parameters appears attenuated under ECMO, where the haemodynamic context fundamentally differs. Systematic RV assessment at the time of tMCS implantation should be considered standard practice to guide device selection, escalation and weaning. Supplementary Material Contributor Information Jan Estoppey, Department of Cardiology, Inselpital, Bern University Hospital, Freiburgstrasse 10, 3010 Bern, Switzerland. Iulia Maia Muresan, Department of Clinical Research, University of Bern, Freiburgstrasse 3, 3010 Bern, Switzerland. Matthias Bossard, Cardiology Division, Heart Center Lucerne, Luzerner Kantonsspital, Kantonsspital 37, 6004 Lucerne, Switzerland; Faculty of Health Sciences and Medicine, Alpenquai 4, University of Lucerne, 6005 Lucerne, Switzerland. Gregor Leibundgut, Department of Cardiology, University Hospital Basel, Petersgraben 4, 4031 Basel, Switzerland. Gregor Fahrni, Department of Cardiology, Stadtspital Zürich, Birmensdorferstrasse 497, 8063 Zürich, Switzerland; University of Basel, Petersplatz 1, 4001 Basel, Switzerland. Stefan Blöchlinger, Department of Cardiology, Inselpital, Bern University Hospital, Freiburgstrasse 10, 3010 Bern, Switzerland; Department of Cardiology, Kantonsspital Winterthur, Brauerstrasse 15, 8400 Winterthur, Switzerland. Roger Ludwig, Department of Intensive Care Medicine, Inselspital, Bern University Hospital, Freiburgstrasse 10, 3010 Bern, Switzerland. Anna S Messmer, Department of Intensive Care Medicine, Inselspital, Bern University Hospital, Freiburgstrasse 10, 3010 Bern, Switzerland. Valentin Köchli, Department of Cardiology, Inselpital, Bern University Hospital, Freiburgstrasse 10, 3010 Bern, Switzerland. Thomas Seiler, Department of Cardiology, Inselpital, Bern University Hospital, Freiburgstrasse 10, 3010 Bern, Switzerland; Cardiology Division, Heart Center Lucerne, Luzerner Kantonsspital, Kantonsspital 37, 6004 Lucerne, Switzerland. Armin Spahr, Cardiology Division, Heart Center Lucerne, Luzerner Kantonsspital, Kantonsspital 37, 6004 Lucerne, Switzerland; Faculty of Health Sciences and Medicine, Alpenquai 4, University of Lucerne, 6005 Lucerne, Switzerland. Alain M Bernheim, Department of Cardiology, Stadtspital Zürich, Birmensdorferstrasse 497, 8063 Zürich, Switzerland. Anna Schmidt, Department of Cardiology, Kantonsspital Winterthur, Brauerstrasse 15, 8400 Winterthur, Switzerland; Pediatric Intensive Unit, Great Ormond Street Hospital NHS Foundation Trust, London, WC1N3JH, UK. Florim Cuculi, Cardiology Division, Heart Center Lucerne, Luzerner Kantonsspital, Kantonsspital 37, 6004 Lucerne, Switzerland; Faculty of Health Sciences and Medicine, Alpenquai 4, University of Lucerne, 6005 Lucerne, Switzerland. Lukas Hunziker, Department of Cardiology, Inselpital, Bern University Hospital, Freiburgstrasse 10, 3010 Bern, Switzerland. Monika Fürholz, Department of Cardiology, Inselpital, Bern University Hospital, Freiburgstrasse 10, 3010 Bern, Switzerland. Supplementary material Supplementary material is available at European Heart Journal Open online. Author contributions Monika Fürholz (Conceptualization [equal], Supervision [lead], Writing—original draft [supporting]), Jan Estoppey (Conceptualization [equal], Writing—original draft [lead]), Iulia Maia Muresan (Formal analysis [lead]), Matthias Bossard (Writing—review & editing [supporting]), Gregor Leibundgut (Writing—review & editing [supporting]), Gregor Fahrni (Writing—review & editing [supporting]), Stefan Blöchlinger (Writing—review & editing [supporting]), Roger Ludwig (Writing—review & editing [supporting]), Anna S. Messmer (Writing—review & editing [supporting]), Valentin Köchli (Data curation [supporting]), Thomas Seiler (Data curation [supporting]), Armin Spahr (Data curation [equal]), Alain M. Bernheim (Writing—review & editing [supporting]), Anna Schmidt (Data curation [supporting]), Florim Cuculi (Writing—review & editing [supporting]), and Lukas Hunziker (Conceptualization, Writing—review & editing [supporting], Supervision [equal]) Funding No funding The other authors report no conflict. References Associated Data Supplementary Materials