Thiele 2005.

Study characteristics
Methods	Study design: RCT Study grouping: parallel group Date of study: August 2000 to December 2003 Number of study centres: 1 Study setting: Department of Internal Medicine/Cardiology, University of Leipzig‐Heart Centre, Strumpellstr. 39, 04289 Leipzig, Germany Total duration of the study: 3 years and 5 months
Participants	Baseline characteristics TandemHeart VAD Age (median): 63 (IQR 57–70) years Male gender, n (%): 16 (76) Diabetes, n (%): 11 (52) Smoking, n (%): 9 (43) pH (median): 7.28 (IQR 7.24–7.36) Lactate prior to intervention (median): 4.5 (3.1–6.5) Lung function: not documented Cause of CS: AMI IABP Age (median): 65 (IQR 59–73) years Male gender, n (%): 15 (75) Diabetes, n (%): 11 (55) Smoking, n (%): 6 (30) pH (median): 7.34 (IQR 7.28–7.38) Lactate prior to intervention (median): 3.8 (3.5–6.7) Lung function: not documented Cause of CS: AMI Baseline clinical characteristics were similar for both groups. Baseline haemodynamic characteristics were similar for both groups except for a higher PCWP in the IABP group (MCS (median): 20.0 (IQR 18–23); IABP 27 (IQR 20–30); P = 0.02). Inclusion criteria: he presence of CS complicating AMI and intention to revascularise the infarcted artery by PCI as first‐line treatment option. CS identified as: persistent systolic blood pressure < 90 mmHg or vasopressors required to maintain blood pressure > 90 mmHg; evidence of end organ failure (e.g. urine output < 30 mL/hour, cold skin and extremities, and serum lactate > 2 mmol/L); evidence of elevated left ventricular filling pressures (pulmonary congestion or PCWP > 15 mmHg); and Cardiac Index > 2.1 L/min/m². Exclusion criteria: aged > 75 years; mechanical complications of AMI; duration of CS > 12 hours; right heart failure; sepsis; significant aortic regurgitation; severe cerebral damage; resuscitation > 30 minutes; severe peripheral vascular disease and other diseases with reduced life expectancy. Number of participants analysed: 41 Number of participants lost to follow‐up: 0 1 participant did not receive the MAD as the patient showed rapid haemodynamic improvement after PCI but was included in the final analysis according to intention‐to‐treat principles. Number of participants randomised to intervention: 41
Interventions	Intervention characteristics MCS Type of MCS: TandemHeart VAD; after transeptal puncture a venous inflow cannula was inserted into left atrium. Oxygenated blood was drawn and returned via a centrifugal pump and an arterial cannula in the femoral artery to the lower abdominal aorta. To avoid limb ischaemia in smaller participants, 2 arterial cannulae of 12‐French in both femoral arteries were recommended. System was capable of delivering flow up to 4.0L/minute at 7500 rpm but the 12‐French cannula limited flow to 3.0 L/minute. Duration (diagnosis to intervention) (median): 11.0 (IQR 6.8–18.80) minutes Duration of treatment (median): 3.5 (IQR 2.0–4.5) days Inotropes or vasopressors (median): 3.0 (IQR 1.0–3.0) days IABP Type of IABP: Datascope Corporation IABP inserted percutaneously, heparin administered intravenously and all participants were initially on a pumping ratio of 1:1 with 100% balloon inflation. Duration (diagnosis to intervention) (median): 10.0 (IQR 5.3–25.5) minutes Duration of treatment (median): 4.0 (IQR 3.5–4.0) days Inotropes or vasopressors (median): 2.0 (IQR 1.0–4.0) days
Outcomes	Haemodynamic outcomes CPI improved by MCS from 0.22 (IQR 0.19–0.30) W/m² to 0.37 (IQR 0.30–0.47) W/m² (P < 0.001) when compared with IABP (from 0.22 (IQR 0.18–0.30) W/m² to 0.28 (IQR 0.24–0.36) W/m²; P = 0.02; P = 0.004 for intergroup comparison). Time of first CPI measurement after device implantation was similar (MCS: 40.0 (IQR 25.5–53.5) minutes; IABP: 33.5 (IQR 19.0–50.5) minutes; P = 0.28). Overall outcomes Survival 30‐day survival: MCS: 57% survival; IABP: 55% survival Survival measured to: transplant: unsupported cardiac function Not described Quality of life Not reported Major adverse cardiovascular events In the MCS group, 7 participants developed limb ischaemia after implantation of a 17‐French arterial cannula and 0 in the IABP group (P = 0.009). Dialysis‐dependent Not reported Length of hospital stay Not reported Length of ICU stay Not reported Major adverse events 13 participants in the MCS group and 3 in the IABP group had signs of DIC. In 8 MCS participants, DIC was severe with subsequent haemorrhagic diathesis. In 3 IABP participants, DIC was mild and could be resolved by substitution with antithrombin III.
Identification	Sponsorship source: Cardiac Assist, Pittsburgh Country: Germany Setting: Department of Internal Medicine/Cardiology, University of Leipzig‐Heart Centre, Germany Comments: Authors names: Holger Thiele*, Peter Sick, Enno Boudriot, Klaus‐Werner Diederich, Rainer Hambrecht, Josef Niebauer, and Gerhard Schule Institution: Department of Internal Medicine/Cardiology, University of Leipzig‐Heart Centre Email: thielh@medizin.uni‐leipzig.de Address: Department of Internal Medicine/Cardiology, University of Leipzig‐Heart Centre, Strümpellstraße 39, 04289 Leipzig, Germany
Notes	Outcomes Cause of death: 4 participants in each group 3 in IABP due to LHF in the first 2 hours then 1 within 24 hours of PCI due to MODS. In the MCS group, 0 died within the 24 hours. However, 4 participants died between days 2 and 4 as a cause of MODS despite active circulatory support. 5 additional participants died in each group after weaning during 30‐day follow‐up, resulting in an overall mortality of 45% in the IABP group and 43% in the MCS group (log‐rank, P = 0.86). In the MCS group, 3 participants died after weaning due to recurrent LHF, and 2 due to sepsis or MODS. The cause of death after IABP explantation was MODS in 3 and acute LHF in 2 participants. There were no 30‐day mortality differences for participants with pre‐PCI assist support (MCS: 44%; IABP: 56%) and for those with post‐PCI support (MCS: 42%; IABP: 36%).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomised by drawing sealed envelopes.
Allocation concealment (selection bias)	Low risk	Study specified that some appropriate safeguards were taken, i.e. sealed envelopes, but did not mention if they were non‐opaque or not sequentially numbered.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Study did not describe whether the participants and personnel were blinded or not.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No blinding of outcomes assessors described.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing data outcomes
Selective reporting (reporting bias)	Low risk	All of the study's prespecified (primary and secondary) outcomes that were of interest in the review were reported in the prespecified way.
Other bias	Low risk	Inherent risk of bias was minimised as much as possible.

See Appendix 2 for a glossary of terms.
AMI: acute myocardial infarction; CS: cardiogenic shock; CPI: Cardiac Power Index; CVS: cardiovascular system; DIC: disseminated intravascular coagulation; IABP: intra‐aortic balloon pump; IAVP: implantable aortic valvo‐pump; ICU: intensive care unit; IQR: interquartile range; LHF: left heart failure; MAD: mechanical assist device; MAP: mean arterial blood pressure; MCS: mechanical circulatory support; MODS: multiorgan dysfunction syndrome; N/A: not applicable; PCI: percutaneous coronary intervention; PCWP: pulmonary capillary wedge pressure; pMCS: percutaneous mechanical circulatory support; pVAD: percutaneous ventricular assist device; RCT: randomised controlled trial; rpm: revolutions per minute; SD: standard deviation; STEMI: ST elevation myocardial infarction; VAD: ventricular assist device.