Cardiogenic shock (CS) remains the leading cause of death among patients hospitalised for acute myocardial infarction (AMI), with mortality rates of approximately 50%.1 Even with the benefit of thrombolytic therapy and intra-aortic balloon pumps (IABPs), the in-hospital mortality rate of patients with CS from AMI was only lowered from 63 to 47% according to a report from the SHould we emergently revascularize Occluded Coronaries in cardiogenic shocK? (SHOCK) registry.2 Despite early revascularisation with primary percutaneous coronary intervention (PCI) or coronary artery bypass surgery (CABG), a 30-day mortality rate of 47% was reported in the SHOCK trial.3 IABPs do not provide good haemodynamic support to those patients with severely low left ventricular function.4 Thus, there has been recent interest in percutaneous left ventricular assist devices (pLVADs) that can provide greater haemodynamic support than the IABP and, hopefully, a greater mortality benefit.
TandemHeart
The TandemHeart percutaneous left ventricular assist device (Cardiac Assist Technology, Inc.) is a left-atrial-to-femoral-artery bypass system that provides short-term haemodynamic support for patients with CS.5 It can provide haemodynamic support for a few days or act as a stopgap until surgical treatment is undertaken.6 Specialised TandemHeart cannulae are used to cannulate the femoral artery and vein. The venous cannula is advanced to the right atrium via the inferior vena cava and then into the left atrium through a transseptal approach. The device’s inflow and outflow are connected after proper positioning of the cannulae in the left atrium and femoral artery. Anticoagulation therapy is used to prevent haemolysis and thromboemboli.7 The most common complications include bleeding, cannula dislodgement, a femoral arteriovenous (AV) fistula, thromboembolism, atrial–septal defect (ASD), limb ischaemia and wound infection.
TandemHeart Randomised Trials
Thiele et al.8 investigated the TandemHeart device in a randomised study against an IABP from August 2000 to December 2003 at a single site in Germany. The inclusion criteria were the presence of CS complicating AMI with the intention to revascularise the infarcted artery by PCI. Glycoprotein IIb/IIIa receptor antagonists were also recommended for thrombolysis in MI (TIMI) flow less than III or evidence of intracoronary thrombus after PCI.
CS was defined by persistent systolic blood pressure <90mmHg or vasopressors required to maintain blood pressure >90mmHg, evidence of end-organ failure, evidence of elevated left ventricular filling pressure and cardiac index <2.1l/minute/m2. The exclusion criteria were age >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; the exclusion criteria eliminated >50% of patients in CS in the study. The haemodynamic parameters measured were cardiac output, mean blood pressure, cardiac power index, mean pulmonary artery pressure, central venous pressure and heart rate. The metabolic parameters measured were standard base excess, serum lactate and pH. The study attempted to avoid limb ischaemia in smaller patients by recommending the use of 12Fr arterial cannulae instead of the usual 17Fr used in most patients, therefore limiting the flow from the system to 3l/minute compared with 4l/minute. The cannulae were removed using a surgical approach in the first eight patients and subsequent removal was manual using a compression system.
The median time for implantation of the TandemHeart was longer than that for the IABP: 25 versus 11.5 minutes. The median duration of cardiac support for the device was four days in the IABP and 3.5 days in the pLVAD group. Ninety-five per cent of patients in the study underwent PCI and one patient in each treatment group underwent CABG. Revascularisation was performed prior to device implantation in 55 and 57% of the IABP and TandemHeart groups, respectively.
The primary end-point was haemodynamic improvement as defined by improvement in cardiac power index (cardiac output x mean arterial pressure x 0.0022) within two hours after device implantation. The primary end-point of cardiac power index was statistically improved for the TandemHeart group compared with the IABP group. Secondary end-points were all other haemodynamic and metabolic parameters as well as mortality at 30 days and device-related complications.
In the TandemHeart group, there was a statistically significant increase in the development of limb ischaemia and transfusion of blood products (packed red blood cells, fresh frozen plasma and platelets). More patients in the TandemHeart group had signs of disseminated intravascular coagulation (DIC) than in the IABP group, and the DIC was mild in all cases in the IABP group while a significant number of patients in the TandemHeart group had severe haemorrhagic diathesis. There was no statistical difference in metabolic parameters between the groups. There was no 30-day mortality difference for patients with assist support pre-PCI versus post-PCI. Overall, 30-day mortality was 45% in the IABP group and 43% in the TandemHeart group.
Burkhoff et al.9 explored the safety and efficacy of the TandemHeart compared with an IABP in a prospective randomised study at 12 sites in the US and Switzerland from April 2002 to April 2004. A major difference compared with the study by Thiele et al. was the inclusion of patients who already had an IABP if they continued to meet haemodynamic criteria for CS. Furthermore, sites not having prior experience with the TandemHeart device were permitted to treat the first patient with the TandemHeart in a roll-in phase before initial randomisation. The study included a total of 42 patients of whom 19 were randomised to the TandemHeart device, nine were rolled in with the TandemHeart device and 14 were randomised to the IABP.
Inclusion criteria were patients ≥18 years of age having cardiac index ≤2.2l/m2 per minute, mean arterial pressure ≤70mmHg, pulmonary capillary wedge pressure ≥15mmHg and evidence of end-organ damage or need for high-dose pressor and/or inotropic support. Exclusion criteria were isolated right heart failure, coagulopathy, sepsis, severe peripheral vascular disease, stroke within six months, 2+ or greater aortic regurgitation and ventricular septal rupture.
The study was stopped early by the recommendation of an independent data safety monitoring board, which concluded that the haemodynamic effects were superior in the TandemHeart group compared with the IABP group, and the slow rate of patient accrual made it unlikely enough patients would be enrolled in a reasonable time to determine mortality benefit. A statistically significant haemodynamic improvement was seen in cardiac index and mean arterial pressure, with a significant decrease in pulmonary capillary wedge pressure in the pLVAD group compared with the IABP group.
The study acknowledged its inability to make conclusions about the haemodynamic effectiveness of the IABP because patients who showed significant haemodynamic support from the IABP would not be enrolled in the study. The TandemHeart group had more adverse effects of limb ischaemia, arrhythmias, bleeding and infection compared with the IABP. There was one case of TandemHeart failure and one device removal due to device-related problems with clotting within the cannula. The overall 30-day mortality rate was 36% in the IABP group and 46% in the TandemHeart pooled group (roll-in and randomised).
Impella LP 2.5
The Impella LP 2.5 (Abiomed, Inc.) is a catheter-based assist device that acts as a microaxial rotary blood pump, bringing blood from the left ventricle to the ascending aorta to provide left ventricular support.10,11 The device is implanted via the femoral artery to the left ventricle. There is still a need for anticoagulation therapy to prevent haemolysis and thromboemboli, but, contrary to the TandemHeart device, there is no transseptal puncture required in the implantation process. There is a larger device, the Impella LP 5.0 system, which can provide 5l/minute compared with the 2.5l/minute provided by the Impella LP 2.5 system; however, it requires femoral artery surgery for placement and thus is mainly used after cardiac surgery.12
Impella LP 2.5 Randomised Trial
Seyfarth et al.13 performed a prospective, randomised, two-centre trial in Germany from September 2004 to January 2007. Inclusion criteria were patients ≥18 years of age who had AMI within 48 hours and CS onset within 24 hours defined by clinical or haemodynamic criteria. The clinical criteria were systolic blood pressure ≤90mmHg or need for positive inotropic drug to maintain blood pressure ≥90mmHg and end-organ damage. The haemodynamic criteria were cardiac index ≤2.2l/m2 per minute and pulmonary capillary wedge pressure ≥15mmHg or angiographically measured left ventricular ejection fraction of ≤30% and left ventricular end-diastolic pressure >20mmHg. The exclusion criteria were prolonged resuscitation (>30 minutes), hypertrophic obstructive cardiomyopathy, definite thrombus in the left ventricle, treatment with an IABP, severe valvular disease or mechanical heart valve, ventricular septal defect, acute 2+ mitral regurgitation, right heart failure, sepsis, known cerebral disease, bleeding requiring surgical intervention, pulmonary embolism, coagulopathy, 2+ aortic regurgitation or pregnancy.
The median time for implantation of the device was longer for the Impella group than for the IABP group: 22±9 and 14±8 minutes, respectively. PCI was performed in 92% of patients in each group. All devices were implanted after PCI.
The primary end-point of improvement in cardiac index was statistically greater in the Impella group over the IABP group 0.49 and 0.11l/minute/m2, respectively. There was also a statistical rise in observed diastolic arterial pressure and a trend towards increased mean arterial pressure in the Impella group on analysis of secondary haemodynamic end-points. There was no device-related technical failure, major bleeding or ischaemia during support with the device. However, there was a case of acute limb ischaemia after Impella removal. There was a significantly higher rate of haemolysis and blood product use (packed red blood cells and fresh frozen plasma) in the Impella group. The overall 30-day mortality rate was 46% in both groups.
Meta-analysis of TandemHeart and Impella LP 2.5
Cheng et al.14 pooled data from controlled trials of pLVAD versus IABP to compare haemodynamic parameters, 30-day mortality and adverse events. The only studies that met the inclusion criteria for the metaanalysis were the two studies described above of the TandemHeart device and the single study of the Impella LP 2.5 system. There were a total of 100 patients pooled for the meta-analysis, of whom 53 had implantation of a pLVAD while 47 had implantation of an IABP. Figure 1 represents the results of a mean difference and random effects model of the pooled studies in the meta-analysis. The consensus is that patients receiving pLVAD compared with IABP had higher cardiac index, higher mean arterial pressure and lower pulmonary capillary wedge pressure. Pooled 30-day mortality relative risk showed no difference between patients receiving a pLVAD compared with an IABP (see Figure 2). Figure 3 indicates the relative risk of adverse events using a random effects model. There was a higher incidence of leg ischaemia with all pLVAD and more frequent bleeding in patients receiving the TandemHeart device. The major limitation of the meta-analysis that the authors acknowledge was possible type II error due to the inclusion of only a small number of trials. Overall, the conclusion of the meta-analysis was that pLVAD provides better haemodynamic support, but it must be weighed against the higher rate of adverse events and the lack of 30-day mortality benefit.
Reitan Catheter Pump
A new device on the horizon is the Reitan catheter pump (Jomed),15,16 a propeller-based pump placed in the high descending aorta that can reach 14,000 revolutions per minute. It has been shown to decrease diastolic coronary flow and carotid flow, but did not improve cardiac output in an animal model of CS with acute mitral regurgitation.17 The advantage of the Reitan catheter pump versus the TandemHeart or Impella system is that it can be placed in the presence of aortic regurgitation or left ventricular thrombus as it is not implanted in the left ventricular cavity.
Conclusion
The American College of Cardiology/American Heart Association (ACC/AHA) ST-elevation MI (STEMI) guidelines18 recommend early revascularisation and IABP support for patients presenting with CS caused by AMI. As high-risk PCIs are attempted there will be a need for greater haemodynamic support than is available with the IABP. pLVADs have shown promising results with haemodynamic support; however, the mortality benefit remains to be determined. More widespread application of timely reperfusion to prevent CS in acute coronary syndrome patients, including pre-hospital triage (electrocardiogram), effective nationwide STEMI networks and round-the- clock door-to-balloon interventional teams, may have the greatest impact on reducing mortality from current levels. Ôûá