Interventional procedure overview of percutaneous insertion of a catheter-based intravascular microaxial flow pump for cardiogenic shock
Closed for comments This consultation ended on at Request commenting lead permission
Evidence summary
Population and studies description
This interventional procedures overview is based on over 950,000 people from multiple studies including 1 randomised controlled trial (Møller 2024), 1 individual patient data meta-analysis (Thiele 2024), 5 systematic reviews and meta-analyses (Sassani 2025, Ardito 2023, Panuccio 2022, Bogerd 2025 and Stub 2025), and 3 registry studies (Movahed 2024, Padberg 2024 and Higuchi 2024). Of these, over 100,000 people had the procedure. There was some overlap in studies included in the systematic reviews, and the studies by Møller 2024 and Padberg 2024 were both included in at least 1 of the reviews. This interventional procedures overview is a rapid review of the literature, and a flow chart of the complete selection process is shown in figure 1. This overview presents 10 studies as the key evidence in table 2 and table 3, and lists 138 other relevant studies in appendix B, table 5, including 20 case reports of adverse events (table 5a).
The evidence included studies that used Impella CP, 2.5 and 5.0 models. Impella 2.5 has lower maximum flow rate than Impella CP and they are both typically used for shorter durations than Impella 5.5.
Evidence was excluded on catheter-based intravascular microaxial flow pumps that were described as being inserted surgically or when Impella 5.5 was used, which is inserted surgically. Evidence was excluded from studies that primarily used microaxial flow pumps as support during high-risk PCI, or for left ventricular unloading during VA-ECMO support.
The evidence included people with cardiogenic shock from various aetiologies, the most common being AMICS. People often had PCI as part of their treatment, either before or after percutaneous insertion of a catheter-based intravascular microaxial flow pump. In most studies, the mean or median age was above 60 years and there was a higher proportion of males than females. Studies included evidence from Europe, North and South America, Asia and Africa.
The randomised controlled trial by Møller (2024) included 355 people with STEMI related cardiogenic shock, mostly from Denmark and Germany, who had a microaxial flow pump (Impella CP) plus standard care or standard care alone. The median age was 67 years and 79% of the study population was male. The median LVEF at baseline was 25%, median arterial lactate level was 4.5 mmol per litre and 56% of people were SCAI-CSWG stage C, 28% were stage D and 16% were stage E. Randomisation to treatment groups was done before or after PCI. The study had strict inclusion and exclusion criteria. Notably, people who had been resuscitated from out-of-hospital cardiac arrest and remained comatose on arrival to the cardiac catheterisation laboratory and people with overt right ventricular failure were excluded. The primary end point was death from any cause at 180 days. A composite safety end point was severe bleeding, limb ischaemia, haemolysis, device failure, or worsening aortic regurgitation.
The individual patient data meta-analysis by Thiele (2024) included data from 1,059 people in 9 randomised controlled trials, 1 of which was Møller (2024). The study compared early routine use of active MCS with best medical therapy in people with AMICS having revascularisation. The evidence included several types of MCS, including VA-ECMO, and about 40% used a microaxial flow pump. People with non-myocardial infarction causes of cardiogenic shock were excluded. The median age was 65 years and 80% of the study population was male. The median lactate at baseline was 5.6 mmol per litre. About half the study population had successful resuscitation before inclusion. The primary outcome was 6-month all-cause mortality.
The systematic review by Sassani (2025) included 2,617 people from 18 studies on cardiogenic shock related to acute coronary syndrome, 4 of which were prospective. The mean age was 64.7 years. In most of the studies, lactate levels at baseline varied from 4.7 to 8.6 mmol per litre. Where reported, the proportion of people with previous resuscitation ranged from 23% to 74%. The evidence included 2 types of microaxial flow pump (Impella 2.5 and Impella CP). The primary outcome was 30-day mortality.
The systematic reviews by Ardito (2023), Bogerd (2025) and Stub (2025) all compare microaxial flow pumps and VA-ECMO. Ardito (2023) included 44,951 people with cardiogenic shock from any cause, 13,848 of whom had a microaxial flow pump. The review included 102 studies, most of which were retrospective and observational. The mean age of the study populations ranged from 31 to 78 years. Most of the papers on microaxial flow pumps used Impella CP (41 papers), followed by Impella 2.5 (25 papers) and Impella 5.0 (20 papers). One paper used Impella 5.5, which is inserted surgically and so not within the remit of this assessment.
Bogerd (2025) included 108,736 people with AMICS, 92,262 of whom had a microaxial flow pump or percutaneous ventricular assist device. The review included 13 studies, describing 12 retrospective cohorts. The mean or median age ranged from 55 to 80.6 years and 73% of the study population was male. Outcomes were reported to hospital discharge. The type of microaxial flow pump used in each study was not specified in the review.
The systematic review by Stub (2025) included 5 propensity score-matched or adjusted studies with 9,871 people who had a microaxial flow pump or VA-ECMO. The matched sample included 984 people who had a microaxial flow pump. The most common aetiology of shock in the included studies was AMICS. The mean age was consistent across studies, ranging from 60.0 to 67.7 years and the proportion of males ranged from 70% to 86%. Of the 5 studies, 3 specified that the microaxial flow pump used was Impella CP or 2.5. Follow-up was in-hospital or 30 days.
The systematic review by Panuccio (2022) compared microaxial flow pump and IABP, and included 7,290 people with cardiogenic shock, most of whom had an acute coronary syndrome as the underlying cause. The review included 33 observational studies, describing 5,203 people who had a microaxial flow pump and 2,087 who had IABP. The main types of microaxial flow pump used were Impella CP (50%) and 2.5 (37%). The primary endpoint was 30-day or in-hospital mortality.
The studies by Movahed (2024), Padberg (2024) and Higuchi (2024) are all multicentre registries. Movahed (2024) used the US Nationwide Inpatient Sample database from 2016 to 2020 and included cardiogenic shock from any cause. A total of 844,020 people were included, 39,645 of whom had a microaxial flow pump alone and 6,790 people had a microaxial flow pump with IABP. The mean age was 66.4 years and 62% of the study population was male. The primary outcome was in-hospital mortality. The type of microaxial flow pumps was not specified but the paper describes the devices as being inserted percutaneously through the femoral artery.
Padberg (2024), which was also included in the review by Bogerd (2025), assessed data from a German registry from 2010 to 2017. It included 39,864 people with AMICS, 776 of whom had a microaxial flow pump. The proportion of males was 61% overall and 69% in those who had temporary MCS. Outcomes for people who had a microaxial flow pump were compared to those who had VA-ECMO, IABP or no temporary MCS. People with out-of-hospital cardiac arrest before admission were not included in the data. The primary endpoint was in-hospital survival, and longer term estimates up to 8 years were reported. The type of microaxial flow pumps was not specified and it was not specified if they were all inserted percutaneously.
Higuchi (2024) included data from a Japanese nationwide registry from 2020 to 2022. It included 5,718 people with drug refractory heart failure who had a microaxial flow pump, most of which were Impella CP. The median age overall was 69 years and the cohort was divided into a younger group (less than 75 years) and an older group (75 years or above). There was a higher proportion of males in the younger group (81% versus 69%, p<0.0001). Most of the devices (93%) were inserted using transfemoral access. Impella combined with ECMO was used in 40% of people and 64% of the cohort had PCI. The primary outcome was 30-day mortality.
In addition to the key evidence summarised in table 2 and table 3, individual case reports describing adverse events associated with the procedure have been listed in table 5a in appendix B.
Table 2 presents study details.
Procedure technique
All catheter-based intravascular microaxial flow pumps were Impella devices (Abiomed). Of the 10 key studies, 7 specified the type of Impella devices used. In the randomised controlled trial by Møller (2024), Impella CP was the main intervention. A small proportion of people in the treatment and control groups had escalation to Impella 5.0, which has a higher flow rate. The other studies that reported the type of Impella mainly used Impella CP or Impella 2.5. In the systematic review by Ardito (2023), 41 papers used Impella CP, 25 used Impella 2.5 and 20 used Impella 5.0. The studies did not describe the technique of inserting the microaxial flow pump in detail, but none of them described surgical implantation as the main technique.
Efficacy
In-hospital or 30-day mortality
In-hospital or 30-day mortality was reported in 9 studies.
In the individual patient data meta-analysis of 9 randomised controlled trials by Thiele (2024), 30-day mortality was 44% (95% CI 40 to 48) for the MCS group, which included VA-ECMO as well as microaxial flow pumps, and 47% (95% 43 to 52%) for the standard care group (HR 0.89, 95% CI 0.74 to 1.06, p=0.10).
In the systematic review of 18 studies on cardiogenic shock related to acute coronary syndrome, the mean 30-day mortality was 45% (range 28 to 67%). The meta-analysis result was 45% (95% CI 43 to 47) but heterogeneity was high (I2=90%) and remained high after sensitivity or leave-out analyses (Sassani 2025). In the systematic review of 13 studies on AMICS, in-hospital mortality was 44% for the microaxial flow pump group and 57% for the VA-ECMO group (RR=0.84, 95% CI 78 to 86, p<0.001, 11 studies, I2=86%; Bogerd 2025). In the systematic review of 5 propensity score matched or adjusted studies on cardiogenic shock from any aetiology, the odds of in-hospital or 30-day mortality was statistically significantly reduced in the microaxial flow pump group compared to VA-ECMO (OR 0.57, 95% CI 0.44 to 0.74, p<0.0001, I2=27%). In absolute terms, the proportion of people who had a microaxial flow pump and died in-hospital or within 30 days (40%) was lower than the proportion of those who had VA-ECMO (54%; Stub 2025). In the systematic review of 33 studies on cardiogenic shock, 30-day or in-hospital mortality for the microaxial flow pump group was 47% (95% CI 44 to 50, I2=80%. Meta-regression analysis showed that older age was associated with a negative impact on the outcome (p=0.01). There was a significant interaction between the percentage of people having higher MCS (Impella CP or Impella 5.0) and short-term mortality (p=0.004), suggesting a larger benefit with higher MCS. Positioning of a microaxial flow pump before starting PCI was associated with lower mortality. In the comparative analysis of 7 studies, short-term mortality was 46% for the microaxial flow pump group and 37% for the IABP group (RR 1.08, 95% CI 0.89 to 1.31, p=0.45, I2=67%; Panuccio 2022).
In the retrospective registry study by Movahed (2024), in-hospital mortality was 34% overall, 34% in people who had no MCS, 41% in people who had Impella support (OR 1.32, 95% CI 1.26 to 1.39, p<0.001), 25% for people who had IABP (OR 0.65, 95% CI 0.62 to 0.67, p<0.001) and 47% for people who had both Impella and IABP (OR 2.12, 95% CI 1.89 to 2.37, p<0.001). Using multivariate analysis adjusting for multiple demographics and patient-specific factors, Impella remained associated with the highest mortality (OR 1.33, 95% CI 1.25 to 1.41, p<0.001), whereas IABP remained associated with the lowest mortality (OR 0.69, 95% CI 0.66 to 0.72, p<0.001). In the retrospective registry study by Padberg (2024), in-hospital mortality was 62% for people who had Impella, 59% for those who had VA-ECMO, 47% for those who had IABP and 59% for those who had no MCS (p<0.001). Compared with no temporary MCS, there was no statistically significant difference in mortality in either the Impella (p=0.09) or the VA-ECMO subgroups (p=0.79). In the registry study by Higuchi (2024), 30-day mortality was 35% overall, 32% in the younger group and 39% in the older group (p<0.0001). When Impella alone was used, 30-day mortality was 24% compared to 50% when Impella combined with ECMO was used (p<0.0001). Age was a statistically significant predictor of mortality (HR for age 75 years or above was 1.83, 95% CI 1.56 to 2.16, p<0.001).
In the systematic review by Ardito (2023) of 102 studies on people with cardiogenic shock, an analysis of qualitatively labelled mortality outcomes was reported (at discharge, in hospital, to the next therapy, to explant, and on device). Mortality was 45% for microaxial flow pumps (95% CI 40 to 49, 39 papers, I2=96%) and 49% for VA-ECMO (95% CI 45 to 53; 40 papers, I2=95%) although there was a high degree of heterogeneity and 95% CIs overlapped.
6-month mortality
Mortality at 6 months was reported in 2 studies.
In the randomised controlled trial of 355 people, death from any cause at 180 days was statistically significantly lower in those who had a microaxial flow pump and standard care (46% [82 out of 179]) compared with those who had standard care alone (58% [103 out of 176]), in the intention to treat population (HR=0.74, 95% CI 0.55 to 0.99, p=0.04). The number needed to treat to avoid 1 death was 8 (Møller 2024). In the as-treated groups, the HR was 0.77, 95% CI 0.57 to 1.03.
In the individual patient data meta-analysis of 9 randomised controlled trials by Thiele (2024), there was no statistically significant difference in 6-month all-cause mortality between the MCS, which included VA-ECMO and Impella, and standard care groups in the intention to treat population (51% for the MCS group [95% CI 46 to 55] and 56% for the standard care group [95% CI 52 to 60]; HR 0.87 [95% CI 0.74 to 1.03], p=0.10). In the trials that used left ventricular unloading MCS, there was also no statistically significant difference in 6-month mortality (47% in the MCS group [95% CI 41 to 54] and 57% [95% CI 51 to 63] in the standard care group (HR 0.80, 95% CI 0.62 to 1.02, p=0.075). For people with STEMI without resuscitation or with only short resuscitation, the 6-month mortality was statistically significantly lower in the MCS group (44% [95% CI 39 to 50]) compared to the standard care group (55% [95% CI 50 to 61]; HR=0.77 [95% CI 0.61 to 0.97], p=0.024).
Longer term survival
Mortality or survival beyond 6 months was reported in 2 studies.
In the systematic review by Ardito (2023) of 102 studies on people with cardiogenic shock, overall mortality at 30 days, 6 months and 1 year combined was reported, although there was a high degree of heterogeneity. Mortality was 44% for microaxial flow pumps (95% CI 39 to 50, 42 papers, I2=91%) and 50% for VA-ECMO (95% CI 43 to 58; 30 papers, I2=94%).
In the retrospective registry study by Padberg (2024), Kaplan-Meier survival estimates for people who had a microaxial flow pump were 29% at 1 year, 28% at 2 years, 22% at 5 years and 11% at 8 years.
Bridge to recovery, durable LVAD or transplantation
In the systematic review by Ardito (2023) of 102 studies on people with cardiogenic shock, the rate of successful weaning was 58% (95% CI 49 to 66, 9 papers, I2=54%) for microaxial flow pumps and 53% (95% CI 47 to 60, 19 papers, I2=74%) for VA-ECMO. The proportion of people bridged to long-term LVAD was the same for the microaxial flow pump and VA-ECMO groups (10%), without statistically significant differences. Bridge to transplant was reported for 4% of people who had microaxial flow pump (95% CI 0 to 12%, 11 papers, I2=92%) and 6% for people who had VA-ECMO (95% CI 3 to 9%, 14 papers, I2=70%).
Composite outcomes
The randomised controlled trial of 355 people who had a microaxial flow pump and standard care or standard care alone reported a composite cardiac endpoint of escalation of treatment to additional MCS, heart transplantation, or death from any cause, whichever came first. This was 52% for the microaxial flow pump group and 64% for the standard care group (HR=0.72, 95% CI 0.55 to 0.95, p=0.04; Møller 2024).
Safety
Bleeding or vascular complications
Bleeding or vascular complications were reported as an outcome in 9 studies.
Moderate or severe bleeding was reported in a higher proportion of people in the microaxial flow pump group (22% [39 out of 179]) compared to the standard care alone group (12% [21 out of 176], RR 2.06, 95% CI 1.15 to 3.66) in the randomised controlled trial of 355 people (Møller 2024).
Moderate or severe bleeding at 30 days was reported in 28% (143 out of 506) of people who had MCS and 13% (67 out of 503) of people who had standard care (OR 2.64, 95% CI 1.91 to 3.65) in the individual patient data meta-analysis of 9 randomised controlled trials by Thiele (2024), which included the study by Møller (2024). The rate of access site bleeding was 16% in the microaxial flow pump group (95% CI 8 to 27, 8 studies, I2=92%) and 19% in the VA-ECMO group (95% CI 14 to 23, 5 studies, I2=26%) and major bleeding was 20% in the microaxial flow pump group (95% CI 12 to 30, 15 studies, I2=95%) and 25% in the VA-ECMO group (95% CI 16 to 34, 11 studies, I2=88%) in the systematic review of 102 studies (Ardito 2023). The odds of bleeding events needing transfusion were statistically significantly lower in the microaxial flow pump group compared to VA-ECMO (OR 0.61, 95% CI 0.46 to 0.80, p=0.0004, I2=0%) in the systematic review of 5 propensity score matched or adjusted studies (Stub 2025). The rate of major bleeding and vascular complications were both 16% in the systematic review of 33 studies, with high heterogeneity between studies. In the comparative analysis, the rate of vascular complications and major bleeding were statistically significantly higher in the microaxial pump group compared to the IABP group. The rate of vascular complications was 11% in the microaxial flow pump group and 3% in the IABP group (RR 3.32, 95% CI 2.54 to 4.33, p<0.001, 7 studies, I2=0%) and the rate of major bleeding was 28% in the microaxial flow pump group and 14% in the IABP group (RR 1.99, 95% CI 1.75 to 2.25, p<0.001, 6 studies, I2=0%; Panuccio 2022). The mean rate of bleeding was 14% (range 3 to 24%) in the systematic review of 18 studies (Sassani 2025).
Post procedural haemorrhage was reported in 2% of people who had a microaxial flow pump and 1% of people who had IABP (OR 1.99, p<0.001) and procedural bleeding was reported in 0.3% and 0.1%, respectively (OR 2.37, 95% CI 1.33 to 4.23, p=0.004) in the retrospective registry study by Movahed (2024). Bleeding was reported in 21% (162 out of 776) of people who had a microaxial flow pump, 38% (317 out of 833) of people who had VA-ECMO, 18% (980 out of 5,451) of people who had IABP and 9% (3,058 out of 32,804) of people who had no temporary MCS (p<0.001) in the retrospective registry study by Padberg (2024). Bleeding was reported in 16% (538 out of 3,409) of people and vascular injury was reported in 1% (30 out of 3,409) of people who had support with Impella alone in the registry study by Higuchi (2024).
Limb ischaemia
Limb ischaemia was reported as an outcome in 5 studies.
Limb ischaemia was reported in a higher proportion of people in the microaxial flow pump group (6% [10 out of 179]) compared to the standard care alone group (1% [2 out of 176], RR 5.15, 95% CI 1.11 to 23.8) in the randomised controlled trial of 355 people (Møller 2024).
Peripheral ischaemic vascular complication at 30 days was reported in 10% (50 out of 518) of people who had MCS and 2% (12 out of 516) of people who had standard care (OR 4.43, 95% CI 2.37 to 8.26) in the individual patient data meta-analysis of 9 randomised controlled trials (Thiele 2024). The rate of limb ischaemia was 6% in the microaxial flow pump group (95% CI 4 to 8, 19 studies, I2=61%) and 10% in the VA-ECMO group (95% CI 7 to 15, 23 studies, I2=90%) in the systematic review of 102 studies (Ardito 2023). The mean rate of ischaemia was 8% (range 3 to 14) in the systematic review of 18 studies (Sassani 2025). Limb ischaemia was reported in 4% (125 out of 3,409) of people who had support with Impella alone in the registry study by Higuchi (2024).
Stroke
Stroke was reported as an outcome in 5 studies.
Stroke was reported in 4% (7 out of 179) of people in the microaxial flow pump group and 2% (4 out of 176) of people in the standard care group (RR 1.75, 95% CI 0.50 to 6.01) in the randomised controlled trial of 355 people (Møller 2024).
Stroke at 30 days was reported in 4% (19 out of 498) of people who had MCS and 3% (13 out of 496) of people who had standard care (OR 1.48, 95% CI 0.72 to 3.04) in the individual patient data meta-analysis of 9 randomised controlled trials (Thiele 2024). The rate of ischaemic stroke was 0% in the Impella group (95% CI 0 to 0, 26 studies, I2=67%) and 7% in the VA-ECMO group (95% CI 5 to 10, 24 studies, I2=79%) in the systematic review of 102 studies (Ardito 2023). Haemorrhagic stroke was reported in 2% (12 out of 776) of people who had microaxial flow pump, 3% (26 out of 833) of people who had VA-ECMO, 1% (53 out of 5,451) of people who had IABP and 1% (249 out of 32,804) of people who had no temporary MCS (p<0.001) in the retrospective registry study by Padberg (2024). In the same study, ischaemic stroke was reported in 4% (31 out of 776) of people who had microaxial flow pumps, 10% (80 out of 833) of people who had VA-ECMO, 5% (289 out of 5,451) of people who had IABP and 4% (1,347 out of 32,804) of people who had no temporary MCS (p<0.001). Cerebrovascular accident was reported in 4% (152 out of 3,409) of people who had support with Impella alone in the registry study by Higuchi (2024).
Acute kidney injury, renal failure or renal replacement therapy
Acute kidney injury, renal failure or renal replacement therapy was reported as an outcome in 6 studies.
Renal replacement therapy was reported in a higher proportion of people in the microaxial flow pump group (42% [75 out of 179]) compared to the standard care alone group (27% [47 out of 176]; RR 1.98, 95% CI 1.27 to 3.09) in the randomised controlled trial of 355 people (Møller 2024).
Renal replacement therapy at 30 days was reported in 24% (116 out of 484) of people who had MCS and 20% (98 out of 482) of people who had standard care (OR 1.29, 95% CI 0.94 to 1.77) in the individual patient data meta-analysis of 9 randomised controlled trials by Thiele (2024). The rate of renal failure was 34% in the microaxial flow pump group (95% CI 26 to 43, 25 studies, I2=96%) and 36% in the VA-ECMO group (95% CI 31 to 42, 23 studies, I2=90%) in the systematic review of 102 studies (Ardito 2023). Post procedural acute kidney failure was reported in 0.2% of people who had a microaxial flow pump and 0.4% of people who had IABP (OR 0.44, 95% CI 0.25 to 0.77, p=0.004) in the retrospective registry study by Movahed (2024). Acute kidney injury was reported in 44% (341 out of 776) of people who had a microaxial flow pump, 53% (440 out of 833) of people who had VA-ECMO, 32% (1,732 out of 5,451) of people who had IABP and 27% (8,835 out of 32,804) of people who had no temporary MCS (p<0.001) in the retrospective registry study by Padberg (2024). In the same study, renal replacement therapy was reported in 30% (237 out of 776) of people who had a microaxial flow pump, 50% (419 out of 833) of people who had VA-ECMO, 25% (1,590 out of 5,451) of people who had IABP and 12% (3,975 out of 32,804) of people who had no temporary MCS (p<0.001). Acute kidney injury was reported in 7% (227 out of 3,409) of people who had support with Impella alone in the registry study by Higuchi (2024).
Sepsis
Sepsis was reported as an outcome in 4 studies.
Sepsis with positive blood culture was reported in a higher proportion of people in the microaxial flow pump group (12% [21 out of 179]) compared to the standard care alone group (4% [8 out of 176], RR 2.79, 95% CI 1.20 to 6.48) in the randomised controlled trial of 355 people (Møller 2024).
Sepsis at 30 days was reported in 17% (80 out of 480) of people who had MCS and 14% (66 out of 474) of people who had standard care (OR 1.28, 95% CI 0.87 to 1.88) in the individual patient data meta-analysis of 9 randomised controlled trials by Thiele (2024). Sepsis was reported in 16% (123 out of 776) of people who had a microaxial flow pump, 22% (187 out of 833) of people who had VA-ECMO, 14% (757 out of 5,451) of people who had IABP and 9% (3,034 out of 32,804) of people who had no temporary MCS (p<0.001) in the retrospective registry study by Padberg (2024). Sepsis was reported in 4% (136 out of 3,409) of people who had support with Impella alone in the registry study by Higuchi (2024).
Pericardial effusion, cardiac tamponade or cardiac perforation
Pericardial effusion was reported in 4% of people in both the microaxial flow pump and IABP groups (OR 1.04, 95% CI 0.91 to 1.19, p=0.54) in the retrospective registry study by Movahed (2024). In the same study, cardiac tamponade was reported in 3% of people who had a microaxial flow pump and 2% of people who had an IABP (OR 1.27, 95% CI 1.07 to 1.51, p=0.007) and cardiac perforation was reported in 1.3% of people who had a microaxial flow pump and 0.9% of people who had an IABP (OR 1.48, 95% CI 1.17 to 1.88, p=0.001). Cardiac tamponade was reported in 1% (47 out of 3,409) of people who had support with Impella alone in the registry study by Higuchi (2024).
Cardiac function
Cardioversion after ventricular tachycardia or fibrillation was reported in 33% (59 out of 179) of people in the microaxial flow pump group and 30% (52 out of 176) of people in the standard care group (RR 1.17, 95% CI 0.75 to 1.83) in the randomised controlled trial of 355 people (Møller 2024).
Intraoperative cardiac functional disturbance was reported in 0.5% of people in the microaxial flow pump group and 0.4% of people in the IABP group (OR 1.41, 95% CI 0.97 to 2.05, p=0.08) in the retrospective registry study by Movahed (2024).
Haemolytic anaemia
Acquired haemolytic anaemia was reported in 0.7% of people in the microaxial flow pump group and 0.1% of people in the IABP group (OR 8.36, 95% CI 4.89 to 14.3, p<0.001) in the retrospective registry study by Movahed (2024).
Disseminated intravascular coagulation
Disseminated intravascular coagulation was reported in 3% of people in the microaxial flow pump group and 2% of people in the IABP group (OR 1.76, 95% CI 1.48 to 2.08, p<0.001) in the retrospective registry study by Movahed (2024).
Composite outcomes
A composite safety end-point event of severe bleeding, limb ischaemia, haemolysis, device failure and worsening of aortic regurgitation was reported in a higher proportion of people in the microaxial flow pump group (24% [43 out of 179]) compared to the standard care alone group (6% [11 out of 176)], RR 4.74, 95% CI 2.36 to 9.55) in the randomised controlled trial of 355 people (Møller 2024). In the microaxial flow pump group, the number needed to harm was 6.
Other adverse events
Case reports describing adverse events associated with the procedure have been listed in table 5a in appendix B. These include left ventricular perforation, aortic valve insufficiency, mitral regurgitation, mitral chordal rupture, functional mitral stenosis, right-to-left arterial shunt through an iatrogenic atrial septal defect, pseudoaneurysm, infected pseudoaneurysm, arteriovenous fistula, intracerebral haemorrhage, aortic saddle embolism, Impella inlet entrapment in the mitral subvalvular apparatus, intra-arterial fibrinous sheath development, coiled Impella drive line in the left ventricle, intravascular device tip fracture and haemolysis.
MHRA Field Safety Notice
An MHRA field safety notice was issued for all Impella heart pumps in April 2024. In summary, information on safe use of Impella pumps had been issued with 2 technical bulletins, but the instructions for use were not updated to include the same level of detail covered in the bulletins and 1 of the bulletins was not distributed to European customers. These included a technical bulletin for operator mishandling of the Impella left-sided devices resulting in iatrogenic ventricular wall perforation and an Impella Product Update for an issue with fibres entrapped in the impeller. The action to mitigate the risk was for users to take note of amendment and reinforcement of instructions for use.
Anecdotal and theoretical adverse events
Expert advice was sought from consultants who have been nominated or ratified by their professional society or royal college. They were asked if they knew of any other adverse events for this procedure that they had heard about (anecdotal), which were not reported in the literature. They were also asked if they thought there were other adverse events that might possibly occur, even if they had never happened (theoretical).
They listed the following anecdotal or theoretical adverse events:
Thrombosis around the device
Thrombocytopenia
Access site haematoma
Aortic dissection or rupture
Heart valve injury
Compartment syndrome
Irreversible neurological impairment
Distal embolisation
Device migration
Closure device failure.
Six professional expert questionnaires for insertion of a catheter-based intravascular microaxial flow pump for cardiogenic shock were submitted. Find full details of what the professional experts said about the procedure in the https://www.nice.org.uk/guidance/indevelopment/gid-ipg10404/documents.
Validity and generalisability
Although the evidence includes a relatively large randomised controlled trial based in Europe that included 1 UK centre (Møller 2024), most studies are non-UK based and might not be generalisable to practice within the UK. Most studies were observational and retrospective, which have more potential for bias than randomised controlled trials.
The randomised controlled trial by Møller (2024) had very strict inclusion and exclusion criteria. People who had been resuscitated from out-of-hospital cardiac arrest and remained comatose on arrival to the cardiac catheterisation laboratory and people with overt right ventricular failure were excluded. The registry study by Padberg (2024) also specified that people with out-of-hospital cardiac arrest before admission were not included in the data. The 2 studies had conflicting results regarding the effect of microaxial flow pumps on mortality, but the outcomes were reported at different timepoints, the registry study lacked baseline data on patient characteristics and treatments were not randomly allocated. Also, the study by Møller (2024) used Impella CP but the study by Padberg (2024) did not specify what type of Impella microaxial flow pump was used.
In the randomised controlled trial by Møller (2024), the finding of excess need for renal-replacement therapy in the microaxial-flow-pump group may be attributable to the fact that more patients died early in the standard-care group, which may have introduced a survival bias owing to a competing risk.
In the randomised controlled trial by Møller (2024), allocation to treatment group was not masked and there may have been differential intensive care treatment between the 2 groups.
Some studies included people who had the procedure more than 10 years ago. There may have been changes in the approach to treating cardiogenic shock over time.
The individual patient data meta-analysis by Thiele (2024) includes different MCS devices analysed together as a group and results are not presented separately for microaxial flow pumps. Earlier trials included in the meta-analysis used IABP as the comparator, whereas in more recent trials the comparator was best medical therapy with MCS escalation in selected patients.
The overall results reported by Thiele (2024) are dominated by the weight of 2 large studies, one of which was the trial on microaxial flow pumps by Møller (2024).
Most of the systematic reviews reported high heterogeneity between studies.
Studies included in the systematic reviews had different patient selection criteria and definition of outcomes, such as bleeding, may have differed.
Safety and efficacy outcomes, such as vascular complication rates, are likely to be affected by the expertise of the study centre.
Registry studies use data that has been collected for other purposes and may not include baseline characteristics such as the degree of cardiogenic shock. Also, they may not distinguish between different models of Impella.
In the registry report by Movahed (2024), the authors note that the complication rate was lower than other publications, which they attributed to under-reporting.
There was a lack of data on quality of life.
The evidence included different devices, with different flow rates. Some of the devices used are no longer available or have been superseded.
The study by Sassani (2025) had some discrepancies between the text and the tables; the results presented in this overview were extracted from the tables.
The randomised controlled trial by Møller (2024) was supported by the Danish Heart Foundation and Abiomed.
Ongoing trials
IMPELLA, Complications and Tolerance (IMPACT; NCT06644963); Observational study; France; n=800; estimated study completion: January 2025
Evaluation of the Efficacy of Early Implantation of a Percutaneous Left Ventricular Assist Devices in Acute Coronary Syndrome Complicated by Cardiogenic Shock Compared to Conventional Therapy: a Prospective, Multicenter, Randomized, Controlled and Open-label Clinical Trial (ULYSS; NCT05366452); randomised controlled trial; France; n=204; estimated study completion: December 2026
Observational Assessment of Support With Impella® Best Practices in Acute Myocardial Infarction Complicated by Cardiogenic Shock (NCT06964685); prospective cohort study; US; n=250; estimated study completion: October 2028
Clinical Outcomes of Contemporary IMPELLA Devices in Cardiogenic Shock and High-risk Percutaneous Coronary Intervention (IMMERGE; NCT06690567); cohort study; Italy; n=700; estimated study completion: December 2024
Cardiogenic Shock Working Group Registry (CSWG; NCT04682483); cohort study; US; n=5,000; estimated study completion: June 2026
How are you taking part in this consultation?
You will not be able to change how you comment later.
You must be signed in to answer questions