Improved clinical outcomes associated with the Impella 5.5 compared to the Impella 5.0 in contemporary cardiogenic shock and heart failure patients

A redesigned surgically implanted heart pump incorporates several design changes from the prior device generation, but no published comparative data demonstrate if these changes translate to improved outcomes. We retrospectively compared clinical characteristics and outcomes, drawn from an FDA-mandated QA database, for contemporary patients treated with the Impella 5.5 or Impella 5.0 for acute myocardial infarction complicated by cardiogenic shock (AMICS), cardiomyopathy, or postcardiotomy cardiogenic shock (PCCS). A total of 1238 patients at 290 US sites were included for analysis. Patients receiving the Impella 5.5 had significantly higher survival through explant (i.e., successfully weaned or bridged to heart replacement therapy) than those receiving the Impella 5.0 in all 3 settings: AMICS (70.5% vs 56.8%; p = 0.005), cardiomyopathy (88.1% vs 76.9%; p = 0.001), and PCCS (76.1% vs 55.7%; p = 0.003). Duration of support was significantly longer for Impella 5.5 patients with AMICS (9.2 vs 6.1 days; p = 0.008) and cardiomyopathy (10.7 vs 8.1 days; p 0.99106/111 (95.5)132/137 (96.4)0.75622/40 (55.0)48/63 (76.2)0.031 Femoral4/76 (5.3)15/148 (10.1)0.3124/111 (3.6)3/137 (2.2)0.7041/40 (2.5)7/63 (11.1)0.146 Ascending aorta2/76 (2.6)2/148 (1.4)0.6061/111 (0.9)1/137 (0.7)>0.9916/40 (40.0)7/63 (11.1)0.001 Other2/76 (2.6)0/148 (0.0)0.1140/111 (0.0)1/137 (0.7)>0.991/40 (2.5)1/63 (1.6)>0.99Aborted placement10/169 (5.9)15/305 (4.9)0.67111/287 (3.8)7/245 (2.9)0.6346/126 (4.8)14/106 (13.2)0.033AMICS, acute myocardial infarction complicated by cardiogenic shock; LVEF, left ventricular ejection fraction; PAC, pulmonary artery catheterization; PCCS, postcardiotomy cardiogenic shock; SD, standard deviation.Categorical data is presented as numerator/denominator (percentage); continuous data as mean ± standard deviation (denominator), and median (range).p values were calculated with Fisher's exact test for categorical data and unpaired t tests for continuous data. Open table in a new tab AMICS, acute myocardial infarction complicated by cardiogenic shock; LVEF, left ventricular ejection fraction; PAC, pulmonary artery catheterization; PCCS, postcardiotomy cardiogenic shock; SD, standard deviation. Categorical data is presented as numerator/denominator (percentage); continuous data as mean ± standard deviation (denominator), and median (range). p values were calculated with Fisher's exact test for categorical data and unpaired t tests for continuous data. Impella 5.5 patients had significantly higher survival through device explant for all 3 indications: AMICS (70.5% vs 56.8%; p = 0.005), cardiomyopathy (88.1% vs 76.9%; p = 0.001), and PCCS (76.1% vs 55.7%; p = 0.003) (Table 2). Cardiomyopathy patients had similar rates of successful weaning off support, but a significantly higher percentage of Impella 5.5 cardiomyopathy patients were successfully bridged to heart replacement therapy (54.4% vs 41.8%; p = 0.005). PCCS patients had similar rates successfully bridged to heart replacement therapy, but Impella 5.5 PCCS patients had a significantly higher rate of successful weaning (70.1% vs 52.3%; p = 0.013). Duration of support was significantly longer in Impella 5.5 AMICS patients (median 9.2 vs 6.1 days; p = 0.002) and Impella 5.5 cardiomyopathy patients (median 10.7 vs 8.1 days; p = 0.0008).Table 2Clinical Outcomes through Device Explant in AMICS, Cardiomyopathy, and PCCS Patients Treated with the Impella 5.5 or 5.0AMICSCardiomyopathyPCCSImpella 5.5(N=156)Impella 5.0(N=278)p-valueImpella 5.5(N=270)Impella 5.0(N=225)p-valueImpella 5.5(N=117)Impella 5.0(N=88)p-valueSuccessfully weaned or bridged to heart replacement therapy110/156 (70.5)158/278 (56.8)0.005238/270 (88.1)173/225 (76.9)0.00189/117 (76.1)49/88 (55.7)0.003 Successfully weaned78/156 (50.0)118/278 (42.4)0.13391/270 (33.7)79/225 (35.1)0.77682/117 (70.1)46/88 (52.3)0.013 Bridged to therapy32/156 (20.5)40/278 (14.4)0.108147/270 (54.4)94/225 (41.8)0.99Hemolysis5/156 (3.2)10/278 (3.6)>0.998/270 (3.0)21/225 (9.3)0.0032/117 (1.7)1/88 (1.1)>0.99CVA5/156 (3.2)3/278 (1.1)0.1436/270 (2.2)2/225 (0.9)0.3012/117 (1.7)1/88 (1.1)>0.99Bleeding1/156 (0.6)5/278 (1.8)0.4263/270 (1.1)5/225 (2.2)0.4783/117 (2.6)6/88 (6.8)0.177Vascular injury1/156 (0.6)0/278 (0.0)0.3590/270 (0.0)1/225 (0.4)0.4550/117 (0.0)0/88 (0.0)>0.99Duration of support, days Mean ± SD (N)13.2 ± 20.2 (156)8.7 ± 9.5 (278)0.00815.1 ± 13.4 (270)11.4 ± 10.6 (225)<0.00110.2 ± 23.5 (117)6.6 ± 8.3 (88)0.127 Median (range)9.2 (0.04-233.3)6.1 (0.01-87.1)10.7 (0.03-71.1)8.1 (0.3-64.1)6.0 (0.0007-245.9)4.4 (0.02-49.2)AMICS, acute myocardial infarction complicated by cardiogenic shock; CVA, cerebrovascular accident; PCCS, post cardiotomy cardiogenic shock.Categorical data is presented as numerator/denominator (percentage); continuous data as mean ± standard deviation (denominator), and median (range).p values were calculated with Fisher's exact test for categorical data and unpaired t tests for continuous data.Patients with aborted Impella placement, unknown outcome, or outcome of “on support” (i.e., unknown outcome) were not included in denominators for clinical outcomes nor calculation of duration of support values (63 patients had aborted placement, 35 had unknown outcome, and 6 remained on support also unknown outcome).Other complications observed on support include hematoma, oozing, thrombus, ischemia, renal failure, systemic organ failure, arrythmia, and other. Open table in a new tab AMICS, acute myocardial infarction complicated by cardiogenic shock; CVA, cerebrovascular accident; PCCS, post cardiotomy cardiogenic shock. Categorical data is presented as numerator/denominator (percentage); continuous data as mean ± standard deviation (denominator), and median (range). p values were calculated with Fisher's exact test for categorical data and unpaired t tests for continuous data. Patients with aborted Impella placement, unknown outcome, or outcome of “on support” (i.e., unknown outcome) were not included in denominators for clinical outcomes nor calculation of duration of support values (63 patients had aborted placement, 35 had unknown outcome, and 6 remained on support also unknown outcome). Other complications observed on support include hematoma, oozing, thrombus, ischemia, renal failure, systemic organ failure, arrythmia, and other. Rates of hemolysis, cerebrovascular accident, bleeding, and vascular injury were similar in AMICS and PCCS patients, with significantly lower hemolysis rates in Impella 5.5 cardiomyopathy patients (3.0% vs 9.3%; p = 0.003). This may owe to the removal of the pigtail catheter and increased ease of repositioning. In this retrospective analysis, we found significantly improved outcomes through device explant with the Impella 5.5 in real-world AMICS, PCCS, and cardiomyopathy patients. This analysis does not allow the attribution of improved bridging and weaning rates to the device redesign, or unobserved differences in the patient populations. The Impella 5.5 patients were supported for about 4 days longer than the 5.0 patients in all 3 settings, without an increased complication rate. We would speculate that the capability to provide robust longer-term support with the Impella 5.5 has resulted in higher successful use of the 5.5 as a bridge to heart replacement therapy, notably, a 14-percentage point increase in in cardiomyopathy patients. Longer support times also provide an increased opportunity for native heart recovery, which may have impacted the significantly higher successful weaning rates in the AMICS and PCCS populations. We aimed to compare similar populations by restricting the analysis to the same time period for both devices and removing patients with concomitant ECMO, as the comparative efficacy of these devices for LV unloading with ECMO is a separate question that should be examined in further study. Importantly, the assessment of LVEF was nearly identical in the study populations. However, the IQ registry collects limited data, and confounding factors affecting clinical outcomes likely exist. In an analysis assessing prior implant experience (any indication) in the 2 cohorts, we found a significantly higher level of prior device experience in the 5.5 centers (mean 6 vs 4.8 years’ experience; p = 0.005). However, this comparison was limited to prior 5.0 experience, as the market release of the 5.5 constituted the start of the study window for this analysis, and comprised the initial experience with the 5.5 for these centers. Any potential, as yet unstudied, learning curve with the 5.5 may also have impacted outcomes. We are further limited by the nature of the IQ database, which only collects outcomes through device explant, and is per field representative reporting with limited oversight. This analysis challenges the heart failure community to conduct a prospective trial evaluating this device. Within the database and analysis limitations, in contemporaneous real-world patient populations, we observed a marked improvement in clinical outcome with the Impella 5.5 device compared to the Impella 5.0 as treatment for AMICS, PCCS, and cardiomyopathy. D.R.–Study design, data analysis and manuscript writing. All co-authors contributed to data analysis and reviewed and approved the final manuscript. D.R. reports honoraria from Abiomed, Medtronic, and Abbott. E.G. Soltesz reports honoraria from Abiomed. S.S. reports consulting fees from Abbott and Abiomed. M.D. reports funding for clinical trial work to the institution from Abiomed, Abbott, and TransMedics; travel support from Abbott and Abiomed for HM3 and Impella 5.5 Users Meetings, respsectively. M.K. reports consulting fees from Abiomed and CareDx. D.A.D'A. reports honoraria from Abiomed and Paragonix. Auxiliary medical writing services were provided by the device manufacturer (acknowledged above). Dana Bentley, MWC, provided medical writing services and is employed by the device manufacturer. JetPub Scientific Communications LLC, supported by Abiomed, assisted in the preparation of this manuscript.