Procedural volume and outcomes of Impella support in cardiogenic shock

The use of Impella for cardiogenic shock has increased substantially, largely replacing intra-aortic balloon pumps; however, clinical evidence demonstrating improved outcomes remains limited, and in-hospital mortality is still high [1, 2]. Most previous reports reflect the early period of Impella adoption. A volume–outcome relationship has been well established for other percutaneous cardiovascular procedures, and has recently been described for venoarterial extracorporeal membrane oxygenation [3], but whether such a relationship exists for Impella remains unclear.

We analysed data from the Japan Registry for Percutaneous Ventricular Assist Device (UMIN000033603), a nationwide prospective registry of Impella use in Japan, including 1,390 patients with cardiogenic shock who received Impella support between February and December 2021 across 182 hospitals. Annual institutional volume was defined as the total number of Impella procedures performed at each institution during the 12 months preceding each case, regardless of indication, and was categorised into quartiles: Q1, 0–7; Q2, 8–12; Q3, 13–18; Q4, 19–48 cases per year. The median volume was 12 cases per year (interquartile range, 7–18). The mean age was 66 years, 76% were men, 49% had acute coronary syndrome, and 20% had out-of-hospital cardiac arrest. Patients treated at Q1 institutions were more likely to have prior heart failure, to receive inotropes or vasopressors before Impella initiation, and to have mechanical complications of myocardial infarction as the indication for Impella use. Conversely, patients treated at Q4 institutions were more likely to have lower systolic blood pressure and arrhythmia as the indication for Impella use. Multivariable logistic regression and restricted cubic spline analyses with multiple imputation were performed to adjust for patient-level confounders including demographic characteristics, comorbidities, clinical presentation, haemodynamic parameters, and treatment factors.

Overall in-hospital mortality was 43.5% (Q1, 43.2%; Q2, 46.9%; Q3, 42.0%; Q4, 40.4%). After adjustment, patients treated at Q4 institutions had significantly lower in-hospital mortality than those treated at Q1 institutions (adjusted odds ratio [OR], 0.67; 95% confidence interval [CI], 0.47–0.96), and lower 30-day mortality (adjusted OR, 0.69; 95% CI, 0.48–0.99). Restricted cubic spline analysis revealed an inverse, nonlinear association between annual institutional volume and in-hospital mortality, with a significant reduction observed at approximately 24 or more cases per year (Fig.1). Conversely, haemolysis (adjusted OR, 4.88; 95% CI, 3.08–7.72), acute kidney injury (adjusted OR, 4.30; 95% CI, 2.52–7.33), and thrombocytopenia (adjusted OR, 2.93; 95% CI, 1.68–5.13) were more frequent at Q4 institutions than at Q1 institutions, whereas major bleeding, stroke, lower limb ischaemia, and vascular injury did not differ significantly across quartiles. The higher rates of device-related complications at high-volume centres likely reflect more vigilant surveillance and earlier detection rather than poorer clinical care. These complications were not associated with excess mortality.

Fig. 1.

Nonlinear associations between annual institutional Impella procedural volume and in-hospital mortality using restricted cubic spline analyses. A bold line shows adjusted odds ratios of in-hospital mortality and dotted lines show the 95% confidence intervals. Gray boxes show the histogram of annual institutional volume. Four annual institutional volumes (2, 8, 14, and 31 cases/year) are used as the knots in the cubic splines. In the cubic spline analyses, we adjusted for the same variables in the multivariable logistic regression analyses, including age, sex, body mass index at admission, prior myocardial infarction, chronic kidney disease, clinical presentation, systolic blood pressure, lactate, inotrope or vasopressor use before Impella initiation, primary indications for Impella use, duration of the first Impella, and extracorporeal membrane oxygenation use. Odds ratios and their 95% confidence intervals for each value of annual institutional volume were calculated relative to the reference value of 12 cases/year.

These findings align with and extend recent reports from the United States [4, 5]. Previous analyses identified a near-linear relationship beginning at five to seven cases per year; however, our data suggest a nonlinear threshold effect, with a significant reduction in mortality emerging at around 24 cases per year. This discrepancy may be partly due to methodological differences: we defined annual institutional volume based on the 12 months preceding each case rather than the calendar year, and included all Impella uses, regardless of indication, in order to clarify the Impella-specific volume–outcome relationship. The Impella volume–outcome relationship may involve at least three interrelated domains: technical skill in large-bore vascular access and device positioning; team-based management across the emergency department, catheterisation laboratory, and intensive care unit; and clinical judgement in patient selection and haemodynamic monitoring. Establishing a structured multidisciplinary shock team and developing a regional hub-and-spoke system that integrates high-volume centres may be key to improving outcomes for patients with cardiogenic shock. Notably, Watanabe et al. revealed that patient-hospital distance was not associated with mortality [5], further supporting the importance of ensuring sufficient Impella volume at each institution rather than simply increasing the number of facilities.

Several limitations should be acknowledged. First, the observational design precludes causal inference, and unmeasured confounders, including heterogeneity in institutional shock protocols, may persist. Second, complications may have been underreported at lower-volume centres, as indicated by the differential ascertainment of haemolysis and acute kidney injury. Long-term outcomes were unavailable. Finally, these findings may not be fully generalisable to other healthcare systems or to later periods with greater device experience.

In conclusion, a higher annual institutional volume was associated with lower in-hospital mortality among patients with cardiogenic shock, with a threshold of approximately 24 cases per year. This may support consolidating Impella use at experienced centres, and cautions against expanding the number of institutions performing these procedures.

Abbreviations

CI

Confidence interval

OR

Odds ratio

Author contributions

YN : Conceptualization, Software, Formal analysis, Investigation, Writing - Original Draft; AK : Software, Investigation, Resources, Supervision; YSat: Software, Investigation, Resources, Supervision; TW : Software, Investigation, Resources, Supervision; TY : Software, Investigation, Resources, Supervision; TH : Software, Investigation, Resources, Supervision; MK : Software, Investigation, Resources, Supervision; YSo : Writing - Review and Editing, Supervision. YSak : Writing - Review and Editing, Supervision. All authors read and approved the final manuscript.

Funding

No direct funding was received for this correspondence. The J-PVAD infrastructure is funded by Abiomed Japan K.K. and the Council for Clinical Use of Ventricular Assist Device–Related Academic Societies, and is additionally supported by certification and renewal fees from participating institutions. No specific grant numbers apply.

Data availability

The data may be available with permission from the Council for Clinical Use of Ventricular Assist Device Related Academic Societies, Impella committee, through the corresponding author.

Declarations

Ethics approval and consent to participate

The Japan Registry for Percutaneous Ventricular Assist Device was approved by the Graduate School of Medicine/Faculty of Medicine, Osaka University Ethics Committee (Approval no. 17232).

Consent for publication

The review board waived the requirement for informed consent because of the anonymous nature of the data. No information describing the individual patients, hospitals, or treating physicians was obtained.

Competing interests

Dr. Nishimoto received lecture fees from Abiomed Japan. Dr. Sotomi has received grants from Abiomed and Johnson & Johnson. All other authors declare that they have no competing interests.