Impella Outcomes in Acute Myocardial Infarction Versus Non-Acute Myocardial Infarction Cardiogenic Shock: A United...

Using the National Inpatient Sample (2016-2022), we identified adult hospitalizations with cardiogenic shock (ICD-10-CM R57.0) and an Impella procedure code (ICD-10-PCS 5A0221D or 5A0211D), excluding elective admissions, transfers out, postcardiotomy shock, and missing core data. Patients were classified as acute myocardial infarction-cardiogenic shock (AMI-CS) or non-AMI-CS by diagnosis. One-to-one nearest-neighbor propensity-score matching without replacement was performed on 64 covariates using a caliper of 0.2 standard deviations of the logit propensity score, with balance evaluated by absolute standardized mean differences <0.10. Co-primary outcomes were in-hospital all-cause mortality and the Saito composite (death or any major Impella-related complication: bleeding requiring transfusion, acute kidney injury requiring dialysis, stroke, vascular complication, limb ischemia, or access-site infection). Secondary outcomes included individual complications, length of stay, inflation-adjusted charges, and discharge disposition. The final analytic cohort comprised 9866 unweighted Impella-supported CS hospitalizations (7507 AMI-CS; 2359 non-AMI-CS). After 1:1 propensity-score matching, 1917 pairs were retained with excellent postmatch balance (median |standardized mean difference| 0.014; maximum 0.053). In the matched cohort, in-hospital mortality was significantly lower in AMI-CS (43.9%) than in non-AMI-CS (47.4%) (absolute risk difference -3.5%; OR 0.87, 95% CI, 0.76-0.99; P = 0.030), as was the Saito composite (61.1% vs 65.2%; OR 0.84, 95% CI, 0.74-0.96; p=0.010). Within non-AMI-CS, in-hospital mortality varied substantially by etiology, lowest in myocarditis-associated CS (25.0%) and highest in ventricular arrhythmia and pulmonary embolism etiologies. National Impella-supported CS hospitalizations rose progressively in both groups over 2016-2022 (annual percent change +12.0% and +11.1% for AMI-CS and non-AMI-CS, respectively).