Belford, P. Matthew

AbstractBackgroundThere are limited national‐level data on the contemporary practices of mechanical circulatory support (MCS) use in acute myocardial infarction‐cardiogenic shock (AMI‐CS).MethodsWe utilized the Healthcare Cost and Utilization Project‐National/Nationwide Inpatient Sample data (2005–2017) to identify adult admissions (>18 years) with AMI‐CS. MCS devices were classified as intra‐aortic balloon pump (IABP), percutaneous left ventricular assist devices (pLVAD), or extracorporeal membrane oxygenation (ECMO). We evaluated trends in the initial device used (IABP alone, pLVAD alone or ≥2 MCS devices), device escalation, bridging to durable LVAD/heart transplantation, and predictors of in‐hospital mortality and device escalation.ResultsAmong 327,283 AMI‐CS admissions, 131,435 (40.2%) had an MCS device placed with available information on timing of placement. IABP, pLVAD, and ≥2 MCS devices were used as initial device in 120,928 (92.0%), 8202 (6.2%), and 2305 (1.7%) admissions, respectively. Most admissions were maintained on the initial MCS device with 1%–1.5% being escalated (IABP to pLVAD/ECMO, pLVAD to ECMO). Urban, medium, and large‐sized hospitals and acute multiorgan failure were significant independent predictors of MCS escalation. In admissions receiving MCS, escalation of MCS device was associated with higher in‐hospital mortality (adjusted odds ratio: 1.56, 95% confidence interval: 1.38–1.75; p < 0.001). Admissions receiving durable LVAD/heart transplantation increased over time in those initiated on pLVAD and ≥2 MCS devices, resulting in lower in‐hospital mortality.ConclusionsIn this 13‐year study, escalation of MCS in AMI‐CS was associated with higher in‐hospital mortality suggestive of higher acuity of illness. The increase in number of durable LVAD/heart transplantations alludes to the role of MCS as successful bridge strategies.

AbstractBackgroundThere has been conflicting reports on the effect of new trainees on clinical outcomes at teaching hospitals in the first training month (July in the United States of America). We sought to assess this “July effect” in a contemporary acute myocardial infarction (AMI) population.MethodsAdult (>18 years) AMI hospitalizations in May and July in urban teaching and urban nonteaching hospitals in the United States were identified from the HCUP‐NIS database (2000–2017). In‐hospital mortality was compared between May and July admissions. A difference‐in‐difference analysis comparing a change in outcome from May to July in teaching hospitals to a change in outcome from May to July in nonteaching hospitals was also performed.ResultsA total of 1,312,006 AMI hospitalizations from urban teaching (n = 710,593; 54.2%) or nonteaching (n = 601,413; 45.8%) hospitals in the months of May and July were evaluated. May admissions in teaching hospitals, had greater comorbidity, higher rates of acute multiorgan failure (10.6% vs. 10.2%, p < 0.001) and lower rates of cardiac arrest when compared to July admissions. July AMI admissions had lower in‐hospital mortality compared to May (5.6% vs. 5.8%; adjusted odds ratio 0.94 [95% confidence interval 0.92–0.97]; p < 0.001) in teaching hospitals. Using the difference‐in‐difference model, there was no evidence of a July effect for in‐hospital mortality (p = 0.19).ConclusionsThere was no July effect for in‐hospital mortality in this contemporary AMI population.

Background and objectives: Primary percutaneous coronary intervention (PCI)-related outcomes in acute myocardial infarction (AMI) have improved over time, but there are limited data on the length of stay (LOS) in relation to in-hospital mortality. Materials and Methods: A retrospective cohort of adult AMI admissions was identified from the National Inpatient Sample (2000–2017) and stratified into short (≤3 days) and long (>3 days) LOS. Outcomes of interest included temporal trends in LOS and associated in-hospital mortality, further sub-stratified based on demographics and comorbidities. Results: A total 11,622,528 admissions with AMI were identified, with a median LOS of 3 (interquartile range [IQR] 2–6) days with 49.9% short and 47.3% long LOS, respectively. In 2017, compared to 2000, temporal trends in LOS declined in all AMI, with marginal increases in LOS >3 days and decreases for ≤3 days (median 2 [IQR 1–3]) vs. long LOS (median 6 [IQR 5–9]). Patients with long LOS had lower rates of coronary angiography and PCI, but higher rates of non-cardiac organ support (respiratory and renal) and use of coronary artery bypass grafting. Unadjusted in-hospital mortality declined over time. Short LOS had comparable mortality to long LOS (51.3% vs. 48.6%) (p = 0.13); however, adjusted in-hospital mortality was higher in LOS >3 days when compared to LOS ≤ 3 days (adjusted OR 3.00, 95% CI 2.98–3.02, p < 0.001), with higher hospitalization (p < 0.001) when compared to long LOS. Conclusions: Median LOS in AMI, particularly in STEMI, has declined over the last two decades with a consistent trend in subgroup analysis. Longer LOS is associated with higher in-hospital mortality, higher hospitalization costs, and less frequent discharges to home compared to those with shorter LOS.