Faculty Bibliography

Several machine learning (ML) algorithms have been increasingly utilized for cardiovascular disease prediction. We aim to assess and summarize the overall predictive ability of ML algorithms in cardiovascular diseases. A comprehensive search strategy was designed and executed within the MEDLINE, Embase, and Scopus databases from database inception through March 15, 2019. The primary outcome was a composite of the predictive ability of ML algorithms of coronary artery disease, heart failure, stroke, and cardiac arrhythmias. Of 344 total studies identified, 103 cohorts, with a total of 3,377,318 individuals, met our inclusion criteria. For the prediction of coronary artery disease, boosting algorithms had a pooled area under the curve (AUC) of 0.88 (95% CI 0.84-0.91), and custom-built algorithms had a pooled AUC of 0.93 (95% CI 0.85-0.97). For the prediction of stroke, support vector machine (SVM) algorithms had a pooled AUC of 0.92 (95% CI 0.81-0.97), boosting algorithms had a pooled AUC of 0.91 (95% CI 0.81-0.96), and convolutional neural network (CNN) algorithms had a pooled AUC of 0.90 (95% CI 0.83-0.95). Although inadequate studies for each algorithm for meta-analytic methodology for both heart failure and cardiac arrhythmias because the confidence intervals overlap between different methods, showing no difference, SVM may outperform other algorithms in these areas. The predictive ability of ML algorithms in cardiovascular diseases is promising, particularly SVM and boosting algorithms. However, there is heterogeneity among ML algorithms in terms of multiple parameters. This information may assist clinicians in how to interpret data and implement optimal algorithms for their dataset.

OBJECTIVE:To describe from a non-interventional registry the short-term ischemic and hemorrhagic outcomes in patients with NSTEMI managed with a loading dose of a P2Y12 inhibitor (P2Y12i) given at least four hours prior to diagnostic angiography and delineation of coronary anatomy. Prior data on the effects of such "upstream loading" have been inconsistent. METHODS:In 53 US hospitals, we evaluated the in-hospital care and outcomes of patients with confirmed NSTEMI managed with an interventional strategy and loaded upstream (at least four hours before diagnostic angiography) with P2Y12 inhibitor therapy. Patients entered into the database were grouped into one of four cohorts for analysis: (1) overall cohort, (2) thienopyridine (clopidogrel or prasugrel) load, (3) ticagrelor load, and (4) ticagrelor-consistent. The fourth cohort is a subset of cohort 3 that received ticagrelor throughout the index hospital stay and at discharge. We evaluated in-hospital clinical course and ischemic and bleeding outcomes in all patients, and also 30-d outcomes in the ticagrelor-consistent cohort. RESULTS:A total of 3,355 patients were enrolled, of whom 1,087 had 30-day follow-up. The mean (+/-SD) age was 63.3+/-12.5 y and 62.6% were male. TIMI and GRACE scores placed these patients in the intermediate risk range and CRUSADE scores were in the moderate risk range. The loading dose in UPSTREAM was clopidogrel in 45.6%, ticagrelor in 53.6%, and prasugrel in 0.8%. The median upstream interval (loading dose to angiography) was 17:27 hours and did not change appreciably over the course of the data collection period (2/15 - 10/19). Access was radial in 48.6% and femoral in 51.4%. Post-angiography management was medical only in 32.3%, PCI in 59.4%, and CABG in 8.3%. Median LOS was 2.7d, and median time from angiography to CABG was 3.6d. In-hospital mortality was 0.51% and major bleeding (TIMI) was 0.24%; the in-hospital MACE rate was 0.7% and stent thrombosis occurred in 0.18%. No significant differences were seen between the ticagrelor and clopidogrel cohorts in hospital, but 16% received more than one P2Y12i in-hospital. On follow-up (93.2% response), 86.7% of patients reported taking ticagrelor as directed. CONCLUSION/CONCLUSIONS:Upstream loading of P2Y12 inhibitors was associated with very low rates of bleeding and short LOS in a large cohort of NSTEMI patients managed invasively.

INTRODUCTION/BACKGROUND:Chronic kidney disease is an independent risk factor for cardiovascular disease. Despite careful preoperative evaluation, there is a risk of acute coronary syndromes after kidney transplant. METHODS:The National Inpatient Sample for the years 2004-2013 was used for this retrospective cohort study. All adult patients undergoing kidney transplantation were identified using the appropriate ICD-9-CM codes. Multivariate logistic regression was used to identify predictors of acute coronary syndromes in the peri-operative period after kidney transplantation. RESULTS:A total of 147,431 kidney transplants were performed from 2004 through 2013 in United States. The average peri-operative in-hospital mortality was 0.5%. Acute coronary syndrome occurred in 1.3% patients in the peri-operative period. Half of patients with acute coronary syndromes had pre-existing coronary artery disease. The strongest predictors of acute coronary syndromes included older age: 45-64 yrs. OR 3.28 95% CI (1.85-5.83), ≥65 yrs. OR 4.84 (2.59-9.05), race: African-American, OR 0.66 (0.47-0.93) and preexisting coronary artery disease OR 3.83 (2.84-5.15). The case fatality rates were 16.9% and 5.3% for STEMI and NSTEMI, respectively. The overall mortality for any ACS event was 7.1%. CONCLUSION/CONCLUSIONS:Acute coronary syndrome in the immediate peri-operative period after kidney transplantation is rare but is associated with high rates of mortality.

BACKGROUND:In the COMPLETE (Complete vs Culprit-only Revascularization to Treat Multi-vessel Disease After Early PCI for STEMI) trial, angiography-guided percutaneous coronary intervention (PCI) of nonculprit lesions with the aim of complete revascularization reduced major cardiovascular (CV) events in patients with ST-segment elevation myocardial infarction (MI) and multivessel coronary artery disease. OBJECTIVES/OBJECTIVE:The purpose of this study was to determine the effect of nonculprit-lesion stenosis severity measured by quantitative coronary angiography (QCA) on the benefit of complete revascularization. METHODS:Among 4,041 patients randomized in the COMPLETE trial, nonculprit lesion stenosis severity was measured using QCA in the angiographic core laboratory in 3,851 patients with 5,355 nonculprit lesions. In pre-specified analyses, the treatment effect in patients with QCA stenosis ≥60% versus <60% on the first coprimary outcome of CV death or new MI and the second co-primary outcome of CV death, new MI, or ischemia-driven revascularization was determined. RESULTS:The first coprimary outcome was reduced with complete revascularization in the 2,479 patients with QCA stenosis ≥60% (2.5%/year vs. 4.2%/year; hazard ratio [HR]: 0.61; 95% confidence interval [CI]: 0.47 to 0.79), but not in the 1,372 patients with QCA stenosis <60% (3.0%/year vs. 2.9%/year; HR: 1.04; 95% CI: 0.72 to 1.50; interaction p = 0.02). The second coprimary outcome was reduced in patients with QCA stenosis ≥60% (2.9%/year vs. 6.9%/year; HR: 0.43; 95% CI: 0.34 to 0.54) to a greater extent than patients with QCA stenosis <60% (3.3%/year vs. 5.2%/year; HR: 0.65; 95% CI: 0.47 to 0.89; interaction p = 0.04). CONCLUSIONS:Among patients with ST-segment elevation MI and multivessel coronary artery disease, complete revascularization reduced major CV outcomes to a greater extent in patients with stenosis severity of ≥60% compared with <60%, as determined by quantitative coronary angiography.

Atherosclerotic heart disease remains a leading cause of morbidity and mortality worldwide. While extensive research supports cardiovascular risk factor reduction in the form of achieving evidence-based blood pressure, lipid, glucose, and body weight targets as a means to improve cardiovascular outcomes, residual risk remains. Emerging data have demonstrated that the intraindividual variability of these risk factor targets potentially contribute to this residual risk. It may therefore be time to define risk factor by not only its magnitude and duration as done traditionally, but perhaps also by the variability of that particular risk factor over time.

Background: Exposure to low-dose ionizing radiation is associated with malignancies. Lead garment specifications in the cardiac catheterization laboratory are not currently regulated, potentially resulting in unprotected areas.
Method(s): Interventional cardiology attendings and fellows wore 7 dosimeters, one externally on the thyroid shield and six inside the lead apron: bilateral axilla, chest wall, and pelvis. Radiation protection included a lower table-mounted lead drape, upper ceiling-mounted lead shield, and use of 7.5 frames per second during fluoroscopy. All procedures were performed with operators standing to the right of the patient. The primary endpoint was operator radiation exposure to the left versus right axilla. Radiation exposures in millirem (mrem) per participant over the study period are shown as median [interquartile range] and compared between left- and right-sided measures using paired Wilcoxon tests.
Result(s): Nine participants (66% female) wore dosimeters during 231 cases. Transradial coronary angiography was selected in 81.1% of cases and PCI was performed in 32.1%. A sterile radiation drape placed on the patient abdomen was used in 18.6% of cases. Median dose area product and fluoroscopy time for the participants ranged from 29.0-60.5 Gy cm² and 6.2-13.5 minutes, respectively. Radiation exposure at the left axilla was higher than the right axilla (5 vs. 0.9 mrem, p=0.018) but did not differ between left or right chest wall and left or right pelvis (Figure).
Conclusion(s): This analysis demonstrates insufficient protection in the left axillary area. The use of additional left axillary protection should be evaluated. (Figure Presented)