López-Jiménez, Francisco

Introduction: Newly developed Systematic Coronary Risk Evaluation-2 (SCORE-2) and the version for older people [(SCORE2-OP) ≥70 years] algorithms for 10-yr incident cardiovascular disease (CVD) have been derived incorporating European and US populations. Its performance has not been externally validated in community-based populations representative of real-world clinical practice. Hypothesis: We tested the hypothesis that the SCORE-2 and SCORE2-OP algorithms will accurately estimate 10-yr risk of first-onset CVD in this population. Methods: Consecutive patients who, between 1998-2000, sought primary care in Olmsted county, MN, and were followed-up using the Rochester Epidemiology Project. Inclusion criteria were set as those used in the derivation of the original algorithm. The composite outcome of CVD was defined as first myocardial infarction, ischemic stroke, or cardiovascular mortality. We compared predicted and observed events and generated C-statistics across predefined risk subgroups. Results: We included 22,858 adults, mean age of 55.11±11.58 (13.50% >70 yr) 54% females. After a mean follow-up of 9.55±1.5 yr, 2,145 events were observed. 6.39% of subjects for SCORE-2 and 59.88% using SCORE2-OP were considered at high risk. Overall, the SCORE-2 had better discrimination than the SCORE2-OP (C-statistic: 0.72 and 0.64, respectively) (Fig B-D). The algorithms performed better in individuals 50-69 yr while underpredicting in younger individuals and overpredicting CVD in ages above 70 (in the low and moderate risk group) (Fig A-C). Prediction was mostly similar between men, women and middle-aged individuals but demonstrated pronounced underprediction in older men (Fig B-D). Conclusions: The SCORE-2 Risk prediction tool performed with marginal results whereas the SCORE2-OP had poor performance when predicting CVD events in a community cohort representing real-world clinical practice in a non-European population.

Abstract Background and purpose A heart age biomarker has been developed using deep neural networks applied to electrocardiograms. Whether this biomarker is associated with cognitive function was investigated. Methods Using 12‐lead electrocardiograms, heart age was estimated for a population‐based sample ( N = 7779, age 40–85 years, 45.3% men). Associations between heart delta age (HDA) and cognitive test scores were studied adjusted for cardiovascular risk factors. In addition, the relationship between HDA, brain delta age (BDA) and cognitive test scores was investigated in mediation analysis. Results Significant associations between HDA and the Word test, Digit Symbol Coding Test and tapping test scores were found. HDA was correlated with BDA (Pearson's r = 0.12, p = 0.0001). Moreover, 13% (95% confidence interval 3–36) of the HDA effect on the tapping test score was mediated through BDA. Discussion Heart delta age, representing the cumulative effects of life‐long exposures, was associated with brain age. HDA was associated with cognitive function that was minimally explained through BDA.

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): UiT The Arctic University of Norway, Northern Norway Regional Health Authority. Background/Introduction Machine learning models have been applied to magnetic resonance images of the brain to estimate brain age gap as a biomarker of biological aging. Because of its association with cognitive performance and neurological disorders, brain age gap has become an indicator of general brain health. Among others, cardiovascular risk factors are associated with brain health. Recently a novel non-invasive biomarker of vascular ageing has been proposed based on the application of a deep-learning algorithm to digitised 12-lead electrocardiograms. The difference between electrocardiogram predicted age and chronological age is defined as vascular age gap and has been shown to be associated with an adverse cardiovascular profile. However, the association of vascular age gap with cognitive health and dementias has not yet been examined. Purpose We apply a causal mediation framework to vascular age gap, brain age gap and cognitive function data obtained in the same individuals who took part in a population-based study in Norway. We hypothesise that cardiovascular risk factors encapsulated in vascular age gap can indirectly affect cognitive function by increasing the brain age gap. Methods We applied a deep learning algorithm (1) to electrocardiograms (N=7779, age 40-85 years) collected from the population-based sample in Norway (mean age 63.8 years, 45.3% men) to obtain a vascular age gap. For a subsample (N=1693) magnetic resonance images of the brain were also available and another deep neural network algorithm (2) was applied to obtain a biomarker of brain age. For 1678 participants with both vascular age gap and brain age gap available we used mediation analysis to test whether brain age gap mediates the effects of vascular age gap on cognitive function (as measured by mini mental status test, word memory test, digit–symbol coding test, tapping test) by estimating total, direct and indirect effects. Results We found a weak correlation (Pearson’s r = 0.12) between vascular age gap and brain age gap. Both age gaps were associated with cognitive scores capturing psychomotor speed, attention and mental flexibility. For a 1 standard deviation (7.3 years) increase in vascular age gap digit-symbol coding test score decreased by 0.63 (95%CI: -0.89, -0.36; p<0.001) points, and tapping test decreased by 0.36 (95%CI: -0.56, -0.14; p=0.001). Only 13% of effect of vascular age gap on cognitive test results was mediated by brain age gap. Conclusion(s) Vascular age gap was weakly associated with poorer cognitive functioning, and only 13% of this association was mediated by brain age gap. This suggests that cognitive decline linked to accumulated damage of the vascular system is not fully captured by structural changes in the brain. Further investigations of novel measures of biological aging in relation to cognitive decline are warranted as these may hold promise for improved understanding of mechanisms and clinical prediction.

Abstract Funding Acknowledgements Type of funding sources: None. Background Cardiac rehabilitation (CR) is a key component in the continuum of care for patients with CVD that reduces mortality, morbidity and rehospitalization. Clinical practice guidelines strongly recommend and encourage CR participation for eligible patients with CVD. Although there are multiple studies on barriers and solutions to improve the initiation rates of CR, a limited number of studies have explored the barriers and solutions to increase the degree of adherence to CR. Purpose The aim of this study was to define those characteristics that predict adherence to CR with special focus on women and employed patients, since these groups are generally understudied. Methods We performed a retrospective electronic medical record review of the Mayo Clinic CR program, Rochester to identify patients who had started the Mayo Clinic program between January 1, 2014 and September 30, 2015. Among patients who met the inclusion criteria, we conducted a telephone survey with the aim of defining barriers to participation and possible solutions proposed by patients to increase CR adherence. The survey included factors previously identified in the literature as possible barriers, as well as other hypothetical factors; including closed and open-ended questions. The study was approved by the Mayo Clinic institutional review board. Results 376 patients met the inclusion criteria. 257 patients completed the survey. 240 (63.8%) returned the HIPAA (Health Information Use Authorization) document. 72 (30%) were women with a mean age of 66.4 ±12.4 years. Forty-seven percent (n = 113) of the participants did not complete the recommended course of 36 sessions of CR. Among the most important barriers, we found that schedule conflicts were more frequently noted to be a barrier to CR participation among employed patients compared to non-employed patients (p=0.04). Lack of motivation and an inadequate awareness of the importance of completing CR were significantly reported as barriers from employed and non-employed patients. Among the proposed solutions, the participants suggested the potential benefit of alternative delivery models for CR programs; as well efforts to improve awareness of the importance of CR adherence among patients and professionals. Conclusions This study has great relevance because it addressed barriers and potential solutions to CR adherence. Barriers such as mental stress and motivation, as well as a lack of knowledge about CR benefits highlight the role of patient counseling and education as keys to help improve CR adherence. Alternative CR delivery models are attractive options to help facilitate participation, with special focus on the employed population and women, as they more frequently report logistical barriers to CR compared to non-employed patients and men. Further research and advocacy work are needed to help identify and implement strategies that address barriers to CR adherence and completion.

Background: We developed a deep learning algorithm that detects coronary artery calcium (CAC) score using 12-lead electrocardiograms (CAC- ECG). We tested the hypothesis that the output from the CAC-ECG algorithm would be associated with incident atherosclerotic cardiovascular disease (ASCVD) events and that the CAC- ECG would refine the AHA/ACC Pooled Cohort Equation’s (PCE) predictive capabilities. Methods: A community-based cohort of consecutive patients seeking primary care in Olmsted County, MN, between 1998-2000 with passive follow-up via record linkage. Inclusion was identical to the PCE. The original CAC-ECG was developed in 43,210 subjects yielding an AUC of 0.83. Herein, we used the CAC-ECG output to predict a high CAC (≥ 300). Primary outcome was ASCVD defined as fatal and non-fatal myocardial infarction and ischemic stroke, secondary outcome was Major Adverse Cardiovascular Events (MACE) further including PCI, CABG, and mortality. Events were validated in duplicate. Cox proportional hazard models adjusted for variables included in the PCE and were stratified to evaluate the effect of the CAC-ECG on PCE-predicted risk. Follow-up was truncated at 10 years for PCE analyses. Results: We included 24,793 subjects, mean ± SD age 53.9 ± 12.1, 52% women, 95% white. After 16.7±3.7 yrs follow-up, 2,366 (9.5%) had ASCVD and 3,401 (13.7%) had MACE. Risk of ASCVD and MACE increased with CAC-ECG probability quintiles, independent of risk factors, p for trend <0.001 ( Fig. A-B ). The CAC-ECG enhanced the predicted capabilities of the PCE across all ASCVD risk groups ( Fig. C ). Net reclassification improved 13.7% with comparable C-statistic from 0.77 vs. 0.78 for PCE and CAC-ECG. Conclusions: CAC-ECG was associated with ASCVD and MACE and improved PCE predicted risk. The CAC-ECG algorithm could identify individuals at risk in primary prevention; Unlike the PCE, the CAC-ECG can be applied without chart review or performing a computer tomography and may be reliably used retrospectively in cohorts with digitally stored ECGs.

Background. Mortality risk assessment before kidney transplantation (KT) is imperfect. An emerging risk factor for death in nontransplant populations is physiological age as determined by the application of artificial intelligence to the electrocardiogram (ECG). The aim of this study was to examine the relationship between ECG age and KT waitlist mortality. Methods. We applied a previously developed convolutional neural network to the ECGs of KT candidates evaluated 2014 to 2019 to determine ECG age. We used a Cox proportional hazard model to examine whether ECG age was associated with waitlist mortality. Results. Of the 2183 patients evaluated, 59.1% were male, 81.4% were white, and 11.4% died during follow-up. Mean ECG age was 59.0 ± 12.0 y and mean chronological age at ECG was 53.3 ± 13.6 y. After adjusting for chronological age, comorbidities, and other characteristics associated with mortality, each increase in ECG age of >10 y than the average ECG age for patients of a similar chronological age was associated with an increase in mortality risk (hazard ratio 3.59 per 10-y increase; 95% confidence interval, 2.06-5.72; P < 0.0001). Conclusions. ECG age is a risk factor for KT waitlist mortality. Determining ECG age through artificial intelligence may help guide risk-benefit assessment when evaluating candidates for KT.

Abstract Aims An artificial intelligence-enabled electrocardiogram (AI-ECG) is a promising tool to detect patients with aortic stenosis (AS) before developing symptoms. However, functional, structural, or haemodynamic components reflected in AI-ECG responsible for its detection are unknown. Methods and results The AI-ECG model that was developed at Mayo Clinic using a convolutional neural network to identify patients with moderate–severe AS was applied. In patients used as the testing group, the correlation between the AI-ECG probability of AS and echocardiographic parameters was investigated. This study included 102 926 patients (63.0 ± 16.3 years, 52% male), and 28 464 (27.7%) were identified as AS positive by AI-ECG. Older age, atrial fibrillation, hypertension, diabetes, coronary artery disease, and heart failure were more common in the positive AI-ECG group than in the negative group (P &lt; 0.001). The AI-ECG was correlated with aortic valve area (ρ = −0.48, R2 = 0.20), peak velocity (ρ = 0.22, R2 = 0.08), and mean pressure gradient (ρ = 0.35, R2 = 0.08). The AI-ECG also correlated with left ventricular (LV) mass index (ρ = 0.36, R2 = 0.13), E/e′ (ρ = 0.36, R2 = 0.12), and left atrium volume index (ρ = 0.42, R2 = 0.12). Neither LV ejection fraction nor stroke volume index had a significant correlation with the AI-ECG. Age correlated with the AI-ECG (ρ = 0.46, R2 = 0.22) and its correlation with echocardiography parameters was similar to that of the AI-ECG. Conclusion A combination of AS severity, diastolic dysfunction, and LV hypertrophy is reflected in the AI-ECG to detect AS. There seems to be a gradation of the cardiac anatomical/functional features in the model and its identification process of AS is multifactorial.

Abstract Aims Current non-invasive screening methods for cardiac allograft rejection have shown limited discrimination and are yet to be broadly integrated into heart transplant care. Given electrocardiogram (ECG) changes have been reported with severe cardiac allograft rejection, this study aimed to develop a deep-learning model, a form of artificial intelligence, to detect allograft rejection using the 12-lead ECG (AI-ECG). Methods and results Heart transplant recipients were identified across three Mayo Clinic sites between 1998 and 2021. Twelve-lead digital ECG data and endomyocardial biopsy results were extracted from medical records. Allograft rejection was defined as moderate or severe acute cellular rejection (ACR) based on International Society for Heart and Lung Transplantation guidelines. The extracted data (7590 unique ECG-biopsy pairs, belonging to 1427 patients) was partitioned into training (80%), validation (10%), and test sets (10%) such that each patient was included in only one partition. Model performance metrics were based on the test set (n = 140 patients; 758 ECG-biopsy pairs). The AI-ECG detected ACR with an area under the receiver operating curve (AUC) of 0.84 [95% confidence interval (CI): 0.78–0.90] and 95% (19/20; 95% CI: 75–100%) sensitivity. A prospective proof-of-concept screening study (n = 56; 97 ECG-biopsy pairs) showed the AI-ECG detected ACR with AUC = 0.78 (95% CI: 0.61–0.96) and 100% (2/2; 95% CI: 16–100%) sensitivity. Conclusion An AI-ECG model is effective for detection of moderate-to-severe ACR in heart transplant recipients. Our findings could improve transplant care by providing a rapid, non-invasive, and potentially remote screening option for cardiac allograft function.

Background There is a scarcity of validated rapid dietary screening tools for patient use in the clinical setting to improve health and reduce cardiovascular risk. The Healthy Eating Index (HEI) 2015 measures compliance with the 2015 to 2020 Dietary Guidelines for Americans but requires completion of an extensive diet assessment to compute, which is time consuming and impractical. The authors hypothesize that a 19‐item dietary survey assessing consumption of common food groups known to affect health will be correlated with the HEI‐2015 assessed by a validated food frequency questionnaire and can be further reduced without affecting validity. Methods and Results A 19‐item Eating Assessment Tool (EAT) of common food groups was created through literature review and expert consensus. A cross‐sectional survey was then conducted in adult participants from a preventive cardiology clinic or cardiac rehabilitation and in healthy volunteers (n=661, mean age, 36 years; 76% women). Participants completed an online 156‐item food frequency questionnaire, which was used to calculate the HEI score using standard methods. The association between each EAT question and HEI group was analyzed by Kruskal‐Wallis test. Linear regression models were subsequently used to identify univariable and multivariable predictors for HEI score for further reduction in the number of items. The final 9‐item model of Mini‐EAT was validated by 5‐fold cross validation. The 19‐item EAT had a strong correlation with the HEI score ( r =0.73) and was subsequently reduced to the 9 items independently predictive of the HEI score: fruits, vegetables, whole grains, refined grains, fish or seafood, legumes/nuts/seeds, low‐fat dairy, high‐fat dairy, and sweets consumption, without affecting the predictive ability of the tool ( r =0.71). Conclusions Mini‐EAT is a 9‐item validated brief dietary screener that correlates well with a comprehensive food frequency questionnaire. Future studies to test the Mini‐EAT's validity in diverse populations and for development of clinical decision support systems to capture changes over time are needed.