Search

Background: Machine learning (ML) employs algorithms that learn from data, building models with the potential to predict events by aggregating a large number of variables and assessing their complex interactions. The aim of this study is to assess ML potential in identifying patients with ischemic heart disease (IHD) at high risk of cardiac death (CD)., Methods: 3987 (mean age 68 ± 11) hospitalized IHD patients were enrolled. We implemented and compared various ML models and their combination into ensembles. Model output constitutes a new ML indicator to be employed for stratification. Primary variable importance was assessed with ablation tests., Results: An ensemble classifier combining three ML models achieved the best performance to predict CD (AUROC of 0.830, F1-macro of 0.726). ML indicator use through Cox survival analysis outperformed the 18 variables individually, producing a better stratification compared to standard multivariate analysis (improvement of ∼20%). Patients in the low risk group defined through ML indicator had a significantly higher survival (88.8% versus 29.1%). The main variables identified were Dyslipidemia, LVEF, Previous CABG, Diabetes, Previous Myocardial Infarction, Smoke, Documented resting or exertional ischemia, with an AUROC of 0.791 and an F1-score of 0.674, lower than that of 18 variables. Both code and clinical data are freely available with this article., Conclusion: ML may allow a faster, low-cost and reliable evaluation of IHD patient prognosis by inclusion of more predictors and identification of those more significant, improving outcome prediction towards the development of precision medicine in this clinical field., Competing Interests: Declaration of competing interest there is no conflict of interest to declare by all authors., (Copyright © 2024 The Authors. Published by Elsevier B.V. All rights reserved.)

Background: Previous studies have suggested a decline in positivity of stress cardiac imaging, suggesting the need for developing better strategies for test selection to achieve acceptable cost-effectiveness balance. The aim of this retrospective study was to assess the rate of positivity of stress echocardiography (SE) over 27 consecutive years., Methods: We assessed the rate of SE positivity in 2007 patients without previous myocardial infarction or coronary revascularization who performed SE in a tertiary care referral center from 1983 to 2009. SE was performed with dipyridamole (1427), dobutamine (136) or exercise (444)., Results: There was a progressive decline over time in the rate of SE positivity from 42% (1983-1991) to 22% (2001-2009), with a relative increase of patients with low pre-test probability of disease (from 5% to 27%). The percentage of patients studied with SE under anti-ischemic therapy increased markedly (from 8% in the first to 61% in the last nine years)., Conclusion: Over 27 consecutive years, we observed a steady decline in SE positivity rate (with >5-fold increase of low probability patients), with almost 8-fold increase in anti-ischemic therapy at testing. We probably need refined criteria of referral for testing and/or better ways to titrate the negative response beyond wall motion abnormalities during SE., (Copyright © 2016 Elsevier Ireland Ltd. All rights reserved.)

Background: Although magnesium (Mg) has recognized cardioprotective properties and hypomagnesemia is common in patients with acute myocardial infarction (AMI), data regarding the role of Mg as prognostic factor for adverse events are scarce, as well as there are conflicting results on the use of Mg as adjuvant therapy in AMI., Aim: To evaluate the role of Mg as predictor for hard events (HE, all cause death, and nonfatal myocardial infarction) in AMI patients., Design and Patients: We studied 406 AMI patients (306 males, age: 67 ± 12 years, mean ± SD). Patient data were collected from the Institute electronic databank which saves demographic, clinical, instrumental, therapeutical and follow-up data of all patients admitted to our Coronary Unit., Results: During a mean follow-up period of 21 ± 18 months, the combined endpoint accounted for 63 HE, 44 (11%) deaths (35 cardiac deaths), 19 (5%) nonfatal MI. The multiple regression model identified glycemia as the only independent determinant of Mg in AMI pts. (T value = - 2.8, standard coefficient = - 0.15, p < 0.01). The Kaplan-Meier survival estimates failed to show a significantly worst outcome in patients presenting low Mg (< 0.783 mmol/L, 25th percentile). Aging (> 67 years-50th percentile), and ejection fraction (< 40%) remained as prognostic factors for HE in the adjusted Cox multivariate proportional hazard model (HR = 2.8, 95% CI = 1.6-5, p < 0.001; HR = 3.2, 95% CI = 1.9-5.3 p < 0.001, respectively)., Conclusion: The present findings do not support a significant role of low Mg as predictor for HE in AMI.

Background: Low radiation doses from diagnostic and therapeutic procedures may increase cancer incidence after decades in cardiac patients. The aim of the study was to evaluate the long-term outcome and the prognostic effect of radiological exposure., Methods: A retrospective cohort study examined 16,311 Italian cardiac patients without cancer consecutively hospitalized between 1970 and 2009 and followed-up for a median of 10 years. Cumulative estimated effective dose was calculated for the imaging procedures. Main causes of death and primary cancer onset were the end points and their relation with radiation exposure was analyzed by Cox model., Results: Over the years, there was an increasing trend of cancer, a decreasing trend of cardiovascular, with stability of other-causes death. Cumulative estimated effective radiation dose was higher in patients who died for cancer (20 ± 19 mSv) compared to cardiovascular (17 ± 17, P < 0.001) and other-causes deaths (15 ± 17, p < 0.001). The cumulative estimated effective dose adjusted HR was 1.004 (95% CI 1.000-1.009, p = 0.045) for cancer death and 1.008 (95% CI 1.005-1.011, p < 0.001) for (fatal and non-fatal) cancer onset. Radiation effective dose was not predictive of cardiovascular mortality (HR = 0.991), nor it was of other types of deaths (HR = 0.995). Smoking was a predictor for cardiovascular mortality, but it was not of other types of deaths., Conclusions: Non-cardiovascular causes of death are responsible for almost half of later deaths in cardiac patients. The radiological exposure from medical imaging procedures is predictive of cancer risk and cancer death., (Copyright © 2015 Elsevier Ireland Ltd. All rights reserved.)

Background: Several epidemiological reports indicate that the body mass index (BMI) is inversely related with mortality, in spite of the notion that obesity is a recognized cardio-metabolic risk factor. The aim of the study was to evaluate the independent impact of overweight and obesity on long-term mortality in a large cohort of patients with heart disease (HD)., Methods: The study included 10,446 patients hospitalized in the last three decades for ischemic (60%) or non-ischemic HD and followed-up for 10 years. The relationship between BMI and total or cardiovascular mortality was analyzed in the whole cohort, and in age-stratified categories (≤ 65 and >65 years). Considering that survival in HD patients has improved after the introduction of revascularization, beta-blockers, ACE inhibitors, and statins, the relationship was re-examined separately in patients hospitalized before and after 1990., Results: Diabetes, hyperuricemia, hypertension, glycaemia, and triglyceridemia increased across BMI groups. During follow-up (73 ± 59 months) there were 1707 all-cause deaths (47% cardiac). Any relationship between BMI and mortality was lost in the ≤ 65 age category and in patients hospitalized before 1990, but it persisted in old patients hospitalized after 1990. Most significant independent predictors of mortality in all groups were hyperuricemia, diabetes and impaired ejection fraction., Conclusions: No independent relationship was found between BMI and mortality in subjects ≤ 65 years of age. This neutral relationship seems to be partly counteracted by treatment, particularly in old patients. A different effect of obesity onset in old vs. young age cannot be ruled out., (Copyright © 2012 Elsevier Ireland Ltd. All rights reserved.)

Background: In some Italian areas, colonoscopic surveillance of first-degree relatives (FDRs) of colorectal cancer (CRC) patients is provided as a part of local population-based faecal occult blood test (FOBT) screening programmes. The objective of the present study was to assess the feasibility and early results of this surveillance model., Methods: Data from district screening centres were used to evaluate the process of identification and selection of eligible FDRs (residence in the Emilia-Romagna Region, age 40-75 years, no recent colonoscopy) of screen-detected CRC patients and the detected prevalence of disease. The probability for an FDR to undergo colonoscopy and to be diagnosed with CRC and advanced adenoma was estimated using the Kaplan-Meier method. The sex- and age-standardised ratio of detected prevalence to that expected based on results from a colonoscopy screening study of the Italian general population was estimated., Results: Between 2005 and 2011, 9319 FDRs of 2437 screen-detected CRC patients (3.8 per patient) were identified and contacted. Their likelihood of being eligible for, and accepting, colonoscopy was 0.11 (95% confidence interval: 0.11-0.12). Among the 926 subjects undergoing colonoscopy, the prevalence of previous negative screening FOBT was 63%. Eleven CRCs (1.2%) and 100 advanced adenomas (10.8%) were detected. The standardised ratio of detected prevalence to that expected was 0.91 (95% confidence interval: 0.19-2.66) for CRC and 1.48 (1.04-2.05) for advanced adenoma., Conclusions: The procedure of selection of FDRs was extremely ineffective. Due to previous negative screening tests, the prevalence of disease was less than expected. A population-based FOBT screening programme is a highly unsuitable setting for the provision of surveillance to FDRs of CRC patients., (Copyright © 2013 Elsevier Ltd. All rights reserved.)

Background: Several biohumoral variables, taken individually, are predictors of prognosis in patients with chronic coronary artery disease (CAD). We hypothesized that taken together, laboratory tests provide prognostic information that is additive to a complete diagnostic work-up., Methods: We prospectively examined 2370 consecutive patients with chronic CAD, as shown by a >50% coronary stenosis (in 95% of patients), previous coronary revascularization (in 31% of patients), and/or previous myocardial infarction (MI, in 54% of patients). We tested the ability of laboratory and clinical variables to predict future cardiac events (cardiac death and non-fatal MI)., Results: During follow-up (median, 46 months), 147 patients (6.2%) died from cardiac causes and 81 (3.4%) experienced a non-fatal MI. Using multivariate analysis, after adjustment for clinical variables (including left ventricular ejection fraction and angiographic extent of coronary stenoses), a high-density lipoprotein cholesterol (HDLc) concentration<35 mg/dL (p<0.0001), a neutrophil-to-lymphocyte ratio >2.4 (p=0.0014), and an fT3 serum level<2.1 pg/mL with normal thyrotropin (low-T3 syndrome) (p=0.0260) showed an independent and incremental prognostic value, and were associated with an increase in the rate of cardiac events of 86%, 57% and 41%, respectively. When these variables were added to clinical and instrumental variables, the prognostic power of the model increased significantly (global chi-square improvement: from 157.01 to 185.07, p<0.0001)., Conclusion: Low HDLc, high neutrophil-to-lymphocyte ratio and low-T3 syndrome, both individually and taken together, provide prognostic information that is independent of and incremental to the main clinical and instrumental findings., (Copyright © 2012 Japanese College of Cardiology. Published by Elsevier Ltd. All rights reserved.)

Introduction: There are no studies on the long-term prognostic role of abnormal thyrotropin value in patients with acute cardiac diseases. Aim of the study was to assess the incidence and persistence of risk of cardiac and overall deaths in patients with acute cardiac diseases., Methods: A total of 1026 patients (mean age: 67.7 years) were divided into 4 groups: (1) euthyroid (EU, n=579); (2) subclinical-like hypothyroidism (SLHYPO, n=68); (3) subclinical-like hyperthyroidism (SLHYPER, n=23) and (4) low-triiodothyronine syndrome (LowT3, n=356). Follow-up started from the day of thyroid hormone evaluation (mean follow-up: 30 months). The events considered were cardiac and overall deaths., Results: Survival rate for cardiac death was lower in SLHYPO and in LT3 than in EU (log rank test; χ(2)=33.6; P < 0.001). Survival rate for overall death was lower in SLHYPO, SLHYPER and LowT3 than in EU (48.3; P < 0.001). After adjustment for several risk factors, the hazard ratio for cardiac death was higher in SLHYPO (3.65; P=0.004) in LowT3 (1.88; P=0.032) and in SLHYPER (4.73; P=0.047). Hazard ratio for overall death was higher in SLHYPO (2.30; P=0.009), in LowT3 (1.63; P=0.017) and in SLHYPER than in EU (3.71; P=0.004). Hazards for SLHYPO, SLHYPER and LowT3 with respect to EU were proportional over the follow-up period., Conclusion: In patients with acute cardiac disease, a mildly altered thyroid status was associated with increased risk of mortality that remains constant during all the follow-up.

Background: Literature on the appropriateness of coronary revascularization in chronic angina is still scanty. The study aimed to compare long-term effects of revascularization with those of medical therapy in stable angina., Methods: In an observational single center study, we assessed 10 year follow-up of 1442 consecutive patients with chronic angina, at least one coronary vessel disease, no previous myocardial infarction, screened for inducible ischemia. Patients>70 years were excluded. The event-free probabilities were estimated by Kaplan-Meier curves; all cause death, cardiac death, non-fatal myocardial infarction were the considered end points., Results: Age was 56±8 yrs. Global left ventricular function was preserved in all. Myocardial ischemia was documented in 1190 patients. Coronary disease was more severe in patients with inducible ischemia as compared to those with negative stress test (p<0.001); 868 patients underwent one revascularization procedure, 511 coronary angioplasty. Median follow-up was 106 months; 13% all cause deaths, 8% cardiac deaths, 6% non-fatal myocardial infarction were registered. When provocative test was negative revascularization did not improve survival (1% per year mortality irrespective of type of treatment). Conversely survival was significantly improved by revascularization when ischemia was documented (0.7% vs 1.8% per year mortality for revascularization vs medical therapy, p<0.05). Incidence of non-fatal myocardial infarction was low and similar in both groups., Conclusion: In low-risk chronic angina coronary revascularization does not improve long-term prognosis unless inducible myocardial ischemia is present. This suggests considering coronary revascularization as an effective tool in treating coronary artery disease only when myocardial ischemia has been documented., (Copyright © 2009 Elsevier Ireland Ltd. All rights reserved.)

Objectives: The aim of the study was to assess the prognostic value of stress echocardiography after surgical revascularization., Methods: We evaluated 500 (100 women) patients who had undergone exercise or pharmacological SE after a median of 69 months after coronary artery by-pass grafting (CABG). Of these, 351 (70%) complained of symptoms suggestive of ischemic origin while 149 (30%) were tested for asymptomatic progression of the disease., Results: SE was positive for ischemia in 196 (39%) patients. During a median follow-up of 25 months, 61 patients died, 33 had a nonfatal myocardial infarction, and 112 underwent late (>3 months) revascularization. Multivariable Cox' regression analysis indicated age (HR=1.04; 95% CI 1.01-1.06; p<0.003), and peak WMSI (HR=3.07; 95% CI 1.96-4.81; p=0.0001) as independent predictors of hard (total mortality and myocardial infarction) events. SE information provided a significant improvement in predictive power of the statistical model (chi-square increase 34%, p<0.0001 for hard and 91%, p<0.0001 for major events, respectively). Survival analysis showed ischemia at SE to be associated with significantly higher hard and major event rate in both symptomatic and asymptomatic patients., Discussion: SE represents an effective tool for the risk stratification of patients with previous CABG independently of the presence of symptoms suggestive of ischemic origin., (Copyright © 2009 Elsevier Ireland Ltd. All rights reserved.)

Objectives: To analyze the impact of ABO groups on coronary heart disease risk factors, coronary involvement and prognosis., Methods: An observational single center study was conducted to examine 4901 consecutive patients with heart disease receiving coronary angiography and ABO group determination at National Research Council Institute of Clinical Physiology between January 1993 and December 2003, with maximum 10 years follow-up. All-cause death and cardiac death, were the considered end points., Results: When compared to the official distribution of ABO groups in the Italian population (O 40%, A 36%, B 17%, AB 7%), a substantially different distribution was observed in the study population (O 43.3%, A 41.4%, B 11.2%, AB 4.1%). In addition, a significant association was found between group non-O and family history of ischemic heart disease, hypercholesterolemia and presence of coronary atherosclerosis. Higher prevalence of A and B alleles was found in patients with myocardial infarction (P<0.05). Group non-O was a powerful predictor of cardiac mortality in patients aged <65 years, particularly in women (HR 1.53, 95% CI 1.06-2.21 and HR 5.29, 95% CI 1.57-17.82, respectively)., Conclusions: Group non-O is associated with increased mortality in patients with ischemic heart disease. Group non-O increases the risk for cardiac death in non-elderly patients, particularly in younger females, and groups A and B prevail in myocardial infarction. ABO group determination might aid in genetic screening for ischemic heart disease and become relevant in the management of risk factor control., (Copyright 2010 Elsevier Ireland Ltd. All rights reserved.)

Background: Depressed heart rate variability at acute myocardial infarction discharge is associated with poor long-term prognosis. However, its early (<48 h) predictive value has not been extensively investigated. Aim of this Multicenter Italian Study was to investigate, during acute myocardial infarction, in-hospital prognostic value of heart rate variability and its short-term evolution., Methods: Twenty-four hour ECG monitoring was prospectively obtained on admission in 413 patients with new-onset acute myocardial infarction and repeated in 349 at discharge. Heart rate variability statistical and frequency domain indices, peak creatine kinase, echocardiographic wall motion score index and risk factors were obtained. The occurrence of cardiac death and resuscitated ventricular fibrillation were the primary end-points; cardiogenic shock, ventricular tachycardia, post-infarction angina and heart failure the secondary end-points., Results: At admission, a marked reduction in heart rate variability indices was evident. Nine patients died during hospitalization and 13 were resuscitated from ventricular fibrillation. Secondary endpoints occurred in other 91 patients. At univariate analysis, low frequencies (LF), mean time interval between consecutive heart beats (RR), wall motion score index and family history of ischemic heart disease were predictive of combined primary and secondary end-points. At multivariate analysis, only LF and family history were predictive with a relative risk of 2.01 and 1.84, respectively (P<0.003). In survivors, heart rate variability indices significantly increased during hospitalization, still remaining below reference values., Conclusions: A depressed heart rate variability was present in the early phase of infarction and improved at discharge. LF power was an independent predictor of the combined unfavorable short-term events.