Heart diseases and stroke are the number one cause of death and disability among people with type 2 diabetes (T2D). Clinicians and health authorities for many years have expressed interest in identifying individuals at increased risk of coronary heart disease (CHD). Our main objective is to develop a prognostic workflow of CHD in T2D patients using a Holter dataset. This workflow development will be based on machine learning techniques by testing a variety of classifiers and subsequent selection of the best performing system. It will also assess the impact of feature selection and bootstrapping techniques over these systems. Among a variety of classifiers such as Naive Bayes (NB), Random Forest (RF), Support Vector Machine (SVM), Alternating Decision Tree (ADT), Random Tree (RT) and K-Nearest Neighbour (KNN), the best performing classifier is NB. We achieved an area under receiver operating characteristics curve (AUC) of 68,06% and 74,33% for a prognosis of 3 and 4 years, respectively.
Funding Acknowledgements Type of funding sources: None. OnBehalf on behalf of the Investigators of " Portuguese Registry of ACS " Introduction Regarding prognosis, acute coronary syndromes (ACS) are heterogeneous. Post-hospitalization (PH) risk stratification is crucial. The Get With The Guidelines Heart Failure score (GWTG-HFS) predicts in-hospital mortality (M) of patients (P) admitted with acute heart failure. Objective To validate GWTG-HFS as predictor of PH early and late M and readmission (RA) rates, in our center population, using real-life data. Methods Based on a single-center retrospective study, data collected from admissions between 1/01/20168 and 11/12/2019. Patients who survived the ACS and were discharged from the hospital were included. Concerning prognosis, we assessed 1-month M and RA (1mM and 1mRA), 6-month M and RA (6mM and 6mRA), 1-year M and RA (1yM and 1yRA). Statistical analysis used non-parametric tests, logistic regression and ROC curve analysis. Results 268 patients with ACS, mean age was 66.4 ± 12.5 years old and 59.7% were male. The diagnosis was unstable angina in 2.6%, non-ST elevation myocardial infarction (NSTEMI) in 66.4% and ST elevation myocardial infarction (STEMI) in 31%. 41.8% of the P were or had been smokers, 68.5% had hypertension, 34.5% were diabetic and 50.9% had dyslipidaemia. Concerning coronary artery disease, 250 were submitted to coronary angiography – 18.8% had no lesions or non-significant lesions (stenosis <50%), 34.8% had one significant lesion, 23.2% had 2 significant lesions and 23.2% had 3 or more. Regarding left ventricle (LV) function, 70.5% of the P had no LV dysfunction, 15.7% had mild LV impairment (LVI), 9.3% moderate LVI and 4.5% had severe LVI. 1mM rate was 1.9% and 1yM rate was 7.8%. Age (p = 0.034), diabetes (p = 0.031), KKC (p < 0.001), BUN (p = 0.003) and LV function (p < 0.001) were predictors of 1mM. Age (p < 0.001), HR (p = 0.009), KKC (p = 0.032), BUN (p < 0.001), sodium (p < 0.001), creatinine (p < 0.001), Hb (p < 0.001), LV function (p < 0.001), de novo AF (p < 0.001) and number of arteries with significant disease (p = 0.044) were predictors of 1yM. Logistic regression and ROC curve analysis showed that GWTG-HFS was able to predict 1mM (Odds ratio (OR) 1.18, p = 0.005, confidence interval (CI) 1.05-1.33; area under curve (AUC) 0.872) and 1yM (OR 1.16, p = 0.001, CI 1.09-1.24, AUC 0.838) with excellent accuracy, and 1mRA (OR 1.10, p = 0.006, CI 1.03-1.18, AUC 0.677) and 1yRA (OR 1.04, p = 0.024, CI 1.01-1.08, AUC 0.580) with poor accuracy. A sub-analysis regarding NSTEMI P showed that GWTG-HFS was able to predict 1mM (OR 1.20, p = 0.010, CI 1.05-1.39, AUC 0.902) and 1yM (OR 1.15, p < 0.001, CI 1.07-1.23, AUC 0.817) with excellent accuracy. On the other hand, sub-analysis regarding STEMI showed that GWTG-HFS was not able to predict 1mM (p = 0.495) but was accurate at predicting 1yM (OR 1.18, p = 0.048, CI 1.00-1.39, AUC 0.881). Conclusion This study confirms that, in our population, GWTG-HFS is a valuable tool in PH risk score stratification in ACS, particularly NSTEMI.
Funding Acknowledgements Type of funding sources: None. OnBehalf on behalf of the Investigators of " Portuguese Registry of ACS " Introduction Heart failure (HF) is a frequent complication of acute coronary syndromes (ACS). Therefore, it is important to access its impact on prognosis and identify patients (pts) with higher risk of HF. Objective To evaluate predictors and prognosis of HF in the setting of ACS. Methods Based on a multicenter retrospective study, data collected from admissions between 1/10/2010 and 4/09/2019. Pts without data on cardiovascular history or uncompleted clinical data were excluded. Pts were divided in 2 groups (G): GA – pts without HF; GB - pts with HF during hospitalization. Results HF occurred in 4003 (15.6%) out of 25718 pts with ACS. GB was older (74 ± 12 vs 65 ± 13, p < 0.001), had more females (36.3% vs 26.2%, p < 0.001), had higher rates of arterial hypertension (78.4% vs 69.3%, p < 0.001), dyslipidaemia (64.4% vs 61.1%. p < 0.001), previous ACS (25.6% vs 19.7%, p < 0.001,), previous HF (16.4% vs 4.1%, p < 0.001), previous stroke (11.9% vs 6.4%, p < 0.001), chronic kidney disease (CKD) (17.1% vs 5.5%, p < 0.001), chronic obstructive pulmonary disease (COPD) (7.8% vs 3.8%, p < 0.001) and longer times from first symptoms to admission (268min vs 238min, p < 0.001). GA had higher rate of smokers (28.4% vs 16.2%, p < 0.001) and higher rate of non-ST-elevation myocardial infarction (MI) (46.5% vs 43.0%, p < 0.001). GB had higher rates of ST-elevation MI (STEMI) (49.2% vs 41.1%, p < 0.001), namely anterior STEMI (58.1% vs 44.9%, p < 0.001). GB had lower blood pressure (130 ± 32 vs 140 ± 28, p < 0.001), higher heart rate (86 ± 23 vs 76 ± 18, p < 0.001), Killip-Kimball class (KKC) ≥2 (63.2% vs 6.7%, p < 0.001), atrial fibrillation (AF) (15.4% vs 5.7%, p < 0.001), left bundle branch block (7.5% vs 3.1%, p < 0.001) and were previously treated with diuretics (39.1% vs 22.1%, p < 0.001), amiodarone (2.2% vs 1.4%, p < 0.001) and digoxin (2.8% vs 0.7%, p < 0.001). GB had higher rates of multivessel disease (66.0% vs 49.5%, p < 0.001) and planned coronary artery bypass grafting (7.3% vs 6.0%, p < 0.001), reduced left ventricle function (72.3% vs 33.4%, p < 0.001) and needed more frequently mechanical ventilation (8.2% vs 0.9%, p < 0.001), non-invasive ventilation (8.7% vs 0.5%, p < 0.001) and provisory pacemaker (4.5% vs 1.0%, p < 0.001). Logistic regression confirmed females (p < 0.001, OR 1.42, CI 1.29-1.58), diabetes (p < 0.001, OR 1.43, CI 1.30-1.58), previous ACS (p < 0.001, OR 1.27, CI 1.10-1.47), previous stroke (p < 0.001, OR 1.35, CI 1.16-1.57), CKD (p < 0.001, OR 1.76, CI 1.50-2.05), COPD (p < 0.001, OR 2.15, CI 1.82-2.54), previous usage of amiodarone (p = 0.041, OR 1.35, CI 1.01-1.81) and digoxin (p < 0.001, OR 2.30, CI 1.70-3.16), and multivessel disease (p < 0.001, OR 1.64, CI 1.67-2.32) were predictors of HF in the setting of ACS. Event-free survival was higher in GA than GB (79.5% vs 58.1%, OR 2.3, p < 0.001, CI 2.09-2.56). Conclusion As expected, HF in the setting of ACS is associated with poorer prognosis. Several features may help predict the HF occurrence during hospitalizations, allowing an earlier treatment.
Funding Acknowledgements Type of funding sources: None. Introduction Regarding prognosis, acute coronary syndromes (ACS) are heterogeneous. Non-ST elevation myocardial infarction (NSTEMI) is a subtype of ACS. In-hospital (IH) and post-hospitalization (PH) risk stratification is crucial. Objective To identify predictors of IH and PH mortality (early and late), as well as predictors of early and late re-admission (RA) in our center population suffering NSTEMI, using real-life data. Methods Based on a single-center retrospective study, data collected from admissions between 1/01/2018 and 11/12/2019. Patients (pts) who survived the ACS and were discharged from the hospital were included. Concerning prognosis, we assessed 1-month M and RA (1mM and 1mRA), 6-month M and RA (6mM and 6mRA), 1-year M and RA (1yM and 1yRA). Results 268 pts with ACS, 59.7% were males and mean age was 66.4 ± 12.5 years old. NSTEMI was the diagnosis in 66.4% and ST elevation myocardial infarction (STEMI) in 31%. Mean creatinine was 1.2 ± 1ml/min, mean sodium was 138 ± 3mmol/L, mean blood urea nitrogen (BUN) was 21 ± 12mg/dL and mean haemoglobin (Hb) was 13.6 ± 1.9g/dL. 88.2% of the pts presented in Killip-Kimball class (KKC) 1, 5.7% in KKC 2, 5.7% in KKC 3 and 0.4% in KKC IV; furthermore, 4.1% of the pts presented de novo AF. Concerning coronary artery disease, 250 were submitted to coronary angiography – 18.8% had no lesions or non-significant lesions (stenosis <50%), 34.8% had one significant lesion, 23.2% had 2 significant lesions and 23.2% had 3 or more. Regarding left ventricle (LV) function, 70.5% of the pts had no LV dysfunction, 15.7% had mild LV impairment (LVI), 9.3% moderate LVI and 4.5% had severe LVI. 8.4% of the patients experienced IH complications, such as auriculoventricular block, heart failure, ventricular tachycardia, stroke, cardiorespiratory arrest and major haemorrhage, during hospitalization. 1mM rate was 1.9% and 1yM rate was 7.8%. KKC (p = 0.001), BUN (p = 0.007), LV function (p= 0.001) and de novo AF (p = 0.46) were predictors of 1mM. Age (p = 0.004), KKC (p = 0.031), BUN (p = 0.002), sodium (p = 0.037), creatinine (p = 0.001), Hb (p = 0.003), LV function (p < 0.001), de novo AF (p < 0.001) and occurrence of IH complications (p < 0.001) were predictors of 1yM. Age (p = 0.010), male gender (p = 0.19), Hb (p = 0.031), de novo AF (p < 0.001) and occurrence of IH complications (p = 0.001) were predictors of 1mRA. Age (p = 0.004), smoking (p = 0.040), hypertension (p = 0.040), glycemia at admission (p = 0.031), Hb (p = 0.004), LV function (p = 0.019), de novo AF (p < 0.001) and occurrence of IH complications (p < 0.001) were predictors of 1yRA. Conclusion This study suggests that de novo AF and occurrence of IH complications are very important prognosis factors regarding early and late mortality and readmission rates.
Funding Acknowledgements Type of funding sources: None. OnBehalf on behalf of the Investigators of " Portuguese Registry of ACS " Introduction Reinfarction (RI) is a potential complication of acute coronary syndromes (ACS) and it is, therefore, important to access its impact on prognosis and identify patients with higher risk of RI in the setting of ACS. Objective To evaluate predictors and prognosis of RI in the setting of ACS. Methods Based on a multicenter retrospective study, data collected from admissions between 1/10/2010 and 4/09/2019. Patients (pts) without data on previous cardiovascular history or uncompleted clinical data were excluded. Pts were divided in 2 groups (G): GA – pts without RI; GB - pts with RI during hospitalization. Logistic regression and survival analysis were performed. Results Between 25718 pts with ACS, RI occurred in 223 (0.87%). Regarding epidemiological factors and past history, GB was older (70 ± 12 vs 67 ± 14, p < 0.001), had higher rates of hypertension (77.4% vs 70.6%, p = 0.028), previous stroke (12.1% vs 7.2%, p = 0.005), peripheric arterial disease (10.0% vs 5.5%, p = 0.004) and chronic obstructive pulmonary disease (8.6% vs 4.4%, p = 0.003). GB had higher rates of non-ST-elevation myocardial infarction (MI) (54.3% vs 45.9%, p = 0.012) and GA had higher rates of ST-elevation MI (42.4% vs 35.9%, p = 0.049). The groups were similar regarding blood pressure (p = 0.285), heart rate (p = 0.796) and Killip-Kimball class at admission, but GB had higher levels of brain natriuretic peptide (392 vs 180, p = 0.005). GB had higher rates of multivessel disease (62.8% vs 51.6%, p = 0.002), left ventricle dysfunction (50.0% vs 39.1%, p = 0.002), higher needs of mechanical ventilation (6.3% and vs 1.9%, p < 0.001) non-invasive ventilation (5.4% vs 1.7%, p < 0.001). Logistic regression confirmed that peripheric arterial disease (p = 0.011, OR 1.93, CI 1.17-3.19), multivessel disease (p = 0.003, OR 1.69, CI 1.20-2.39) and lower left ventricle function (p < 0.001, OR 2.42, CI 1.69-3.47) were predictors of RI in the setting of ACS. Event-free survival was similar between groups (p = 0.399). Conclusion RI in the setting of ACS was associated multivessel disease and left ventricle disfunction, however, 1-year prognosis was similar to pts who didn’t suffer RI.
Funding Acknowledgements Type of funding sources: None. OnBehalf on behalf of the Investigators of " Portuguese Registry of ACS " Introduction Sustained ventricular tachycardia (SVT) complicates up to 20% of acute coronary syndromes (ACS) and it is, therefore, important to access its impact on prognosis and identify patients with higher risk of SVT. Objective To evaluate predictors of early onset (<48h) and late onset (≥48h) SVT. Methods Based on a multicenter retrospective study, data collected from admissions between 1/10/2010 and 4/09/2019. Patients (pts) were divided in two groups (G): A – pts that presented early onset SVT (ESVT), and B – pts that presented late onset SVT (LSVT). Pts without data on previous cardiovascular history or uncompleted clinical data were excluded. Logistic regression was performed to assess predictors of SVT in ACS. Results Between 29851 pts with ACS, 364 (1.2%) presented SVT. ESVT – 251 pts (69%); LSVT – 91 pts (25%). LSVT G was older (74 ± 13 vs 68 ± 14, p = 0.003), was admitted directly to cat lab less frequently (10.1% vs 24.8%, p = 0.003), had longer times from first symptoms to admission (440min vs 261 min, p < 0.001) and had higher rates of previous stroke (14.4% vs 6.8%, p = 0.028). LSVT G had higher rates of non-ST-elevation myocardial infarction (MI) (35.2% vs 23.1%, p = 0.025) and lower rates of ST-elevation MI (53.8% vs 71.7%, p = 0.002), although both G were similar regarding MI location (anterior – p = 0.135, inferior – p = 0.097). LSVT G had higher systolic blood pression (130 ± 33 vs 122 ± 33, p = 0.050), presented more frequently in Killip-Kimball class ≥2 (52.5% vs 35.5%, p = 0.005) and with atrial fibrillation (21.2% vs 12.4%, p = 0.045), and had higher brain-natriuretic peptide (1075 vs 329, p < 0.001). LSVT G was treated more frequently with diuretics (80.0% vs 47.8%, p < 0.001), amiodarone (62.2% vs 48.8%, p = 0.029), digoxin (8.9% vs 2.4%, p = 0.013) and levosimendan (11.1% vs 2.8%, p = 0.004). ESVT G had higher rates of performed coronarography (88.4% vs 79.1%, p = 0.028) but lower rate of 3 vessels disease (58.5% vs 70.8%, p = 0.017). LSVT G had higher rates of severe (<30%) left ventricle dysfunction (32.9% vs 15.4%, p < 0.001) and need to non-invasive ventilation (23.1% vs 6.8%, p < 0.001). Regarding in-hospital complications, ESVT G had higher rates of heart failure (34.7% vs 19.1%, p = 0.006), atrioventricular block (15.7% vs 1.1%, p < 0.001), atrial fibrillation (20.4% vs 7.7%, p = 0.006) and major haemorrhage (5.2% vs 0.0%, p = 0.024). LSVT G had higher rates of in-hospital death (44.4% vs 20.9%, p < 0.001) and in-hospital stay (14 days vs 7 days, p < 0.001). The G were similar regarding re-infarction (p = 0.216), shock (p = 0.179), mechanical complications (p = 1.00), cardiac arrest (p = 0.097) and stroke (0.348) rates. Logistic regression confirmed ESVT was predictive in-hospital heart failure (p = 0.010, OR 2.67) and de novo AF (p = 0.001, OR 5.56), whether LSVT was predictive of in-hospital death (p = 0.002, OR 2.70). Conclusion LSVT was associated with higher rates of in-hospital complications, but ESVT was associated with higher in-hospital mortality.
Funding Acknowledgements Type of funding sources: None. OnBehalf Portuguese Registry of Acute Coronary Syndromes Background The presence of atrioventricular block (AVB) in ST-elevation myocardial infarction (STEMI) is more frequently registered when is identified in the inferior leads. However, AVB maybe occurs in anterior STEMI, yet the AVB and STEMI localization maybe had different implications. Objective Evaluate the impact and prognosis of AVB according to the STEMI localization. Methods Multicenter retrospective study, based on the Portuguese Registry of Acute Coronary Syndrome between 1/10/2010-3/05/2020. Patients were divided into two groups: A – patients with anterior STEMI, and B – patients with inferior STEMI. Were excluded patients without a previous cardiovascular history or clinical data regarding AVB occurrence. Logistic regression was performed to assess AVB as a prognostic marker in STEMI patients. Results From 32157 patients, was identified 462 with AVB, 72 in group A (15.6%) and 390 in group B (84.4%). Both groups were similar regarding gender (p = 0.710), age (p = 0.068), body mass index (p = 0.535), admitly directly to cat lab (p = 0.635), initial symptons until first medical contact (p = 0.561), smoker status (p = 0.483), diabetes mellitus (p = 0.331), coronary artery disease (p = 0.053), previous stroke (p = 0.332), peripheral artery disease (p = 0.348), chronic kidney disease (p = 0.425), systolic blood pressure (p = 0.057), multivessel diasease (p = 0.235), new-onset of atrial fibrillation (p = 0.582), cardiac arrest (p = 0.062) and stroke complication (p = 0.685). Group B had higher left ventricular ejection fraction (LVEF) >50% (16.9 vs 60.7%, p < 0.001). On the other hand, group A had more arterial hypertension (79.7 vs 66.2%, p = 0.027), dislipidaemia (58.2 vs 54.4%, p = 0.038), heart rate at admission (81 ± 20 vs 59 ± 23, p < 0.001), Killip-Kimball class > I (45.7 vs 29.6%, p = 0.008), sinus rhythm at admission (84.5 vs 72.6%, p = 0.035), heart failure complication (65.3 vs 37.1%, p < 0.001), cardiogenic shock complication (42.3 vs 24.7%, p < 0.001), ACS mechanical complication (8.3 vs 3.1%, p = 0.047), sustained ventricular tachycardia during ACS hospitalization (19.4 vs 8.5%, p = 0.005) and hospitalization death (52.9 vs 44.7%, p < 0.001). Logistic regression revealed that AVB in inferior STEMI was a predictor of new-onset of atrial fibrillation (odds ratio (OR) 3.817, p = 0.038, confidence interval (CI) 1.123-12.975), with a R2 Nagelkerke 24.4. Also, revealed that AVB in anterior STEMI was a predictor of death (OR 0.111, p < 0.001, CI 0.034-0.366), with a R2 Nagelkerke 55.2. Conclusions AVB in inferior STEMI was a predictor of new-onset of atrial fibrillation and AVB in anterior STEMI was a predictor of death.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.