Deep learning-based NT-proBNP prediction from the ECG for risk assessment in the community.

OBJECTIVES: The biomarker N-terminal pro B-type natriuretic peptide (NT-proBNP) has predictive value for identifying individuals at risk for cardiovascular disease (CVD). However, it is not widely used for screening in the general population, potentially due to financial and operational reasons. This study aims to develop a deep-learning model as an efficient means to reliably identify individuals at risk for CVD by predicting serum levels of NT-proBNP from the ECG. METHODS: A deep convolutional neural network was developed using the population-based cohort study Hamburg City Health Study (HCHS, n=8,253, 50.9 % women). External validation was performed in two independent population-based cohorts (SHIP-START, n=3,002, 52.1 % women, and SHIP-TREND, n=3,819, 51.2 % women). Assessment of model performance was conducted using Pearson correlation (R) and area under the receiver operating characteristics curve (AUROC). RESULTS: NT-proBNP was predictable from the ECG (R, 0.566 [HCHS], 0.642 [SHIP-START-0], 0.655 [SHIP-TREND-0]). Across cohorts, predicted NT-proBNP (pNT-proBNP) showed good discriminatory ability for prevalent and incident heart failure (HF) (baseline: AUROC 0.795 [HCHS], 0.816 [SHIP-START-0], 0.783 [SHIP-TREND-0]; first follow-up: 0.669 [SHIP-START-1, 5 years], 0.689 [SHIP-TREND-1, 7.3 years]), comparable to the discriminatory value of measured NT-proBNP. pNT-proBNP also demonstrated comparable results for other incident CVD, including atrial fibrillation, stroke, myocardial infarction, and cardiovascular death. CONCLUSIONS: Deep learning ECG algorithms can predict NT-proBNP concentrations with high diagnostic and predictive value for HF and other major CVD and may be used in the community to identify individuals at risk. Long-standing experience with NT-proBNP can increase acceptance of such deep learning models in clinical practice.

Risk Factors, Subsequent Disease Onset, and Prognostic Impact of Myocardial Infarction and Atrial Fibrillation.

Background Although myocardial infarction (MI) and atrial fibrillation (AF) are frequent comorbidities and share common cardiovascular risk factors, the direction and strength of the association of the risk factors with disease onset, subsequent disease incidence, and mortality are not completely understood. Methods and Results In pooled multivariable Cox regression analyses, we examined temporal relations of disease onset and identified predictors of MI, AF, and all-cause mortality in 108 363 individuals (median age, 46.0 years; 48.2% men) free of MI and AF at baseline from 6 European population-based cohorts. During a maximum follow-up of 10.0 years, 3558 (3.3%) individuals were diagnosed exclusively with MI, 1922 (1.8%) with AF but no MI, and 491 (0.5%) individuals developed both MI and AF. Association of sex, systolic blood pressure, antihypertensive treatment, and diabetes appeared to be stronger with incident MI than with AF, whereas increasing age and body mass index showed a higher risk for incident AF. Total cholesterol and daily smoking were significantly related to incident MI but not AF. Combined population attributable fraction of cardiovascular risk factors was >70% for incident MI, whereas it was only 27% for AF. Subsequent MI after AF (hazard ratio [HR], 1.68; 95% CI, 1.03-2.74) and subsequent AF after MI (HR, 1.75; 95% CI, 1.31-2.34) both significantly increased overall mortality risk. Conclusions We observed different associations of cardiovascular risk factors with both diseases indicating distinct pathophysiological pathways. Subsequent diagnoses of MI and AF significantly increased mortality risk.

Alcohol consumption, cardiac biomarkers, and risk of atrial fibrillation and adverse outcomes.

AIMS: There is inconsistent evidence on the relation of alcohol intake with incident atrial fibrillation (AF), in particular at lower doses. We assessed the association between alcohol consumption, biomarkers, and incident AF across the spectrum of alcohol intake in European cohorts. METHODS AND RESULTS: In a community-based pooled cohort, we followed 107 845 individuals for the association between alcohol consumption, including types of alcohol and drinking patterns, and incident AF. We collected information on classical cardiovascular risk factors and incident heart failure (HF) and measured the biomarkers N-terminal pro-B-type natriuretic peptide and high-sensitivity troponin I. The median age of individuals was 47.8 years, 48.3% were men. The median alcohol consumption was 3 g/day. N = 5854 individuals developed AF (median follow-up time: 13.9 years). In a sex- and cohort-stratified Cox regression analysis alcohol consumption was non-linearly and positively associated with incident AF. The hazard ratio for one drink (12 g) per day was 1.16, 95% CI 1.11-1.22, P < 0.001. Associations were similar across types of alcohol. In contrast, alcohol consumption at lower doses was associated with reduced risk of incident HF. The association between alcohol consumption and incident AF was neither fully explained by cardiac biomarker concentrations nor by the occurrence of HF. CONCLUSIONS: In contrast to other cardiovascular diseases such as HF, even modest habitual alcohol intake of 1.2 drinks/day was associated with an increased risk of AF, which needs to be considered in AF prevention.

Outcome of pregnancy in a contemporary cohort of adults with congenital heart disease-a 10-year, single-center experience.

BACKGROUND: Pregnancy may be associated with adverse outcome in women with congenital heart disease (CHD). However, data regarding the outcome of pregnancy in women with CHD who receive care in cardiac-obstetric expert units are limited. METHODS: We retrospectively analyzed baseline characteristics and outcome of pregnancy in 67 females with CHD who received medical care in our tertiary center for 61 singleton and 6 twin pregnancies between 2009 and 2018. RESULTS: According to the modified World Health Organization (mWHO) risk scale for pregnancy, CHD lesions in 39 enrolled women (58%) were classified as mWHO class I or II, and in 28 females (42%) as mWHO class III or IV. Preterm births were more frequent in mWHO classes III or IV (P=0.003). Cardiac signs and complications occurred more often in mWHO classes III or IV than in women with cardiac lesions assigned to mWHO classes I or II (42.9% vs. 7.7%, P=0.002). N-terminal pro B-type natriuretic peptide (NT-proBNP) levels during pregnancy were higher in mWHO classes III or IV than in mWHO classes I or II (median 269.0 vs. 115.5 pg/mL, P=0.019). Presence of functional NYHA class III [odds ratio (OR) per standard deviation (SD) 8.8, 95% confidence interval (CI): 2.2-57.2, P=0.008] and mWHO classes III/IV (OR per SD 3.4, 95% CI: 1.2-9.9, P=0.018) prior to pregnancy were identified as independent predictors of adverse cardiac outcome of pregnancy. CONCLUSIONS: Adverse cardiac events and preterm deliveries should be anticipated in pregnant women with CHD, especially in those with mWHO classes III or IV. Therefore, these pregnancies should be under close surveillance and managed in specialized, multidisciplinary tertiary referral centers. Preconception counseling including individualized risk assessment is strongly recommended in women with CHD.

Pancreatic resection for intraductal papillary mucinous neoplasm- a thirteen-year single center experience.

BACKGROUND: The purpose of this study is to review our results for pancreatic resection in patients with intraductal papillary mucinous neoplasm (IPMN) with and without associated carcinoma. METHODS: A total of 54 patients undergoing pancreatic resection for IPMN in a single university surgical center (Medical University of Graz) were reviewed retrospectively. Their survival rates were compared to those of patients with pancreatic ductal adenocarcinoma. RESULTS: Twenty-four patients exhibit non-invasive IPMN and thirty patients invasive IPMN with associated carcinoma. The mean age is 67 (+/-11) years, 43 % female. Surgical strategies include classical or pylorus-preserving Whipple procedure (n = 30), distal (n = 13) or total pancreatectomy (n = 11), and additional portal venous resection in three patients (n = 3). Median intensive care stay is three days (range 1 - 87), median in hospital stay is 23 days (range 7 - 87). Thirty-day mortality is 3.7 %. Median follow up is 42 months (range 0 - 127). One-, five- and ten-year overall actuarial survival is 87 %; 84 % and 51 % respectively. Median overall survival is 120 months. Patients with non-invasive IPMN have significantly better survival than patients with invasive IPMN and IPMN-associated carcinoma (p < 0.008). In the subgroup of invasive IPMN with associated carcinoma, a positive nodal state, perineural invasion as well as lymphovascular infiltration are associated with poor outcome (p < 0.0001; <0.0001 and =0.001, respectively). Elevated CA 19-9(>37 U/l) as well as elevated lipase (>60 U/l) serum levels are associated with unfavorable outcome (p = 0.009 and 0.018; respectively). Patients operated for pancreatic ductal adenocarcinoma show significantly shorter long-term survival than patients with IPMN associated carcinoma (p = 0.001). CONCLUSIONS: Long-term outcome after pancreatic resection for non-invasive IPMN is excellent. Outcome after resection for invasive IPMN with invasive carcinoma is significantly better than for pancreatic ductal adenocarcinoma. In low- and intermediate risk IPMN with no clear indication for immediate surgical resection, a watchful waiting strategy should be evaluated carefully against surgical treatment individually for each patient.

Portal vein resection in advanced pancreatic adenocarcinoma: is it worth the risk?

INTRODUCTION: Portal vein resection represents a viable add-on option in standard pancreaticoduodenectomy for locally advanced ductal pancreatic adenocarcinoma, but is often underused as it may set patients at additional risk for perioperative and postoperative morbidity and mortality. We aimed to review our long-term experience to determine the additive value of this intervention for locally advanced pancreatic adenocarcinoma. PATIENTS AND METHODS: Single, university surgical center audit over a 13-year period; cohort comprised 221 consecutive patients undergoing pancreatic resection; in 47 (21 %) including portal vein resection. Predictors for short- and long-term survival were assessed via multivariate logistic and Cox regression. RESULTS: Baseline and perioperative characteristics were similar between the two groups. However, overall skin-to-skin times, intraoperative transfusion requirements as the need for medical inotropic support were higher in patients undergoing additional portal vein resection (p < 0.0001; p = 0.001 and p = 0.03). Postoperative complication rates were 34 vs. 35 % (p = 0.89), 14 patients (5 % vs. 11 %; p = 0.18) died in-hospital. An American Society of Anesthesiologists Score >2 was the only independent predictor for in-hospital mortality (OR 10.66, 95 % CI 1.24-91.30). Follow-up was complete in 99.5 %, one-year survival was 59 % vs. 70 % and five-year overall survival 15 % vs. 12 % with and without portal vein resection, respectively (Log rank: p = 0.25). For long-term outcome, microvascular invasion (HR 2.03, 95 % CI 1.10-3.76) and preoperative weight loss (HR 2.17, 95 % CI 1.31-3.58) were independent predictors. CONCLUSION: Despite locally advanced disease, patients who underwent portal vein resection had no worse perioperative and overall survival than patients with lower staging and standard pancreaticoduodenectomy only. Therefore, the feasibility of portal vein resection should be evaluated in every potential candidate at risk.

Pancreas Cancer Surgery in Octogenarians - Should We or Should We Not?

BACKGROUND/AIM: In this study we aimed to determine if advanced age represents a risk factor for negative perioperative and long-term outcome in patients undergoing curative surgery ductal pancreatic adenocarcinoma surgery. PATIENTS AND METHODS: Two-hundred-twenty-one consecutive patients, twelve (6%) patients ≥80 years were included in the study. We assessed perioperative and long-term outcome and independent predictors for in-hospital mortality with Cox regression analysis. RESULTS: Advanced age was not a predictor for in-hospital mortality (6.3% in non-octogenarian versus 8.3% in octogenarians; p=0.55) nor for morbidity (31% vs. 32%; p=0.69). An ASA score >II was the only predictor for in-hospital mortality (odds ratio (OR)=10.10, 95%CI=1.28-79.60; Hosmer-Lemeshow: p=0.86). No significant difference was observed in one- and five-year survival rates (68 and 58% vs. 16 and 14%; log-rank p=0.61). CONCLUSION: Advanced age is not a risk factor for negative outcome in curative pancreatic cancer surgery. Therefore, this single curative option should be considered in octogenarians at risk.