Archives of Cardiovascular Diseases最新文献

Background: Patients resuscitated after out-of-hospital cardiac arrest (OHCA) have a poor prognosis, with high death rates. Multiple scoring systems have been developed to predict survival in all-comers with OHCA. Acute coronary syndromes and ST-segment elevation myocardial infarction (STEMI) are the primary causes of OHCA. Recently, a specific prognostic score (Tran risk model) was developed for patients with STEMI-related OHCA.

Aim: To compare the accuracy of established non-STEMI-specific prognostic scores (OHCA, modified CAHP and NULL-PLEASE) with the Tran risk model in predicting in-hospital death among patients with STEMI-related OHCA.

Methods: This was an observational single-centre study including 315 consecutive patients treated for STEMI-related OHCA. The OHCA score was calculated for 310 patients (98.4%), the NULL-PLEASE and modified CAHP (mCAHP) scores were calculated for 308 patients (97.8%) and the Tran risk model score was calculated for 306 patients (97.1%). A C-statistic analysis was performed to determine score performance.

Results: The area under the curve (AUC) for the Tran risk model was 0.75 (95% confidence interval [CI] 0.69-0.79). The AUCs for the OHCA, mCAHP and NULL-PLEASE scores were 0.74 (95% CI 0.69-0.80), 0.74 (95% CI 0.69-0.80) and 0.76 (95% CI 0.71-0.82), respectively. There was no significant difference in AUCs between the Tran risk model and the mCAHP score (P=0.95), the NULL-PLEASE score (P=0.42) or the OHCA score (P=0.93). Similarly, no significant difference was observed between the mCAHP, NULL-PLEASE and OHCA scores. Predictors of death were no-flow duration, diabetes, blood lactate, femoral access and age>75 years.

Conclusions: The OHCA, NULL-PLEASE and mCAHP scores and the Tran risk model showed moderate to good performance in predicting in-hospital death in patients with STEMI-related OHCA. No differences in accuracy were found between non-STEMI-specific scores and the Tran risk model developed for patients with STEMI-related OHCA.

Diagnosing cancer therapy-related cardiovascular toxicities may be a challenge. The interplay between cancer and cardiovascular diseases, beyond shared cardiovascular and cancer risk factors, and the increasingly convoluted cancer therapy schemes have complicated cardio-oncology. Biomarkers used in cardio-oncology include serum, imaging and rhythm modalities to ensure proper diagnosis and prognostic stratification of cardiovascular toxicities. For now, troponin and natriuretic peptides, multimodal cardiovascular imaging (led by transthoracic echocardiography combined with cardiac magnetic resonance or computed tomography angiography) and electrocardiography (12-lead or Holter monitor) are cornerstones in cardio-oncology. However, the imputability of cancer therapies is sometimes difficult to assess, and more refined biomarkers are currently being studied to increase diagnostic accuracy. Advances reside partly in pathophysiology-based serum biomarkers, improved cardiovascular imaging through new technical developments and remote monitoring for rhythm disorders. A multiparametric omics approach, enhanced by deep-learning techniques, should open a new era for biomarkers in cardio-oncology in the years to come.

Background: Transcatheter aortic valve implantation may be associated with significant haemorrhagic complications.

Aims: To evaluate the timing, incidence, predictors and clinical impact of bleeding events after transcatheter aortic valve implantation, according to the updated Valve Academic Research Consortium (VARC)-3 criteria, compared with the VARC-2 criteria.

Methods: A retrospective observational study involving 487 consecutive patients who underwent transcatheter aortic valve implantation between July 2017 and May 2019 was performed. Bleeding events were classified according to the VARC-2 and VARC-3 definitions.

Results: Bleeding events occurred in 17.6% of patients, with early bleeding (in-hospital) in 12.5% and late bleeding (occurring after discharge) in 6.1%. The primary vascular access site was the most common source of early bleeding, whereas gastrointestinal bleeding was predominant in late events. Significant predictors of early VARC-3-defined bleeding included active cancer, previous implantable cardioverter-defibrillator, history of mitral valve surgery, a non-transfemoral approach and occurrence of an in-hospital major vascular complication or new-onset atrial fibrillation. Late bleeding was independently associated with a history of myocardial infarction and treatment with vitamin K antagonists at discharge. Early bleeding events were not associated with increased late all-cause mortality. No significant difference was observed based on the VARC-2 and VARC-3 bleeding definitions.

Conclusions: Bleeding events occurred in one sixth of patients undergoing transcatheter aortic valve implantation without significant difference in their incidence between the VARC-2 and VARC-3 classifications. Early bleeding events were not associated with poorer long-term survival, regardless of the classification used. Larger studies with greater statistical power, including more contemporary patients, are needed to confirm these findings.

Background: Although recreational drug use is a strong risk factor for acute cardiovascular events, systematic testing is currently not performed in patients admitted to intensive cardiac care units, with a risk of underdetection. To address this issue, machine learning methods could assist in the detection of recreational drug use.

Aims: To investigate the accuracy of a machine learning model using clinical, biological and echocardiographic data for detecting recreational drug use in patients admitted to intensive cardiac care units.

Methods: From 07 to 22 April 2021, systematic screening for all traditional recreational drugs (cannabis, opioids, cocaine, amphetamines, 3,4-methylenedioxymethamphetamine) was performed by urinary testing in all consecutive patients admitted to intensive cardiac care units in 39 French centres. The primary outcome was recreational drug detection by urinary testing. The framework involved automated variable selection by eXtreme Gradient Boosting (XGBoost) and model building with multiple algorithms, using 31 centres as the derivation cohort and eight other centres as the validation cohort.

Results: Among the 1499 patients undergoing urinary testing for drugs (mean age 63±15 years; 70% male), 161 (11%) tested positive (cannabis: 9.1%; opioids: 2.1%; cocaine: 1.7%; amphetamines: 0.7%; 3,4-methylenedioxymethamphetamine: 0.6%). Of these, only 57% had reported drug use. Using nine variables, the best machine learning model (random forest) showed good performance in the derivation cohort (area under the receiver operating characteristic curve=0.82) and in the validation cohort (area under the receiver operating characteristic curve=0.76).

Conclusions: In a large intensive cardiac care unit cohort, a comprehensive machine learning model exhibited good performance in detecting recreational drug use, and provided valuable insights into the relationships between clinical variables and drug use through explainable machine learning techniques.