Use of Electrocardiogram Monitoring in Adult Patients Taking High-Risk QT Interval Prolonging Medicines in Clinical Practice: Systematic Review and Meta-analysis

Risk Stratification of QTc Prolongations in Hospitalized Cardiology and Gastroenterology Patients Using the Tisdale Score-A Retrospective Analysis

Background/Objectives: QTc prolongation can result in lethal arrhythmia. Risk scores like the Tisdale score can be used for risk stratification for targeted pharmaceutical interventions. However, the practical usability across different medical specialties has not been sufficiently investigated. The aim of this study was to compare relevant risk factors for QTc prolongation and to investigate the use of the Tisdale score in cardiology and gastroenterology patients. Methods: For patients on a cardiology and a gastroenterology ward receiving a weekly pharmaceutical electronic chart review, risk factors for QTc prolongation, QTc-prolonging drugs, and electrocardiograms (ECGs) were retrospectively collected for a four-month period (07-10/2023), and the Tisdale score and its sensitivity and specificity were calculated. Results: A total of 627 chart reviews (cases) (335 cardiology, 292 gastroenterology) were performed. The median age was 66 (range 20-94) years, and 39% (245) of patients were female. The presence of established risk factors (hypokalemia, renal impairment, age ≥ 68 years, cardiac diseases) differed significantly between the specialties. A median of 2 (range 0-5) QTc-prolonging drugs were prescribed in both groups. Baseline and follow-up ECG were recorded in 166 (50%) cardiology cases, of which prolonged QTc intervals were detected in 38 (23%) cases. In the 27 (9%) gastroenterology cases with baseline and follow-up ECG, no QTc prolongations were detected. Across both specialties, the Tisdale score achieved a sensitivity of 74% and a specificity of 30%. Conclusions: The presence of established risk factors for QTc prolongation differed significantly between cardiology and gastroenterology cases. The Tisdale score showed acceptable sensitivity for risk stratification; however, the limited availability of ECGs for gastroenterology cases was a limiting factor.

Determining sensitivity and specificity of risk scores for QTc interval prolongation in hemato-oncology patients prescribed systemic antifungal therapy: a retrospective cross-sectional study

BACKGROUND

QTc interval prolongation can result in potentially lethal arrhythmias. One risk factor is QTc-prolonging drugs, including some antifungals often used in hemato-oncology patients. Screening tools for patients at risk have not yet been investigated in this patient population.

AIM

Our aim was to evaluate the sensitivity and specificity of five QTc risk scores in hemato-oncology patients receiving systemic antifungal therapy.

METHOD

Data were retrieved from an internal study database including adult hemato-oncology patients prescribed systemic antifungal therapy. Data on QTc-prolonging medication, risk factors for QTc prolongation, and electrocardiograms (ECG) were collected retrospectively for a period of 12 months. The QTc risk scores according to Tisdale, Vandael, Berger, Bindraban, and Aboujaoude as well as their sensitivity and specificity were calculated.

RESULTS

During the evaluated period, 77 patients were prescribed systemic antifungals resulting in 187 therapy episodes. Regarding therapy episodes, median age was 56 years (IQR 44-68), 41% (77) were female, and a median of 3 QTc-prolonging drugs were prescribed (range 0-6). ECGs were available for 45 (24%) of the therapy episodes 3-11 days after initiation of the antifungal therapy, 22 of which showed QTc prolongation. Regarding these 45 therapy episodes, sensitivity and specificity of the risk scores were calculated as follows: Tisdale 86%/22%, Vandael 91%/35%, Berger 32%/83%, Bindraban 50%/78%, Aboujaoude 14%/87%.

CONCLUSION

The QTc risk scores according to Tisdale and Vandael showed sufficient sensitivity for risk stratification in the studied patient population. In contrast, risk scores according to Berger, Bindraban, and Aboujaoude cannot be considered suitable due to poor sensitivity.

Enhancing Cardiovascular Risk Prediction: Development of an Advanced Xgboost Model with Hospital-Level Random Effects

BACKGROUND

Ensemble tree-based models such as Xgboost are highly prognostic in cardiovascular medicine, as measured by the Clinical Effectiveness Metric (CEM). However, their ability to handle correlated data, such as hospital-level effects, is limited.

OBJECTIVES

The aim of this work is to develop a binary-outcome mixed-effects Xgboost (BME) model that integrates random effects at the hospital level. To ascertain how well the model handles correlated data in cardiovascular outcomes, we aim to assess its performance and compare it to fixed-effects Xgboost and traditional logistic regression models.

METHODS

A total of 227,087 patients over 17 years of age, undergoing cardiac surgery from 42 UK hospitals between 1 January 2012 and 31 March 2019, were included. The dataset was split into two cohorts: training/validation (n = 157,196; 2012-2016) and holdout (n = 69,891; 2017-2019). The outcome variable was 30-day mortality with hospitals considered as the clustering variable. The logistic regression, mixed-effects logistic regression, Xgboost and binary-outcome mixed-effects Xgboost (BME) were fitted to both standardized and unstandardized datasets across a range of sample sizes and the estimated prediction power metrics were compared to identify the best approach.

RESULTS

The exploratory study found high variability in hospital-related mortality across datasets, which supported the adoption of the mixed-effects models. Unstandardized Xgboost BME demonstrated marked improvements in prediction power over the Xgboost model at small sample size ranges, but performance differences decreased as dataset sizes increased. Generalized linear models (glms) and generalized linear mixed-effects models (glmers) followed similar results, with the Xgboost models also excelling at greater sample sizes.

CONCLUSIONS

These findings suggest that integrating mixed effects into machine learning models can enhance their performance on datasets where the sample size is small.