A carregar...
Publicação
Time Domain Analysis of Heart Rate Variability Metrics to Assess Coronary Artery Disease in Patients Participating in a Long-term Cardiac Rehabilitation Program - An Exploratory Study
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 8.14 MB | Adobe PDF |
Orientador(es)
Resumo(s)
This study evaluated whether time-domain heart rate variability (HRV) metrics can noninvasively distinguish coronary artery disease (CAD) patients in a Cardiac Rehabilitation (CR) program from apparently healthy older adults at rest. Twenty three CAD patients (mean age 62.9 ± 11.5 years), enrolled for at least 6 months in CR, underwent an adapted Ewing protocol with continuous ECG Holter recording: five minutes of supine rest, deep breathing, a 20-second Valsalva maneuver, and two minutes of isometric handgrip. Nineteen healthy controls (mean age 74.1 ± 3.9 years) were selected from the Fantasia database, which provides 120-minute ECG and respiratory data during film viewing. Comparative analysis focused on the initial five-minute resting ECG common to both cohorts. Preprocessing included bandpass filtering (0.05-40Hz), validated R-peak detection, and manual removal of ectopic and artifactual beats. Time-domain HRV features - meanRR, SDNN, RMSSD, NN50, and pNN50 - were extracted. At rest, CAD patients exhibited significantly reduced HRV compared to controls: SDNN (36.90 ± 20.10 ms vs. 55.79 ± 27.82 ms, p < 0.01), RMSSD (23.37 ± 15.33 ms vs. 46.77 ± 36.49 ms, p < 0.001), and pNN50 (4.45 ± 6.46 % vs. 7.78 ± 8.01 %, p < 0.05). Coefficients of variation (CV) revealed greater interindividual variability among CAD patients, particularly in pNN50 (CV = 1.45 vs. 1.03), suggesting more heterogeneous parasympathetic modulation. Principal component analysis followed by Ridge Logistic Regression yielded strong classification performance (AUC: 0.943 ± 0.023; sensitivity: 0.951 ± 0.029; specificity: 0.947 ± 0.007). RMSSD and SDNN were the most discriminative features. These findings support the use of time-domain HRV as a noninvasive marker of autonomic dysfunction in CAD. Future work should integrate frequency-domain and nonlinear metrics and balance demographics.
Descrição
Tese de Mestrado, Engenharia Biomédica e Biofísica, 2025, Universidade de Lisboa, Faculdade de Ciências
Palavras-chave
CAD HRV Autonomic Function RMSSD SDNN