O Nasario-Junior1, P R Benchimol-Barbosa, G A Trevizani, M Marocolo, J Nadal. 1. Laboratório de Processamento de Sinais, Programa de Engenharia Biomédica, COPPE, Universidade Federal do Rio de Janeiro, PO box: 68510, Rio de Janeiro, RJ 21941-972, Brazil. Electronic address: olivasse@hotmail.com.
Abstract
BACKGROUND: The athlete's heart represents a reversible structural and functional adaptations of myocardial tissue developed through physical conditioning. Surface electrocardiogram (ECG) has the capability to detect myocardial hypertrophy but has limited performance in monitoring physical conditioning-induced myocardial remodeling. The aim of this study was to develop an ECG-derived test for detecting incipient myocardial hypertrophy in well-conditioned athletes based on a principal components (PC) analysis. METHODS: Two groups of study composed of 14 sedentary healthy volunteers (CONTROL GROUP) and 14 professional long distance runners (Athlete group) had their maximal metabolic equivalents (MET) estimated (mean ± SD: CONTROL GROUP: 9 ± 2 METs vs. Athlete group: 20 ± 1 METs, p<0.05). All participants had their high-resolution ECG (HRECG) recorded, and a 120 ms segment starting at the QRS complex onset and ending in the ST segment was extracted to build a data matrix for PC analysis. The Mahalanobis distance was evaluated by a logistic regression model to determine the optimal separation threshold between groups. HRECG was also analyzed using the classical time domain approach. The comparison of areas under the receiver operating characteristic curve (c-statistic) in 10,000 bootstrap re-samplings measured how well each method detected physical conditioning (α<0.05). RESULTS: Average bootstrap c-statistic for PC analysis and time domain approaches were 0.98 and 0.79 (p<0.05), respectively. PC analysis and maximal oxygen consumption exhibited comparable performances to distinguish between groups. DISCUSSION: The PC analysis method applied to HRECG signals appropriately discriminates well-conditioned athletes from healthy, sedentary subjects.
BACKGROUND: The athlete's heart represents a reversible structural and functional adaptations of myocardial tissue developed through physical conditioning. Surface electrocardiogram (ECG) has the capability to detect myocardial hypertrophy but has limited performance in monitoring physical conditioning-induced myocardial remodeling. The aim of this study was to develop an ECG-derived test for detecting incipient myocardial hypertrophy in well-conditioned athletes based on a principal components (PC) analysis. METHODS: Two groups of study composed of 14 sedentary healthy volunteers (CONTROL GROUP) and 14 professional long distance runners (Athlete group) had their maximal metabolic equivalents (MET) estimated (mean ± SD: CONTROL GROUP: 9 ± 2 METs vs. Athlete group: 20 ± 1 METs, p<0.05). All participants had their high-resolution ECG (HRECG) recorded, and a 120 ms segment starting at the QRS complex onset and ending in the ST segment was extracted to build a data matrix for PC analysis. The Mahalanobis distance was evaluated by a logistic regression model to determine the optimal separation threshold between groups. HRECG was also analyzed using the classical time domain approach. The comparison of areas under the receiver operating characteristic curve (c-statistic) in 10,000 bootstrap re-samplings measured how well each method detected physical conditioning (α<0.05). RESULTS: Average bootstrap c-statistic for PC analysis and time domain approaches were 0.98 and 0.79 (p<0.05), respectively. PC analysis and maximal oxygen consumption exhibited comparable performances to distinguish between groups. DISCUSSION: The PC analysis method applied to HRECG signals appropriately discriminates well-conditioned athletes from healthy, sedentary subjects.