Session PB1.1

Analysis of T Wave Morphology Parameters with Signal Averaging during Ischemia Induced by Percutaneous Transluminal Coronary Angioplasty

FH Baglivo, PD Arini*, JP Martínez, P Laguna

Instituto Argentino de Matemática
Buenos Aires, Argentina

Electrocardiographic repolarization abnormalities detected by Principal Component Analysis Ratio: the ratio of the 2nd to the 1st eigenvalue (PCAr) and the absolute and relative T-wave Residuum (TWRa and TWRr) have been proposed as risk markers for severe myocardial arrhythmia (Kesek et al. 2004, Pac. Clin. Electrophysiol). The aim of this work is to examine the effect of signal averaging on the mean value and reproducibility of PCAr, TWRa and TWRr as measures of T wave morphology. This is motivated in the fact that non-dominant eigenvalues are usually contaminated by noise, thus not reflecting properly the T wave morphology. For that purpose, we used a study group consisting of 93 patients receiving PTCA in one of the coronary arteries. A 5-min control ECG record and the ECG record during the PTCA procedure were used. The 8 algebraically independent leads (I, II, V1-V6) were used. The PCAr, TWRa and TWRr temporal series were measured at different times during the procedure: just before the start of occlusion taken as control (c), one (o1), two (o2) and three (o3) minutes after the occlusion start. All values of TWRa and TWRr obtained when averaging 10 consecutive beats (10cx) are significantly lower than those obtained from single beats (1cx) (*p<0.0001), while PCAr does not show significant differences when averaging. The results, mean±SEM, are: TWRa in ‘1cx: (c)=87661, 4±7134, 7 uV2; o1=184464, 4±26410, 9 uV2; o2= 211693, 3±29896, 4 uV2; o3=200833, 4±23807, 2 uV2 and in ’10cx’ (c) =12726, 2±955, 2* uV2; o1= 37618, 4±4798, 9* uV2; o2=46609, 0±4776, 9* uV2; o3= 40192, 3±4281, 9* uV2. TWRr in ‘1cx’: (c) = 0, 29±0, 04 %; o1=0, 65±0, 10 %; o2=0, 59±0, 10 %; o3=0, 61±0, 10 % and in ‘10cx’ (c) =0, 05±0, 01* %; o1= 0, 10±0, 01 *%; o2=0, 17±0, 02* %; o3=0, 15±0, 02* %. We also computed the intra-subject standard deviation (i.e. reproducibility) in control situation. The values were: TWRa 27339 uV2 (1cx) vs. 3808* uV2 (10cx), TWRr 0, 09 % (1cx) vs. 0, 01* %(10cx) and PCA 1, 27 %(1cx) vs. 0, 99* % (10cx) (NS). The obtained results indicate that: 1) TWRa and TWRr computed from 10cx show lower values, and higher reproducibility than those obtained from 1cx. This suggests that a big amount of TWRa and TWRr measured on one beat is due to noise. 2) PCAr with averaging did not show statistical significant with respect to PCAr on single beats; possibly due to the numerator of PCAr (2nd eigenvalue) is not so much affected by noise as it is the numerator of TWR; obtained from the smaller 4-to-8 eigenvalues. 3) Results observed during ischemia are concordant with other study from healthy subject (Batdorf et al. 2005, J. of Electrocardiology).

(Abstract Control Number: 2)