13th International Congress
THE "NEW FRONTIERS"
OF ARRHYTHMIAS 1998

January 24-31, 1998
Marilleva, Trento, Italy

RT-43

A new computer programme for Holter ECG (3 orthogonal leads) to evaluate dynamically the QT dispersion, as well as its correlation to RR intervals

Antonio Conti, Antonio Michelucci*, Simone Mininni*, Mauro Lenuzza*, Andrea A. Conti*, Chiara Lazzeri**, Giuseppe Petrogalli***, Giuseppe Catto***, Franco Franchi**, Gian Franco Gensini*.
Dept. of Clinical Physiopathology, *Dept. of Internal Medicine and Cardiology, **Dept. of Internal Medicine and Gastroenterology, University of Florence, ***ELA Medical, Italy.

Material and methods

The starting point of this study were the Holter ECG recordings with 3 orthogonal leads. The Holter recorder used was the Ela Medical ELATEC (solid state) recorder. ELATEC-HRV and -QT (Ela Medical) are 2 dedicated software programmes which permit the extraction of many different parameters from a Holter ECG (solid state recorders) with 3 leads. The 3 leads were positioned in an orthogonal or pseudo-orthogonal way⁸. Among the data available in the thesaurus are the 2880 mean values of the RR intervals, and of the QT apex and QT end (tangent method) intervals in each of the 3 leads (all mean values are averaged over 30 seconds).

A customized programme has been written which (i) exports the data in ASCII format, (ii) computes the QT dispersion (defined as QTmax-QTmin) for the 3 leads, (iii) computes simple descriptive statistical parameters for desired day intervals, (iv) computes linear and non-linear regressions, and (v) allows many different graphical presentations of the data. The programme also has some routines to study both the time- and the frequency-domain. Finally some further routines are present to study possible concomitant variations of different parameters, as for example the QT dispersion and the RR interval, and to eventually define new indices which could be useful in clinical studies.

In the first part of the study the main attention has been devoted to the values of the QT dispersion (only the QT end values have been considered), of the RR intervals, and of their possible correlations and clusters. There have been computed, among other values, the percentage of the available data, the correlation between the QT dispersion and the RR intervals for the whole 24 hr period, as well as for the day and for the night (9.00 p.m.-6 a.m.) periods. The correlation was established by a linear regression. Non-linear regressions were also computed (data not presented), but these methods did not present a different behaviour from the linear one. Possible clusters were searched for using usual clustering techniques.

In a preliminary clinical study, the Holter ECG of 51 subjects have been studied, with special reference to the QT dispersion. The subjects examined were 'controls' (pts affected neither by cardiovascular nor by neuropathic disorders), cardiovascular subjects and neuropathic ones. The statistical analysis performed was the study of the linear correlation, applying the usual criteria.