[Role of the conditions of recording and analysis of the atrial signal on the results of ECG signal averaged P wave].

Several studies have shown that high amplification and averaging of the electrocardiographic P wave may soon enable better identification of people at high risk of paroxysmal atrial fibrillation. The results of this technique are closely dependent on the conditions of recording and analysis of the atrial signal but this factor has not been widely appreciated. The authors, therefore, undertook a study of the reliability of the manual method of measuring P wave duration, the influence of the filters used and the level of background noise on the 3 parameters of analysis: the duration of the P wave (DUR), the root mean square of the amplitude of the terminal 20 milliseconds (RMS20) and the integral under the P curve (IN). The recorder was a Predictor II 5.0 (Corazonix) with a special programme to identify and analyse the P wave in 3 bipolar non-orthogonal leads. DUR was determined manually after amplification and averaging; RMS20 and IN were calculated automatically. In order to assess the reliability of manual determination of the P wave, 48 consecutive patients were studied: two analyses were made with a least a 2 months' interval by entering the data through two filters (unidirectional and FIR). The role of the filters was then studied by passing the preceding 48 recordings to an analysis with three filters: unidirectional, bidirectional and FIR, and comparing the results.(ABSTRACT TRUNCATED AT 250 WORDS)