Comparison of gastric aspiration and continuous pH monitoring with antimony electrode: methodological remarks

In an attempt to eliminate the noise affecting continuous 24-hour intragastric pH metry curves in order to obtain a more appropriate identification of hourly intraluminal pH values to which the corresponding aspirate ones were compared, the raw data of one of our previous experiments were treated using a smoothing technique (removal of noise peaks was followed by moving averages on three adjacent points, 1 min apart, which were applied both singly and combined with Fourier's filtration employing the first 32 harmonics) A gradual improvement of all parameters of regression analysis from original data to those achieved with the two above-mentioned methods was found. In particular, the regression coefficient r increased from 0.83 to 0.88 with the singly moving averages technique and even to 0.95 when Fourier's filtration was combined. Both techniques provided also similar improvement as far as slope and elevation values were concerned: the slope was almost equal to 1 and the elevation was very low, so that paired data might be assumed to be proportional. It can be concluded that hourly pH values of continuous intraluminal monitoring and those of simultaneous gastric aspiration appear better correlated if the elimination of noise disturbing the in vivo pH metry curves is obtained. With regard to this, the use of a combined smoothing-filtering technique, such as moving averages plus Fourier's filtration after removal of the noise peaks, seems a very advantageous modality.