Electrocardiographic alternans in coronary artery disease patients

The aim of this study was to quantify cardiovascular risk by means of a risk index, the electrocardiographic alternans, on a coronary artery disease population to be able to distinguish, whether there is a different risk in the case of comorbidities with respect to coronary artery disease, specifically diabetes mellitus. To quantify cardiovascular risk by means of electrocardiographic alternans, a population with coronary artery disease was considered in this study (Intercity Digital Electrocardiogram Alliance (IDEAL) database). Of this population, two subgroups were studied. The first consisting of 36 patients with both CAD and diabetes (CAD+DM group) and the second composed by 36 subjects with only CAD (CAD group). A further distinction can be made in the diabetic subgroup between the patients with type 1 diabetes (9 subjects, CAD+T1DM group) and those with type 2 diabetes (27 subjects, CAD+T2DM). The ECG signals of the duration of 24 hours were acquired with Holter ECG and analysed with MATLAB R2022b program. The algorithm used to detect ECGA is the Enhanced Adaptive Matched Filter (EAMF). 11 features were obtained from each ECG window. The features considered in the statistical analysis are the median amplitude of PWA, QRSA, TWA including any null values; the median amplitude of PWA, QRSA, TWA excluding any null values and the number of alternating beats for PWA, QRSA, TWA. the median was performed among all subjects belonging to the three subgroups, respectively. Polynomial curve fitting was applied to the median trends obtained to appreciate possible differences of the baseline trend, neglecting the local variations. Statistical comparison was assessed through the Wilcoxon rank-sum test. Statistical significance was set at p<0.05. The trends faithfully follow the circadian rhythm. From the results obtained from the Wilcoxon rank-sum test performed on the median trends of the studied features, it can be observed that between CAD subjects and CAD+T1DM the features statistically different are: the median amplitude of the PWA (p=0.03), the median amplitude of the QRSA (p=0.02), the number of beats affected by PWA (p<10-3) and the number of beats affected by QRSA (p=0.01). Regarding the comparison between the CAD and CAD+T2DM, it is noticeable that there is statistical difference in all the features considered (p<0.05). In the case of the Wilcoxon rank-sum test performed on the median trends obtained by polynomial curve fitting, the statistically different features between the CAD subgroup and the CAD+T1DM subgroup are: the median amplitude of the PWA (p=0.03), the median amplitude of the PWA excluding any null values (p=0.01), the number of beats affected by PWA (p<10-3) and the number of beats affected by QRSA (p=0.01). In this case also there is statistical difference in all the features considered (p<0.05) in the comparison between CAD and CAD+T2DM. The CAD+T1DM subgroup is the one that would appear to have a higher cardiovascular risk than the others.