Detection of endo-epicardial atrial low-voltage areas using unipolar and omnipolar voltage mapping

Abstract

Background: Low-voltage areas (LVA) can be located exclusively at either the endocardium or epicardium. This has only been demonstrated for bipolar voltages, but the value of unipolar and omnipolar voltages recorded from either the endocardium and epicardium in predicting LVAs at the opposite layer remains unknown. The goal of this study was therefore to compare simultaneously recorded endo-epicardial unipolar and omnipolar potentials and to determine whether their voltage characteristics are predictive for opposite LVAs. Methods: Intra-operative simultaneous endo-epicardial mapping (256 electrodes, interelectrode distances 2 mm) was performed during sinus rhythm at the right atrium in 93 patients (67 ± 9 years, 73 male). Cliques of four electrodes (2 × 2 mm) were used to define maximal omnipolar (V_omni,max) and unipolar (V_uni,max) voltages. LVAs were defined as V_omni,max ≤0.5 mV or V_uni,max ≤1.0 mV. Results: The majority of both unipolar and omnipolar LVAs were located at only the endocardium (74.2% and 82.0% respectively) or epicardium (52.7% and 47.6% respectively). Of the endocardial unipolar LVAs, 25.8% were also located at the opposite layer and 47.3% vice-versa. In omnipolar LVAs, 18.0% of the endocardial LVAs were also located at the epicardium and 52.4% vice-versa. The combination of epicardial V_uni,max and V_omni,max was most accurate in identifying dual-layer LVAs (50.4%). Conclusion: Unipolar and omnipolar LVAs are frequently located exclusively at either the endocardium or epicardium. Endo-epicardial LVAs are most accurately identified using combined epicardial unipolar and omnipolar voltages. Therefore, a combined endo-epicardial unipolar and omnipolar mapping approach is favoured as it may be more indicative of possible arrhythmogenic substrates.

Keywords: atrial fibillation; endo-epicardial mapping; low-voltage areas; omnipolar mapping; sinus rhythm; unipolar voltage; voltage mapping.

FIGURE 1

Construction of unipolar, bipolar and omnipolar voltages in 2 × 2 mm cliques. (A) Two high-density electrode arrays consisting of 128 unipolar electrodes are fixed together. One array was placed on the epicardium and one array is introduced into the RA using the incision for venous cannulation to map the endocardium and epicardium simultaneously. The RA was mapped with the tip of the electrode arrays toward the inferior vena cava (inferior), the superior vena cava (superior) and in between, toward the terminal crest (mid). For each square area, enclosed by four electrodes, four unipolar EGMs and matched bipolar and omnipolar EGMs were derived from two electrode orientations [along the vertical y-axis (green) and horizontal x-axis (red)] as indicated by the dotted lines. (B) Examples of a unipolar, horizontal bipolar-x, vertical bipolar-y and omnipolar EGM recorded from both the epicardium (upper) and endocardium (lower). The two bipolar EGMs differed considerably, illustrating the electrode orientation dependence of bipolar mapping. Omnipolar mapping provides electrode orientation-independent voltages that are larger than the bipolar with the largest measurable peak-to-peak voltage, in both cases the horizontal bipolar-x EGMs. (C) LVAs can be either located at solely the endocardium or epicardium (green) or at both sides (red). The corresponding clique at the exact opposite side is highlighted (grey). The other cliques are then indicated as normal voltage (light grey). (D) Example of endo-epicardial activation time maps with isochrones drawn at every 10 ms. Arrows indicate the main direction of the propagation wavefront and thick black lines indicate areas of conduction block (time difference between adjacent electrodes ≥12 ms). (E–G) Peak-to-peak voltages of corresponding EGMs are used to create different voltage maps. Bipolar voltage map illustrates the maximal bipolar voltage in both horizontal and vertical orientations within one clique. LVAs are highlighted by a white line and areas of EEA are indicated by a white X. In this example, endocardial bipolar and omnipolar LVA are present in respectively 20.0% and 16.2% of the cliques. EEA was present in 9.5% of the cliques at both the endocardium and epicardium. EEA = endo-epicardial asynchrony; EGM = electrogram; (EE-) LVA = (endo-epicardial-) low-voltage area.

FIGURE 2

Relation between endocardial and epicardial unipolar and omnipolar voltage. (A) and (C) Kernel density plots of V_uni,max (A) and V_omni,max (C) voltages between the endocardium and epicardium. The colors indicate the data density. A black line indicates the ordinary least squares prediction. Statistical significance is indicated by an asterisk (p < 0.001). V_uni,max and V_omni,max are subdivided according to the 25th, 50th, and 75th percentiles of the endocardial voltages, and are indicated by dashed vertical lines. (B) and (D) Bland-Altman plots of endocardial versus epicardial V_uni,max (B) and V_omni,max (D) voltages. The colors indicate the data density. A black line indicates the median per one means mV. The mean difference and 95% confidence intervals are indicated by the dashed lines. V_omni,max = omnipolar clique voltage; V_uni,max = unipolar clique voltage.

FIGURE 3

ROC-curves of the prediction of overlapping endocardial (left) and epicardial (right) LVAs based on all parameters recorded from the opposite layer. AUC = area under the curve; CV = conduction velocity; EEA = endo-epicardial asynchrony.