P‐wave parameters and their association with thrombi and spontaneous echo contrast in the left atrial appendage

The aim of this study was to examine the prevalence of abnormal P‐wave parameters in patients with thrombus and/or spontaneous echo contrast (SEC) in the left atrial appendage (LAA), and to identify P‐wave parameters particularly associated with thrombus and SEC formation.


| INTRODUCTION
Alterations in P-wave parameters can reflect structural, functional, as well as electrical remodeling of the atria. 1 In the past, associations of abnormal P-wave parameters with clinical outcomes, particularly atrial fibrillation (AF) and ischemic stroke, have been reported. 1 In addition, P-wave parameters are useful in the diagnosis of atrial cardiomyopathy which is defined as as "any complex of structural, architectural, contractile or electrophysiological changes affecting the atria with the potential to produce clinically-relevant manifestations." 1,2 Patients with embolic stroke of undetermined source (ESUS) and patients with AF are particularly interesting because atrial cardiomyopathy is often presumed to be the underlying disease. 1 Thromboembolic events resulting from thrombus formation, predominantly in the left atrial appendage (LAA), may occur as a result of both AF and atrial cardiomyopathy, independent of AF. Thus, the development of thrombus and spontaneous echo contrast (SEC) in the LAA may be the causal link between AF and/or atrial cardiomyopathy and ischemic stroke.
Thrombi and SEC in the LAA are relatively common findings on transesophageal echocardiography, particularly in patients with AF. 3 Besides the increased risk of thromboembolic events, both phenomena are associated with higher morbidity and mortality. [4][5][6][7][8][9] Furthermore, the occurrence of thrombus or SEC is considered as an additional diagnostic criterion for the presence of atrial cardiomyopathy, characterized by functional and structural atrial dysfunction. 10,11 However, there is limited evidence on the association of electrical dysfunction of the atrium and the occurrence of thrombi and SEC on transoesophageal echocardiography.
Consequently, because thrombus formation and SEC could be the link between AF, atrial cardiomyopathy, and ischemic stroke, the question arises whether there are P-wave indices associated with thrombus and SEC. Therefore, the aim of this study was to examine the prevalence of abnormal P-wave parameters in patients with thrombus/SEC, and to identify P-wave parameters particularly associated with thrombus and SEC formation.

| METHODS
This study examined the P-wave indices of patients with evidence of thrombus/SEC in the LAA compared with the P-wave indices of patients with comparable stroke risk and exclusion of thrombus or SEC. For the analysis of P-wave indices, it was necessary that the studied patients were in sinus rhythm.

| Inclusion and exclusion criteria
The study population was composed of a thrombus/SEC group and a control group. For the thrombus/SEC group, we examined retrospectively all transoesophageal echocardiographies performed between 2015 and 2020. 12 All patients with sinus rhythm, in whom a thrombus or a SEC in the LAA were detected were included in the study. The indications for transoesophageal echocardiography were planned pulmonary vein isolation, in cases of ESUS, and for evaluation of valvular heart diseases, or unclear infections. Patients with other rhythm than sinus rhythm (e.g., AF, atrial flutter), thrombi in other cardiac cavities, cardiac tumors and previous interventional or surgical occlusion of the LAA were excluded.
To assess the significance of P-wave indices for association with thrombus or SEC, we identified a control group with the following characteristics: (1)

| Echocardiographic analysis
Transthoracic and transoesophageal echocardiography were performed with a digital ultrasound scanner (Vivid E9; General Electric).
Transoesophageal echocardiography was conducted by experienced physicians according to current recommendations. 13 Thrombus in the LAA was defined as a solid echo dense structure that was detectable in multiple image planes in the LAA. 14 SEC in the LAA was diagnosed when swirling dynamic echoes were detected without evidence of thrombus 14 (Supporting Information: Figure 1).

| Electrocardiographic analysis
All patients received a 12-lead ECG performed within 24 h before transoesophageal echocardiography. The standard 12-lead surface ECG was recorded at a rate of 50 mm/s and a voltage of 10 mm/mV. All evaluations were conducted by a single observer who was blinded to the patients' group. The focus of the ECG analysis was on the examination of the P-wave.
The duration of the P-wave was determined in all 12 leads. The resulting arithmetic average represented the mean P-wave duration. P-wave dispersion was calculated by subtracting the shortest P-wave duration from the longest P-wave duration. 1 A P-wave dispersion of 40 ms is associated with worse clinical outcome, such as the occurrence of AF, and was therefore considered as abnormal. 15 A prolonged P-wave duration ≥120 ms is pathological and represents a partial IAB which indicates delayed excitation conduction between the right and left atrium via the Bachmann bundle. 1 If the Bachmann bundle is completely blocked, there is an advanced IAB that consists not only of a prolonged P-wave duration ≥120 ms, but also of an additional biphasic morphology in the inferior leads due to retrograde excitation of the left atrium. 1 A biphasic P-wave is often found in lead V1. The terminal negative portion represents excitation in the left atrium. Left atrial enlargement may be reflected by an abnormal P-wave terminal force in lead V1 (PTFV1). It is calculated by multiplying the terminal portion of the P-wave in V1 by the duration of this portion. A PTFV1 ≤ −4000 µV*ms is considered pathological. 1 The patients' ECG were examined for a P-wave voltage ≤0.1 mV in lead I, which is defined as abnormal. 1 The P-wave area was measured by multiplying half of the P-wave duration in lead II by the P-wave voltage in this lead, and is considered pathological at a value of at least 4 ms*mV. 1 The P-wave axis was determined. The normal P-wave axis is a value between 0°and +75°. Deviations from this axis were defined as abnormal P-axis 1 (Supporting Information: Figure 2).

| STATISTICS
The statistical software SPSS 26 was used for statistical analysis.
Numerical values are expressed as mean ± SD. Continuous variables were compared between groups using an unpaired t-test (for normally distributed variables) or Mann-Whitney U-test (for nonnormally distributed variables). χ 2 analysis was used to compare categoric variables. Receiver operating characteristic curves were generated to define cutoff values for the mean P-wave duration, P-wave dispersion and PTFV1. The odds ratio was measured for risk calculation. A p < .05 was considered significant. All probability values reported are 2-sided.

| Patients' clinical characteristics
Patients with thrombus/SEC were significantly more likely to have diabetes mellitus and to have a history of stroke/transient ischemic attack (TIA) than the control group. While 20 patients (74%) in the thrombus/SEC group had a history of AF, all patients (100%) in the control group presented with a history of AF (p < .001). Accordingly, it is not surprising that the control group was significantly more likely to take oral anticoagulation.
Otherwise, there were no significant differences between the patient groups regarding laboratory parameters (e.g., creatinine, hemoglobin), left atrial diameter, left ventricular ejection fraction or CHA 2 DS 2 -VASc score (Table 1).

| Prevalence of abnormal P-wave indices in the study population
Both groups were examined for the presence of abnormal P-wave indices. In the thrombus/SEC group, 13 patients (48%) had partial, and 11 patients (41%) had advanced IAB. A P-wave dispersion of at least 40 ms was present in 19 patients (70%), and a PTFV1 ≤ −4000 µV*ms in 20 patients (74%). Also 20 patients (74%) had a P-wave area of at least 4 ms*mV, and a P-wave voltage ≤0.1 mV in lead I was detected in 21 patients (78%), respectively. In addition, 3 patients (11%) exhibited a pathological P-wave axis (Table 2, Figure 1). In the thrombus/SEC group, all patients had at least one combination of two abnormal P-wave indices.
In the control group, 30 patients (38%) exhibited partial, and 12 patients (15%) advanced IAB. 20 patients (25%) had a P-wave dispersion of at least 40 ms, and 42 patients (53%) had a PTFV1 ≤ −4000 µV*ms. In 67 patients (85%), there was a P-wave area of at least 4 ms*mV, and in 52 patients (66%) a P-wave voltage ≤0.1 mV in lead I was found. Furthermore, 10 patients (13%) displayed a pathological P-wave axis (Table 2, Figure 1). In the control group, there was only one patient who had no abnormal P-wave indices.

| Comparison of electrocardiographic characteristics of the thrombus/SEC group and the control group
Several P-wave indices showed significant differences between the thrombus/SEC group and the control group. Patients with thrombus/SEC were significantly more likely to present with a KREIMER ET AL. | 399 greater mean P-wave duration, a greater P-wave dispersion, and a more negative PTFV1 than the control group. Furthermore, these patients had also more often an advanced IAB. All other ECG parameters, in particular those related to ventricular excitation conduction and repolarization, demonstrated no significant differences (Table 1).

| Subgroup analysis of patients with oral anticoagulation
For a subgroup analysis, we included all patients with oral anticoagulation (n = 91). The thrombus/SEC group (n = 13) exhibited a significant lower left ventricular ejection fraction than the control group (n = 78), and higher creatinine and lower hemoglobin levels (Supporting Information: Table 1). However, significant differences in the P-wave indices were also observed in the subgroup analysis.
Patients with a thrombus/SEC had a significant longer mean P-wave duration than the control group, a greater P-wave dispersion, and a more negative PTFV1. There was no significant difference in the presence of a partial or advanced IAB between both groups (Supporting Information: Table 2).

| DISCUSSION
In the present study, ECG parameters and their association with thrombi/SEC in the LAA were investigated. The main findings of our study are that in a group of high-risk patients for stroke, some P-wave indices are associated with the presence of thrombus/SEC.
The risk for the occurrence of a thrombus/SEC increased significantly in the presence of a P-wave duration >118 ms, a P-wave dispersion >40 ms, and an advanced IAB (Table 3). Even in the subgroup analysis of patients with oral anticoagulation, significant differences in P-wave indices were observed between the thrombus/SEC group and the control group (Supporting Information: Table 2).
In contrast, there were no significant differences regarding ECG parameters representing ventricular excitation propagation and regression (Table 1).

| Diabetes mellitus and the presence of thrombus/spontaneous echo contrast
In the past, an association of diabetes mellitus and the occurrence of thrombus in the left atrium has been described by several studies. [16][17][18][19][20][21] In our study, the group with thrombi or SEC presented significantly more patients with diabetes mellitus, despite an overall similar CHA 2 DS 2 -VASc score. This observation may support previous studies (Table 1). Hyperinsulinemia and hyperglycemia, particularly coexisting, have been demonstrated to cause a prothrombotic state.
In addition, proinflammatory and proatherogenic processes induced by hyperglycaemia and hyperinsulinemia play an important role in T A B L E 2 Prevalence of abnormal P-wave parameters in the thrombus/SEC group (n = 27) and in the control group (n = 79).

F I G U R E 1
Prevalences of abnormal P-wave indices in the thrombus/spontaneous echo contrast group and in the control group. IAB, interatrial block; PTFV1, P-wave terminal force in V1. KREIMER ET AL. | 401 thrombogenesis. 22,23 Furthermore, diabetes mellitus, accompanied by metabolic changes, insulin resistance, and hyperglycaemia, is associated with increased platelet activation and decreased response to antiplatelet agents. 23

| QRS width and the presence of left atrial appendage thrombus
Recently, Wegner et al. 24 were able to demonstrate an association of QRS width and the presence of LAA thrombus and reduced LAA flow in 512 patients receiving transoesophageal echocardiography before cardioversion. However, in this study, left ventricular ejection fraction differed between the group with normal LAA flow and LAA thrombus (55% vs. 35%). Reduced ejection fraction is often associated with ventricular excitation propagation abnormalities, which may be reflected in widened QRS duration. 25 In our study population, there was no significant difference regarding left ventricular ejection fraction (Table 1). A possible explanation might be the heterogeneity of the two study cohorts.

| Prevalences of abnormal P-wave indices in different populations
The P-wave reflects the electrical excitation propagation of first the right and then the left atrium. For this reason, the P-wave is particularly interesting regarding the evaluation of left atrial enlargement as well as intra-or interatrial conduction delays.
In the past, an association of abnormal P-wave parameters with clinical outcomes (e.g., AF, thromboembolic events) has been described. 1,[26][27][28][29] For this reason, it is not surprising that abnormal P-wave parameters are common in our study cohort, consisting of a very large proportion of patients with a history of AF (Table 1).
Interestingly, however, for many pathological P-wave indices, a higher prevalence was seen in patients with thrombus/SEC (Table 2).
Particularly in comparison to the general population, these prevalences are remarkable: epidemiologic findings indicate a prevalence of advanced IAB in the general population at 0.1%-0.5%, with an incidence of 2.27 per 1000 person-years. 27,30 The prevalence of partial IAB is most likely even higher. In the large primary care population of the Copenhagen ECG Study, a P-wave F I G U R E 2 Receiver operating characteristics analysis for indication of thrombus/spontaneous echo contrast; (A) cutoff value P-wave duration >118 ms (sensitivity 82%, specificity 63%); (B) cutoff value P-wave dispersion >40 ms (sensitivity 70%, specificity 75%); (C) cutoff value PTFV1 <−4000 µV*ms (AUC 0.671, sensitivity 63%, specificity 63%). AUC, area under the curve; PTFV1, P-wave terminal force in V1.
≥120 ms was found in 20% of subjects. 30,31 In a representative Finnish population sample the prevalences of abnormal PTFV1, partial, and advanced IAB were 4.9%, 11.8%, and 1.9%. 32 A Japanese study also demonstrated a prevalence for PTFV1 of 4.9%. 33 In another Finnish cohort, incidence rates for prolonged P-wave duration, abnormal PTFV1, left P-wave axis deviation, and right Pwave axis deviation were found to be 16.0%, 7.4%, 3.4%, and 2.2%, respectively. 34 In the large Atherosclerosis Risk in Communities (ARIC) study, the prevalence of abnormal P-wave axis was comparable at 8.3%. 35

| P-wave indices and the risk of thrombus/ spontaneous echo contrast
In this study cohort, P-wave parameters could be derived from ECG analysis as indicators for the presence of thrombus/SEC. These findings suggest that abnormal ECG parameters as an expression of electrical dysfunction may be accompanied by concomitant mechanical dysfunction. However, it is not possible to conclude how the respective contribution of electrical and mechanical dysfunction to the thrombus formation is.
Nevertheless, our findings address a gap in the causal chain between AF and/or atrial cardiomyopathy, thrombus development, and ischemic stroke, because only an association of P-wave indices and AF and/or atrial cardiomyopathy and ischemic stroke has been demonstrated so far.
More interestingly, several P-wave indices may indicate an increased risk of thrombus formation and thus may be helpful in risk stratification of patients with ESUS, and/or atrial cardiomyopathy.
Notably, in our study, the left atrial diameter as measured in the transthoracic echocardiography was not significantly different between the thrombus/SEC group and the control group (Table 1).
On the one hand, this observation could indicate that a structural change of the left atrium (in the sense of enlargement) has no hemodynamic effects and thus does not lead to thrombus formation.
On the other hand, it is striking that in both patient groups the left atrial diameters are enlarged, and thus it could be that both groups are similarly limited in left atrial hemodynamic. Therefore, the results of the ECG analysis are particularly interesting, as they emphasize the importance of the electrical excitation conduction in the atria for the development of a thrombus. In our study, a P-wave duration >118 ms, a P-wave dispersion >40 ms, and an advanced IAB were independent risk factors for the presence of thrombus/SEC in the transoesophageal echocardiography. However, there was an association between the PTFV1 and the presence of thrombus/SEC, but the risk calculation did not reveal a significantly increased risk. Okin et al. 36 demonstrated that an abnormal PTFV1 correlated strongly with incident stroke in 7778 patients with hypertension. The authors suggested that an underlying subclinical atrial cardiomyopathy in the absence of AF might cause thrombus formation in the left atrium. 36 Instead, our findings may imply a smooth transition because this parameter was also pathological in both groups. Furthermore, this observation reinforces the hypothesis that electrical remodeling of the atrium has a greater impact on thrombus formation than structural remodeling, because PTFV1 is considered as an important parameter for atrial enlargement. 1

| P-wave duration
Normally, the P-wave duration is less than 120 ms. In our study, the cutoff value with the highest sensitivity and specificity to discriminate between thrombus/SEC group and control group was 118 ms, which is close to the definition of a partial IAB (≥120 ms) 1 (Table 3). Interestingly, LAA thrombus was found in in 11 patients (35.5%) after wide-area LAA isolation, but in no patient after LAA ligation. 38 Moreover, the combination of wide-area LAA isolation and LAA ligation reduced the P-wave duration significantly. 38 The study by Kawamura et al. 39 supports these observations. Fifteen Patients with LAA ligation were analyzed for changes in P-wave parameters after the procedure. P-wave duration also decreased significantly in these patients. 39 Huang et al. 40 examining 128 patients with AF at risk for LAA thrombus, exhibited that the amplified P-wave duration predicts left atrial thrombogenesis. They found an optimal threshold of 165 ms for amplified P-wave duration with an AUC of 0.90, sensitivity of 88.5%, and specificity of 75.5%, respectively. 40 Müller-Edenborn et al. 41 conducted a case-control study dealing with amplified P-wave duration for atrial cardiomyopathy and thromboembolism risk stratification. Stages of atrial cardiomyopathy were defined using amplified P-wave duration. They demonstrated an OR for LAA thrombus of 24.6 (p < .001) per atrial cardiomyopathy stage. 41 Prolonged P-wave duration often indicates left atrial electrical dysfunction (the so-called "P-sinistroatriale"). The described studies as well as ours could show that mechanical dysfunction may also accompany it, increasing the risk for thromboembolism.
T A B L E 3 P-wave indices with significant association with thrombus or spontaneous echo contrast.

| P-wave dispersion
In the past, studies demonstrated an association of abnormal P-wave dispersion with incident AF and AF recurrence after cardioversion. 42,43 More recently, the only independent predictor of AF detection was P-wave dispersion of 40 ms in a study cohort of patients with cryptogenic stroke which were monitored by implantable loop recorders. 15 In our study, we were able to demonstrate an association between an abnormal P-wave dispersion ≥40 ms and the presence of thrombus/SEC.

| Advanced interatrial block
Advanced IAB is present if the P-wave exhibits a duration ≥120 ms and simultaneously a biphasic morphology in the inferior leads. 1 This occurs during complete block of interatrial excitation propagation via the Bachmann bundle. 1 The presence of an advanced IAB is associated with worse clinical outcomes, for example, incident AF or all-cause mortality after ischemic stroke. 27,46 In 2001, Goyal et al. 47 48 Left atrial thrombi and/or SEC were found in 6 patients with IAB (15%), but in no patient without IAB. 48 In our study, advanced IAB was an independent risk factor for the presence of thrombus or SEC on transoesophageal echocardiography (Table 3). However, it remains unclear why interatrial excitation delay is associated with mechanical dysfunction and thrombus formation in the left atrium.

| LIMITATIONS
The main limitation of the present study is its retrospective and monocentric character. Therefore, it was not possible to evaluate the patients' atrial hemodynamic or laboratory parameters, which may also have influenced the outcome. Furthermore, the prevalence of abnormal P-wave indices in our study cohort is high compared with the general population. Moreover, the thrombus/SEC group and the control group are heterogenous regarding the intake of oral anticoagulants. However, for this reason, we performed a subgroup analysis that included only patients with oral anticoagulation.
Another limitation is the relatively small number of included patients.

| CONCLUSION
Our findings address a previously existing gap between the known association of abnormal P-wave indices with AF, atrial cardiomyopathy, and ischemic stroke. P-wave parameters, which have the advantage of objectivity and broad, routine clinical application, may be suitable to predict the presence of thrombi and SEC on transoesophageal echocardiography. We were able to demonstrate a very high prevalence of abnormal P-wave parameters in patients with thrombus/SEC. Even compared with high-risk patients, the prevalence was higher for multiple parameters.
At the same time, P-wave indices are markers of atrial cardiomyopathy. P-wave indices could therefore help to identify patients at risk who suffer from atrial cardiomyopathy and are therefore particularly at risk for thrombus formation in the LAA and consequent thromboembolic events. In addition, the findings could allow risk stratification in patients with AF to differentiate which patients may be at particular risk for thrombus development and cardioembolic stroke.
Hence, the identified parameters could be of particular clinical importance for risk stratification of patients with AF, atrial cardiomyopathy, and ESUS. In general, patients with AF receive outpatient cardiological care, and patients with ESUS and atrial cardiomyopathy also receive cardiological evaluation. Herein, more attention should be paid to P-wave parameters in routine ECG and changes in these parameters should also be monitored. Patients without previously diagnosed AF with atrial cardiomyopathy or ESUS who have risk factors according to the CHA 2 DS 2 -VASc score in addition to abnormal P-wave parameters could potentially benefit from therapy with oral anticoagulants. However, this has not yet been demonstrated in any large, randomized trial. In contrast, patients with known AF who already receive anticoagulation because of an increased CHA 2 DS 2 -VASc score but who exhibit abnormal P-wave parameters might benefit from closer echocardiographic monitoring and, if a thrombus is detected, from modification or escalation of therapy. This is particularly important in patients who present with a progressively lower left ventricular ejection fraction and thus are at greatly increased risk for cardiac thrombus formation, atrial as well as ventricular.

DATA AVAILABILITY STATEMENT
Data available on request due to privacy/ethical restrictions.