Interatrial block (IAB) is an electrical abnormality due to delayed conduction in the Bachmann’s bundle region between the two atria. There are two types of IAB, partial (pIAB) and advanced (aIAB) [1]. In the pIAB, the stimulus is delayed but can cross the septum via the Bachmann’s bundle [2, 3]. Advanced IAB is produced when electrical stimulus conduction is totally blocked in that region; the activation of the left atrium (LA) is then from the right atrium in the caudo-cranial retrograde direction. The pIAB is characterized by a P-wave duration ≥120 milliseconds (ms) on a 12-lead electrocardiogram (ECG) (Fig. 1a) [2, 3]. The diagnosis of aIAB is made by a P-wave duration ≥120 ms and a biphasic or +/− morphology of the P-wave in leads II, III, and aVF of the ECG (Fig. 1b) [2, 3]. Interatrial block is a marker of electromechanical dysfunction of the LA [4] and a substrate for the development of supraventricular arrhythmias in different scenarios [5, 6], mainly atrial fibrillation (AF), which occurs in a condition referred to as Bayés’ Syndrome [7]. Currently, a multicenter registry is tracking these cases to demonstrate this assumption [8].

Several authors have studied alterations in LA wall deformity using speckle tracking echocardiography (STE). Patients undergoing STE before ablation of paroxysmal AF (PAF) have shown reduced LA function [9]. Additionally, decreased peak longitudinal strain rate (SR) during the early reservoir period (SRs) of the LA is related to AF recurrence after electrical cardioversion [10]. In addition, decreased peak longitudinal SR during the period of atrial booster pump function (SRa) of the LA is a predictor of new-onset AF [11].

Changes in atrial function and remodeling and their predisposition to present with new-onset AF [12] have been observed using transthoracic echocardiography (TTE) in patients with IAB [2]. However, to date, no studies have described possible alterations in the LA wall using STE. Our hypothesis was that LA remodeling and altered period function could be observed in patients with IAB in sinus rhythm without a history of supraventricular arrhythmias. Our objective was to determine whether there is an association between IAB and atrial remodeling, with possible alterations in the deformation and functionality of the LA wall, determined using STE.

Of 105 patients initially included, four were excluded due to AF and two because of poor-quality TTE images. The final sample consisted of 99 patients: 21 with pIAB, 22 with aIAB, and 56 were patients with no IAB. ECGs were analyzed by one observer with low intra-observer variability (mean 0.6%; 95% CI 0.3–0.9%). The STE was analyzed in 95% of the acquired segments. Bland-Altman analysis showed good intra- and inter-observer agreement with a non-significant bias. Intra-observer variabilities were mean change 1.20% (95% CI, 1.10–1.31%) for LA global longitudinal strain and 1.36% (95% CI, 1.12–1.51%) for LA global longitudinal SR, and inter-observer variabilities were 1.32% (95% CI, 1.15–1.62%) and 1.6% (95% CI, 1.25–1.92%), respectively. Clinical and TTE characteristics of the patients included in this study are shown in Table 1 and Additional file 1: Table S1. Patients with aIAB were older and had a higher prevalence of diabetes. On TTE, patients with aIAB and pIAB had larger LAVImax compared to the patients with no IAB; Doppler echocardiography showed that the ASCTMF decreased and the E wave to A wave (E/A) ratio increased in the group of patients with pIAB and even more in the group with aIAB, compared to the patients with no IAB group. The values of strain and SR are shown in Table 2, with significant differences between the groups. We observed that the greater the degree of IAB, the lower the absolute values of LA wall deformation during the CT and R period of LA.

Table 3 shows the standardized regression coefficients of the unadjusted simple linear regression model and the adjusted multiple linear regression model for potential confounders’ variables. The variables that showed the strongest associations with P-wave duration were LAVImax (direct association) and ASCTMF, εs LAmax, absolute SRa, and SRs (inverse associations). With these variables, we performed a backward stepwise analysis adjusting for age, gender, arterial hypertension, diabetes mellitus, LAVImax, and ASCTMF. The sample size allowed supporting all covariates from the linear model. In the first model, the SRs was associated with P-wave duration (standardized β − 0.345, p = 0.004). In the second model, LAVImax (standardized β 0.269, p = 0.013) and SRa (standardized β 0.420, p <  0.001) remained associated. In the third model, LAVImax (standardized β 0.319, p = 0.004) and εs LAmax (standardized β − 0.290, p = 0.008) remained associated.

Table 4 shows the variables significantly associated with both types of IAB after univariate multinomial regression analysis. Several multinomial logistic sub-models were constructed to avoid saturation and collinearity (Additional file 2: Table S2) of the regression model. The reference model comprised age and diabetes mellitus, which were included in successive models with the following variables introduced one by one: LAVImax, ASCTMF, εs LAmax, SRa, and SRs (Table 5). Comparing the overall differences between five nested models, the models which included LAVImax and SRa proved statistically superior to the others (Fig. 3). The results of the linear and logistic models are similar, suggesting in a robust way the association between P-wave of ECG duration or type of IAB and the mean peak of SRa and SRs. According to the ROC curve analysis (Fig. 4), the best cut-off that identified patients with IAB versus patients with no IAB was absolute SRa ≤ − 1.20 S^− 1 (AUC 0.85, CI 0.77–0.93, p <  0.001), with a sensitivity of 70% (CI: 54.6–81.9) and specificity of 78% (CI: 64.8–87.2). On grouping the total sample of SRa by tertiles and observing the distribution, according to the presence or absence of IAB, a linear trend was demonstrated, such that the lower the absolute value of SRa, the greater the number of patients with IAB (Mantel Haenszel Test for tendencies p <  0.001) (Additional file 3: Table S3).

Our main finding was that a longer duration of the P-wave in the ECG, belonging to the group of patients with pIAB and aIAB, correlated directly with the increase in LAVImax on 2D echocardiogram (2DE) and inversely with the decrease in the absolute value of LA SRa and SRs on STE (Fig. 1d). To our knowledge, our study is the first to describe LA wall deformation alterations determined using STE in patients with IAB. The SRa was especially useful for correlation with the presence or absence of IAB. This indicates the reliable correlation between ECG and STE in patients with IAB.

Several studies using TTE have shown increased LA size and slow and poor contractility in patients with IAB [3, 4, 12]. Similar findings were observed in our patients with IAB, in relation to the LA volume increase, especially in those with aIAB, where ASCTMF was significantly decreased compared to in patients with no IAB. Although IAB and enlarged LA share a similar ECG pattern, a recent study has indicated that they are independent entities [3].

The European Association of Cardiovascular Imaging/European Heart Rhythm Association Expert Consensus for the evaluation of patients with AF indicates that, in this context, measuring LA strain using STE is very promising [21], even in patients with permanent AF [16]. Sun et al. analyzed the effect of aging on the strain of the LA using 2DE STE, noting that, as age increases, the absolute value of SRa increases, regardless of gender. This is due to coupling between LA CT and decreased diastolic function with older age [20]. In our series, patients with aIAB were older than the subjects in the other two groups, despite the fact that their absolute SRa values were pathologically diminished (Tables 1 and 2), probably due to age-related atrial fibrosis [22]. However, not only age must be considered, because the regression analyses adjusted for this and the other confounding variables showed correlations with both types of IAB. In addition, absolute values of SRa showed a linear decrease in the groups of patients with pIAB and aIAB.

Speckle tracking echocardiography findings of LA deformation similar to ours have been reported in patients with permanent or PAF [16]. Hirose T et al. reported that subjects who develop new-onset non-valvular AF have lower absolute values of SRa and SRs than those who do not [11]. Other authors have evaluated the role of LA deformation imaging as a predictor of a successful second cardiac ablation in PAF, noting that strain and SR during the CT and R periods of the LA were significantly lower in the second cardiac ablation group compared to healthy volunteers and patients undergoing first cardiac ablation without AF recurrence (p <  0.001, for all) [23]. It has also been shown that a decrease in the LA global longitudinal strain is an independent predictor of the progression of PAF towards persistent or permanent AF [24].

Atrial fibrosis can be visualized and quantified using cardiac magnetic resonance imaging (CMR) with late gadolinium enhancement (LGE), but it can also manifest as atrial enlargement, decrease in deformation during CT and R periods, determined using LA 2DE STE, even before the occurrence of atrial dilation, as well as higher daily AF load [25]. All these findings have been associated with stroke in patients with AF [26]. One possible application of determining atrial strain could be to predict the risk of AF. This has recently been demonstrated by Mahnkopf et al., who assessed LA fibrosis with LGE-CMR and atrial function using STE before cardiac ablation, creating the Utah classification. They correlated AF recurrence directly with decreased εs LAmax and the highest degree of atrial fibrosis on LGE-CMR [27]. Therefore, the STE 2DE finding of decreased LA deformation in our patients with IAB is identical to that predicting AF in patients with no IAB [2, 22]. Given the direct relationship between decreased deformation and LA fibrosis in PAF, this same relationship could exist between decreased deformation and LA fibrosis in our patients with IAB. The latter, associated with the LA enlargement in the patients with IAB, would indicate that these patients present the two factors that define atrial structural remodeling [28]. The Expert Consensus on Atrial Cardiomyopathy proposed the following as a working definition of this: “any complex of structural, architectural, contractile, or electrophysiological changes affecting the atria with the potential to produce clinically-relevant manifestations.” It could, therefore, be considered that the contractile alterations of the LA wall and the occurrence of supraventricular arrhythmias in patients with IAB would be a form of atrial cardiomyopathy [28]. The atrial myopathy hypothesis proposes that atrial fibrosis increases the risk of thromboembolism independently of atrial rhythm [29]. Identifying these STE parameters in patients with IAB is important because they indicate alteration of atrial structure and function, even before severe enlargement of the LA occurs, as we have detected in pIAB. It remains to be clarified whether this decrease, as determined using STE in IAB, has a similar ability to predict the development of new-onset AF.

To our knowledge, this is the first study to describe LA wall deformation alterations determined using 2D STE in patients with IAB. It demonstrated that SRa and SRs of LA are pathologically diminished in patients with IAB. This finding could indicate the existence of an atrial cardiomyopathy in patients with IAB.