The role of left atrio-ventricular coupling index and left atrial ejection fraction in predicting onset of atrial fibrillation and adverse cardiac events in hypertrophic cardiomyopathy

Abstract

Background

Several predictors of atrial fibrillation (AF) onset in patients with hypertrophic cardiomyopathy (HCM) have been proposed, however, all of them showed limited accuracy. This study aims to assess the role of new echographic parameters in predicting AF onset and major adverse cardiovascular outcomes (cardiovascular death or heart transplantation).

Methods

Clinical and imaging data from 141 patients with HCM and without a history of AF were retrospectively analyzed over a 5-year period. Patients who developed AF during the study were compared to those who did not. The analysis focused on key atrial parameters, including the Left Atrial Contraction Index (LACI) and Left Atrial Ejection Fraction (LAEF). LACI was defined as the ratio of left atrial end-diastolic volume to left ventricular end-diastolic volume. Echocardiographic measurements were standardized using cardiac magnetic resonance (CMR) as the reference. Regarding statistical analysis, each significant continuous variable was categorized by identifying a cut-off value using the Youden index. Independent associations with outcomes and cumulative survival were assessed using Cox regression analysis.

Results

Thirty-five patients developed AF, at a mean time of 4 years. The HCM-AF group had significantly higher values of LACI, left atrial diameter (LAD), and left atrial minimum volume (LAVmin). A LACI > 43% on echocardiography and LACI > 44% on CMR showed the best performance in identifying patients at risk for AF. In multivariate analysis, an echocardiographic LAEF < 43% was independently associated with the occurrence of AF (HR 2.9, 95% CI: 1.2–6.9). Additionally, a LAD > 40.5 mm was independently associated with AF onset, with a hazard ratio of 2.5 (95% CI 1.1–5.5). Eleven patients experienced the composite outcome of cardiovascular death or heart transplant, and a LACI > 60% was associated with this outcome.

Conclusion

In patients with HCM, both LACI and LAEF were significantly associated with the occurrence of AF over a 4-year period, demonstrating higher sensitivity and specificity compared to other parameters. A LACI > 60% was also found to be associated with cardiovascular death or heart transplant in this population.

Graphical Abstract

Evaluation of left atrial morpho-functionl parameter using multimodality imaging, and  their association with atrial fibrillation in HCM patients. The figure summarizes the main findings of the present article. AF: atrial fibrillation; CMR: cardiac magnetic resonance; HCM: hypertrophic cardiomyopathy; TTE: transthoracic echocardiography; LACI: left atrioventricular coupling index; LAEF: left atrium ejection fraction.

Background

Atrial fibrillation (AF) is the most common arrhythmia in patients with hypertrophic cardiomyopathy (HCM), with an estimated prevalence of 25%, rising to over 40% in those older than 70 years [1]. A HCM patient has a 4 to sixfold increased lifetime risk of developing AF, with the first episode typically occurring at a younger age (around 55 years) compared to the general population [2]. Left ventricular (LV) diastolic dysfunction, mitral insufficiency, and left ventricular outflow tract obstruction (LVOTO) are key factors to atrial dilation and wall stress, promoting left atrial (LA) fibrosis and mechanical dysfunction, and finally creating a favorable substrate for AF [3]. In HCM, AF has significant prognostic implications, being associated with a threefold increased risk of heart failure (HF) and stroke [4]. Identifying patients at higher risk for AF is crucial to ensure early therapeutic intervention and prevent AF-related complications, ultimately improving their prognosis. Over the years, various indicators have been proposed to predict the development of AF, but none have demonstrated reliable sensitivity or specificity. Among these, the most commonly used is the anteroposterior left atrium diameter (LAD) ≥ 45 mm [5, 6]. Recently, the left atrio-ventricular coupling index (LACI) has been proposed as early marker of left atrium remodeling and impaired ventricular compliance. This parameter was defined as the ratio between left atrial end-diastolic volume and left ventricular end-diastolic volume (LVEDV) [7, 8]. Meucci et al., in a large population of patients with HCM, demonstrated that LACI is a promising predictor of AF [8]. However, to date, no previous study has compared LACI to other established and more commonly used indicators. Furthermore, the standardization of LACI echocardiographic measurements has yet to be evaluated. Additionally, while the prognostic implications of LACI have been investigated in other populations, they remain unexplored in the context of HCM [9].

Among other imaging parameters, left atrial ejection fraction (LAEF) has also been evaluated by Maron et al., who demonstrated that LAEF, assessed by cardiac magnetic resonance (CMR), was associated with a higher risk of developing AF. However, only a few other studies have investigated this correlation [10].

Our study aims to analyze the morpho-functional parameters of the left atrium and investigate their association with AF in HCM, compared to more commonly used parameters such as LAD. We also sought to enhance the objectivity and standardization of echocardiographic measurements by using CMR as a reference. Furthermore, we assessed the correlation between these parameters and major adverse cardiovascular outcomes, including heart transplantion (HTx) and cardiovascular death (CV death).

Materials and methods

Study population and design

The data of 141 HCM patients who underwent their first outpatient evaluation between 2011 and 2018 at the Mediterranean Institute for Transplantation and Advanced Specialized Therapies (ISMETT) were retrospectively retrieved from our digital archive and analyzed. Initially, the population consisted of 206 patients; however, 64 patients were excluded based on the following criteria: 1. Significant mitral valve disease that could influence the analysis of morpho-functional atrial parameters (defined as mitral regurgitation greater than moderate or the presence of mitral stenosis); 2. reduced ejection fraction (EF), defined as EF below 50%; 3. Initial LV dilation (LVEDV > 150 ml for males, > 106 ml for females); 4. History of AF or supraventricular tachycardia. The diagnosis of HCM was based on the presence of left ventricular hypertrophy, after excluding other possible cardiac and extracardiac causes [11]. Both obstructive and non-obstructive HCM patients were included. AF was defined according to guidelines [12]. Patients were followed up with annual continuous Holter monitoring and outpatient visits, including a 12-lead electrocardiogram (ECG) every 6 months or whenever symptoms occurred. Data from cardiac implantable devices were also included in the analysis. Indications for HTx were based on guidelines [13].

Echocardiographic data

All transthoracic echocardiograms (TTE) were analyzed offline with specialized software (EchoPAC). Conventional measurements were conducted in accordance with the recommendations of the European Society of Echocardiography and based on established reference values [14]. The end-systolic (LAVmax) and end-diastolic (LAVmin) volumes of the left atrium were measured in both the apical 4-chamber and 2-chamber views. LAEF was calculated as follows:

$$\text{LAEF}=(\text{LAVmax}-\text{LAVmin})/\text{LAVmax}$$

and expressed as a percentage. The LACI was defined as the ratio of LAVmin to left ventricular end-diastolic volume (LVEDV):

$$\mathrm{LACI}\;=\;\mathrm{LAVmin}/\mathrm{LVEDV}$$

and expressed as a percentage. Both volumes were measured during the same end-diastolic phase, at the time of mitral valve closure [7].

Cardiac magnetic resonance protocol

Ninety (90) patients from our cohort underwent CMR within 6 months of the indexed echocardiographic evaluation. CMR was performed using a 1.5 T scanner (Signa Excite HDxt platform, GE Healthcare, Milwaukee, WI, U.S.A.) with prospective ECG gating and surface coils positioned anteriorly and posteriorly on the patient's chest. In all patients, scout images were acquired in the transaxial, coronal, and sagittal planes, followed by breath-hold bSSFP (balanced steady-state free precession) cine images in long axis, and in the 4-, 3-, and 2-chamber views, as well as short-axis views covering the entire left ventricle from the atrioventricular plane to the apex. T2-weighted images were also obtained. Late gadolinium enhancement (LGE) images were acquired 10 min after intravenous infusion of gadolinium-based contrast medium (DTPA—diethylenetriaminepentaacetic acid) at a dosage of 0.1 mmol/kg, in the short axis and the 4-, 3-, and 2-chamber long-axis views, using the same orientations as the bSSFP cine images. The slice thickness was set to 8 mm with no inter-slice gap, a 224 × 224 matrix, a 50° flip angle, TI of 300 ms, and 30 cardiac phases. The acquired images were then analyzed.

Estimation of left ventricular volumes and mass was performed by manually tracing the endocardial and epicardial borders in short-axis bSSFP cine sequences. Atrial volumes were evaluated using the 4- and 2-chamber long-axis cine bSSFP sequences, with the subendocardial border traced manually [15]. LAEF and LACI were calculated according to the same percentage ratios described for TTE. Finally, a comparison was made between the CMR data and those obtained from TTE to corroborate the results.

Statistical analysis

Continuous variables were tested for normality with the Shapiro–Wilk test, and subsequently analyzed with Student's t-test for independent samples or, when appropriate, with the Mann–Whitney test. The variance of the groups under analysis was found to be homogeneous with the Levene test. Results are reported as mean ± standard deviation. The discriminative capacity of the continuous variables found to be significant was tested through the analysis of the ROC curves. AUCs were compared between each other with the DeLong test to determine whether there was a significantly better performing indicator. Each significant continuous variable was made categorical by identifying a cut-off value with the Youden index. Categorical variables, reported as absolute count and percentage, were analyzed in contingency tables using the chi-square test. Categorical variables demonstrating a statistically significant association with the event in univariate analysis were subsequently included in multivariate Cox regression models to adjust for potential confounding factors. Only one variable per category (e.g., atrial dimensions: LAD, LAV, LAVi, LACI and LAEF) was chosen from those significant in the univariate analysis. The final model was selected based on the comparison of c-statistics for each model using the DeLong test. The strength of the association was reported as odds ratio (OR) or Hazard ration (HR) with a 95% confidence interval. For all analyses, statistical significance was defined as a two-sided p value < 0.05. To define whether there was concordance between echocardiographic and CMR measurements, Spearman correlation between echo and CMR measurements was performed. The statistical analysis was performed using IBM SPSS Statistics version 24 (IBM Corporation, Armonk, NY, USA), IBM SPSS Statistics version 29 (IBM Corporation, Armonk, NY, USA), and R version 4.0.5 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Between 2011 and 2018, 141 consecutive patients with HCM and no history of AF (mean age 51 ± 17 years) were included in our registry. All patients underwent outpatient evaluations, including ECG and TTE, at our institution and were followed for a mean period of 5 ± 3 years. During follow-up, 35 patients developed AF and were assigned to the HCM-AF group, while the remaining 106 patients were assigned to the HCM-non-AF group. No statistically significant differences were observed between the two groups regarding age, comorbidities, or common cardiovascular risk factors. AF occurred after a mean period of 4 ± 3 years, at a mean age of 55 ± 17 years. Baseline characteristics of the two study groups are summarized in Table 1.

Table 1 Baseline characteristics of the two HCM-AF and HCM-no-AF groups

Echocardiographic parameters

HCM-AF group showed thicker LV wall (maximum thickness 20 ± 5 vs. 18 ± 3 in the HCM-AF group vs HCM-non-AF group respectively, p = 0.04). LAD, LAVmax, LA area, LAVmin and LACI showed to be greater in HCM-AF compared to HCM-non-AF group, respectively: 41 ± 5 vs. 38 ± 5, p = 0.01; 86 ± 23 vs. 77 ± 27, p = 0.01; 25 ± 4 vs. 23 ± 5, p = 0.004; 55 ± 18 vs. 40 ± 20, p < 0.001; 62 ± 20 vs 43 ± 19%, p < 0.001. LAEF and A wave velocity were greater in the HCM-non-AF group (37 ± 9% vs. 49 ± 11% for HCM-AF and HCM-non-AF group, respectively; p < 0.001 and 0.61 ± 28 vs. 0.74 ± 30 for HCM-AF and HCM-non-AF group, respectively; p < 0001). No significant differences emerged regarding ventricular volumes and left and right ventricular function (Table 2).

Table 2 Evaluation of the echocardiographic variables of the two groups

Cardiac magnetic resonance

Ninety (90) patients, 74 male, mean age 53 ± 18 years, underwent CMR. The HCM-AF group had significantly greater values of LAVmax, LAVImax, LAVmin, LAVImin, and LACI (109 ± 25 vs. 87 ± 25, p = 0.001; 59 ± 13 vs. 47 ± 12, p < 0001; 69 ± 21 vs. 50 ± 17, p < 0001; 37 ± 11 vs. 27 ± 9, p < 0001; 56 ± 16 vs. 39 ± 13, p < 0001; respectively). LAEF was significantly higher in the HCM-non-AF group (35 ± 10 vs. 42 ± 8, p = 0.001). Table 3 illustrates in detail the CMR parameters analyzed in the two groups. Among the parameters analyzed, LACI showed the highest concordance between the two methods, while LVEF showed the lowest. Results are reported in Table 4.

Table 3 Evaluation of variables at cardiac magnetic resonance

Table 4 Correlation between TTE and CMR parameters

Indicators comparison

In the ROC analysis, LACI, LAVmin, and LAD demonstrated the strongest discriminative ability. Specifically, a LACI greater than 43% on TTE demonstrated 85% sensitivity and 58% specificity (AUC 0.75) for identifying patients who would develop AF. On CMR, LACI > 44% showed 85% sensitivity and 70% specificity (AUC 0.81) in identifying the same patients.

LAD demonstrated its best performance with a cutoff value of 40.5 mm. Compared to a LAD cutoff of 40.5 mm, LACI exhibited superior accuracy, with an AUC of 0.71 (Fig. 1A, B, C, D, E). Overall, LACI proved to be the parameter with the best performance (AUC 0.8 at TTE, and 0.79 at CMR) (Supplementary Table 1, Supplementary Fig. 1). However, no statistically significant differences were observed when the AUCs were compared using DeLong’s Test. Detailed results of the DeLong Test are provided in the Supplementary Table 2. LAEF > 43% was instead associated with the non-occurrence of AF (sensitivity 76%, specificity 71%, AUC 0.81) (Fig. 1F).

Fig. 1

ROC curves of echocardiographic variables in predicting AF occurrence (A-E) or non AF occurrence during the study period (F). A ROC curve LACI echo, B ROC curve LAVmin echo, C ROC curve LAD, D ROC curve LACI cmr, E ROC curve LAVmin cmr, F ROC curve LAEF echo. LACI: left atrioventricular coupling index, LAD: left atrium diameter, LAEF: left atrium ejection fraction; LAV min: minimal left atrium volume

The univariate analysis highlighted how female sex, LVThick > 19.5 mm, LAD > 40.5 mm, LAVmax > 35 ml, LAVmin > 43 ml, LAEF < 43%, and LACI > 43% were all factors significantly associated with the development of AF. The results of the univariate and multivariate analyses are reported in Table 5. After selecting only one variable per category, as previously described, four models were developed: (1) LV thick > 19.5 mm, sex, LAVmax > 35 ml, LAEF < 43%, LACI > 43%; (2) LV thick > 19.5, female sex, LAD > 40.5 mm, LAEF < 43%, LACI > 43%; (3) LV thick > 19.5 mm, female sex, LACI > 43%, LAEF < 43%, LAVmin > 43 ml; and (4) LV thick > 19.5 mm, female sex, LAVmin > 43 ml, LAEF < 43 ml, LAD > 40.5 mm. The discriminative ability of each model was compared using DeLong’s test for C-statistics. Since no statistically significant differences were observed, the model with the highest C-statistic was selected: Model 2 (LV thick > 19.5 mm, female sex, LAD > 40.5 mm, LAEF < 43%, LACI > 43%). The C-statistic for each model and the details of the DeLong test are provided in Supplementary Materials.

Table 5 Univariate and multivariate analysis of variables associated with the development of AF

Cox regression analysis revealed that LAD > 40.5 mm, LAEF < 43%, and female sex were independently associated with the occurrence of AF, with HR of 2.5 (95% CI: 1.1–5.5), 2.9 (95% CI: 1.2–6.9), and 2.3 (95% CI: 1.1–5.0), respectively.

These factors were also linked to reduced AF-free survival (Fig. 2). The cumulative risk of AF was as high as 5% at two years and 24% after 10 years.

Fig. 2

Survival Curve free from AF for LAEF (on the left) and for LAD (on the right). LAD: left atrium diameter; LAEF: left atrium ejection fraction

Figure 2A and B represent AF free survival in patients with and without LAEF<43% and LAD>40.5mm.

Eleven patients (5 males, mean age at event 53 ± 20 years) either underwent HTx (6) or died for CV causes (5). The events occurred at a mean time of 3 ± 2 years from recruitment. Only three of them had AF onset prior to HTx or death. There was no statistically significant association between LVEF and the composite outcome of HTx or CV death. Conversely, higher values of LACI, LAVmin, and lower values of LAEF, and E wave deceleration time (DecT) were found to be associated with the composite outcome of HTx or death (Supplementary Table 3). LACI revealed good accuracy in identifying patients who would have met the composite outcome of HTx or CV death (AUC 0.74) (Fig. 3). More specifically, a LACI > 60% showed the best discriminative accuracy (sensitivity 64%, specificity 78%). In contrast, a LAEF > 45% and a DecT > 170 ms showed sensitivity of 55% and specificity of 91% (AUC 0.71), and sensitivity of 75% and specificity of 81% (AUC 0.79), respectively, in identifying patients who would be free from HTx or CV death at 5 year (Supplementary Table 4).

Fig. 3

ROC Curves for LACI and LAVmin in predicting HTx or death. LACI: left atrioventricular coupling index, LAV min: minimal left atrium volume

Discussion

Several studies have focused on identifying parameters capable of predicting AF in patients with HCM, but despite numerous proposals, there is a lack of strong evidence supporting any parameter, except for LAD, which remains the most commonly used. In our analysis, we evaluated the association between AF and the morpho-functional parameters of the left atrium. The main findings were: (i) LACI strongly correlates with onset of AF; this result is confirmed by CMR data. (ii) LAEF is independently associated with the development of AF. (iii) LACI was significantly associated with a worse prognosis.

LACI and LAEF in the development of AF

LACI was first evaluated by Pezel et al. in individuals free of clinically recognized HF and cardiovascular disease at baseline, using CMR. They found that LACI provided incremental prognostic value for predicting cardiovascular events, beyond left atrial or left ventricular parameters alone [8].

In our study, LACI, assessed by both TTE and CMR, demonstrated superior performance in identifying patients at risk of developing AF at univariate analysis, superior to LAD, which is the current reference parameter recommended by the European Society of Cardiology (ESC) [5]. Our results are consistent with those of Meucci et al., who identified a similar cut-off (LACI > 40%) in HCM patients with AF, further supporting the reliability of our findings [8].

The strength of LACI in the context of HCM likely lies in its intrinsic ability to reflect early impairment of left ventricular diastolic function and atrial remodeling. Indeed, left atrial emptying is closely associated with left ventricular filling pressures; pathological atrioventricular coupling affects left atrial pressure, contributing to the development and worsening of atrial myopathy (Fig. 4).

Fig. 4

Booster pump phase illustration. In this phase, the LA emptying (white dashed arrow) depends mainly on the strength of LA contraction (white arrows) and atrial afterload. These elements are sensibly expressed in the LACI

In Fig. 5, we show two patients from our cohort: one from the HCM no AF group, exhibiting a low LACI, and the other from the HCM AF group, who developed arrhythmia after 18 months, showing a high LACI. This illustration highlights the differences in atrial function and atrioventricular coupling, which are clearly observed on TTE (Fig. 5).

Fig. 5

Comparison of two patients from our cohort. On the left, a 56-year-old male with obstructive HCM, who remained free of AF until the most recent follow-up (3 years). His LACI was low, at 34%. On the right, a 54-year-old female with non-obstructive HCM, who developed her first episode of AF 18 months after baseline TTE. Her LACI was elevated, at 49%

Notably, in our study, the findings from TTE and CMR imaging are consistent. A strong concordance was observed in the LACI measured with the two different methods, further supporting the standardization of this parameter. Regarding the identification of patients at risk of developing AF, we demonstrated the superior performance of LACI measured with CMR compared to TTE, likely due to CMR's superior precision in volume assessment (LACI CMR AUC 0.8 vs. LACI TTE AUC 0.75).

It is important to note that our cohort presented with early-stage atrial disease, characterized by only mildly dilated atria. Therefore, we can conclude that the superiority of LACI was evident even before severe atrial dilation and arrhythmia occurred, making it a sensitive parameter in the early stages.

On the other hand, LAEF was the parameter independently associated with the development of AF in multivariate analysis. An LAEF < 43% identified patients with a 2,ninefold increased risk of developing AF compared to other HCM patients, who already have an inherent 4- to sixfold higher risk of AF than the general population [2]. In these patients, given the very high risk of AF and the elevated stroke risk, more intensive arrhythmic monitoring may be warranted. This could include periodic 48-h Holter ECG or even the implantation of a loop recorder. Furthermore, an LAEF > 43% suggests a lower risk of developing AF. Several previous studies have already explored the role of LAEF as a potential predictor of AF in patients with HCM. Our results are consistent with those of Tuluce et al., who demonstrated that LAEF, assessed by TTE, was associated with a higher risk of arrhythmia, and that a cut-off value of 49% predicted its development with good specificity [16]. Compared to LACI, LAEF is a more direct marker of intrinsic atrial disease, which explains its stronger association with the development of AF at multivariate analysis.

In our study, LAD was found to have a good diagnostic accuracy for development of AF compared to the other variables. Our results are consistent with the literature, which identifies LAD as current reference parameter for screening the development of AF, although it is often criticized for its relatively low specificity [5, 16, 17].

The role of LA strain may also provide additional insights, as suggested by Debonnaire et al. In a cohort of 242 patients, a strain < 23% was shown to be associated with an increased risk of AF, even in cases with non-dilated LA [18]. However, despite its elegant approach, LA strain analysis requires specialized software and is time-consuming, which limits its applicability in routine clinical practice. Therefore, we opted to focus on parameters that are more feasible for everyday use, providing a more practical approach to arrhythmia screening.

LACI and LAEF as predictors of HTx and death

In our study, we found that LACI was significantly associated with HTx or CV death. Previous studies have reported an association between LACI and worse outcomes in patients with heart failure and reduced LVEF [8, 11, 19]. To the best of our knowledge, we are the first to report this association in the context of HCM. In this setting, where LVEF remains preserved until the later stages of the disease, LACI plays an even more critical role in early prognostic stratification. Atrioventricular uncoupling appears to be a subclinical marker of disease progression toward LVEF reduction and LV dilation. Identifying patients who are more susceptible to this progression may enable closer follow-up and help prevent future hospitalizations.

In our analysis, no statistically significant correlation was found between LAEF and major adverse cardiovascular outcomes. This is likely due to the fact that LAEF is a parameter that primarily reflects atrial function.

Limitations

The main limitations of our study are its retrospective design and small sample size. A prospective registry with a larger population is needed to confirm and strengthen our preliminary findings. Moreover, due to the retrospective nature of the study, advanced echocardiographic parameters were often unavailable, preventing their inclusion in our analysis. Finally, AF detection was based solely on clinical evaluation, which may have led to the underdiagnosis of asymptomatic and subclinical episodes.

Conclusions

In HCM patients, both LACI and LAEF were functional indices associated with a higher risk of AF, a finding further confirmed by CMR data. Additionally, in HCM patients, LACI was associated with the composite outcome of HTx or CV death.