Introduction

Atrial arrhythmias (AAs) and heart failure (HF) have been found to interact synergistically, leading to hemodynamic deterioration [1,2,3]. Recently, due to emerging evidence, the paradigm regarding AA management has shifted markedly to emphasize effective rhythm rather than rate control [4], highlighting the importance of rhythm control interventions in the management of AAs. Among these strategies, electrical cardioversion (ECV) has emerged as the foremost therapeutic option for restoring sinus rhythm (SR) in patients with AA [5, 6]. Nevertheless, the timing of ECV that is ideal for sustaining SR and ensuring cardiac function recovery and associated clinical outcomes remains ambiguous in patients hospitalized for acute decompensated HF (ADHF) with concurrent AAs. The aim of this study was to retrospectively evaluate the impact of ECV timing on SR maintenance, hospitalization duration, and change in cardiac function in patients with ADHF and AAs and to identify factors associated with good outcomes after ECV.

Materials and methods

Study design

Data were obtained from SAKURA HF REGISTRY-2 (UMIN 000043852), a single-center, prospective, observational cohort registry. Among the 1,792 patients with HF who were included in the registry between October 2017 and December 2022, 601 were hospitalized for ADHF with concomitant AAs, including atrial fibrillation (AF) and atrial flutter (AFL). In all cases, according to guidelines, ECV was generally indicated for strongly suspected AAs contributing to the worsening of HF symptoms despite optimal medical treatment [3]. However, ECV was not performed in patients with possible left atrial (LA) or left atrial appendage (LAA) thrombus on transesophageal echocardiography (TEE), spontaneous SR restoration before the initiation of ECV, severe hemodynamic compromise requiring intensive care management, or the inability to give informed consent. In total, ECV was attempted in 73 patients (62 with AF and 11 with AFL) during hospitalization in accordance with the current treatment guidelines (Fig. 1) [7]. The patients who died in the hospital or were transferred to another hospital were excluded. Clinical presentation was categorized as (1) typical AA symptoms (palpitation with or without other concomitant symptoms); (2) atypical AA symptoms (shortness of breath without palpitations or chest pain); or (3) others (leg edema without any other symptoms) [8].

Fig. 1
figure 1

Flow chart of the present study. AA atrial arrhythmias, ADHF acute decompensated heart failure, ECV electrical cardioversion, NUIH Nihon University Itabashi Hospital

Full size image

The timing of ECV was determined on the basis of the patient’s hemodynamic tolerance and anticoagulation before ECV. All patients underwent transesophageal echocardiography (TEE) before ECV to confirm the absence of LA or LAA thrombi. ECV was performed using short sedation with intravenous midazolam, thiopental, or propofol under continuous blood pressure and electrocardiogram (ECG) monitoring. The cardioversion protocol included adhesive pads with the anterior–posterior electrode position and a biphasic shock waveform with up to four cardioversion attempts [9]. Energy level of cardioversion was gradually increased to a maximum of 200 J. A 12-lead ECG trace was recorded before and after the procedure. Periprocedural antiarrhythmic drugs (AADs) were administered within 24 h after ECV. The AAD regimen was administered at the discretion of the physician. Each patient’s diagnosis of HF was based on Framingham criteria [3].

Patients were classified into the following two groups based on the median duration from hospitalization to ECV: early and delayed ECV [10]. Catheter ablation (CA) was not performed during the study period. Before admission, frailty was assessed using the Clinical Frailty Scale (CFS) during the stable phase. The CFS score was assessed based on interviews with patients and their families. This study complied with the principles of the Declaration of Helsinki. The use of patient information was approved by the Nihon University Itabashi Hospital Ethics Committee (RK-180612-2).

Endpoints

The primary evaluated endpoints included both very short-term and short-term ECV failure. Very short-term failure was defined as a combination of unsuccessful cardioversion and AA recurrence during hospitalization. Short-term failure was defined as the same combination occurring within a month after ECV [9, 11, 12]. Successful ECV was defined as a restored SR lasting at least 1 min after ECV. Secondary endpoints included the (1) acute success of ECV, (2) number of ECVs attempted before SR restoration, (3) the maximum energy delivered for successful ECV, (4) periprocedural complications, (5) changes in transthoracic echocardiographic parameters within 2 months after successful ECV, and (6) duration of hospitalization. Periprocedural complications included sedation-related complications, bradycardia (heart rate [HR] < 40 bpm), hypotension requiring treatment, thromboembolism, worsening HF, and critical arrhythmia such as ventricular fibrillation or cardiac arrest (> 5 s) requiring treatment [11, 13, 14]. After ECV, ECG was continuously monitored throughout hospitalization in all patients. Furthermore, outpatient hospital visits were scheduled within 2–3 weeks of discharge to check the 12-lead ECG. When patients experienced symptoms such as palpitations or shortness of breath, they visited the emergency department to detect AAs. AAs lasting > 30 s on the ECG monitor, 24-h Holter, or 12-lead ECG were defined as recurrent AAs [7].

Statistical analysis

Categorical variables are reported as counts and percentages and were compared using Pearson’s χ2 or Fisher’s exact tests, where appropriate. Distributions of continuous variables were assessed using the Shapiro–Wilk test and are presented as a mean ± standard deviation [SD] or median and IQR. Continuous variables were compared using Student’s t-test or the Mann–Whitney U test, as appropriate. Kaplan–Meier cumulative survival curves were constructed with group differences compared using the log-rank test to estimate the short-term failure of ECV. The relationship between the duration of hospitalization and time to ECV after hospitalization was tested using Spearman’s rank correlation test.

To investigate factors with the potential to affect the duration of hospitalization, we defined a long hospital stay as one that was longer than the median hospital stay duration of the group, as was done in previous analyses [10]. Among the study population, the median duration of hospitalization was 16 days. Univariate and multivariate logistic regression analyses were performed to identify factors associated with a long hospital stay. In the multivariable analysis, all variables with a single-variable value of P < 0.05 were adjusted. Furthermore, univariate and multivariate Cox proportional hazards regression analyses were used to identify predictors of short-term ECV failure. To satisfy assumptions of the model, a natural transformation (ln) was applied to N-terminal pro-brain natriuretic peptide (NT-proBNP) data. Statistical analyses were performed using JMP Pro 16.1.0 (SAS Institute, Cary, NC, USA).

Results

Patient characteristics

Between October 2017 and December 2022, 1792 patients hospitalized for HF had data included in the SAKURA HF REGISTRY-2 (UMIN 000043852). Among them, 601 were hospitalized for ADHF with concomitant AAs. After inclusion and exclusion criteria were applied, 73 patients (62 with AF and 11 with AFL) who underwent ECV were included in this study. Patients were classified into the following two groups based on the median duration from hospitalization to ECV: early ECV (ECV within 8 days of hospitalization, n = 38) or delayed ECV (ECV performed at least 9 days after hospitalization, n = 35) [9]. The median number of days to ECV was 4 (interquartile range [IQR] = 3–7) for those of the early ECV group and 13 (IQR = 11–16) for those of the delayed ECV group.

Baseline characteristics of included patients are shown in Table 1. The prevalence of stroke history, CHA2DS2-VASc score, and hemoglobin A1c levels were significantly increased in the early ECV group versus those in the delayed ECV group (all P < 0.05). Other medical history-related factors, laboratory findings, transthoracic echocardiographic parameters, and heart failure etiologies of the two groups were similar. Regarding clinical presentation on admission, 33% (24), 63% (46), and 3% (2) of the patients presented with typical, atypical, and other AA symptoms, respectively. The clinical characteristics on the day of ECV are shown in Table 2. The frequency of New York Heart Association (NYHA) class IV and HR values just before ECV were significantly higher in early ECV group than in the delayed group (all P < 0.05). In addition, AFL was more frequent (27% vs. 3%, respectively, P = 0.008). AF occurred less frequently (73% vs. 97%, respectively, P = 0.004) in the early ECV group than in the delayed ECV group. Although the delayed ECV group more frequently received mineral corticoid receptor antagonists (P = 0.020) and less frequently received bepridil (P = 0.039), no between-group differences regarding other HF medications or periprocedural AADs were observed.

Table 1 Baseline characteristics
Full size table
Table 2 Clinical variables on the day of ECV
Full size table

Endpoints

The clinical outcomes are presented in Table 3. ECV successfully restored SR in 62 of 73 patients (85%) at a mean energy level of 100 J. In 10 patients (14%), multiple ECV attempts (≥ 3) were needed. Periprocedural complications occurred in six patients (8%). Among them, the majority were temporal bradycardia and hypotension, which were resolved in all patients without pacemakers. Furthermore, no stroke occurred and worsening HF occurred in only one patient (1%). No significant between-group differences in endpoints were observed. Very short-term failure of ECV (unsuccessful cardioversion and AAs recurrence during hospitalization) occurred in 39 patients (53%), without any significant between-group differences (51% early ECV vs. 63% delayed ECV, P = 0.87) (Fig. 2A). The incidence of short-term ECV failure (unsuccessful cardioversion and AAs recurrence within 1 month) also did not differ significantly between the groups (61% early ECV vs. 72% delayed ECV, P = 0.43) (Fig. 2B).

Table 3 Clinical outcomes
Full size table
Fig. 2
figure 2

Kaplan–Meier curves for very short-term (A) and short-term (B) failures of electrical cardioversion. ECV electrical cardioversion

Full size image

Among the 62 patients with acute successful ECV, 37 underwent transthoracic echocardiography (TTE) before and after cardioversion. The median days to follow-up TTE after ECV was 11 days (IQR = 5–30 days). Among the 37 patients, the comparison of TTE findings on admission and those at follow-up are shown in Fig. 3. Left ventricular ejection fraction (LVEF) values significantly improved (38% [31–52] to 51% [39–63], P = 0.008), and left atrial volume index decreased (LAVI) (48 ml/m2 [37–57] to 38 ml/m2 [29–61] ml/m2, P = 0.07) from admission to follow-up. There were no differences in the echocardiographic parameters when values of early and delayed TTE groups were compared (Table 3). There were no between-group differences in TTE-related parameters observed when recurrence and non-recurrence groups were compared, except follow-up TTE was earlier in the non-recurrence versus recurrence group (6 days [3–21] vs. 30 days [8–42] days, P = 0.008; Table 4).

Fig. 3
figure 3

The change of transthoracic echocardiographic findings following cardioversion. Values are expressed as the median (interquartile range) or number (%). LAVI left atrial volume index, LVEF left ventricular ejection fraction, LVEDV left ventricular end-diastolic volume, LVESV left ventricular end-systolic volume, MR mitral regurgitation

Full size image
Table 4 Comparison of echocardiographic findings between patients with and without recurrence of atrial arrhythmias after electrical cardioversion
Full size table

The hospitalization duration was shortened in the early ECV group versus that of the delayed group (14 days [11–21] vs. 17 days [15–26], P < 0.001) (Table 3). The length of hospital stay was also correlated with days to ECV during hospitalization (Spearman’s ρ = 0.47, P < 0.001) (Fig. 4). After adjusting for all covariates associated with hospitalization duration in univariate analyses, early ECV, heart rate just before ECV, and tolvaptan administration were identified as independent determinants of a shortened duration of hospitalization (Table 5).

Fig. 4
figure 4

Correlation between the length of hospital stay and time until electrical cardioversion during hospitalization. ECV electrical cardioversion

Full size image
Table 5 Univariate and multivariate logistic regression analyses of factors associated with a longer hospital stay (more than 16 days)
Full size table

After adjusting for all covariates with univariate values of P < 0.05, AF was identified as an independent positive predictor of the short-term failure of ECV (OR 11.0, 95% CI 1.40–87.3, P = 0.023) (see Online Resource Table S1). Conversely, AFL was an independent negative predictor of short-term ECV failure via multivariate analysis (OR 0.09, 95% CI 0.01–0.72, P = 0.023).

Discussion

This study revealed that among the patients hospitalized due to ADHF with concomitant AAs, (1) ECV successfully restored SR in 85% with a median number of ECV attempts of 1 and periprocedural complication rate of 8%; (2) short-term failure of ECV, defined as the combination of unsuccessful cardioversion and AA recurrence within 1 month after ECV, occurred in 67% of patients; (3) LVEF improvement and LAVI decrease were observed during a median follow-up period of 11 days; and (4) early ECV resulted in the shortest duration of hospitalization.

ECV is effective rhythm control therapy for AF/AFL management in the restoration of SR, with an overall high success rate of 70–90% [3, 5, 15]. However, AA recurrence is common, especially in those with congestive HF, which is reported to be a powerful predictor of unfavorable short-term outcomes after ECV [11, 16]. In clinical practice, the optimal timing of ECV in patients with ADHF and AAs has not been clarified. A previous multicenter randomized trial investigating the impact of ECV timing on the safety and effectiveness of the procedure in patients with recent-onset symptomatic AF without HF has been performed [12]. In the study, the wait-and-see strategy was non-inferior to early cardioversion for restoring SR at 4 weeks after the index visit. As for AF patients with left ventricular dysfunction (LVEF ≤ 45%), the ECV acute success rate was reported to be strongly associated with the degree of HF pharmacological therapy before ECV [15]. Considering these findings, the optimal timing of ECV in patients with HF and AAs might be different from that of those without HF. Our data suggest that early ECV is associated with a short hospitalization duration. Several previous studies have demonstrated the usefulness of ECV in terms of its ability to rapidly improve LV function, quality of life, and NYHA functional class [5, 17]. Consistent with findings from previous studies, we showed that LVEF significantly improved and LAVI tended to decrease within 2 months of ECV in our population. The present study also demonstrated that the AA recurrence rate within 1 month after ECV was high (61%) despite a high acute ECV success rate (84%). Given these findings, CA should be recommended after ECV to facilitate the long-term maintenance of SR and improve HF prognosis [4]. Meanwhile, CA is associated with a potential risk of periprocedural complications, especially during the acute phase of HF. Patients who undergo CA have a higher volume overload burden during the procedure, causing periprocedural HF [18]. Congestive HF itself was also reported to be an independent predictor of periprocedural stroke [19]. In our population, critical periprocedural complications, including worsening HF and thromboembolism, were also rare, and the success rate of acute ECV was high. Therefore, early ECV may be useful for achieving temporary rhythm control, facilitating smooth and safe CA subsequently, by improving cardiac function and shortening hospital stay.

Our additional analysis showed that AF is an independent positive predictor of short-term ECV failure, but vice versa in AFL. AF and AFL commonly overlap due to their shared precipitants and risk factors [2, 20]. In addition, AF/AFL is a common arrhythmia associated with a cause or consequence of HF, which leads to a worsened prognosis [2]. Data related to AF/AFL are scarce, especially regarding AFL treatment strategies. Further, a serious issue in which HR control with landiolol was difficult in those with AFL [2, 3, 20]. The present study showed that physicians attempted ECV earlier in patients with AFL than in those with AF, possibly due to difficulty controlling HR. Furthermore, another study reported that ECV more effectively controlled rhythm in new-onset AFL than in new-onset AF [16]. These results indicate that ECV is a reasonable strategy for treating patients with HF and AFL. Our study also showed that HR just before ECV was a determinant of long hospital stay, a finding that aggresses with the suggestion of a previous study [20, 21]. Among patients with HF with uncontrolled rapid HR regardless of AAs, early ECV may be more beneficial than continuing to rely on medication.

This study had several limitations. First, this was a retrospective and single-center study; therefore, ECV were not randomly assigned. Physician discretion limits the comparability of early and delayed ECV. However, application of the same ECV protocol to all patients with HF and AAs excluded as much bias as possible. In addition, using a consecutive series of patients in long-term observational ECV practice (> 5 years) minimizes selection bias and reflects contemporary clinical practice. Further randomized prospective trials are warranted to determine the optimal timing of ECV. Secondly, hospitalization duration variability may have been affected by unmeasured confounding variables. Third, the recurrence rate of AAs may have been underestimated due to failing to consider patients with asymptomatic recurrence. Nonetheless, we monitored ECG continuously after ECV as long as possible, that was possible because our population underwent ECV during hospitalization.

Conclusions

Early ECV was a major determinant of a short hospitalization duration, along with a significant increase in LVEF. Additionally, AFL was an independent negative predictor of short-term ECV failure. As a more effective treatment strategy for acute HF and AAs, early ECV in combination with standard pharmacological therapy may benefit patients, particularly in those with uncontrolled rapid HR due to rapid AFL.