Introduction

Currently, the prevailing classification of atrial fibrillation (AF) is based on the duration and spontaneous termination of AF episodes. Paroxysmal AF is defined as AF episodes that terminated spontaneously within 7 days of onset, while persistent AF describes AF episodes lasting longer than 7 days and less than 12 months1. Paroxysmal AF is considered the early stage of the natural history of AF. Most patients inevitably progress from brief, rare episodes of AF to long-term, frequent episodes, which are associated with risk factors, including age, left atrial size, hypertension, diabetes, and heart failure (HF), even with drug control2. This progression is frequently characterized by deteriorating atrial remodeling and is associated with adverse cardiovascular events, hospitalizations, and death3. Moreover, both medical therapy and radiofrequency ablation are significantly less effective in persistent AF than in paroxysmal AF. Therefore, delaying AF progression is a highly attractive management strategy to improve the prognosis of patients with paroxysmal AF.

Angiotensin receptor-neprilysin inhibitor (ARNI, sacubitril/valsartan), a novel single co-crystal, is composed of sacubitril and valsartan in a ratio of 1:14. Sacubitril is a prodrug that is converted to an active metabolite, LBQ657, which can inhibit the activity of neutral endopeptidase, thereby elevating the levels of natriuretic peptides with antihypertensive and organ-protective effects in vivo4. Valsartan, a traditional angiotensin II type 1 receptor inhibitor, also has antihypertensive and anti-cardiac remodeling effects. The co-crystal structure of sacubitril/valsartan ensures synchronization in the absorption and elimination of sacubitril and valsartan, which generates a synergistic effect of cardiovascular benefits5. In the last decade, numerous well-designed clinical studies have been conducted to verify whether sacubitril/valsartan with dual effect is superior to conventional renin–angiotensin–aldosterone system (RAAS) inhibitors. Sacubitril/valsartan further reduced the risk of hospitalization for HF and death in HF patients with left ventricular ejection fraction (LVEF) ≤ 40%6. Compared with olmesartan, sacubitril/valsartan significantly reduced 24-h ambulate blood pressure in patients with hypertension7,8. In addition, sacubitril/valsartan appears to improve outcomes in patients with myocardial infarction (MI)9,10,11. However, few studies have reported the effects of sacubitril/valsartan on AF. Herein, we examined whether sacubitril/valsartan could inhibit the progression of paroxysmal AF.

Results

Characteristics of the study population

A total of 1083 patients diagnosed with paroxysmal AF were identified, of which 170 patients were eligible for analysis (Fig. 1). Of the 170 patients, 113 (66.5%) patients received angiotensin receptor blocker (ARB) and 57 (33.5%) received ARNI. Before propensity-score matching (PSM), baseline characteristics such as age, total cholesterol (TC), low density lipoprotein cholesterol (LDL-c), brain natriuretic peptide (BNP), left atrial diameter (LAD), LVEF, and history of HF significantly differed between ARB and ARNI groups (Table 1). However, no notable difference in these baseline characteristics was detected between the two groups after PSM. Of the 47 ARB-treated patients, the average age was 64.2 years, 66.0% were males, and the average body mass index (BMI) was 24.9. In the ARNI group, the average age was 64.2 years, 68.1% were males, and the average BMI was 24.7. Moreover, 12 (27.3%) patients in the ARB group and 10 (22.7%) in the ARNI group reached the target dose during follow-up. Types of drugs used in the ARB cohort are shown in Table 2. The time interval for each Holter monitoring from the both groups are shown in Supplementary Tables S1 and S2.

Figure 1
figure 1

Study flow diagram. AF atrial fibrillation, ARB angiotensin receptor blocker, ARNI angiotensin receptor-neprilysin inhibitor, PSM propensity score matching.

Full size image
Table 1 Baseline characteristics of the study population.
Full size table
Table 2 Type of ARB prescribed in the matched cohort.
Full size table

Primary endpoint

Before PSM, 33 (29.2%) patients in the ARB cohort and 6 (10.5%) patients in the ARNI had persistent AF after a median follow-up of 684 (interquartile range [IQR] 487–884) and 726 (IQR 544–910) days, respectively. Figure 2A shows the Kaplan–Meier curve of AF progression. It was found that compared with ARB, ARNI treatment significantly reduced the risk of AF progression (hazard ratio [HR] 0.34; 95% confidence interval [CI] 0.14–0.81; P = 0.015; Table 3). After PSM, the median follow-up time and the occurrence of persistent AF were 743 (IQR 529–936) days and 17 (36.2%) and 709 (IQR 543–886) days and 5 (10.6%) in ARB and ARNI groups, respectively. Relative to those patients with ARB, the HR for AF progression in those patients with ARNI was 0.32 (95% CI 0.12–0.88; P = 0.016; Fig. 2B and Table 3). The changes in LVEF and LAD of the two groups before and after treatment are shown in Supplementary Table S3.

Figure 2
figure 2

Kaplan–Meier curves for AF progression before PSM (A) and after PSM (B). AF atrial fibrillation, ARB angiotensin receptor blocker, ARNI angiotensin receptor-neprilysin inhibitor, PSM propensity score matching, CI confidence interval.

Full size image
Table 3 Risk of AF progression in the original cohort and the matched cohort.
Full size table

Discussion

The present study explores for the first time the efficacy of ARNI in patients with paroxysmal AF. It was found that compared with ARB, ARNI treatment substantially reduced the risk of progression from paroxysmal to persistent AF. This finding may be valuable in guiding AF management.

Currently, antiarrhythmic drugs (AADs) and energy ablation are considered the primary rhythm control strategy for patients with paroxysmal AF. According to the 2020 European Society of Cardiology (ESC) AF guidelines, AADs are recommended for the ‘general’ paroxysmal AF population, while ablation is recommended for paroxysmal AF patients with HF with reduced ejection fraction1. However, both AADs and ablation therapies are faced with a high risk of AF recurrence and progression in the future due to advance atrial remodeling and persistent risk factors12. A previous study involving 1219 patients showed that progression of AF occurred in 178 (15%) patients at 12 months13. Another large cohort study demonstrated that the risk of progression was 8.6% and 24.7% at 12 months and 5 years, respectively14. Other studies reported that the transition from paroxysmal to persistent AF is highly variable, ranging from 2 to 34% in 1 year12.

Substantial evidence indicates that sacubitril/valsartan is a promising drug against HF and hypertension. Compared with traditional RAAS inhibitors, it exhibited a better anti-ventricular remodeling effect and antihypertensive effect due to its dual synergistic effect. Logically, patients with MI or AF are also likely to benefit from sacubitril/valsartan. However, the recent PARADISE-MI study found that sacubitril/valsartan did not achieve statistical significance in reducing cardiovascular death or overall HF in patients with acute MI15. Nevertheless, several clinical studies have shown that sacubitril/valsartan is superior to traditional RAAS inhibitors in improving atrial structural remodeling and reducing AF recurrence in patients with AF after catheter ablation16,17.

In the era of precision medicine, AF progression is a concern in AF management. AF-, age-, and other disease-related remodeling all promote AF progression. Cardiac electrical remodeling together with structural remodeling was implicated in this incremental process. The key electrical remodeling mainly involves alterations in calcium handling that cause triggered activity, changes in sodium channel function that lead to slowed conduction, and alterations in ionic current that facilitate reentry by shortening the action potential duration (APD)/refractory period18. Previous studies have indicated that pulmonary vein ectopic activity is associated with the occurrence of early afterdepolarizations (EADs) and delayed afterdepolarizations (DADs)19. EADs predominantly occur in the context of prolonged APD, and the prolongation of APD occurs as a result of both increased inward current of depolarization (persistent Na+ current, late Na+ current, and L-type Ca2+ current ([ICaL]) and decreased outward K+ current of repolarization20,21. DADs are generally derived from Ca2+-handling abnormalities, including anomalous sarcoplasmic reticulum (SR) Ca2+ release and spontaneous diastolic SR Ca2+ leakage, and are enhanced by SR Ca2+ overload and ryanodine receptor 2 (RyR2) dysfunction22. Increased cytosolic Ca2+ levels during diastole can result in sodium-calcium exchanger (NCX) hyperfunction that leads to a transient-inward current, which can cause membrane depolarization, triggering arrhythmias22. A recent study showed that sacubitril/valsartan could ameliorate the dysfunction of the RyR2 complex and NCX1 complex, which suggesting that sacubitril/valsartan may improve SR Ca2+ mishandling and help reduce AF vulnerability23. Moreover, sacubitril/valsartan was shown to reduce the expression of phosphorylated calmodulin-dependent protein kinase II (CaMKII), which is an established pro-arrhythmic molecule24.

Cardiac conduction velocity is closely related to voltage-dependent sodium currents, cardiomyocyte–cardiomyocyte gap junctional coupling, and muscle bundle anatomic structure18,25. Lowered sodium currents, reduced cardiomyocyte electric coupling, and atrial muscle bundle disorganization caused by fibrosis all reduce cardiac conduction and facilitate reentry. In addition, lowered ICaL, increased inward-rectifier K+ currents, and slow delayed rectifier K+ currents shorten APD and promote reentry26,27,28. A recent study demonstrated that sacubitril/valsartan increased ICaL density in a rapid atrial pacing-induced rabbit AF model, which may contribute to inhibit the formation of reentry29.

Structural remodeling in AF progression is characterized primarily by increased atrial fibrosis and atrial enlargement. The major profibrotic molecules include angiotensin II and transforming growth factor β1 (TGFβ1), and the signaling pathways mainly involve Jun N-terminal kinase (JNK), mitogen-activated protein kinase (MAPK), and extracellular signal-related kinases30,31. In a rodent study, sacubitril/valsartan was found to suppress TGFβ1-Smad2/3, p-p38, and p-JNK signaling pathways, and reverse atrial fibrosis, thereby inhibiting AF progression32. Similarly, in a clinical setting, improved left atrial size was observed in HF AF patients17,33.

Over the past few decades, despite tremendous advances in our understanding of the electropathology of AF, the mechanisms underlying AF progression remain elusive. In addition to the general concern of fibrosis, fat accumulation, amyloidosis, and other still unidentified factors may be important for AF progression34,35. Besides, animal models induced by a specific single stimulus in the short term may limit the observation of complex mechanisms. Interventions that are successful in animal models often fail in clinical practice. In fact, clinical AF is usually the result of long-term complex pathophysiology. Therefore, the inhibition of AF progression by sacubitril/valsartan observed in the real-world has important clinical significance.

This study has several limitations. Firstly, this is a retrospective study, which inevitably includes bias in patient selection; our analysis should be considered hypothesis generating, and therefore further prospective studies are required. Secondly, subgroup analysis could not be conducted, due to limited subjects. Thirdly, the proportions of patients who achieved the target dose were 27.3% and 22.7% in sacubitril/valsartan and ARB groups, respectively. Although these numbers are not ideal, they reflect a real world clinical setting. Lastly, we used 7-days Holter monitoring to record the transition from paroxysmal to persistent AF. However, long-term (> 7 days) continuous electrocardiograph (ECG) monitoring can obtain more accurate information.

In summary, sacubitril/valsartan may be superior to ARB in reducing the risk of AF progression from paroxysmal to persistent in patients with paroxysmal AF who did not receive catheter ablation.

Methods

Study design and participants

This retrospective cohort study was conducted on consecutive patients with paroxysmal AF admitted at the Second Affiliated Hospital of Nanchang University between January 2017 and January 2022. Patients with paroxysmal AF were reviewed from the hospital’s electronic database. Exclusion criteria were: (1) patient who had received ablation therapy; (2) did not receive ARB or ARNI; (3) did not receive 7-day Holter monitoring; (4) previous history of cardiomyopathy or valvular; (5) chronic inflammatory disease; (6) connective tissue disease; (7) hyperthyroidism and (8) malignancy.

Drug therapy

The first-line antiarrhythmic drug was propafenone (600 mg per day). However, amiodarone (200 mg per day) or sotalol (160 mg per day) was administered when propafenone was contraindicated. Rate control drugs, including beta receptor blockers, calcium channel blockers, and digoxin, were administered as necessary. The use of ARB or ARNI was based on the recommended guideline and physician’s choice, and the dosages were adjusted according to blood pressure, and individual tolerance36,37. The estimated risk of thromboembolism was calculated for each patient based on the CHA2DS2-VASc. Oral anticoagulants, including warfarin, dabigatran, and rivaroxaban, were recommended to prevent ischemic stroke in patients with a CHA2DS2-VASc score greater than 1 in males or greater than 2 in females.

Data collection

General information on age, sex, BMI, duration of AF, and comorbidities was collected. Moreover, the blood pressure values of all patients were recorded at the time of the first outpatient visit or admission. Laboratory values, such as TC, triglyceride (TG), high density lipoprotein cholesterol (HDL-c), LDL-c, estimated glomerular filtration rate (eGFR), and BNP were also collected. Echocardiographic cardiac parameters such as LAD, and LVEF were measured. Other clinical data included the New York Heart Association functional classification, medications, and long-term ECG record.

Patients follow-up and clinical outcomes

During the first year, outpatient follow-up was conducted every 1–3 months and every 6 months thereafter. Patients were advised to seek immediate clinic follow-up if AF-related symptoms occurred36,37. Twenty-four-hour Holter monitoring was conducted at each visit. Seven-day Holter monitoring was performed annually or whenever necessary. The risk of progression from paroxysmal to persistent AF was compared between paroxysmal patients treated with ARNI and those who received an ARB. Paroxysmal AF was defined as AF that spontaneously terminated or with intervention within 7 days, and persistent AF was defined as AF that lasted ≥ 7 days. Patients were censored if they discontinued ARNI or ARB therapy during the period of follow-up.

Statement of ethics

This study was approved by the Medical Ethics Committee of the Second Affiliated Hospital of Nanchang University and met the standards of the Declaration of Helsinki. Informed consent was waived by the Medical Ethics Committee of the Second Affiliated Hospital of Nanchang University because of the retrospective nature of this study.

Statistical analysis

Variables were expressed as means (standard deviations, [SD]) or median (IQR), and frequencies (proportions [%]). Between-group comparisons were conducted using Student's t- or the rank-sum test, and χ2- or the Fisher’s exact test. PSM was conducted using the nearest-neighbor method with a caliper width of 0.2 to reduce potential bias between ARNI and ARB. A standardized differences < 0.2 was considered to indicate acceptable balanced groups on a given covariate38. Survival analysis was performed in the matched cohorts to compare the risk of AF progression between groups. The HR, and 95% CI were computed. The proportional hazards assumption was checked using the Schoenfeld residuals. All data analyses were performed using RStudio version 1.1.414 (Boston, MA, USA) and Empower (http://www.empowerstats.com; X&Y Solutions, Inc., Boston, MA).