Cardiac Resynchronization as an Enabler of ARNI Efficacy: Insights From a Real‐World Cohort

ABSTRACT

Background

Sacubitril/valsartan (ARNI) improves outcomes in heart failure with reduced ejection fraction (HFrEF). In patients undergoing cardiac resynchronization therapy (CRT), the potential synergy between device‐based and pharmacological strategies remains underexplored.

Methods

The aim of this study was to assess the effect of ARNI therapy on reverse remodeling, clinical response, and cardiovascular events at 1 year in CRT recipients, in order to clarify the potential synergistic role of pharmacological and device‐based strategies. A retrospective cohort of 194 consecutive CRT patients was analyzed, of whom 65 (33.5%) received ARNI. Echocardiographic response was defined as an increase in LVEF > 5% or ≥ 15% reduction in LV end‐systolic volume at 12 months; clinical response as an improvement of at least one NYHA class. The primary endpoint was echocardiographic response. Device‐detected remote monitoring heart failure (RM‐HF) events were pre‐specified as the key time‐to‐event endpoint, while mortality and HF hospitalization were exploratory.

Results

Baseline characteristics were comparable, although systolic blood pressure was higher in ARNI recipients. At 12 months, ARNI therapy was associated with greater reverse remodeling (ΔLVEF 14.4% ± 10.2% vs. 10.8% ± 11.0%, p = 0.04). Rates of echocardiographic (68% vs. 67%) and clinical response (68% vs. 74%) were similar. Mortality and HF hospitalization did not differ significantly. In contrast, ARNI recipients showed fewer RM‐HF events, with higher event‐free survival among echocardiographic responders (95.7% vs. 82.6%; log‐rank p = 0.021, q = 0.042) and clinical responders (93.4% vs. 85.3%; p = 0.040, q = 0.040).

Conclusions

In this real‐world CRT cohort, ARNI therapy enhanced reverse remodeling and reduced subclinical congestion events detected by remote monitoring. These findings support a synergistic interaction between CRT and ARNI, where resynchronization provides the substrate for pharmacological efficacy.

CRT provides the mechanical substrate that enables sacubitril/valsartan to maximize reverse remodeling. In this real‐world cohort, ARNI therapy was associated with greater improvement in left ventricular function and fewer clinically validated device‐detected HF events, particularly among CRT responders, supporting a synergistic interaction between device‐based resynchronization and pharmacological therapy.

1. Introduction

Neprilysin inhibitors, when combined with angiotensin receptor antagonists (ARNI), represent a major advance in the treatment of heart failure with reduced ejection fraction [1, 2]. Current international guidelines recommend ARNI as a fundamental therapy in this setting. It is now regarded as a cornerstone of heart failure (HF) management [3]. However, several aspects remain to be clarified, including its impact on specific patient subgroups, the optimal timing of initiation, and its potential interaction with device‐based or advanced therapeutic strategies. Alongside pharmacological therapy, cardiac resynchronization therapy (CRT) plays a central role in patients with electrical dyssynchrony, aiming to restore coordinated ventricular contraction and promote reverse remodeling [4, 5]. Traditionally achieved through biventricular pacing (BiVp), CRT has demonstrated significant benefits in symptoms and outcomes [5]. More recently, conduction system pacing (CSP) has emerged as a physiologic alternative or as a bailout strategy when conventional CRT is not feasible or effective, providing a direct means of engaging the His–Purkinje system to achieve resynchronization [6, 7]. Because worsening HF is typically preceded by a phase of subclinical haemodynamic deterioration, device‐based multiparametric remote monitoring (RM‐HF) may identify patients at increased risk before overt decompensation requiring hospitalization. Evaluating these clinically actionable episodes may therefore provide complementary information to conventional hard endpoints by capturing an earlier stage of HF progression [8, 9]. The aim of this study was to evaluate the effect of ARNI therapy on echocardiographic reverse remodeling, clinical response, and its relationship with cardiovascular events at 1‐year follow‐up in patients undergoing cardiac resynchronization therapy, in order to better define the role of ARNI in this population, given the potential synergistic effects between pharmacological and device‐based strategies. These were analyzed on all population study and in the responder and non‐responder group.

2. Methods

Between October 2020 and May 2024, all consecutive patients with a guideline‐based indication for CRT who underwent device implantation at our institution were retrospectively screened for study eligibility. A total of 194 patients constituted the study population. Baseline data included medical history, baseline ECG, pharmacological therapy, NYHA functional class, echocardiographic parameters, and the percentage of ventricular pacing. Sacubitril/valsartan was titrated according to blood pressure, renal function, serum potassium levels, and individual tolerability. At the time of CRT implantation, all patients receiving ARNI were treated with the maximum individually tolerated dose. At 12‐month follow‐up, all these parameters were reassessed, and the change in left ventricular ejection fraction (ΔEF, defined as EF at 12 months minus baseline EF) was calculated. Echocardiographic responders were defined as patients with a ΔEF > 5% or a reduction in LV end‐systolic volume > 15% at 1 year. Clinical responders were defined as those with an improvement of at least one NYHA class [4, 5].

2.1. Echocardiographic Exam

Echocardiographic evaluation was performed using standard parameters, including chamber diameters, ventricular volumes and left ventricular ejection fraction assessed by the biplane Simpson's method. All echocardiographic measurements were performed offline by two independent readers who were blinded to ARNI status and clinical outcomes.

2.2. Clinical Evaluation

At baseline and at 12‐month follow‐up, patients underwent a HF outpatient clinic visit, including physical examination, 12‐lead ECG, and assessment of functional status according to the NYHA classification.

2.3. Biventricular Pacing Implant Description

In the standard group, after placement of the right ventricular (RV) and atrial leads, a quadripolar left ventricular (LV) lead was advanced into the coronary sinus (CS) under venographic guidance. Whenever a lateral or posterolateral vein was suitable, the interval between the onset of the RV electrogram and that recorded from the different LV dipoles (RVs–LVs) was measured using the PSA system. If the paced QRS morphology was unsatisfactory, the interventricular conduction delay was < 100 ms, or based on operator judgment, BiVp was not pursued and CSP was performed instead.

2.4. CSP Description

In patients in whom BiVp was not feasible or deemed suboptimal, CSP was performed as a bailout strategy. The pacing lead was advanced through a dedicated delivery sheath to the basal interventricular septum. Under fluoroscopic and electrogram guidance, the lead was progressively screwed deep into the septal myocardium until criteria for left bundle branch capture were achieved, including the presence of a terminal r′ in lead V1 and a short stimulus‐to‐left ventricular activation time, according to CSP consensus document [6, 7]. Once left bundle branch pacing was confirmed, pacing thresholds and sensing parameters were systematically evaluated to ensure stability and adequate lead performance before final fixation. CSP was performed exclusively by means of left bundle branch area pacing (LBBAP). LBBAP was adopted only as a bailout strategy when conventional CS lead implantation was unsuccessful, anatomically unfeasible, or considered suboptimal according to the operator's judgment. No patients underwent left bundle branch optimized cardiac resynchronization therapy (LOT‐CRT). Therefore, LBBAP was used as an alternative to CS lead implantation rather than in combination with it.

2.5. Twelve Months Follow Up

At 12‐month follow‐up, patients underwent a comprehensive re‐evaluation including transthoracic echocardiography, clinical assessment with NYHA class evaluation, 12‐lead ECG, and device interrogation. Echocardiographic and clinical parameters were compared with baseline values to assess ventricular remodeling and treatment response.

2.6. Clinical Endpoint

Device‐detected remote monitoring heart failure (RM‐HF) events were defined as alerts generated by vendor‐specific diagnostic algorithms (e.g., Medtronic TriageHF, Boston Scientific HeartLogic, Biotronik HeartInsight, Abbott CorVue). Although these algorithms differ in the physiological parameters included and their diagnostic performance, each alert was systematically reviewed through a structured outpatient assessment before being considered a clinically actionable RM‐HF event. All RM‐HF alerts triggered a structured outpatient evaluation, including clinical review, adjustment of diuretic therapy, and ancillary tests such as chest radiography when indicated, ensuring that alerts reflected clinically actionable episodes of incipient decompensation. RM‐HF events were generated according to the manufacturer‐specific heart failure diagnostic algorithm implemented in the implanted CRT device. Specifically, all Medtronic CRT‐P and CRT‐D devices used the TriageHF algorithm, all Boston Scientific CRT‐D devices used HeartLogic, Abbott CRT‐P and CRT‐D devices used CorVue, and Biotronik CRT‐D devices used HeartInsight. Therefore, all patients within each device platform were monitored using the same manufacturer‐specific heart failure diagnostic algorithm [10]. RM‐HF events were selected to evaluate clinically actionable episodes of subclinical HF deterioration, whereas hospitalization for HF and mortality were analyzed as conventional clinical endpoints. Analyses of all‐cause mortality, HF hospitalization, and the composite endpoint were exploratory and are reported in the Supporting Information.

2.7. Statistical Analysis

Continuous variables were expressed as mean ± standard deviation and compared between groups using analysis of variance (ANOVA). Within‐group changes from baseline to 12‐month follow‐up were assessed using paired t‐tests for normally distributed variables or Wilcoxon signed‐rank tests for non‐normally distributed variables. Categorical variables were presented as counts and percentages, with comparisons performed using the chi‐square test or Fisher's exact test, as appropriate. For paired categorical data, McNemar's test was applied. Univariable logistic regression analyses were performed to evaluate the association between baseline demographic, clinical, and echocardiographic parameters and the probability of response (echocardiographic, clinical). Variables associated with the outcome at a significance level of p < 0.10 in univariable analyses were considered candidates for multivariable modeling. Multivariable logistic regression models were constructed to identify independent predictors of response. The number of covariates included was restricted relative to the study population to avoid model overfitting and preserve statistical power, following conventional events‐per‐variable criteria. Results were expressed as odds ratios (OR) with 95% confidence intervals (CI). Time‐to‐event analyses were performed for clinical outcomes. Prespecified subgroup analyses were performed according to baseline QRS morphology (left bundle branch block, right bundle branch block, or other conduction disturbances) and pacing strategy (BiVp or CSP). Interaction testing was applied to assess whether treatment effects were consistent across subgroups. A two‐sided p‐value < 0.05 was considered statistically significant. Statistical analyses were conducted using IBM SPSS Statistics, version 29.0 (IBM Corp., Armonk, NY, USA). The primary endpoint was echocardiographic reverse remodeling. RM‐HF events were pre‐specified as the key time‐to‐event endpoint, with multiplicity controlled using the false discovery rate (FDR) method; the Benjamini–Hochberg procedure (two‐sided FDR 5%). Analyses of all‐cause mortality, HF hospitalization, and the composite endpoint were exploratory, underpowered, and are reported with nominal p‐values only.

3. Results

3.1. Baseline Characteristic

Among the 194 patients included, 65 (33.5%) received ARNI therapy at the time of CRT implantation, whereas 129 (66.5%) did not. Patients receiving ARNI were slightly younger and had higher systolic blood pressure than non‐ARNI patients, whereas renal function and baseline echocardiographic characteristics were comparable between groups. As expected, ACEI/ARB treatment was absent in ARNI recipients. These parameters were shown in Table 1. Among the 65 patients receiving ARNI at CRT implantation, 35 (53.8%) were treated with 24/26 mg twice daily, 24 (36.9%) with 49/51 mg twice daily, and 6 (9.2%) with 97/103 mg twice daily.

TABLE 1.

Baseline demographic, clinical, electrocardiographic, echocardiographic, pharmacological, and device characteristics according to ARNI treatment.

	CRT with ARNI 65/194	CRT without ARNI 129/194	p
Age, years	68.9 ± 10.6	71.7 ± 11.3	0.09
Female sex, n (%)	18 (27.7%)	38 (29.5%)	0.9
BSA, m2	2.0 ± 0.90	2.0 ± 0.97	0.1
NYHA class	2.1 ± 0.53	2.3 ± 0.53	0.09
SBP (mmHg)	109 ± 9.1	101 ± 11.4	0.01
Creatinine (mg/dL)	1.21 ± 1.8	1.24 ± 1.3	0.54
eGFR (mL/min/1.73 m2)	45.3 ± 11.5	47.4 ± 4.2	0.09
QRS duration, ms	147.3 ± 21.9	148.52 ± 22.51	0.7
Echocardiographic parameters
LVEDDi, mm/m2	32.14 ± 5.89	30.84 ± 5.15	0.09
LVESDi, mm/m2	26.11 ± 6.52	23.57 ± 4.61	0.1
LVEDVi, mL/m2	105.12 ± 31.75	92.60 ± 27.8	0.06
LVESVi, mL/m2	75.6 ± 33.77	71.67 ± 23.38	0.1
LVEF, %	27.5 ± 8.73	29.54 ± 8.7	0.2
ECG characteristic
Ischemic cardiomyopathy, n (%)	22 (33.8%)	57 (44.2%)	0.2
LBBB, n (%)	45 (69.2%)	85 (65.9%)	0.8
Non LBBB, n (%)	15 (23.1%)	30 (23.6%)	0.9
CSP implantation, n (%)	11 (15.4%)	29 (22.5%)	0.2
Medical therapy
ACEI/ARB, n (%)	0 (0.0%)	49 (38.0%)
Beta‐blockers, n (%)	31 (47.7%)	50 (38.8%)	0.300
SGLT2 inhibitors, n (%)	20 (30.8%)	37 (28.7%)	0.09
Loop diuretics, n (%)	30 (46.2%)	72 (55.8%)	0.263
MRA n (%)	18 (27.7%)	37 (28.6%)	0.1
ARNI dose, twice daily
ARNI 24/26 mg	35 (53.8%)	—
ARNI 49/51 mg	24 (36.9%)	—
ARNI 97/103 mg	6 (9.3%)	—
Device manufacturer
Medtronic (CRT/D) n. / %	19 (29.2%)	35 (27.1%)	0.76
Medtronic (CRT/P)	9 (13.8%)	18 (13.9%)	0.98
Boston Scientific	16 (24.6%)	33 (25.6%)	0.88
Abbott (CRT/D)	7 (10.8%)	12 (9.3%)	0.75
Abbott (CRT/P)	5 (7.7%)	9 (6.9%)	0.86
Biotronik	9 (13.8%)	18 (13.9%)	0.98

Note: Values are expressed as mean ± standard deviation or number (percentage), as appropriate.

Abbreviations: ACEI, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; ARNI, angiotensin receptor–neprilysin inhibitor; BSA, body surface area; CRT, cardiac resynchronization therapy; CRT‐D, cardiac resynchronization therapy with defibrillator; CRT‐P, cardiac resynchronization therapy with pacemakerCSP, conduction system pacing; eGFR, estimated glomerular filtration rate; LBBB, left bundle branch block; LVEDDi, left ventricular end‐diastolic diameter indexed to body surface area; LVEDVi, left ventricular end‐diastolic volume index; LVEF, left ventricular ejection fraction; LVESDi, left ventricular end‐systolic diameter indexed to body surface area; LVESVi, left ventricular end‐systolic volume index; MRA, mineralocorticoid receptor antagonist; NYHA, New York Heart Association; SBP, systolic blood pressure; SGLT2, sodium–glucose cotransporter 2.

3.2. Follow Up

At follow‐up, NYHA functional class improved in both groups, with a trend toward a lower class in ARNI patients (1.5 ± 0.55 vs. 1.6 ± 0.70, p = 0.09). The degree of NYHA improvement (ΔNYHA) was comparable (0.6 vs. 0.7, p = 0.60). Final QRS duration after CRT did not differ between groups, either overall or when stratified by CSP and BiVp. Echocardiographic reverse remodeling was evident in both cohorts, with final LV volumes and diameters showing no significant differences. A trend toward lower LVEDVi was observed in the ARNI group (80.1 vs. 81.4 mL/m², p = 0.06). Final LVEF was slightly higher in ARNI patients (43.4% vs. 40.6%, p = 0.14). Crucially, the absolute increase in LVEF (ΔEF) was significantly greater in patients receiving ARNI (14.4% ± 10.2% vs. 10.8% ± 11.0%, p = 0.04), confirming the synergistic contribution of ARNI therapy to reverse remodeling in CRT recipients, as reported in Table 2. A total of 65 patients were treated with ARNI. Among them, 44 were classified as echocardiographic responders, 44 as clinical responders. Regarding medical therapy, none of the patients receiving ARNI were treated with ACEI/ARB, whereas 38% of those without ARNI received ACEI/ARB. Beta‐blocker use was similar between groups (47.7% in the ARNI group vs. 38.8% in the non‐ARNI group, p = 0.300). Conversely, loop diuretics were more commonly used in the non‐ARNI group (55.8% vs. 46.2%, p = 0.263). Of these, 20.1% (39/194) underwent CSP with left bundle branch pacing as a bailout strategy compared to BiVp. Out of the entire study population (n = 194), 131 patients (67.5%) met the criteria for echocardiographic response, 140 patients (72.2%) as clinical response as reported in Table 3. These findings indicate a high overall rate of CRT response across different definitions. Subgroup analyses according to QRS morphology (LBBB vs. non‐LBBB) and pacing modality (BiVp vs. CSP) did not demonstrate significant interactions with the effect of ARNI on echocardiographic or clinical response (all interaction p > 0.10).

TABLE 2.

Clinical, electrocardiographic, and echocardiographic findings at 12‐month follow‐up according to ARNI treatment.

12 months	CRT with ARNI 65/194	CRT without ARNI 129/194	p
NYHA class	1.5 ± 0.55	1.63 ± 0.70	0.09
ΔNYHA	0.6 ± 0.67	0.71 ± 0.85	0.6
QRS duration, ms	112.75 ± 11.67	114.23 ± 13.78	0.7
Echocardiographic parameters
LVEDDi, mm/m2	27.49 ± 3.87	28.51 ± 3.70	0.1
LVESDi, mm/m2	20.49 ± 4.6	20.87 ± 5.1	0.9
LVEDVi, mL/m2	80.15 ± 36.3	81.38 ± 28.35	0.06
LVESVi, mL/m2	46.36 ± 31.75	48.25 ± 25.35	0.12
LVEF, %	43.4 ± 12.32	40.60 ± 10.71	0.14
ΔEF	14.38 ± 10.21	10.75 ± 10.96	0.04

Note: Values are expressed as mean ± standard deviation. ΔNYHA indicates the change in NYHA functional class from baseline to 12 months; ΔEF indicates the change in left ventricular ejection fraction from baseline to 12 months.

Abbreviations: ARNI, angiotensin receptor–neprilysin inhibitor; CRT, cardiac resynchronization therapy; LVEDDi, left ventricular end‐diastolic diameter indexed to body surface area; LVEDVi, left ventricular end‐diastolic volume index; LVEF, left ventricular ejection fraction; LVESDi, left ventricular end‐systolic diameter indexed to body surface area; LVESVi, left ventricular end‐systolic volume index; NYHA, New York Heart Association.

TABLE 3.

Clinical outcomes according to ARNI treatment among echocardiographic and clinical responders.

	ARNI (65/33.5)	No‐ARNI (129/66.5)	p
ECHO Responder n/% (131/67.5)	44/67.7	87/67.4	1.0
Dead (n)	0/0	6/6.9	0.19
HF hospitalization (patients) n/%	1/9.1	2/2.3	0.26
HF hospitalization (events)	4/9.1	1/1.1	0.17
RM HF (n)	1/2.3	13/14.9	0.02
Clinical Responder n/% (140/72.2)	44/67.7	96/74.4	0.41
Dead (n)	2/4.5	9/9.3	0.52
HF hospitalization (patients)	3/9.1	6/6.25	0.92
HF hospitalization (events)	4/9.1	8/8.3	0.62
RM HF (n)	3/6.8	16/16.6	0.04

Note: Data are presented as number or number (percentage), as appropriate. Echocardiographic response was defined as an increase in LVEF > 5% or a reduction in LV end‐systolic volume > 15% at 12 months; clinical response was defined as an improvement of at least one NYHA class.

Abbreviations: ARNI, angiotensin receptor–neprilysin inhibitor; HF, heart failure; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; RM‐HF, device‐detected remote monitoring heart failure event.

3.3. Univariate and Multivariate Analysis

In univariate analysis, ischaemic etiology was inversely associated with echocardiographic response, whereas higher baseline NYHA class predicted clinical response. In multivariate models, both associations remained significant as reported in Table 4. No other baseline variables were independently related to response.

TABLE 4.

Univariable and multivariable logistic regression analyses of predictors of echocardiographic and clinical response.

	Echo response OR (95% CI)	p univariate	p multivariate	Clinical response OR (95% CI)	p univariate	p multivariate
Age (per year)	0.98 (0.95–1.01)	0.16	—	0.99 (0.97–1.02)	0.53	—
Ischaemic etiology	0.45 (0.21–0.95)	0.039	0.030	1.24 (0.53–2.93)	0.63	—
Baseline NYHA class	1.02 (0.5–2.1)	0.865	—	2.59 (1.25–5.36)	0.010	0.001
QRS duration (per 10 ms)	1.03 (0.95–1.11)	0.51	—	1.04 (0.94–1.15)	0.42	—
LBBB morphology	1.53 (0.65–3.6)	0.325	—	0.91 (0.29–2.9)	0.87	—
Sex (male)	2.08 (0.75–5.8)	0.155	—	1.40 (0.65–3.0)	0.38	—

Note: Results are expressed as odds ratios (OR) with 95% confidence intervals (CI). Only variables with p < 0.10 in univariable analysis were entered into the multivariable model.

Abbreviations: CI, confidence interval; LBBB, left bundle branch block; multi, multivariate; NYHA, New York Heart Association; OR, odds ratio; uni, univariate.

3.4. Cardiovascular Events

Event rates are summarized in Table 3. Mortality and HF hospitalization did not differ significantly between ARNI and non‐ARNI patients across responder groups. By contrast, the incidence of device‐detected RM‐HF events were consistently lower in the ARNI group, reaching statistical significance in both echocardiographic and clinical responders. At 1 year freedom from RM‐HF events, pre‐specified as the key time‐to‐event endpoint, was consistently higher in patients receiving ARNI. In echocardiographic responders, event‐free survival was 95.7% with ARNI versus 82.6% without ARNI (log‐rank p = 0.021; q = 0.042 after FDR adjustment). In clinical responders, corresponding rates were 93.4% versus 85.3% (p = 0.040; q = 0.040). These findings indicate that ARNI use was associated with a reduced burden of subclinical events, with significantly higher event‐free survival in echocardiographic and clinical responders (Figure 1). Mortality, HF hospitalization and composite endpoint curves did not differ significantly between groups and are presented as exploratory analyses in the Supporting Information (Figure S1a–f).

FIGURE 1.

Kaplan–Meier curves in echo‐responder and clinical responder. (a) Freedom from subclinical heart failure events (RM HF index alerts) was 95.7% in the ARNI group compared with 82.6% in the non‐ARNI group (log‐rank p = 0.021; q = 0.042 after FDR adjustment). (b) Freedom from subclinical heart failure events (RM HF index alerts) was 93.4% in the ARNI group compared with 85.3% in the non‐ARNI group (log‐rank p = 0.04; q = 0.040 after FDR adjustment). Numbers at risk are reported below each Kaplan–Meier curve.

4. Discussion

Patients receiving ARNI exhibited a greater improvement in left ventricular ejection fraction compared with those on conventional therapy, consistent with the established role of ARNI in promoting reverse remodeling [11]. The proportion of patients treated with ARNI in this cohort is consistent with contemporary real‐world reports, where uptake typically ranges between 25%–40%, and younger patients are more likely to tolerate and maintain ARNI therapy [2, 12]. Similarly, the distribution of ARNI doses reflected routine clinical practice in a CRT population, where dose uptitration is frequently limited by haemodynamic status, blood pressure, renal function, and overall tolerability. Compared with contemporary randomized heart failure trials, the implementation of GDMT was lower in our cohort. This likely reflects the real‐world characteristics of patients referred for CRT, who are typically evaluated for device therapy after optimization of medical treatment but frequently remain unable to undergo further pharmacological intensification because of haemodynamic profile, blood pressure, renal function, or overall drug tolerability. Therefore, our study reflects contemporary routine CRT practice rather than the highly selected populations enrolled in randomized pharmacological trials. Accordingly, the present findings should be interpreted as representative of contemporary real‐world CRT practice rather than of the highly selected populations enrolled in randomized pharmacological trials. Importantly, in our series the overall response rate was relatively high, with a substantial increase in ΔEF. Patients receiving ARNI exhibited a greater improvement in left ventricular ejection fraction compared with those on conventional therapy, consistent with the established role of ARNI in promoting reverse remodeling. Our findings are in line with pivotal trials such as PARADIGM‐HF, which established ARNI as superior to ACE inhibitors in HFrEF [12]. However, data specifically addressing patients undergoing CRT remain scarce. A few observational studies have suggested improved LV remodeling and clinical outcomes when ARNI is combined with CRT, but results have been heterogeneous [13, 14, 15]. In line with previous observations, ischaemic etiology emerged as a negative predictor of echocardiographic response, while a more advanced baseline NYHA class was associated with greater clinical improvement. These findings suggest that underlying substrate and symptom severity remain key determinants of CRT response, against which the additional effect of ARNI should be interpreted. Importantly, once mechanical resynchronization was achieved, the association between ARNI use and improved outcomes appeared consistent across QRS morphologies and pacing modalities, suggesting that the benefit of ARNI becomes evident irrespective of baseline electrical substrate.

4.1. Role of ARNI Therapy

The most striking finding relates to the impact of ARNI therapy [13, 16]. When stratified by responder status, ARNI consistently reduced the event rate across subgroups, but the effect was particularly evident in responders. Importantly, responders on ARNI exhibited a greater increase in ΔEF compared with responders not receiving ARNI, suggesting a synergistic effect between reverse remodeling induced by resynchronization and neurohormonal modulation. From a mechanistic standpoint, ARNI is known to enhance reverse remodeling by combining neprilysin inhibition with RAAS blockade. In patients with electrical dyssynchrony and pacing indications, CRT restores atrioventricular, intraventricular, and interventricular synchrony, and improves cellular coupling at the myocyte level. This resynchronization may provide a more favorable haemodynamic substrate, facilitating drug tolerance and efficacy. It is plausible that the observed increase in LVEF among responders on ARNI reflects both improved forward stroke volume and enhanced tolerance to vasodilatory and neurohormonal effects, thereby allowing ARNI to exert its remodeling potential more effectively [14, 17]. Conversely, in non‐responders with persistent low LVEF and limited stroke volume reserve, the benefit of ARNI may be attenuated, explaining why its incremental impact appears less striking in this subgroup compared with HFrEF patients without CRT [13, 14, 15, 16]. Taken together, our data suggest that CRT may provide the mechanical substrate enabling ARNI to exert its therapeutic potential. In contrast, in patients with HFrEF without an indication to CRT, ARNI alone exerts a more pronounced effect on reverse remodeling and outcomes. This duality highlights an important clinical implication: in candidates for CRT, it may be unwise to delay device implantation while awaiting reverse remodeling on medical therapy alone (including ARNI), as the likelihood of achieving it is limited without resynchronization. The addition of ARNI therapy may modulate epigenetic mechanisms involved in adverse cardiac remodeling. In particular, ARNI has been shown to upregulate miR‐18 and miR‐145 and downregulate miR‐181, molecular changes associated with reduced inflammation, attenuation of fibrosis and hypertrophy, and ultimately with improved left ventricular function [14]. While CRT primarily acts by correcting electromechanical dyssynchrony and thereby optimizing residual systolic performance, ARNI targets distinct biological pathways that favor reverse remodeling. This complementary action may explain the conversion of a substantial proportion of initial non‐responders into responders and is consistent with the greater ΔEF observed in our cohort, suggesting a synergistic effect of CRT and ARNI on different but converging targets [14]. The lower incidence of positive RM HF index alerts in patients treated with ARNI suggests that this therapy may attenuate subclinical congestion episodes, likely by improving haemodynamic stability and reverse remodeling. HF hospitalization represents the final stage of a progressive process, whereas RM‐HF alerts are designed to identify earlier phases of subclinical haemodynamic deterioration. Therefore, evaluating RM‐HF events provides complementary information to conventional hard endpoints by capturing clinically actionable episodes before overt decompensation requiring hospitalization. In non‐responders, the impact of ARNI was attenuated, underscoring the central role of resynchronization as a prerequisite for maximizing pharmacological benefit [16]. From a clinical perspective, the reduction in RM‐HF index events is of particular relevance, as it reflects early decompensation detected by device‐based monitoring, often preceding overt HF hospitalization. Thus, ARNI appears to provide an additional layer of protection against congestion once CRT has achieved effective ventricular resynchronization. Importantly, although hospitalization rates for overt HF did not differ between groups, this may be explained by the fact that subclinical RM HF alerts were recognized and treated earlier in ARNI patients, thus preventing progression to clinically manifest decompensation. These findings support the role of ARNI in reducing the burden of subclinical events, aligning with its established effects on reverse remodeling and overall clinical stability.

4.2. Events in Responder Population

In our study, the clinical effect of ARNI was most evident on subclinical congestion, as captured by device‐detected RM‐HF events. These events were pre‐specified as the key time‐to‐event endpoint and remained significantly reduced in ARNI recipients after multiplicity adjustment. By contrast, mortality and HF hospitalization did not differ significantly between groups, reflecting the limited number of hard events and their exploratory nature. The Kaplan–Meier analyses highlight that in both echocardiographic and clinical responders, ARNI use was consistently associated with improved event‐free survival, driven by a lower incidence of RM‐HF alerts. These alerts reflect early pulmonary congestion or haemodynamic deterioration, typically managed in the outpatient setting before progression to overt decompensation [18, 19, 20, 21]. Furthermore, all alerts underwent clinical validation before therapeutic intervention, reducing the potential impact of false‐positive device notifications. This mechanism may explain why differences in hospitalization rates were not observed: early recognition and treatment of subclinical episodes likely prevented progression to clinically manifest events. From a prognostic perspective, responders to CRT already exhibited fewer adverse outcomes, consistent with the clinical relevance of reverse remodeling [20, 21]. Within this context, ARNI appears to provide additional benefit by stabilizing haemodynamics and reducing recurrent subclinical events. Residual mortality in CRT recipients is largely attributable to progressive HF rather than arrhythmias, and sudden cardiac death rates are relatively low [22]. Thus, by further enhancing reverse remodeling and reducing early decompensation, ARNI may complement CRT in addressing the remaining burden of disease progression [23]. Nevertheless, RM‐HF events were derived from manufacturer‐specific algorithms with different diagnostic performance. Although every alert prompted a standardized clinical reassessment before therapeutic intervention, this variability should be considered when interpreting the present findings.

4.3. Study Limitations

This study has several limitations. First, it was a retrospective, single‐center analysis, and despite the use of multivariable models, residual confounding cannot be excluded. Second, although our cohort reflects real‐world clinical practice, the relatively small sample size and the modest number of hard events limit the statistical power, particularly for mortality and hospitalization. In this context, multiplicity adjustment was applied only to RM‐HF events, which were pre‐specified as the key time‐to‐event endpoint. Analyses of mortality, hospitalization, and the composite endpoint were underpowered and therefore reported with nominal p‐values as exploratory findings. Finally, the study population included both BiVp and CSP recipients, with CSP mainly used as a bailout strategy; this heterogeneity reflects contemporary practice but may have influenced outcomes. In addition, RM‐HF events were generated by manufacturer‐specific remote monitoring algorithms with different diagnostic characteristics. Although all alerts underwent standardized outpatient clinical validation before therapeutic intervention, differences among algorithms may have influenced event detection and should be considered when interpreting the results. Although the inclusion of both conventional BiV‐CRT and bailout LBBAP introduces procedural heterogeneity and may limit the generalizability of the findings, this strategy reflects contemporary clinical practice. LBBAP was used exclusively as a bailout strategy to avoid ineffective CRT due to unsuccessful coronary sinus lead implantation or suboptimal left ventricular lead positioning, thereby ensuring delivery of resynchronization therapy whenever feasible.

5. Conclusion

In this real‐world cohort, CRT was associated with high response rates and marked improvements in LVEF, likely reflecting the combined effect of optimal lead positioning and the adoption of bailout CSP strategies. The interplay between device therapy and pharmacological treatment was most evident for ARNI. While ARNI consistently reduced the event rate, its effect was particularly strong in responders, who also demonstrated greater functional recovery. These findings suggest that CRT provides the substrate that enables ARNI to express its full potential, through improved synchrony, haemodynamics, and drug tolerance. Conversely, in non‐responders or in patients awaiting CRT, the incremental benefit of ARNI appears attenuated, reinforcing the concept that reverse remodeling in dyssynchronous HF is unlikely to occur with pharmacological therapy alone. Overall, our data support an integrated approach in which CRT and ARNI act synergistically. In patients with an indication for resynchronization, delaying CRT in favor of prolonged medical optimization may be inappropriate, as the greatest benefit of ARNI emerges once effective resynchronization has been achieved. Our analysis contributes novel evidence by highlighting that the prognostic effect of ARNI may be contingent upon achieving CRT response, thus supporting an integrated approach where device‐based resynchronization and pharmacological therapy act synergistically.

Author Contributions

A.S. conceived the study, collected the data, performed the statistical analysis, drafted the manuscript, and critically revised the paper. L.S. and M.C. contributed to data collection, manuscript review, and revision. C.B., B.M.N., S.L., A.P., and M.F. critically revised the manuscript. M.F. provided supervision and contributed to manuscript review. A.P. reviewed the manuscript, provided supervision, and contributed to the final interpretation of the conclusions. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work; the manuscript or revised it critically for important intellectual content, approved the final version, and agree to be accountable for all aspects of the work.

Funding

The authors have nothing to report.

Ethics Statement

The study was conducted in accordance with the Declaration of Helsinki and approved by the local ethics committee.

Consent

All patients provided written informed consent for participation and for the use of anonymized data for research purposes.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Figure S1: Kaplan–Meier curves for exploratory outcomes in ARNI vs. non‐ARNI patients. Numbers at risk are reported below each Kaplan–Meier curve.

Table S1: Exploratory outcomes (Kaplan–Meier log‐rank p‐values).

Data Availability Statement

The data underlying this article are available from the corresponding author upon reasonable request.