FormalPara Key Summary Points

Why carry out this study?

Cardiac resynchronization therapy (CRT) fails to improve echocardiographic parameters and clinical outcome in up to 40% of patients.

Sacubitril/valsartan (SV) applied as a replacement for angiotensin-converting enzyme 2 inhibitor (ACE2i) or angiotensin receptor blocker (ARB) resulted in left ventricular (LV) reverse remodeling and improved clinical outcomes in patients with heart failure with reduced ejection fraction (HFrEF).

Herein, we investigated whether SV would also result in cardiac remodeling in the specific subgroup of CRT non-responders (CRT-NRs), similarly to what was demonstrated in general patients with HFrEF.

What was learned from the study?

This study demonstrated that a minimum 6-month treatment with SV induced LV reverse remodeling in CRT-NRs, evidenced by a significant increase in left ventricular ejection fraction (LV EF), a reduction in LV end-systolic diameter, and a decrease in plasma N-terminal pro-B-type natriuretic peptide (NT-proBNP) concentration.

The extent of improvement in CRT non-responders was similar to what was observed in general patients with HFrEF with no CRT.


Cardiac resynchronization therapy (CRT) is an established treatment for heart failure (HF) with reduced ejection fraction (HFrEF), intraventricular delay, and mild to moderate HF symptoms despite guideline-dictated medical treatment [1,2,3]. CRT has been demonstrated to induce reverse remodeling, improve quality of life, exercise capacity, HF-related mortality, and hospitalization rates [4, 5]. However, in up to 40% of patients, CRT fails to improve echocardiographic parameters and clinical outcome [6,7,8]. These patients are referred to as CRT non-responders (CRT-NR) based on various criteria used in different studies to describe this clinical entity in the absence of a widely accepted definition. Functional improvement (NYHA (New York Heart Association) functional class, 6-min walk test, and quality of life assessments), the rate of HF hospitalizations, changes in echocardiographic parameters—including a reduction in left ventricular end-systolic volume (LVESV), an improvement in left ventricular ejection fraction (LV EF), or the combination of these measures—have been proposed to assess the response to CRT at different time points, usually at the 6-month or at the 1-year follow-up after device implantation [9]. The rate of non-response is related to the criteria used: a definition of response based on echocardiographic parameters usually results in higher rates of non-responders as compared to a definition based on functional or quality of life assessment. Despite extensive clinical research focusing on patient selection, left ventricular lead placement, and device programming, the phenomenon of CRT non-response is still not fully understood, and the prognosis of CRT-NR patients remains poor [6,7,8,9,10]. In addition, data from the ADVANCE CRT registry indicated that many CRT-NR patients are managed passively in clinical practice with no effort to maximize available therapeutic options—including medical treatment—to improve life expectancy and quality of life [7].

The introduction of sacubitril/valsartan (SV), a new therapeutic class of agents acting on the renin–angiotensin system (RAS) and the neutral endopeptidase system (ARNI, angiotensin receptor neprilysin inhibitor) is considered a highly significant development in HF therapy. SV is recommended as part of the baseline therapy with class I indication according to recent guidelines [1, 3]. These recommendations were based on the significant improvement in both cardiovascular mortality and HF hospitalization with SV applied as a replacement for angiotensin-converting enzyme 2 inhibitor (ACE2i) or angiotensin receptor blocker (ARB) demonstrated in the PARADIGM study [11]. Available data suggest that the outcome benefit with SV is related to left ventricular (LV) remodeling. The improvement in LV contractility and the decrease in LV diameters are accompanied by a significant reduction in the plasma level of NT-pro-B-type natriuretic peptide (NT-proBNP) which can be detected within 6–12 months after the initiation of SV therapy in patients with HFrEF [12, 13].

However, the efficacy of SV in the specific subgroup of patients with HFrEF and previous CRT implantation—and specifically in those who did not respond to CRT—remains ill defined. In a few observational studies which enrolled a limited number of CRT-NR patients and had relatively short follow-up durations, the echocardiographic signs of cardiac remodeling, along with the improvement in functional status, quality of life, as well as reduced rates of hospitalization and mortality have been demonstrated after the initiation of SV [14,15,16].

Herein, we compared the short-term changes in echocardiographic parameters and in NT-proBNP levels in three groups of patients: 1. CRT-NR patients treated with SV; 2. CRT-NR patients treated with ACEi or ARB; 3. patients with HFrEF with no indication for CRT (general HFrEF cohort) treated with SV. The rationale behind this study design was to evaluate the potential benefit of switching ACEi/ARB medication to SV in CRT-NR patients and to compare the extent of improvement with SV treatment in a CRT-NR versus in a general HFrEF cohort.


The clinical database for patients with heart failure in our department was searched to identify patients with HFrEF with LV EF < 40% treated as an inpatient or outpatient between January 2018 and June 2021 as well as consecutive patients who had CRT implantation based on standard indications (pacemaker or implantable cardioverter-defibrillator (ICD) at the discretion of the treating physician) during the same period. Management of patients including medical treatment was guided by their treating physician. Only patients receiving guideline-dictated medical therapy including ACEi/ARB or SV were considered for this analysis. The responder status of CRT patients was determined by comparing the LV EF obtained before and at least 6 months after the implantation. CRT non-responder status was defined as a less than 10% increase in LV EF measured on two-dimensional (2D) echocardiography using Simpson’s biplane method. CRT patients with biventricular capture rate below 95% were excluded from this study. The database was searched for further information on the medical treatment of these patients collected at routine follow-up visits every 3–6 months during the study period. The study was approved by the Medical Research Council (Ministry of Interior, P.O. Box 314, Budapest, 1903, Hungary) File no. BMEÜ/4388-1/2022/EKU. This study was performed in accordance with the Helsinki Declaration of 1964 and its later amendments. Use of the clinical database was approved by the head of department.

Patients were enrolled and assigned to one of three treatment groups based on the following criteria (Fig. 1):

Fig. 1
figure 1

Group assignments of patients. HFrEF heart failure with reduced ejection fraction, SV sacubitril/valsartan, CRT cardiac resynchronization therapy, CRT-NR non-responder to cardiac resynchronization therapy, ACEi/ARB angiotensin-converting enzyme inhibitors/angiotensin receptor blockers

Group I:

CRT-NR patients initiated on SV.

SV was started during the study period as a replacement for ACEi/ARB therapy and the results of 2D echocardiography and plasma NT-proBNP concentration obtained before the initiation of SV and after at least 6 months on SV were available.

Group II:


Patients were receiving evidence-based medical treatment including ACEi/ARB during the whole study period, and the results of 2D echocardiography and plasma NT-proBNP concentration from two time points (minimally 6 months apart) were available.

Group III:

Patients with HFrEF initiated on SV.

These patients represented a general HFrEF cohort with no indication for CRT implantation based on LV EF or QRS (representing ventricular depolarization) duration criteria. SV was started during the study period as a replacement for ACEi/ARB therapy and the results of 2D echocardiography and plasma NT-proBNP concentration obtained before the initiation of SV and after at least 6 months on SV were available.

In addition to the treatment with ACEi/ARB or SV, patients received beta-blockers (BB), mineralocorticoid receptor antagonists (MRA), and diuretics as baseline therapy. SV was always initiated after the discontinuation of ACEi/ARB with a daily starting dose of 100 mg (24/26 BID = two times a day) or 200 mg (49/51 BID) and titrated to the maximum tolerated dose up to 400 mg daily (97/103 mg BID).


The primary endpoints of the study were the changes in LV EF and in NT-proBNP levels between baseline and the final measurements at the end of follow-up in each group. In groups I and III, baseline measurements were performed before the initiation of SV and the final measurements at the end of follow-up on S/V treatment for at least 6 months. In group II, patients had been on continuous ACEi/ARB therapy even before device implantation and therefore measurements used for comparison were not related to a change in therapy. The first and second measurements used for comparison were performed at least 6 months apart on steady ACEi/ARB therapy. The extent of changes obtained for these primary endpoints in groups I and III was also compared. Changes in other echocardiographic parameters including end-diastolic and end-systolic diameters (LVEDD, LVESD), left ventricular outflow tract velocity time integral (LVOT VTI), stroke volume and dP/dt (representing the ratio of pressure change in the ventricular cavity during the isovolumetric contraction period) were considered as secondary endpoints. Systolic and diastolic blood pressure, estimated glomerular filtration rate (eGFR), and potassium levels were evaluated as safety endpoints.

Statistical Analysis

Statistical calculations were performed using IBM SPSS 26 and STATA V17 (StataCorp. 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC). Continuous data were described as means ± standard deviation (SD) or as median [IQR], and discrete variables were reported as case counts and percentages. Missing data were replaced using the last observation carried forward (LOCF) method. The distribution of continuous variables was characterized by the Kolmogorov–Smirnov test. Most of the data series showed a non-normal distribution, and the tests were selected accordingly. To compare treatment groups, we used the Kruskal–Wallis test, the Mann–Whitney test, and Student’s t test. The Wilcoxon test was used to compare data obtained at different time points. For discrete variables, we used the Pearson’s chi-squared test and the Fisher’s exact test. Multiple linear regression or multiple robust regression models were created to adjust for potential confounders. A value of p < 0.05 was considered statistically significant.


Baseline Patient Data

A total of 275 patients were enrolled, including 70 patients in group I, 70 patients in group II, and 135 patients in group III. Baseline patient data are listed in Table 1. Patients in both CRT groups (groups I and II) were older and had more comorbidities as compared with the general HFrEF cohort (group III), with no difference in the etiology of HF. The majority of patients were in NYHA functional class III in all three groups. LV EF was significantly higher in group II as compared to the other two groups. Mean levels of NT-proBNP were similar in the three groups (Table 1).

Table 1 Baseline patient characteristics

Medical Therapy

Maximum tolerated doses of SV are displayed in Fig. 2. Maximum tolerated daily doses were 100 mg (16 patients, 22.8%), 200 mg (35 patients, 50%), and 400 mg (19 patients, 27.1%) in group I and 100 mg (39 patients, 28.8%), 200 mg (50 patients, 37.1%), and 400 mg (46 patients, 34.1%) in group III. No statistical difference was found between the highest tolerated SV doses between groups I and III (p = 0.587).

Fig. 2
figure 2

Maximum doses of sacubitril/valsartan (SV) after titration in groups I and III

The basic principle of uptitrating all HF medications to the maximum tolerated dose was also applied to MRAs and beta-blockers in all three groups throughout the study. MRA doses were increased in 21 (30%), 12 (17%), and 28 (40%) while beta-blocker doses were increased in 25 (35.7%), 28 (40%), and 49 (36.2%) patients during the whole follow-up period in groups I, II and III, respectively. ACEi dose was increased in 13 patients (18.5%) in group II.

Primary Endpoints

The time duration between the baseline and the final measurements was 6–9 months (mean ± SD 7.54 ± 1.8 months for the whole patient cohort; 7.45 ± 1.6 for group I; 7.75 ± 2.2 for group II; 7.51 ± 1.5 for group III). LV EF increased between the baseline (prior to SV initiation) and the final (on SV for > 6 months) measurements in groups I and III (from 25.2 ± 5.7% to 29.4% ± 6.7%, p < 0.001; and from 26.6 ± 6.4% to 29.9 ± 6.7%, p < 0.001, respectively). With a multiple regression model adjusted for age, sex, LV EF before the initiation of SV, etiology, comorbidities (hypertension, diabetes, AF (atrial fibrillation), hyperlipidemia, chronic kidney disease) SV remained an independent predictor of the increase in LV EF both in group I (Coeff = 2.57; p = 0.010) and in group III (Coeff = 2.19; p = 0.014). No significant change in LV EF was demonstrated between the baseline and the final measurements in group II (28.3 ± 5.9% versus 29 ± 6.8%, p = 0.106; Fig. 3). A significant decrease in the serum levels of NT-proBNP (pg/mL) was demonstrated in groups I and III (from 2058.86 [1041.07–4502.51] to 1121.55 [545–2541], p < 0.001 and from 2223.35 [1233.03–4795.96] to 1123.09 [500.38–2651.27], p < 0.001, respectively). With a robust multiple regression model adjusted for age, sex, EF before the initiation of SV, etiology, comorbidities (hypertension, diabetes, AF, hyperlipidemia, chronic kidney disease) SV remained an independent predictor of the decrease in NT-proBNP both in group I (Coeff = − 763.66; p = 0.004) and in group III (Coeff = − 812.38; p = 0.001). No significant change in NT-proBNP was detected in group II (1474.57 [655.8–5273] versus 1986.3 [1025.3–3359.1], p = 0.807; Fig. 3). The extent of improvement between groups I and III was similar with no significant differences either for LV EF (p = 0.161) or for NT-proBNP (p = 0.850).

Fig. 3
figure 3

Primary endpoints. Left ventricular ejection fraction (LV EF; upper panel) and N-terminal pro-B-type natriuretic peptide (NT-proBNP) level (lower panel) changes during follow-up. SV (sacubitril/valsartan)

Echocardiographic parameters considered as secondary endpoints are displayed in Table 2. A significant decrease in the left ventricular end-systolic diameter (LVESD) was demonstrated on SV therapy from 56.6 ± 8.9 mm to 54.3 ± 8.7 mm; p = 0.004 and from 55.9 ± 9.9 mm to 54.3 ± 11.2 mm; p = 0.021 in groups I and III, respectively. No significant changes were detected in any other parameters. Systolic and diastolic blood pressure values decreased on SV therapy in both groups, while eGFR (estimated glomerular filtration rate) decreased in group I and potassium levels were elevated in group III (Table 3). The discontinuation of SV was not required in any patient. The potential clinical implications of the data in Tables 2 and 3 are discussed.

Table 2 Echocardiographic parameters
Table 3 Safety parameters


Main Findings

In this observational study a significant LV reverse remodeling evidenced by an increase in LV EF and a decrease in LVESD and in the level of NT-proBNP were detected in CRT-NR patients after 6–9 months of SV therapy. The extent of improvement was similar to what was found in general patients with HFrEF on SV therapy, while no improvement was demonstrated in CRT non-responders who remained on evidence-based HF therapy including ACEi/ARB. Importantly, improvements in the CRT-NR cohort were observed despite these patients being older and having lower LV EF, higher NT-proBNP values, and more comorbidities (hypertension, diabetes, atrial fibrillation, and dyslipidemia) as compared with the other groups.

SV therapy was associated with a significant decrease in systolic and diastolic blood pressures in both patient groups. eGFR decreased in CRT-NR patients, while potassium levels increased in patients with HFrEF with no need to stop the therapy in any of them. Although statistically significant, these changes related to SV treatment had no clinical relevance, as no therapy discontinuation was required in any patient.

Clinical Implications

More than two decades after the introduction of CRT into clinical practice, non-response to resynchronization remains a problem [17]. CRT-NR patients show high hospitalization and less than 50% survival rates free of assist device or cardiac transplantation at 5 years after the implantation [10]. Unlike patients with other chronic diseases (e.g., malignancies), many of these patients with HF are less willing to seek medical support despite the substantial evidence on poor prognosis [8]. Moreover, data from the ADVANCE CRT registry indicate that in the absence of a widely accepted consensus on the definition of this condition, patients may not be categorized properly as non-responders to CRT [7]. Furthermore, tighter follow-up and therapy intensification are not offered by many physicians even to those identified as non-responders. Consequences of this inertia include deteriorating functional status, quality of life and reduced life expectancy in these patients despite recent developments in the medical therapy of HF.

To our knowledge, this is the first study which evaluated shorter-term changes both in the echocardiographic parameters and plasma levels of NT-proBNP in CRT-NRs in response to SV by comparing the results not only to those who were kept on ACE/ARB therapy but as well as to a general HFrEF cohort treated with SV. This double comparison design including two control groups allowed us to prove that the improvement in CRT-NR patients was truly due to SV therapy and did not simply reflect a natural fluctuation in echocardiographic parameters and biomarker levels. It was also confirmed that the extent of improvement induced by SV therapy was similar in CRT-NR and in general patients with HFrEF. The statistically significant improvements in echocardiographic and NT-proBNP measurements observed on SV therapy in groups I and III are also of clinical relevance as they suggest LV reverse remodeling, a known predictor of favorable outcome. Whether hospitalization and mortality endpoints in CRT-NRs on SV treatment will also be similar to what was demonstrated by the large-scale randomized PARADIGM trial in patients with HFrEF [11] should be answered by further investigations. The results of a few observational studies on the effects of SV therapy in CRT-NRs also support these expectations [14,15,16]. In a retrospective analysis, lower cardiac mortality was proven after 6-month follow-up in 22 CRT-NR patients who were started on SV treatment as compared to those 28 who remained on ACEi/ARB medication [14]. Improvement in quality-of-life indicators and a reduction in hospitalization rates were demonstrated after 6-month treatment with SV in the RESINA (Resynchronization plus an Inhibitor of Neprilysin/Angiotensin) registry [15]. These observations need to be confirmed by multicenter randomized clinical trials involving a sufficient number of CRT non-responders and a longer follow-up period.

Of note, CRT patients in our study were non-responders and therefore our results may not apply to all CRT patients. Indeed, the absence of CRT was a predictor of reverse LV remodeling after the initiation of SV in a registry study on patients with HFrEF including 43% with an implanted CRT device [18]. The authors’ proposed explanation for this finding was that the potential myocardial reserve was already realized by cardiac resynchronization, thereby leaving no room for further improvement with SV. However, the information on whether these patients included CRT responders or non-responders was not disclosed and therefore these observations may not apply to CRT-NR patients.

Although 2D echocardiography represents the gold standard to assess patients with HF, significant interobserver variations pose a well-known shortcoming of this diagnostic modality. The correlation between the echocardiographic features of LV remodeling and the reduction of plasma NT-proBNP concentrations has been substantiated in many studies [13, 19,20,21,22]. Moreover, the level of this biomarker is a known predictor of long-term outcome [19, 21, 22]. Importantly, improvements in the echocardiographic parameters observed in our study were validated by a significant decrease in plasma NT-proBNP concentrations also to a similar extent as in the CRT-NR and in the general (group III) HFrEF cohort.

Limitations and Strengths

This is a single-center observational study with a relatively short observation period including a limited number of patients and a low female representation; nevertheless, the population size was still larger in our analysis than in the few reports published so far on the efficacy of SV in CRT-NR patients. Significant differences were found in the baseline parameters of the three groups. However, improvements on SV therapy were observed in the CRT-NR cohort despite these patients being older and having lower LV EF, higher NT-proBNP values, and more comorbidities as compared with the other groups. In the absence of widely accepted standard criteria, the definitions of CRT non-responders were arbitrary in our study. We decided to use LV EF as a readily available parameter measured routinely in daily practice. To our knowledge, our work is the only one so far to include NT-proBNP measurements to validate the echocardiographic assessment of reverse LV remodeling in CRT non-responders treated with SV. Although this was a registry study, the brief period of data collection and the uniform principles applied for HF management at our institute ensured that the baseline characteristics of the patients assigned to the different groups were comparable and showed only minor differences. Our CRT-NR patients demonstrated a significant improvement on SV despite more comorbidities and lower baseline LV EF values as compared with the other two groups. With the inclusion of two control cohorts, we were able to demonstrate that the favorable changes detected in CRT-NR patients on SV were indeed the result of the therapy, and the extent of improvements was similar to what was detected in general patients with HFrEF treated with SV. In the future, most patients will likely undergo CRT implantation while on treatment with SV and an SGLT2 inhibitor (sodium-glucose co-transporter 2 inhibitor). However, as of today, our findings are still relevant to many patients who demonstrate no significant improvement after CRT implantation.


Cardiac remodeling as evidenced by improvements in echocardiographic parameters and in NT-proBNP levels were demonstrated to a similar extent in CRT non-responders who were started on SV therapy as well as in a general HFrEF cohort. Our results support the early replacement of ACEi/ARB with SV at the largest tolerated dose in CRT-NRs in the absence of contraindication or intolerance to the medication.